Advances in Production Management Systems. Production Management Systems for Responsible Manufacturing, Service, and Logistics Futures: IFIP WG 5.7 ... and Communication Technology, 689) [1st ed. 2023] 303143661X, 9783031436611

Citation preview

IFIP AICT 689

Erlend Alfnes Anita Romsdal Jan Ola Strandhagen Gregor von Cieminski David Romero (Eds.)

Advances in Production Management Systems Production Management Systems for Responsible Manufacturing, Service, and Logistics Futures

IFIP WG 5.7 International Conference, APMS 2023 Trondheim, Norway, September 17–21, 2023 Proceedings, Part I

123

IFIP Advances in Information and Communication Technology

689

Editor-in-Chief Kai Rannenberg, Goethe University Frankfurt, Germany

Editorial Board Members TC 1 – Foundations of Computer Science Luís Soares Barbosa , University of Minho, Braga, Portugal TC 2 – Software: Theory and Practice Michael Goedicke, University of Duisburg-Essen, Germany TC 3 – Education Arthur Tatnall , Victoria University, Melbourne, Australia TC 5 – Information Technology Applications Erich J. Neuhold, University of Vienna, Austria TC 6 – Communication Systems Burkhard Stiller, University of Zurich, Zürich, Switzerland TC 7 – System Modeling and Optimization Lukasz Stettner, Institute of Mathematics, Polish Academy of Sciences, Warsaw, Poland TC 8 – Information Systems Jan Pries-Heje, Roskilde University, Denmark TC 9 – ICT and Society David Kreps , National University of Ireland, Galway, Ireland TC 10 – Computer Systems Technology Achim Rettberg, Hamm-Lippstadt University of Applied Sciences, Hamm, Germany TC 11 – Security and Privacy Protection in Information Processing Systems Steven Furnell , Plymouth University, UK TC 12 – Artificial Intelligence Eunika Mercier-Laurent , University of Reims Champagne-Ardenne, Reims, France TC 13 – Human-Computer Interaction Marco Winckler , University of Nice Sophia Antipolis, France TC 14 – Entertainment Computing Rainer Malaka, University of Bremen, Germany

IFIP Advances in Information and Communication Technology The IFIP AICT series publishes state-of-the-art results in the sciences and technologies of information and communication. The scope of the series includes: foundations of computer science; software theory and practice; education; computer applications in technology; communication systems; systems modeling and optimization; information systems; ICT and society; computer systems technology; security and protection in information processing systems; artificial intelligence; and human-computer interaction. Edited volumes and proceedings of refereed international conferences in computer science and interdisciplinary fields are featured. These results often precede journal publication and represent the most current research. The principal aim of the IFIP AICT series is to encourage education and the dissemination and exchange of information about all aspects of computing. More information about this series at https://link.springer.com/bookseries/6102

Erlend Alfnes Anita Romsdal Jan Ola Strandhagen Gregor von Cieminski David Romero Editors •

•

•

•

Advances in Production Management Systems Production Management Systems for Responsible Manufacturing, Service, and Logistics Futures IFIP WG 5.7 International Conference, APMS 2023 Trondheim, Norway, September 17–21, 2023 Proceedings, Part I

123

Editors Erlend Alfnes Norwegian University of Science and Technology Trondheim, Norway Jan Ola Strandhagen Norwegian University of Science and Technology Trondheim, Norway

Anita Romsdal Norwegian University of Science and Technology Trondheim, Norway Gregor von Cieminski ZF Friedrichshafen AG Friedrichshafen, Germany

David Romero Tecnológico de Monterrey Mexico City, Mexico

ISSN 1868-4238 ISSN 1868-422X (electronic) IFIP Advances in Information and Communication Technology ISBN 978-3-031-43661-1 ISBN 978-3-031-43662-8 (eBook) https://doi.org/10.1007/978-3-031-43662-8 © IFIP International Federation for Information Processing 2023 This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed. The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. The publisher, the authors, and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors give a warranty, expressed or implied, with respect to the material contained herein or for any errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. This Springer imprint is published by the registered company Springer Nature Switzerland AG The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland Paper in this product is recyclable.

Preface

The year 2023 has undoubtedly been a year of contrasts. We are experiencing stunning developments in technology, and creating new products, services, and systems that are changing the way we live and work. Simultaneously, we are experiencing multiple conflicts around the world and the brutal effects of climate change. While many experience success and improved standards of living, others face threats to their lives and even loss. A Scientific Conference cannot change this but can be seen as a symbol for aiming for a different future. We create new knowledge and solutions, we share all our achievements, and we meet to create new friendships and meet people from all over the world. The International Conference on “Advances in Production Management Systems” (APMS) 2023 is the leading annual event of the IFIP Working Group (WG) 5.7 of the same name. At the Conference in Trondheim, Norway, hosted by the Norwegian University of Science and Technology (NTNU), more than 200 papers were presented and discussed. This is a significant step up from the first APMS Conference in 1980, which assembled just a few participants. The IFIP WG5.7 was established in 1978 by the General Assembly of the International Federation for Information Processing (IFIP) in Oslo, Norway. Its first meeting was held in August 1979 with all its seven members present. The WG has since grown to 108 full members and 25 honorary members. After 43 years, APMS has returned to the city where it started. The venue in 1980 was Lerchendal Gård, and the topic marked the turn of a decade: “Production Planning and Control in the 80s”. The papers presented attempted to look into the future – a future which at that time was believed to be fully digitalized. One foresaw that during the coming decade, full automation and optimization of complete manufacturing plants, controlled by a central computer, would be a reality. The batch processing of production plans would be replaced by online planning and control systems. No other technology can show a more rapid development and impact in industry and society than Information and Communication Technology (ICT). The APMS 2023 program shows that the IFIP WG5.7 still can make and will continue to make a significant contribution to production and production management disciplines. In 2023, the International Scientific Committee for APMS included 215 recognized experts working in the disciplines of production and production management systems. For each paper, an average of 2.5 single-blind reviews were provided. Over two months, each submitted paper went through two rigorous rounds of reviews to allow authors to revise their work after the first round of reviews to guarantee the highest scientific quality of the papers accepted for publication. Following this process, 213 full papers were selected for inclusion in the conference proceedings from a total of 224 submissions. APMS 2023 brought together leading international experts from academia, industry, and government in the areas of production and production management systems to discuss how to achieve responsible manufacturing, service, and logistics futures. This

vi

Preface

included topics such as innovative manufacturing, service, and logistics systems characterized by their agility, circularity, digitalization, flexibility, human-centricity, resiliency, and smartification contributing to more sustainable industrial futures that ensure that products and services are manufactured, servitized, and distributed in a way that creates a positive effect on the triple bottom line. The APMS 2023 conference proceedings are organized into four volumes, covering a large spectrum of research addressing the overall topic of the conference “Production Management Systems for Responsible Manufacturing, Service, and Logistics Futures”. We would like to thank all contributing authors for their quality research work and their willingness to share their findings with the APMS and IFIP WG5.7 community. We are equally grateful for the outstanding work of all the International Reviewers, the Program Committee Members, and the Special Sessions Organizers. September 2023

Erlend Alfnes Anita Romsdal Jan Ola Strandhagen Gregor von Cieminski David Romero

Organization

Conference Chair Jan Ola Strandhagen

Norwegian University of Science and Technology, Norway

Conference Co-chair Gregor von Cieminski

ZF Friedrichshafen AG, Germany

Conference Honorary Chair Asbjørn Rolstadås

Norwegian University of Science and Technology, Norway

Program Chair Erlend Alfnes

Norwegian University of Science and Technology, Norway

Program Co-chairs Heidi Carin Dreyer Daryl Powell Bella Nujen Anita Romsdal David Romero

Norwegian University of Science and Norway Norwegian University of Science and SINTEF Manufacturing, Norway Norwegian University of Science and Norway Norwegian University of Science and Norway Tecnológico de Monterrey, Mexico

Technology, Technology/ Technology, Technology,

Organization Committee Chair Anita Romsdal

Norwegian University of Science and Technology, Norway

Doctoral Workshop Chair Hans-Henrik Hvolby

Aalborg University, Denmark

viii

Organization

Doctoral Workshop Co-chair David Romero

Tecnológico de Monterrey, Mexico

List of Reviewers Federica Acerbi Luca Adelfio Natalie Cecilia Agerskans El-Houssaine Aghezzaf Rajeev Agrawal Carla Susana Agudelo Assuad Kosmas Alexopoulos Kartika Nur Alfina Erlend Alfnes Antonio Pedro Dias Alves de Campos Terje Andersen Joakim Andersson Dimitris Apostolou Germán Arana Landín Simone Arena Emrah Arica Veronica Arioli Nestor Fabián Ayala Christiane Lima Barbosa Mohadese Basirati Mohamed Ben Ahmed Justus Aaron Benning Aili Biriita Bertnum Belgacem Bettayeb Seyoum Eshetu Birkie Umit Sezer Bititci Klas Boivie Alexandros Bousdekis Nadjib Brahimi Greta Braun Gianmarco Bressanelli Jim J. Browne Patrick Bründl Kay Burow Jenny Bäckstrand Jannicke Baalsrud Hauge Robisom Damasceno Calado Luis Manuel Camarinha-Matos Violetta Giada Cannas

Ayoub Chakroun Zuhara Chavez Ferdinando Chiacchio Steve Childe Chiara Cimini Florian Clemens Beatrice Colombo Federica Costa Catherine da Cunha Flávia de Souza Yüksel Değirmencioğlu Demiralay Enes Demiralay Tabea Marie Demke Mélanie Despeisse Candice Destouet Slavko Dolinsek Milos Drobnjakovic Eduardo e Oliveira Malin Elvin Christos Emmanouilidis Hakan Erdeş Kristian Johan Ingvar Ericsson Victor Eriksson Adrodegari Federico Matteo Ferrazzi Jannick Fiedler Erik Flores-García Giuseppe Fragapane Chiara Franciosi Susanne Franke Enzo Frazzon Stefano Frecassetti Jan Frick Paolo Gaiardelli Clarissa A. González Chávez Jon Gosling Danijela Gračanin Daniela Greven Eric Grosse

Organization

Zengxu Guo Christopher Gustafsson Petter Haglund Lise Lillebrygfjeld Halse Trond Halvorsen Robin Hanson Stefanie Hatzl Theresa-Franziska Hinrichsen Maria Holgado Christian Holper Djerdj Horvat Karl Anthony Hribernik Hans-Henrik Hvolby Natalia Iakymenko Niloofar Jafari Tanya Jahangirkhani Tim Maximilian Jansen Yongkuk Jeong Kerstin Johansen Björn Johansson Bjørn Jæger Ravi Kalaiarasan Dimitris Kiritsis Takeshi Kurata Juhoantti Viktor Köpman Nina Maria Köster Danijela Lalić Beñat Landeta Nicolas Leberruyer Ming Lim Maria Linnartz Flavien Lucas Andrea Lucchese Egon Lüftenegger Ugljesa Marjanovic Julia Christina Markert Melissa Marques-McEwan Antonio Masi Gokan May Matthew R. McCormick Khaled Medini Jorn Mehnen Joao Gilberto Mendes dos Reis Hajime Mizuyama Eiji Morinaga Sobhan Mostafayi Darmian

Mohamed Naim Farah Naz Torbjørn Netland Phu Nguyen Kjeld Nielsen Ana Nikolov Sang Do Noh Antonio Padovano Julia Pahl Martin Perau Margherita Pero Mirco Peron Fredrik Persson Marta Pinzone Fabiana Pirola Adalberto Polenghi Daryl John Powell Rossella Pozzi Vittaldas Prabhu Hiran Harshana Prathapage Moritz Quandt Ricardo Rabelo Mina Rahmani Slavko Rakic Mario Rapaccini R. M. Chandima Ratnayake Eivind Reke Daniel Resanovic Ciele Resende Veneroso Irene Roda David Romero Anita Romsdal Christoph Roser Nataliia Roskladka Monica Rossi Martin Rudberg Roberto Sala Jan Salzwedel Adrian Sánchez de Ocaña Kszysztof Santarek Biswajit Sarkar Claudio Sassanelli Laura Scalvini Maximilian Schacht Bennet Schulz Marco Semini

ix

x

Organization

Sourav Sengupta Fabio Sgarbossa Vésteinn Sigurjónsson Marcia Terra Silva Katrin Singer-Coudoux Ivan Kristianto Singgih Lars Skjelstad Riitta Johanna Smeds Selver Softic Per Solibakke Vijay Srinivasan Kenn Steger-Jensen Oliver Stoll Jan Ola Strandhagen Jo Wessel Strandhagen Nick B. Szirbik Endre Sølvsberg Iris D. Tommelein Mario Tucci Ebru Turanoglu Bekar Ioan Turcin Arvind Upadhyay Andrea Urbinati

Mehmet Uzunosmanoglu Bruno Vallespir Ivonaldo Vicente da Silva Kenneth Vidskjold Vivek Vijayakumar Gregor von Cieminski Paul Kengfai Wan Piotr Warmbier Kasuni Vimasha Weerasinghe Shaun West Stefan Alexander Wiesner Joakim Wikner Magnus Wiktorsson Heiner Winkler Jong-Hun Woo Thorsten Wuest Lara Popov Zambiasi Matteo Zanchi Yuxuan Zhou Iveta Zolotová Anne Zouggar Mikael Öhma

Contents – Part I

Lean Management in the Industry 4.0 Era Enablers Identification to Support the Combined Implementation of Lean and Industry 4.0 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Ilse Urquia, Anne Zouggar Amrani, and Bruno Vallespir Lean and Digitalization Status in Manufacturing Companies Located in Norway . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Natalia Iakymenko, Daryl Powell, Eivind Reke, Marte Daae-Qvale Holmemo, Eirik Bådsvik Hamre Korsen, Signe Sagli, Sigrid Eliassen Sand, and Sunniva Økland Effects of Lean and Industry 4.0 Technologies on Job Satisfaction: A Case-Based Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Matteo Zanchi, Andrea Lorenzi, Matteo Prezioso, Daryl Powell, and Paolo Gaiardelli

3

15

27

Lean Supply Chain and Industry 4.0: A Study of the Interaction Between Practices and Technologies. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Matteo Rossini, Stefano Frecassetti, and Alberto Portioli-Staudacher

39

A Design Science – Informed Process for Lean Warehousing Implementation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Anna Corinna Cagliano, Giovanni Zenezini, Carlo Rafele, Sabrina Grimaldi, and Giulio Mangano

54

Digitally Enhancing Kanban Lean Practice in Support of Just-in-Time Reconfigurable Supply: A Case Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Christina Papadimitropoulou, Anne Zouggar Amrani, Daryl Powell, Helena Macedo, and David Romero

69

Sociotechnical Approach to Self-reporting in PMM Systems for HSE and Digital Security. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Jarle Nyberg and Sverre Sørbye Larsen

84

The Productivity Leap: Effects of an Industry Program for Norwegian SMEs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Eivind Reke, Natalia Iakymenko, and Mette Holmriis Bugger

97

xii

Contents – Part I

Integrating Smart Manufacturing to Lean: A Multiple-Case Study of the Impact on Shop-Floor Employees’ Autonomy and Empowerment . . . . . . . . . . 109 Thomas Bortolotti, Stefania Boscari, Etta Morton, and Daryl Powell Applying the Value Stream Map to Streamline Energy Consumption: Analysis of an Italian Company . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125 Matteo Ferrazzi and Alberto Portioli-Staudacher Crossroads and Paradoxes in the Digital Lean Manufacturing World A Systematic Literature Review on Combinations of Industry 4.0 and Lean Production . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139 Kristian Ericsson and Antonio Maffei Lean and Digital Strategy Role in Achieving a Successful Digital Transformation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157 Stefano Frecassetti, Anna Presciuttini, Matteo Rossini, and Alberto Portioli-Staudacher Tying Digitalization to the Lean Mindset: A Strategic Digitalization Perspective . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171 Victor Eriksson, Sourav Sengupta, Ann-Charlott Pedersen, Elsebeth Holmen, Heidi Carin Dreyer, Marte Daae-Qvale Holmemo, Signe Sagli, Sigrid Eliassen Sand, Sunniva Økland, Daryl Powell, Natalia Iakymenko, Serkan Eren, and Eirin Lodgaard Characterization of Digitally-Advanced Methods in Lean Production Systems 4.0 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 184 Simon Schumacher, Roland Hall, Michael Hautzinger, Jan Schöllmann, and Thomas Bauernhansl Synergies Between Industry 4.0 and Lean on Triple Bottom Line Performance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 200 Thomas Bortolotti, Stefania Boscari, Willem Grob, and Daryl Powell Driving Sustainability Through a VSM-Indicator-Based Framework: A Case in Pharma SME. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213 Zuhara Zemke Chavez, Mayari Perez Tay, Mohammad Hasibul Islam, and Monica Bellgran Design and Application of a Development Map for Aligning Strategy and Automation Decisions in Manufacturing SMEs . . . . . . . . . . . . . . . . . . . . 228 Malin Löfving, Peter Almström, Caroline Jarebrant, and Magnus Widfeldt

Contents – Part I

xiii

Using the Lean Approach for Improving Eco-Efficiency Performance: A Case Study for Plastic Reduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 242 Matteo Ferrazzi and Alberto Portioli-Staudacher Work Pattern Analysis with and without Site-Specific Information in a Manufacturing Line. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 253 Takeshi Kurata, Rei Watanabe, Satoki Ogiso, Ikue Mori, Takahiro Miura, Karimu Kato, Yasunori Haga, Shintaro Hatakeyama, Atsushi Kimura, and Katsuko Nakahira Digital Transformation Approaches in Production Management Digital Transformation Towards Industry 5.0: A Systematic Literature Review . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 269 Jelena Crnobrnja, Darko Stefanovic, David Romero, Selver Softic, and Ugljesa Marjanovic Industry 5.0 and Manufacturing Paradigms: Craft Manufacturing - A Case from Boat Manufacturing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 282 Bjørnar Henriksen and Maria Kollberg Thomassen Industry 4.0 Readiness Assessment of Enterprises in Kazakhstan . . . . . . . . . . 297 Dinara Dikhanbayeva, Malika Aitzhanova, Yevgeniy Lukhmanov, Ali Turkyilmaz, Essam Shehab, and Idriss El-Thalji Critical Factors for Selecting and Integrating Digital Technologies to Enable Smart Production: A Data Value Chain Perspective . . . . . . . . . . . . . . 311 Natalie Agerskans, Mohammad Ashjaei, Jessica Bruch, and Koteshwar Chirumalla Business Process Reengineering in Agile Manufacturing – A Mixed Method Research . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 326 Khadija Lahlou, Khaled Medini, Thorsten Wuest, and Qussay Jarrar Service-Oriented Architecture for Driving Digital Transformation: Insights from a Case Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 339 Omid Maghazei, Marco Messerli, Thomas Gittler, and Torbjørn Netland Application of Digital Tools, Data Analytics and Machine Learning in Internal Audit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 357 Jelena Popara, Milena Savkovic, Danijela Ciric Lalic, and Bojan Lalic Consumer Engagement in the Design of PLM Systems: A Review of Best Practices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 372 Uchechukwu Nwogu and Richard Evans

xiv

Contents – Part I

A Distributed Ledger Technology Solution for Connecting E-mobility Partners . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 386 Radu Ungureanu, Selver Softic, Emil St. Chifu, and Ioan Turcin Managing Digitalization of Production Systems Leveraging Advanced Digital Technology Practices to Enhance Information Quality in Low-Volume Product Introduction and Manufacturing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 401 Siavash Javadi and Koteshwar Chirumalla Evaluating Augmented Reality, Deep Learning and Paper-Based Assistance Systems in Industrial Manual Assembly . . . . . . . . . . . . . . . . . . . . . . . . . . . . 417 Alexander Riedel, Johanna Gerlach, Maximilian Dietsch, Frank Engelmann, Nico Brehm, and Tobias Pfeifroth Reinforcing the Closing of the Circular Economy Loop Through Artificial Intelligence and Robotics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 432 Waleska Sigüenza Tamayo, Naiara Uriarte-Gallastegi, Beñat Landeta-Manzano, and Germán Arana-Landin A New Generation? A Discussion on Deep Generative Models in Supply Chains . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 444 Eduardo e Oliveira and Teresa Pereira Business Context-Based Approach for Managing the Digitalization of Biopharmaceutical Supply Chain Operational Requirements . . . . . . . . . . . . . . 458 Elena Jelisic, Milos Drobnjakovic, Boonserm Kulvatunyou, Nenad Ivezic, and Hakju Oh Volunteering Service Engineering in Non-profit Organizations . . . . . . . . . . . . 471 Michael Freitag and Oliver Hämmerle Workforce Evolutionary Pathways in Smart Manufacturing Systems The Role of Organizational Culture in the Transformation to Industry 4.0 . . . . 487 Rogerio Queiroz de Camargo, Márcia Terra da Silva, Ana Lucia Figueiredo Facin, and Rodrigo Franco Gonçalves A Reflective Framework for Understanding Workforce Evolutionary Pathways in Industry 5.0 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 501 Alexandra Lagorio, Chiara Cimini, and David Romero

Contents – Part I

xv

Managing Change Towards the Future of Work - Clustering Key Perspectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 513 Katrin Singer-Coudoux, Greta Braun, and Johan Stahre Development of a Task Model for Artificial Intelligence-Based Applications for Small and Medium-Sized Enterprises . . . . . . . . . . . . . . . . . . 528 Florian Clemens, Fabian Willemsen, Susanne Mütze-Niewöhner, and Günther Schuh Indoor Positioning-based Occupational Exposures Mapping and Operator Well-being Assessment in Manufacturing Environment . . . . . . . . . . . . . . . . . 543 Gergely Halász, Tibor Medvegy, János Abonyi, and Tamás Ruppert Next Generation Human-Centered Manufacturing and Logistics Systems for the Operator 5.0 Human in Command in Manufacturing . . . . . . . . . . . . . . . . . . . . . . . . . . . . 559 Doris Aschenbrenner and Cecilia Colloseus Toward a Framework for Human-Technology Cooperation in Manufacturing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 573 Jannick Fiedler, Omid Maghazei, Arne Seeliger, and Torbjørn Netland The Role of Human Factors in Zero Defect Manufacturing: A Study of Training and Workplace Culture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 587 Foivos Psarommatis, Gökan May, and Victor Azamfirei Modeling Human Problem-Solving Behavior in Complex Production Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 602 Susanne Franke and Ralph Riedel Human-Centric Industrial Augmented Reality: Requirements and Design Guidelines for Usability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 617 Tiberiu Florescu, Sabine Waschull, and Christos Emmanouilidis Investigating Human Factors Integration into DT-Based Joint Production and Maintenance Scheduling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 633 Chiara Franciosi, Salvatore Miranda, Ciele Resende Veneroso, and Stefano Riemma Fostering Human-AI Collaboration with Digital Intelligent Assistance in Manufacturing SMEs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 649 Stefan Wellsandt, Mina Foosherian, Alexandros Bousdekis, Bernhard Lutzer, Fotis Paraskevopoulos, Yiannis Verginadis, and Gregoris Mentzas

xvi

Contents – Part I

Metaverse-Based Softbot Tutors for Inclusive Industrial Workplaces: Supporting Impaired Operators 5.0 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 662 Lara Popov Zambiasi, Ricardo José Rabelo, Saulo Popov Zambiasi, and David Romero Bridging the Hype Cycle of Collaborative Robot Applications . . . . . . . . . . . . 678 Omkar Salunkhe, David Romero, Johan Stahre, Björn Johansson, and Anna Syberfeldt Considering Gripper Allocations in Balancing of Human-Robot Collaborative Assembly Lines. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 691 Yüksel Değirmencioğlu Demiralay and Yakup Kara A Smart Work Cell to Reduce Adoption Barriers of Collaborative Robotics . . . 702 Elias Montini, Lorenzo Agbomemewa, Fabio Daniele, Vincenzo Cutrona, Matteo Confalonieri, Andrea Ferrario, Paolo Rocco, and Andrea Bettoni Optimizing Performance-Allocation Trade-Off: The Role of Human-Machine Interface Technology in Empowering Multi-skilled Workers in Industry 4.0 Factories . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 716 Federica Costa, Alireza Ahmadi, and Alberto Portioli-Staudacher Towards Industry 5.0: Empowering SMEs with Blockchain-Based Supplier Collaboration Network. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 730 Prince Waqas Khan, Imene Bareche, and Thorsten Wuest A Stochastic-Based Model to Assess the Variability of Task Completion Times of Differently Aged and Experienced Workers Subject to Fatigue . . . . . 745 Andrea Lucchese, Salvatore Digiesi, and Giovanni Mummolo A Proposal for Production Scheduling Optimization Method with Worker Assignment Considering Operation Time Uncertainty . . . . . . . . . . . . . . . . . . 760 Daiki Nagata, Toshiya Kaihara, Daisuke Kokuryo, Toyohiro Umeda, and Houei Mizuhara The Impact of the Design Decisions of an Order Picking System on Human Factors Aspects of the Order Pickers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 775 Vivek Vijayakumar and Fabio Sgarbossa

Contents – Part I

xvii

SME 5.0: Exploring Pathways to the Next Level of Intelligent, Sustainable, and Human-Centred SMEs From Surviving to Thriving: Industry 5.0 at SMEs Enhancing Production Flexibility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 789 Zuhara Zemke Chavez, Ala Arvidsson, Jannicke Baalsrud Hauge, Monica Bellgran, Seyoum Eshetu Birkie, Patrik Johnson, and Martin Kurdve Challenges in Designing and Implementing Augmented Reality-Based Decision Support Systems for Intralogistics: A Multiple Case Study . . . . . . . . 803 Moritz Quandt, Hendrik Stern, Markus Kreutz, and Michael Freitag Data at the Heart of the Industry of the Future: New Information Issues from an Information and Communication Sciences Perspective . . . . . . . . . . . . 818 Nathalie Pinède and Bruno Vallespir Author Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 831

Lean Management in the Industry 4.0 Era

Enablers Identification to Support the Combined Implementation of Lean and Industry 4.0 Ilse Urquia , Anne Zouggar Amrani(B)

, and Bruno Vallespir

CNRS, IMS, UMR 5218, University Bordeaux, 33405 Talence, France {ilse-denisse.urquia-ortega,anne.zouggar, bruno.vallespir}@ims-bordeaux.fr

Abstract. Companies are in a constant search for continuous improvement of their processes to achieve better performance. At the same time, the adoption of new technologies in conjunction with Lean tools is a topic that has increased in recent years. However, knowing the combinations that are potentially beneficial for performance is not enough for a successful implementation. There are internal and external elements that can affect the conjoint implementation of Lean tools and Industry 4.0 technologies. These aspects are defined as enablers. This article is focused on the study of enablers for the implementation of Lean tools and Industry 4.0 technologies. The results of a survey is provided in order to rank the importance of these enablers according to the experience of industrials on Lean and Industry 4.0 implementation. Keywords: Industry 4.0 technologies · Lean tools · Enablers · Survey

1 Introduction Industry 4.0 (I4.0), also known as the Fourth Industrial Revolution, is a term used to describe the integration of advanced technologies, such as the Industrial Internet of Things (IIoT), robotics, and data analytics (Kolberg and Zühlke 2015; Łupicka and Grzybowska 2018). On the one hand, I4.0 has transformed the way products are designed, produced, and delivered to customers. Industry 4.0 technologies allow manufacturers to optimize their operations, reduce costs, and improve productivity. On the other hand, Lean is a set of principles, tools, and techniques that aim to reduce waste in the process, improving quality, and increasing the efficiency in manufacturing operations. Lean Manufacturing focuses on maximizing value and minimizing waste by constantly improving processes and eliminating non-value-added activities.This can be achieved by applying different tools such as Kanban, Total Productive Maintenance (TPM) and Value Stream Map (VSM). With the advent of Industry 4.0 technologies, the question arises as to whether these approaches would continue to be relevant for process improvement. Today it can be agreed that the two topics work in synergy and help each other (Sanders et al. 2016). © IFIP International Federation for Information Processing 2023 Published by Springer Nature Switzerland AG 2023 E. Alfnes et al. (Eds.): APMS 2023, IFIP AICT 689, pp. 3–14, 2023. https://doi.org/10.1007/978-3-031-43662-8_1

4

I. Urquia et al.

The combination of Industry 4.0 technologies and Lean tools can create a powerful cooperation, allowing manufacturers to achieve even greater efficiency and productivity gains. By integrating real-time data from IIoT sensors and using advanced analytics, manufacturers can identify areas of improvement and optimize their processes for maximum efficiency. This can help to reduce waste, minimize downtime, and improve quality, ultimately leading to increased profitability and customer satisfaction. Both of these concepts are essential to achieving success in today’s competitive business environment. Therefore, the discussion on the combination of these two paradigms has increased and has been discussed in several articles (Satoglu et al. 2018; Valamede and Akkari 2020; Tortorella et al. 2021). Studying the enablers for a conjoint implementation could increase the probability of success and creates a positive contextual environment to succeed. Hence, focusing on what enablers companies need to consider before implementing Lean or Industry 4.0 is becoming relevant. This article aims to propose the elements for a conjoint implementation of Lean and Industry 4.0. In the following section, the concepts and elements found in literature considering enablers for Lean, Industry 4.0 and for both are explained.

2 Literature Background The literature background was conducted searching the critical success factors or key success factors of Lean and Industry 4.0 (meaning enablers in this article). The consulted databases were Scopus and Web of Science. Documents discussing barriers and study cases on the conjoint implementation of Lean and Industry 4.0 were also analyzed, to fully understand the global elements for a successful implementation. The purpose of the literature background is not to go into specific topics according to the Industry, size of company or tool to be implemented. It is rather to identify the global elements to consider for Lean and Industry 4.0 implementation. The articles found were browsed to identify the aspects that helped with the implementation of Industry 4.0 and Lean (separately and in combination).

3 Lean Enablers Lean is an approach that focuses on standardizing processes and continuous improvement. The successful implementation of Lean tools requires the identification and application of enablers. The discussion done is on which enablers have to be considered for the implementation of Lean or Industry 4.0 in different articles. In the case of Lean, Blijleven et al. (2019) conducted interviews with 36 IT outsourcing companies identifying 30 enablers for the implementation of Lean, among them it can be noticed, the management leadership, the existence of a financial compensation and close communication between employees to identify and eliminate waste. However, since the companies interviewed were IT outsourcing, some factors are unique to their context. For this reason, after having analyzed the literature, some enablers of Lean implementation are mentioned below:

Enablers Identification to Support the Combined Implementation

5

Management Commitment and Support: Management must be committed to the implementation of Lean and should support it fully (Puram et al. 2022). Leaders must be actively involved in driving the transformation and provide the necessary resources to make it happen. The managers have to choose the right moment to introduce Lean concept. It can be seen in Mirzaei (2011) how this situation is beneficial to have less resistance to change. To Have a Setup of Clear Roles: Having clear roles ensure that everyone in the company understands its responsibilities and the scope of the work. The assignment of clear roles aids on the identification of bottlenecks and waste in the process to reinforce and assign resources (Hopkins 2021; Sanders et al. 2016; Netland 2016). Start by a pilot project: Implementing Lean tools involves making significant changes in the organization. By starting with a pilot project, this risk can be mitigated. This allows the company to test and validate Lean concepts in a controlled environment.

4 Industry 4.0 Enablers The literature on the enablers of Industry 4.0 often refers to the implementation of technologies as a success factor. Sony and Naik (2019) enlist 10 enablers for Industry 4.0 implementation. Having smart products that collect information about the product and process is considered as an enabler for Industry 4.0. On the strategic side, having a short, medium and long-term plan is listed as the first enabler. Successful implementation of Industry 4.0 needs careful planning, investment, and execution. These are some enablers to consider. Robust IT Infrastructure: Industry 4.0 relies heavily on IT infrastructure. Companies must have a robust and secure IT infrastructure in place to support the new technologies. Part of this infrastructure is also to have a data security protocol. This data often contains sensitive information about processes and customers. Implementing a data security protocol ensures that privacy regulations are met (Alexopoulos et al. 2018). Collaboration and Partnerships: Successful implementation of Industry 4.0 requires collaboration and partnerships. Companies must work with technology providers, suppliers, and other stakeholders to develop and implement new technologies (Devi et al. 2021; Glass et al. 2018). To avoid overloading the reader, here are just a few examples of the enablers found in literature on Lean and Industry 4.0 implementation. The complete synthesis of the enablers on Lean, Industry 4.0 and those found in common can be observed in Table 1. Once the literature was analyzed, the concepts found in the literature were classified according to four dimensions defined by the authors. Technology infrastructure, Strategic, Workforce and Organizational are the dimensions helping to understand the aspect that should be considered in the company. Then, the statements that explain which elements are part of each dimension were listed in the next column. Although the literature focused on Lean or Industry 4.0 enablers separately, elements were common. This is marked in the table by a check mark.

6

I. Urquia et al. Table 1. Enablers for Industry 4.0 and Lean: Author’s compilation. Dimension Technology infrastructure

Strategic

Statement Company should have a standard protocol for communication (e.g., OPC UA) Infrastructure and policy restrictions Data security (protocols, standard) Waste in the business process must be removed before digitalization Have a business focus Continuous evaluation during the process Be clear of the industrial application Professional training development Defined Budget Customer involvement Choose the moment of introduction Start by a pilot project

Workforce

To have the setup of clear roles Appropriation of the Lean approach Take on account the operators Define a leading role Employeesí mindset/Engaged staff

Organizational

External consultancy

Lean

I4.0

Reference (Sanders,2017) (Thun,2021) (Thun,2021) (Sanders,2017) (Flores,2018) (Thun, 2021) (Flores,2018) (Devi,2020) (Devi,2020) (Glass,2018) (Sanders,2016) (Mirzaei,2011) (Scherrer-Rathje et al.,2009) (Anvari et al., 2011) (Maginnis,2019) (Bhasin,2012) (Scherrer-Rathje et al., 2009) (Flores,2018) (Maginnis,2020) (Flores, 2018) (Maginnis,2020) (Sanders et al.,2016) (Hopkins,2021) (Netland,2016)

Networking (helps teams work together more efficiently)

(Devi,2020)

Top management commitment

(Devi,2020) (Antony & Banuelas,2002)

Since Lean has a more organizational focus, it lends itself to the necessary elements of a culture change approach in the company. Contrary to this, the information found in literature on Industry 4.0 enablers tends to be more technical, leaving the organizational part aside (Samaranayake et al. 2017). Although there are case studies talking about the combination of Lean and Industry 4.0, these articles focus on whether there is a relationship or not between the two paradigms, rather than clarifying what elements enable the implementations. In Ciano et al. (2020), an interview was conducted with companies that have worked with Lean tools and Industry 4.0 technologies, focusing on the relationship between these two concepts, leaving aside what facilitated their implementation in these companies. For this reason, grouping which elements are important for a conjoint implementation of Lean and Industry 4.0 allows to see it as a whole and not as two separate paradigms. Identifying the common success factors is a step up to consider to converge towards some enablers positively impacting the combination of Lean and Industry 4.0 implementation.

Enablers Identification to Support the Combined Implementation

7

5 Research Methodology Empirical data was collected through a survey. This was sent to industrials with experience in Industry 4.0 and Lean implementation. The survey is composed of 10 questions to validate the following elements: The combinations between Lean tools and Industry 4.0 technologies, a ranking of the enablers and the operational objectives that are priority for them. These elements were included in the survey in order to better understand the relationship between Lean and Industry 4.0. As this article focuses on the enablers, for a combined implementation between Lean and Industry 4.0, only the question related to this is presented in this article. The survey was sent to 200 industrials and a total of 60 answers were received, this giving a 30% response rate. The survey was anonymous, which means that the company details were not solicited. Once the information found in the literature was classified, a second table was created (Table 2). The 4 dimensions presented in Table 1 were re-categorized into 6 in order to be more specific in their application. Table 2. Enablers for Industry 4.0 and Lean retained for the study.

The reclassification of the enablers allowed to be more specific in order to cover the most recurring aspects of Lean and Industry 4.0. Subsequently, this classification was used to create questions mentioned in the survey to lead the priority analysis of enablers.

8

I. Urquia et al.

6 Enablers for Industry 4.0 and Lean To harness the benefit of both approaches it becomes important to have general enablers for a conjoint implementation (Sanghavi et al. 2019). The dimensions of the enablers (Table 2) were used to ask to the industrials their opinion according to experience. The question from the survey according to the scope of the article is: According to your experience, which factors are most important for the conjoint implementation of Lean and Industry 4.0 technologies? It was requested to rank them according to Likert scale (1 least important to 5 most important). The information about the industry sector and the position of the respondents is presented in Figs. 1 and 2. Knowing this information allowed to segment the respondents’ priorities according to their industry sector and position.

Fig. 1. Industry sectors covered by the interviews.

Of the people surveyed, 20% belong to the aeronautics and aerospace industry, followed by 15% people working in the food sector and 12% working in Automotive. The opinions of users from different industries presented in this article allow to analyze the priorities of each user according to the line of business for each company. Regarding the role of the interviewees in the company, 36% are product or project managers. The other 45% is constituted by consultants, directors, managers and production managers. The remaining 19% is a mix of logistics, quality or automation managers. The positions with the highest density of responses are positions that are invested in shop-floor areas. This information about the profile of the interviewees allows to analyze the ranking given and each of the enablers. However, it also allows to observe the differences or similarities in the answers according to their type of industry and job position. The following section presents the ranking given to the enablers by the respondents and the results of the study.

Enablers Identification to Support the Combined Implementation

Engineer 2%

9

N=60

Fig. 2. Position of the interviewees.

7 Analysis of the Enablers’ Ranking The analysis of the responses obtained in the survey are presented below. Once the ranking is performed, the results are analyzed. The analysis is presented in Fig. 3. It is shown in the first bar (Blue), the number of people who gave a value of 4 or 5 to enablers. The second bar (Orange) shows the number of people giving 3 as a value (neutral) to the enablers. The last bar (Grey) is the number of people giving a value of 1 or 2 to the enablers which is less important.

Fig. 3. Enablers’ Ranking: Interviewees responses.

10

I. Urquia et al.

It can be observed that having a good Financial capacity was considered as one of the main enablers for industrials when implementing Lean and Industry 4.0. The Top management support is mentioned in literature as a primary requirement for both Lean and Industry 4.0 (O. Connor and Cormican 2022; Pozzi et al. 2021; Mustapha et al. 2019). Observing the “Most important” bar in the graph IT infrastructure is in second place along with Top management support. Both are considered by at least half of the interviewees as an enabler for the conjoint implementation. However once observing the average of the responses on Table 3, the conjoint implementation of Lean and Industry 4.0 Top management is placed second. This corroborates what has been found in the literature that top management is important for disruptive changes such as Lean or Industry 4.0 (Ghobakhloo and Fathi 2019). There seems to be a discrepancy among the interviewees. In Fig. 3, half of the respondents (30 interviewees) considered Top management support as an important aspect. At the same time when looking at the grey bar, it can be seen that 24 respondents gave it a score between 1 and 2 meaning is the second less important. In order to have an overall and fair view of the ranking given to each enabler, the average of the responses was used, then ordered from the highest to the lowest to perceive the difference better. The means of every enabler is shown in Table 3. Here it can be seen that with an average of 3.34 the “Financial capacity” was considered important since a total of 31 people selected this aspect as very important (Likert scale 5 and 4) in their ranking. The sectors where these people work are mostly food and aeronautics. In second place there is the “Top management support” with a mean of 3, 28, 30 people gave it a raking of 5. Financial capacity is an aspect that was not highlighted in the first literature background. However, in the enablers that were retained for the study, once the case studies were consulted, it was indicated that financial capacity is a fundamental aspect and this is explained in more detail with the concepts that encompass it. Table 3. Enablers ranking according to the response average Enablers code

E01

E02

E03

E04

E05

E06

Enablers name

Financial capacity

Top Management support

IT Infrastructure

External support

Internal expertise

Clearance in the project scope

Mean

3,34

3,28

3,27

3,09

3,02

2,91

S. D

1,54

1,71

1,52

1,55

1,40

1,59

“IT infrastructure” is in third place with an average of 3.27. As can be seen, the averages of these first three enablers are above 3 and are very close to each other. What can be summarized is that these three concepts in general were considered the most important by the interviewers.

Enablers Identification to Support the Combined Implementation

11

The enablers below an average of 3 can be considered as not so important for the conjoint implementation of Lean and Industry 4.0. Based on the opinions of the interviewees, the only enabler that falls into this category is the “Clarity in the scope of the project.” This aspect was not considered important by most of the interviewees. It can be interpreted that the surveyed companies had well defined the scope of their implementation as a whole or that this aspect is simply not enabling the implementation to be successful or not. Due to the existent discrepancies about the classification of some concepts, the average of the answers given was calculated to find out whether the proposed enabler’s vision was positive or negative. On the other side, it can be observed that the highest standard deviation (1,71) belongs to “Top management support,” this represents a high variance between respondents’ answers. Perhaps by increasing the panel of respondents, this discrepancy will decrease consolidating Top management as an indispensable enabler for Lean and Industry 4.0 implementation. In the case of Internal expertise which has a standard deviation (1,40) indicates a higher agreement between the respondents. Once the averages were calculated, only the average of less than 3 (Neutral) is not considered key for the implementation of Lean and Industry 4.0 implementation. This, according to the opinion of the interviewees.

8 Conclusion The conjoint implementation of Lean and Industry 4.0 has increased in attractiveness. In the literature, there is still a gap in information on how to approach the combined implementation of Lean and Industry 4.0. This article clarifies the literature found on the enablers of these two approaches. It classifies them into different axes to cover the necessary elements for a conjoint implementation to eventually rank them according to the industrial’s feedback. To know the necessary elements to implement Lean and Industry 4.0 helps to better introduce this concept to the company and prepare the companies before the implementation. The results of the ranking allow to identify the top three enablers for the respondents. Financial capacity, Top management support and IT infrastructure are ranked with an average above 3. The analysis of the enablers of both Lean and Industry 4.0 also identify that Clearance in the project scope is not considered as an important enabler, its average was below 3. However, successful implementation requires careful planning paying attention to various factors. This allows not only to have an overview of conjoint implementations but also to validate it with industry experience. Companies that are able to integrate these factors into their implementation by performing them in an effective way would be those who achieve competitive positioning. Validating the importance of enablers with the industry is an advantage to provide interesting managerial insights in order to be careful during the common implementation.

12

I. Urquia et al.

References Ahmad, M.O., Markkula, J., Oivo, M.: Insights into the perceived benefits of Kanban in software companies: practitioners’ views. Presented at the Agile Processes, in Software Engineering, and Extreme Programming: 17th International Conference, XP 2016, Edinburgh, UK, 24–27 May 2016, Proceedings 17 (2016) Alexopoulos, K., Sipsas, K., Xanthakis, E., Makris, S., Mourtzis, D.: An industrial Internet of things-based platform for context-aware information services in manufacturing. Int. J. Comput. Integr. Manuf. 31, 1111–1123 (2018). https://doi.org/10.1080/0951192X.2018.1500716 Alhuraish, I., Robledo, C., Kobi, A.: Key success factors of implementing lean manufacturing and six sigma. Presented at the Liverpool (2014): 17th Toulon-Verona Conference” Excellence in Services (2014) Antony, J., Gupta, S.: Top ten reasons for process improvement project failures. Int. J. Lean Six Sigma. 10, 367–374 (2019) Anvari, A., Zulkifli, N., Yusuff, R.M., Hojjati, S.M.H., Ismail, Y.: A proposed dynamic model for a lean roadmap. Afr. J. Bus. Manage. 5, 6727 (2011) Basl, J.: Analysis of industry 4.0 readiness indexes and maturity models and proposal of the dimension for enterprise information systems. Presented at the Research and Practical Issues of Enterprise Information Systems: 12th IFIP WG 8.9 Working Conference, CONFENIS 2018, Held at the 24th IFIP World Computer Congress, WCC 2018, Poznan, Poland, September 18–19, 2018, Proceedings 12 (2018) Bhasin, S.: Prominent obstacles to lean. Int. J. Product. Perform. Manag. 61, 403–425 (2012) Bhattacharya, I., Ramachandran, A.: Lean manufacturing techniques–implementation in Indian MSMEs and benefits realized thereof (2021) Blijleven, V., Gong, Y., Mehrsai, A., Koelemeijer, K.: Critical success factors for Lean implementation in IT outsourcing relationships. Inf. Technol. People 32, 715–730 (2019). https://doi. org/10.1108/ITP-01-2016-0002 Ciano, M., Dallasega, P., Orzes, G., Rossi, T.: One-to-one relationships between Industry 4.0 technologies and Lean Production techniques: a multiple case study. Int. J. Product. Res. 1, 25 (2020). https://doi.org/10.1080/00207543.2020.1821119 Devi, K.S., Paranitharan, K.P., Agniveesh, A.I.: Interpretive framework by analyzing the enablers for implementation of Industry 4.0: an ISM approach. null. 32, 1494–1514 (2021). https://doi. org/10.1080/14783363.2020.1735933 Flores, M., Maklin, D., Golob, M., Al-Ashaab, A., Tucci, C.: Awareness towards Industry 4.0: key enablers and applications for internet of things and big data. In: Camarinha-Matos, L.M., Afsarmanesh, H., Rezgui, Y. (eds.) PRO-VE 2018. IAICT, vol. 534, pp. 377–386. Springer, Cham (2018). https://doi.org/10.1007/978-3-319-99127-6_32 Ghobakhloo, M., Fathi, M.: Corporate survival in Industry 4.0 era: the enabling role of leandigitized manufacturing. J. Manufac. Technol. Manage. (2019) Glass, R., Meissner, A., Gebauer, C., Stürmer, S., Metternich, J.: Identifying the barriers to Industrie 4.0. Procedia Cirp. 72, 985–988 (2018) Hopkins, J.L.: An investigation into emerging industry 4.0 technologies as drivers of supply chain innovation in Australia. Comput. Indust. 125, 103323 (2021) Jedynak, P.: Lean management implementation: determinant factors and experience. Jagiellonian J. Manage. 1, 51–64 (2015) Kolberg, D., Zühlke, D.: Lean automation enabled by industry 4.0 technologies. IFAC- PapersOnLine. 48, 1870–1875 (2015) Łupicka, A., Grzybowska, K.: Key managerial competencies for industry 4.0-practitioners’, researchers’ and students’ opinions. Logistics Transport 39 (2018)

Enablers Identification to Support the Combined Implementation

13

Maginnis, M.A., Hapuwatte, B.M., Keown, D.: The integration of true lean and Industry 4.0 to sustain a culture of continuous improvement. Presented at the Product Lifecycle Management in the Digital Twin Era: 16th IFIP WG 5.1 International Conference, PLM 2019, Moscow, Russia, July 8–12, 2019, Revised Selected Papers 16 (2019) Mirzaei, P.: Lean production: introduction and implementation barriers with SMEs in Sweden (2011) Mustapha, M.R., Abu Hasan, F., Muda, M.S.: Lean Six Sigma implementation: multiple case studies in a developing country. Int. J. Lean Six Sigma 10, 523–539 (2019) Netland, T.H.: Critical success factors for implementing lean production: the effect of contingencies. Int. J. Prod. Res. 54, 2433–2448 (2016) O. Connor, D., Cormican, K.: Leading from the middle: How team leaders implement lean success factors. Int. J. Lean Six Sigma 13, 253–275 (2022) Pozzi, R., Rossi, T., Secchi, R.: Industry 4.0 technologies: Critical success factors for implementation and improvements in manufacturing companies. Product. Plan. Control 1–21 (2021) Punnakitikashem, P., Buavaraporn, N., Chen, L.: An Investigation of Factors Affecting Lean Implementation Success of Thai Logistics Companies, vol. 10 (2010) Puram, P., Sony, M., Antony, J., Gurumurthy, A.: A conceptual framework for a systemic understanding of barriers during lean implementation. TQM J. 34, 1469–1484 (2022) Samaranayake, P., Ramanathan, K., Laosirihongthong, T.: Implementing industry 4.0—a technological readiness perspective. Presented at the 2017 IEEE International Conference on Industrial Engineering and Engineering Management (IEEM) (2017) Sanders, A., Elangeswaran, C., Wulfsberg, J.: Industry 4.0 implies lean manufacturing: research activities in industry 4.0 function as enablers for lean manufacturing. J. Indust. Eng. Manage. 9, 811–833 (2016). https://doi.org/10.3926/jiem.1940 Sanders, A.K., Subramanian, K.R., Redlich, T., Wulfsberg, J.P.: Industry 4.0 and lean Management–synergy or contradiction? A Systematic interaction approach to determine the compatibility of Industry 4.0 and lean Management in manufacturing environment. In: Advances in Production Management Systems. The Path to Intelligent, Collaborative and Sustainable Manufacturing: IFIP WG 5.7 International Conference, APMS 2017, Hamburg, Germany, September 3–7, 2017, Proceedings, Part II. pp. 341–349. Springer (2017) Sanghavi, D., Parikh, S., Raj, S.A.: Industry 4.0: tools and implementation. Manage. Prod. Eng. Rev. (2019) Satoglu, S., Ustundag, A., Cevikcan, E., Durmusoglu, M.B.: Lean production systems for Industry 4.0. In: Ustundag, A., Cevikcan, E. (eds.) Industry 4.0: Managing The Digital Transformation, pp. 43–59. Springer International Publishing, Cham (2018). https://doi.org/10.1007/978-3319-57870-5_3 Scherrer-Rathje, M., Boyle, T.A., Deflorin, P.: Lean, take two! Reflections from the second attempt at lean implementation. Bus. Horiz. 52, 79–88 (2009) Shah, M.N., Dixit, S., Kumar, R., Jain, R., Anand, K.: Causes of delays in slum reconstruction projects in India. Int. J. Constr. Manag. 21, 452–467 (2021) Sony, M., Naik, S.: Key ingredients for evaluating Industry 4.0 readiness for organizations: a literature review. Benchmark. Int. J. 27, 2213–2232 (2019) Thorat, R., Mahesha, G.: Improvement in productivity through TPM implementation. Mater. Today Proc. 24, 1508–1517 (2020) Thun, S., Bakås, O., Storholmen, T.C.B.: Development and implementation processes of digitalization in engineer-to-order manufacturing: enablers and barriers. AI Soc. 37, 725–743 (2022)

14

I. Urquia et al.

Tortorella, G.L., Fogliatto, F.S., Cauchick-Miguel, P.A., Kurnia, S., Jurburg, D.: Integration of Industry 4.0 Technologies into Total Productive Maintenance Practices (2021). https://doi.org/ 10.1016/j.ijpe.2021.108224 Valamede, L.S., Akkari, A.C.S.: Lean 4.0: A new holistic approach for the integration of lean manufacturing tools and digital technologies. International J. Math. Eng. Manage. Sci. 5, 851 (2020)

Lean and Digitalization Status in Manufacturing Companies Located in Norway Natalia Iakymenko1(B) , Daryl Powell1,2 , Eivind Reke1 , Marte Daae-Qvale Holmemo2 , Eirik Bådsvik Hamre Korsen2 , Signe Sagli2 , Sigrid Eliassen Sand2 , and Sunniva Økland2 1 SINTEF Manufacturing, S.P. Andersens vei 3, 7031 Trondheim, Norway

[email protected] 2 Department of Industrial Economics and Technology Management, Norwegian University of

Science and Technology, Alfred Getz vei 3, 7034 Trondheim, Norway

Abstract. Manufacturing companies are always on the lookout for methods to boost productivity, improve quality, and enhance their service offerings to remain competitive. Two methods that are extensively debated in both industry and academia are digitalization and lean. The main objective of this study discussed is to identify the current status of lean and digitalization use in manufacturing companies located in Norway and to uncover their synergies and potential influence on each other. Specifically, the study aims to explore the history of lean and digitalization implementation in these companies, the lean practices and digital solutions currently in use, the influence of lean on strategic and operational results and people, the difficulties faced in lean implementation, and the potential synergies between lean and digitalization in improving operational excellence in manufacturing companies. Keywords: Lean · digitalization · manufacturing

1 Introduction In order to be competitive manufacturing companies are constantly looking for ways to increase productivity, quality, and the level of services. Digitalization and lean are among the ways discussed widely in industry and academia. The lean concept has been verified empirically and there is available extensive literature on empirical studies proving a positive impact on operational excellence in manufacturing companies [1–3]. In recent years, the concept of digitalization has gained significant attention in the manufacturing industry, and it has become increasingly clear that it has the potential to revolutionize the way manufacturers operate [4, 5]. Digitalization refers to the use of digital technologies and systems to improve operational efficiency, quality, and overall performance. However, despite the potential benefits of digitalization, it is still a relatively new concept with limited industrial applications and measurable positive results. Combining Lean and digital solutions seems to be a necessary evolutionary step to further raise levels of operational excellence. Lean management remains the fundamental © IFIP International Federation for Information Processing 2023 Published by Springer Nature Switzerland AG 2023 E. Alfnes et al. (Eds.): APMS 2023, IFIP AICT 689, pp. 15–26, 2023. https://doi.org/10.1007/978-3-031-43662-8_2

16

N. Iakymenko et al.

approach to operational excellence within Norwegian industry. Digitalization should not replace lean management. Rather industrial companies are seeking to understand how the two approaches can be utilized synergistically. This study attempted to identify the current status of lean and digitalization use in manufacturing companies located in Norway, as well as to uncover their synergies and potential influence the two have on each other.

2 Theoretical Background 2.1 Lean Up until recently lean were seen as mechanistic identification and elimination of waste in production processes by replicating lean tools and practices such as value stream mapping, 5S, kanban, poka-yoke, A3, etc. [1]. However, now researchers see that this approach is faulty. They argue that the scope of lean goes much further. According to Powell (from [1]), lean is a meta-theory – a system to develop better products, time after time. The core of Lean thinking is continuous improvement through shared understanding of problems, cross-functional communication, and employee participation. Lean is an education system that emphasizes deliberate learning through the scientific method of Plan-Do-Study-Act and structured experiments. When viewed as a learning paradigm, Lean tools and techniques can be regarded as enhancers of learning, instead of just a means of achieving operational excellence. They help identify areas for improvement and the actual improvement is carried out by the people involved in the system, rather than solely by the tools. However, adopting Lean best practices for static optimization without understanding its theoretical and enterprise-wide underpinnings may hinder extraordinary business results [1–3]. 2.2 Digitalization For this study we employ description of digitalization in manufacturing as provided by [6]. They break it down into shop floor digitalization, technologies for vertical and horizontal integration and organizational IT competence. Shop floor digitalization has cyber-physical production in the center to collect and control real-time production data. Vertical integration involves integrating IT systems at different hierarchical levels within a factory, such as production sensors, enterprise systems, and product development, and is a key aspect of Industry 4.0. Horizontal integration, on the other hand, refers to integrating IT systems across different stages of manufacturing and business planning processes. Organizational IT competence is an organization’s understanding and effective utilization of IT. According to the literature review by [7], production scheduling and control is the process that is most investigated when it comes to Industry 4.0 implementation. They found that not all technologies are equally discussed in the literature with IoT, Big Data Analytics and Cloud at the forefront of the research. Furthermore, even though the topic of Industry 4.0 and digitalization is much debated in the literature, it remains at a conceptual level, without examples of real-life applications. The issue of digitalization

Lean and Digitalization Status in Manufacturing Companies Located in Norway

17

and Industry 4.0 often remains at an abstract level, making it challenging for practitioners to understand where and how to effectively utilize the new technologies, and what are the benefits of technology implementation [6, 7]. 2.3 Interplay Between Lean and Digitalization The emergence of digitalization and Industry 4.0 technologies has sparked a renewed discussion on the introduction of new technologies into established managerial systems that are centered around human-focused philosophies like lean management [9]. As pointed out by [10] in their recent systematic literature review, the papers on lean and digitalization interplay can be divided into three main categories: lean supporting/influencing digitalization, mutual support of lean and digitalization, and digitalization supporting/influencing lean. In their paper they have also found that lean paves the way for the adoption of digitalization at a strategic level, while digitalization technologies enhance lean practices at the operational level (the authors are talking about lean supply chain management in this paper). Research on the strategic level examines the interplay between lean management and digitalization from a system perspective, focusing on long-term implementation paths. In contrast, research on the operative level analyzes the interplay between the two paradigms from a single implementation point of view, with a short-term perspective linked to specific practices or technology implementation (such as big data, augmented reality, blockchain, etc.) in a particular context. [10] take a perspective of using lean for solving digitalization problems. They identified and listed possible types of waste from the digital industry and argue that lean can helps in identifying and eliminating wastes. In their paper authors take an “old school” perspective of lean being a set of waste reduction tools. [11] studied impacts of lean on Industry 4.0. Their conclusion is that lean thinking facilitates the implementation of Industry 4.0 by simplifying processes and eliminating waste in a way that prevent waste from reoccurring, reduce the risk of depleting scarce resources, and enhance transparency in work processes and organizational practices. Their argument is in line with statement “lean first, then digitalize”. Study conducted by [12] studies impact of digitalization on lean. Through the survey of academic community, they found that digitalization impacts lean operations practices. Specifically, just-in-time, visual management, total production maintenance, continuous improvement, and poka-yoke. These authors take perspective of lean being a set of practices as opposed to the view suggested in the recent research of Lean being an educational system to develop better products. Similar to [12, 13] concluded that digitalization and Industry 4.0 technologies impact lean by giving lean tools a more dynamic way of working, accelerating information sharing processes and improving production manager’s and operator’s decision making. They argue that there is little practical and theoretical contribution of lean to Industry 4.0 and thus the actual contribution is still blurred.

18

N. Iakymenko et al.

3 Method Qualitative research method was chosen to study the current status of lean and digitalization in Norway. Qualitative research methods are valuable in providing rich descriptions of complex phenomena. It can be used to gather in-depth insights into a problem. Semistructured interviews were chosen for data collection. Semi-structured interviews are open-ended, allowing for flexibility, but follow a predetermined thematic framework, giving a sense of order. Semi-structured interview guide was created to cover four main topics: • Background information about the company/interview participant (background information was collected to group and analyze companies, understand the context information and control variables) • Lean management in the company (why, when and how did the lean implementation started; implementation process; outcomes at strategic and operational levels) • Digital manufacturing technologies in the company (why, when and how did the work with digitalization started; implementation process; outcomes at strategic and operational levels) • Interplay between Lean management and Digital manufacturing technologies (how specifically are the two used together; influence on each other – advantages and disadvantages; synergetic influence at strategic and operational levels) 3.1 Data Collection The interview guide was designed to help in conducting interviews. In order to find interview participants, researchers and research assistants involved in this study created an invitation that was sent to manufacturing companies located in Norway. In total, 19 interviews were conducted. In some of the interviews, more than one participant were involved (see Table 1). The participants received interview guides in advance to be able to prepare for the interview. Each interview lasted for 1–2 h. The participants were interviewed by the research assistants. 1–3 research assistants were present during each interview. Interviews are usually conducted by a single investigator, but as [14] points out, the use of multiple investigators can have advantages. They can enhance the creative potential of the teams and convergence of observations increases confidence in the findings. The participants were asked, to the extent possible, the same questions to increase the reliability of the collected interview data. The interviews were audiorecorded with the permission of the participants. Interview transcripts were written and sent to the participants for verification. 3.2 Data Analysis Recommendations of [14] were followed for data analysis. The interview transcripts were coded – descriptive codes were assigned both deductively and inductively to data chunks. Several categories of codes were assigned: 1. Codes describing company’s characteristics: size, products, degree of customization, degree of repetitiveness of production processes,

Lean and Digitalization Status in Manufacturing Companies Located in Norway

19

2. Codes describing interview participant’s experience and competences: position in the company, experience working with lean and/or digitalization, years of experience, 3. Codes describing company’s lean practices: the history of lean implementation in the company, lean practices used now, influence of lean on strategic and operational results, lean implementation difficulties, 4. Codes describing company’s digitalization practices: the history of digital solutions implementation in the company, digital solutions used now, influence of digitalization on strategic and operational results, digitalization implementation difficulties. Based on the coding, a cross-case analysis was performed. Using the methods suggested by [14], the authors looked for the presence of same factors across multiple cases and examined whether familiar themes emerged in multiple settings. To aid the analysis at this stage, all cases were combined in a meta-matrix created by assembling each case in a common format and displaying them together in one large table. Reformatting and resorting the cells and rows in the table helped the authors to identify patterns in the cases and determine whether new observations can be constructed.

4 Results This chapter presents the results of the study without interpreting them. 4.1 Background Information About the Company and Interview Participants In total, representatives from 19 companies were interviewed. In some instances, several participants from one company were interviewed (see information about interview participants in Table 1). In the event where several representatives were present, interviews were carried out simultaneously. Interview participants had different degrees of experience working with lean and digitalization – from several month to more than 20 years. The participating companies vary depending on their size, the type of produced products and types of production processes. Companies’ size varies from small to large: between 50 and 3500 employees and turnover from 40 mill. NOK to 4.5 mrd. NOK. Both Norwegian companies and companies located in Norway, but belonging to international conglomerates participated in the study. Table 1 shows products and production processes of the studied companies. 4.2 Lean Status The History of Lean Implementation in the Companies and Lean Practices Used Now. The studied companies all have been working with lean management during the last 2– 19 years. One of the companies stated that they have been working with lean since 1972. Only one company explicitly emphasized that they see lean as a philosophy rather than a set of tools. The rest of the companies listed tools they are working with as a part of lean management. The most mentioned tools are: 5S, work with whiteboards, improvement work meetings and work on improving the flow. Other tools mentioned

20

N. Iakymenko et al. Table 1. Information about interview participants and manufacturing companies

Comp

Interview participants

Product or service

Product’s degree of Degree of customization repetitiveness in production processes

1

Lean coordinator

Electronics

Products are customized based on customers’ specification

Mix of mass- and project-based production

2

Production mgr

Roofs

Products are customized based on customers’ specification

Processes are similar, but vary depending on the product

3

Operations developer

Milk and milk products

Products are standard

Mass production, repetitive

4

Lean department mgr

Services within insulation, scaffolding and surface treatment

Products are customized

Processes are similar, but vary depending on the product

5

Chairman of the board

Doors

Make-to-order, high customization degree

Processes are repetitive, but not mass production. Customization starts at surface treatment

6

Staff and support mgr

Construction

Products are customized

Project-based production

7

Operations developer

Cheese and yoghurt

Products are standard

Mass production, repetitive

8

IT/digitalization department leader

Industrial lighting

Products are customized

Project-based production

9

Coordinator for cont. Improvement and business devt., Chief information officer

Service of procurement, management and disposal of material for the military

-

Project-based procurement

10

Quality and HMS leader

Steel component and Products are structures customized

Project-based production

11

Lean coordinator

Windows and doors

N/A

Products are customized

(continued)

Lean and Digitalization Status in Manufacturing Companies Located in Norway

21

Table 1. (continued) Comp

Interview participants

Product or service

Product’s degree of Degree of customization repetitiveness in production processes

12

Corporate mgmt., responsible for digitization and operational mgmt.

Furniture

Make-to-order (millions of combinations of furniture)

Processes are repetitive, variation in the end of production

13

Production Fastening materials coordinator and bolts Organizational developer and improvement leader

Products are customized depending on dimensions and materials

Processes are repetitive

14

Technology mgr Research and sustainability mgr

Composite gas containers, propane-butane mix

Products are standard, the only customization possible is product appearance

Processes are repetitive

15

General mgr., technical mgr., IT mgr., Lean mgr

Doors and windows

Products are customized

N/A

16

Head of commercial Radioactive drugs production

Products are standard

Processes are repetitive

17

Lean and quality mgr

Construction

Products are customized

Processes are non-repetitive

18

Production responsible

Ammunition products and rocket engines

Many product variants

Processes are non-repetitive

19

Business systems developer

Thruster, dynamic positioning equipment

Products are customized

Processes are similar, vary depending on the product size

include teamwork, reading circles, takt, poka-yoke, standardisation, waste reduction, Hoshin Kanri, problem solving, zero defect, VSM, PDCA, Lean Six Sigma (please note that we report the tools the way they were named by the participants, without interpreting whether it can be interpreted as a lean tool or not). 4 companies mentioned that either lean course or lean programme was purchased from external consultants, 2 companies used lean programmes developed in the company’s headquarters abroad. The rest of the companies attempted lean implementation on their own or it was not clear (or not known to participants) how the lean implementation started in the company.

22

N. Iakymenko et al.

Influence of Lean on Strategic and Operational Results and People. 9 companies answered that it was difficult to assess lean influence on operational results, 10 companies were able to give examples of the lean influence on operational results, and 4 companies were able to give quantified assessment of operational results. Following impacts on operations were mentioned: • • • • • • • • • •

less material waste (6 companies), cleaner workstations (6 companies). reduced production time (4 companies), increased uptime (4 companies), better quality (4 companies), increased profit (2 companies), increased lead time (2 companies), increased level of service (2 companies), reduced inventories (1 company), increased production volume (1 company), Quantified improvements mentioned by the companies are:

• • • • •

Time used for one of the production operations went down from 4 days to 4 h, 57 million NOK more profit from lean implementation (across the conglomerate), Reduced throughput time from several days to several hours, 80% reduction in internal fails, Increased production volume from 55 to 130 items.

Only one company stated that lean had a clear influence on company’s strategy and that projects’ effectivization was included in their strategy. 13 companies stated that lean had positive effect on employees. Following positive effects were mentioned: • • • • • • • • • •

Calmer, more predictable work (7 companies), Better work culture (7 companies), Increased interest and pride in work, better ownership of own work (6 companies), Increase engagement in work (6 companies), Generation of new ideas and improvement suggestions (6 companies), Tidy workplace (4 companies), Better collaboration between the teams (4 companies), Better HSE (2 companies), Better understanding of business tasks and processes (2 companies), Reduced manual work (2 companies).

Lean Implementation Difficulties. 16 companies mentioned different lean implementation difficulties: • • • • •

Lack of time (8 companies), Disruption in lean implementation due to COVID pandemic (8 companies), Lack of commitment (8 companies), Inability to ensure continuous focus and commitment (8 companies), Lean work is dependent on individual managers (6 companies),

Lean and Digitalization Status in Manufacturing Companies Located in Norway

23

• • • • •

Lack of managerial support (5 companies), Resistance from employees (4 companies), Little competence and understanding of lean (3 companies), Difficult to get out of comfort zones and habitual patterns (3 companies). Difficult to practice lean in external projects (geographically distant projects) (1 company), • Some resources are difficult to measure (in case of Lean Six Sigma) (1 company). 4.3 Digitalization Status The History of Digitalization Implementation in the Companies and Digitalization Now. The studied companies have been working with digitalization during the last 2–30 years. The companies mentioned both hardware and software as part of their digitalization practices. The most mentioned ones were: • Sensors to measure machine performance (5 companies), • Screens/boards with visualization of the status of workstations/production/the entire company (5 companies), • Power BI (6 companies), • ERP systems (12 companies), • Robots (2 companies). Other digital solutions mentioned: SharePoint, Office Dynamics, automated registration of deviations, automatic reporting system, digital training system, digital HR system, digital checklists, Salesforce, Optimizely, product development software, production planning systems, FO365, CNC machines, KPI-dashboards, Hololens, etc. Influence of Digitalization on Strategic and Operational Results and People. Eleven companies stated that it is difficult to answer whether digitalization influence operational results. None of the companies were able to quantify the operational results. 8 companies provided following examples of positive influence of digitalization on operational results: • • • • • • • • •

Improved machine uptime (4 companies), Reduced waste (4 companies), Reduced production deviations (3 companies), Resource savings (2 companies), Better quality (2 companies), Improved in the database (1 company), Better information flow (1 company), Better process control (1 company), Better service (1 company).

All of the participants found it difficult to answer whether digitalization influence company’s strategy. Digitalization Implementation Difficulties. 15 companies gave following examples of digitalization implementation difficulties: • There are no ready-to-use digital tools, everything must be adapted and customized to fit the company’s needs (8 companies),

24

N. Iakymenko et al.

• Resistance from employees (fear of changes or fear that digital tools are there to evaluate employees’ work) (7 companies), • Economy (7 companies), • People may see digitization as unnecessary work (e.g., data must be recorded, which requires additional effort) (6 companies), • Lack of competence (5 companies), • Lack of time (5 companies), • Lack of structure in the development and implementation of digital solutions (4 companies), • Lack of support from management (4 companies), • Difficult to get different IT systems to “talk to each other” (4 companies), • Low level of competence in the company (2 companies), • Reluctance to change or lack of capacity to absorb changes (2 companies), • Cyber-security (1 company). 4.4 Lean and Digitalization Synergies The last part of the interview was dedicated to studying how lean influenced digitalization (and vice versa). This part of the interview had the least answers. When asked if Lean influenced digitalization, most of the interview participants answered “yes”, but struggled to give a clear answer how. 5 companies, however, explained that lean helps to find out which data is required to work further with lean, based on that, the companies can decide which technologies are needed to collect and process the needed data. 3 companies explained that problems identified through lean work are sometimes solved with the help of digital solutions. When asked if and how digitalization influences lean, thirteen companies explained that digitalization (specifically, data collection and analysis) helps to find problems and bottlenecks for further lean work. When it comes to negative influences, some companies mentioned that digitalization influences lean somewhat negatively by: • Digitization can cause some lean principles to not fit in with the new tools -> digitization makes simple tasks that everyone can do into something difficult. Here, no specific example was given by the interview participants, except maybe an example given in the next bullet point below. • Easier to use whiteboards than digital boards. The interview participants think that digitalization of work with boards creates unnecessary barrier for lean work, • Digitalization can be a distraction from improvement work, • Difficult to choose digital platforms that support lean work, • Digitalization efforts take away time and resources from lean implementation work.

5 Discussion and Conclusions There are several important observations that can be made based on the available results: 1. It is clear that manufacturing companies are lagging behind in current understanding of lean. Most of the companies participating in our study see lean as a set of predefined tools. It is understandable since even now many researchers see lean merely as a set

Lean and Digitalization Status in Manufacturing Companies Located in Norway

2.

3.

4.

5.

25

of tools that should be replicated by companies to improve their performance. It is important to spread new knowledge, educate and update companies on new research. It should be stated that the new research does not cancel the usefulness of existing lean tools, but rather encourages people across the organization to develop solutions suitable for specific problems, rather than impose existing solutions. The interview scope is quite limited, and it was not possible to see specifics of lean implementation. The idea for further research can be in-depth case studies with 5-10 companies that have been working with lean. Very few companies are able to see and measure the real influence of lean and digitization on operational performance, corporate strategy, and people. The results of digitization are especially difficult to see. This can be explained by the fact that digitalization is a relatively new initiative in manufacturing companies (except from old technologies, such as CNC-machines) and it needs to mature to see real results. The impacts of lean are somewhat clearer, but still quantitatively limited, even though lean initiatives started quite a long time ago in most companies. This can be explained by the fact that it is difficult to isolate the impact of lean if several initiatives are being implemented in the company. However, the companies should try to estimate real impacts of both lean and digitalization for assessment and learning. This is also a task for future research. It is not clear how manufacturing companies define digitalization. This can be a limitation of the interview guide since we did not give a definition of what we mean under digitalization, but it was done intentionally to see how the companies would answer. Most of the digitalization examples were given are related to information flow digitalization, much less is about digitalization of material flows. This can be because of how companies define digitalization or because information flow digitalization is more advanced. This too requires further research. No clear answers on how lean influences digitalization were given, even though people intuitively say that the influence is positive. Most of the participants were able to clearly say that digitalization supports lean work by helping to find problems and bottlenecks for further lean work. This is clearly a research and practice area that should be worked on. The quality and depth of answers depend on who was interviewed. Many interview participants were working with lean, not digitalization. Hence, their knowledge on digitalization can be somewhat limited. This is a limitation of our research and should be addressed in further studies. Additionally, it is difficult to find experts that apply both lean and digitalization in their firms because of the novelty of the paradigm combination.

Study Limitations This study is subjected to limitations. First, it does not constitute either survey or in-depth case study. Since only interviews were conducted with company representatives, it does not allow for data triangulation and results depend on who was interviewed as we stated above. This considerably weakens the reliability of the research and the results should be treated with care. It is important to note that this are preliminary explorative results and case studies with several of the companies were conducted after these interviews and will be published later.

26

N. Iakymenko et al.

References 1. Pär, Å., et al.: Is lean a theory? Viewpoints and outlook. Int. J. Oper. Product. Manage. (2021) 2. Daryl, P., Reke, E.: No lean without learning: rethinking lean production as a learning system. Advances in Production Management Systems. Production Management for the Factory of the Future: IFIP WG 5.7 International Conference, APMS 2019, Austin, TX, USA, September 1–5, 2019, Proceedings, Part I. Springer (2019) 3. Michael, R., et al.: A self-driving car architecture in ROS2. 2020 International SAUPEC/RobMech/PRASA Conference. IEEE (2020) 4. Matt, D.T., Pedrini, G., Bonfanti, A., Orzes, G.: Industrial digitalization. A systematic literature review and research agenda. Euro. Manage. J. 41(1), (2023), 47–78 (2023) 5. Zhong, R.Y., Xu, X., Klotz, E., Newman, S.T.: Intelligent manufacturing in the context of industry 4.0: a review. Engineering 3(5), 616–630 (2017) 6. Jo Wessel, S., et al.: Digitalized manufacturing logistics in engineer-to-order operations. Advances in Production Management Systems. Production Management for the Factory of the Future: IFIP WG 5.7 International Conference, APMS 2019, Austin, TX, USA, September 1–5, 2019, Proceedings, Part I. Springer (2019) 7. Zheng, T., et al.: The applications of Industry 4.0 technologies in manufacturing context: a systematic literature review. Int. J. Product. Res. 59.6, 1922–1954 (2021) 8. Andrea, Z., et al.: Moving towards digitalization: a multiple case study in manufacturing. Product. Plan. Control 31.2–3, 143–157 (2020) 9. Abhijeet, G., et al.: The impact of Industry 4.0 implementation on supply chains. J. Manufac. Technol. Manage. 31.4, 669–686 (2020) 10. Rossini, M., Powell, D.J., Kundu, K.: Lean supply chain management and Industry 4.0: a systematic literature review. Int. J. Lean Six Sigma 14(2), 253–276 (2023). https://doi.org/ 10.1108/IJLSS-05-2021-0092 11. Victor, B., et al.: Contributions of lean thinking principles to foster Industry 4.0 and sustainable development goals. Lean Eng. Global Dev. 129–159 (2019) 12. Haddud, A., Khare, A.: Digitalizing supply chains potential benefits and impact on lean operations. Int. J. Lean Six Sigma 11(4), 731–765 (2020) 13. Salvadorinho, J., Teixeira, L.: Stories told by publications about the relationship between Industry 4.0 and Lean: dystematic literature review and future research agenda. Publications 9(3), 29 (2021). https://doi.org/10.3390/publications9030029 14. Miles, M.B., Huberman, A.M., Johnny, S.: Qualitative data analysis. A methods courcebook (2019)

Effects of Lean and Industry 4.0 Technologies on Job Satisfaction: A Case-Based Analysis Matteo Zanchi1(B)

, Andrea Lorenzi1 , Matteo Prezioso1 , Daryl Powell2,3 and Paolo Gaiardelli1

,

1 University of Bergamo, Bergamo, Italy {matteo.zanchi,paolo.gaiardelli}@unibg.it, {a.lorenzi5, m.prezioso}@studenti.unibg.it 2 SINTEF Manufacturing, Raufoss, Norway [email protected] 3 University of South-Eastern Norway, Kongsberg, Norway

Abstract. The scarce presence of studies on the role of Digital Lean on the social dimension of sustainability calls for the need to conduct further studies and research to better understand how the integration of the aspects belonging to the Lean Management and Industry 4.0 fields can cope with the issue raised. Combining the notions arising from a review of the scientific literature on the mentioned topics with those resulting from a business case where a Digital Lean program has been implemented, this article provides an analysis of the main effects regarding the integration between the Lean paradigm and digital technologies from a social perspective, assessed through job satisfaction determinants. The results, mapped in an ad-hoc created reference model which merges the main characteristic dimensions of Lean Manufacturing and Industry 4.0, show that the adoption of digital technologies promotes autonomy and upskilling process of operators, just as much as it improves feedback, thus contributing to greater job satisfaction. Nevertheless, the research underlines that such benefits can be dampened whenever digital wastes, high technology costs and a different sensitivity between management and shopfloor operators occur, leading to a strong discouragement among personnel. Keywords: Lean Manufacturing · Industry 4.0 · Sustainability Job satisfaction · Case study

1 Introduction The industrial world is experiencing a significant increase in the number of companies pursuing growth objectives that include environmental and social dimensions in addition to the economic perspective [1]. This transition is often stimulated by national and transnational governments that subsidize studies and fund companies to build and adapt their plants, processes and organizational structures with an eye to sustainability [2]. Lean Manufacturing (LM) methodology can be considered as an important lever for achieving economic, environmental and social sustainability goals [3]. Through the © IFIP International Federation for Information Processing 2023 Published by Springer Nature Switzerland AG 2023 E. Alfnes et al. (Eds.): APMS 2023, IFIP AICT 689, pp. 27–38, 2023. https://doi.org/10.1007/978-3-031-43662-8_3

28

M. Zanchi et al.

implementation of simple operational measures aimed at eliminating waste and conducted with the involvement of people in a process of continuous training [4], LM helps companies to achieve significant improvements in operational performance [5], as well as to reduce their environmental impacts and increase personnel benefits by means of minimizing repetitive tasks, increasing skills and problem-solving abilities, developing safer and healthier environments, and more engaging work contexts [6]. Recently, with the advent of Digital Lean, in which digital technologies have become an integral part of LM methods, principles and techniques [7], new opportunities have arisen for companies in pursuing sustainability goals consistent with the triple bottom line framework [8], thus stimulating an increasing amount of research dealing with this integration [9]. If, from an environmental point of view, researches are numerous and mostly concern the reduced consumption of material and energy resources [10, 11], making the discussion often mono-thematic, the subject of social sustainability, on the other hand, presents a still limited number of works [12], although concerning a wider set of dimensions, including (but not limited to) decision-making processes, training of operators and communication aspects. Moreover, the scarcity of in-depth studies on the effect of Digital Lean on the social sustainability aspect is mainly due to the difficulty in finding suitable indicators of a field which is often difficult to quantify [13], thus requiring the adoption of a qualitative evaluation [14]. In order to assess social sustainability by means of a qualitative study, capable of being effectively adapted to each reality, the criterion of job satisfaction can be considered effective to the purpose, as also highlighted by [15]. Job satisfaction constitutes, in fact, one of the main qualitative criteria for determining whether a working environment is healthy or not [16]. Besides being crucial for good productivity results, it also avoids hidden costs that are often not. This aspect has also been shown by [17], according to which job satisfaction is directly related to a company’s performance. Under these premises, this paper proposes a study aimed at understanding how the integration of I4.0 digital technologies into the LM techniques promotes the improvement of social conditions on the shopfloor, as expressed through the determinants of job satisfaction. By merging the knowledge available from the existing literature and the results of a case study carried out in a company that recently embedded Industry 4.0 technologies into LM initiatives, the study proposes a descriptive model on the effects resulting from the integration between LM and I4.0 technologies on job satisfaction dimensions. Following a brief description of the main characteristics of the topics under observation, Sect. 3 provides a presentation of the research methodology. Then, the reference model and the main evidence emerging from the literature analysis are presented in Sect. 4. Section 5 introduces the case study used for the refinement of the model, while a discussion of the results follows in Sect. 6. Finally, conclusions, research limitations and future developments are reported in last part of the paper.

2 Background Analysis A brief description of the main topics characteristic of this study is provided below.

Effects of Lean and Industry 4.0 Technologies on Job Satisfaction

29

2.1 Lean Manufacturing Lean Manufacturing can be interpreted as a multidimensional approach encompassing different management practices integrated into a single system, whose goals are to align the production with demand and to produce high-quality products avoiding any waste [18]. Several models have been proposed in literature to describe LM. Among the others, [19] describe a LM system through four dimensions: Just In Time (JIT), Total Quality Management (TQM), Total Productive Maintenance and Human Resource Management (HRM). Advocates of LM suggest that the four perspectives should be interpreted as a unified system and that such effective implementation precisely results from a synergetic implementation [20]. JIT encompasses the set of tools to create a system pulled by customer demand, in order to reduce or eliminate waste [21]. TQM encompasses tools designed to avoid errors and promote quality improvement in production. Total Productive Maintenance (TPM) encompasses practices that maximize plant efficiency by adopting predictive and preventive maintenance, supported by rules of order and cleanliness (5S). HRM involves the development of soft skills within organizations, i.e. the promotion of a long-term learning strategy through the achievement of short-term goals [22], continuous improvement developed through kaizen [23], respect to operators [24], job enlargement and job enrichment [25], and teamworking. Table 1 distinguished the main characteristics of LM. Table 1. LM and its distinctive characteristics (based on [19]) LM dimensions

Main techniques/approaches

Just in Time (JIT)

Pull flow/kanban; Heijunka; SMED; Cellar manufacturing; Value Stream Mapping (VSM)

Total Quality Management (TQM)

Poka-yoke; Andon boards; Jidoka

Total Productive Maintenance (TPM)

5S standards; TPM practices

Human Resource Management (HRM)

Long term programs; teamworking; job enlargement and job enrichment; respect to operator

2.2 Industry 4.0 Industry 4.0 (I4.0) can be seen as the integration of information and communication technology within an industrial landscape. Through 4.0 technologies, it is possible to create a Cyber Physical System (CPS), i.e. a network capable of acquiring and processing data, making machines and people communicate with each other. The model proposed by [26] suggests breaking down the CPS paradigm into three main categories: data acquisition and processing, machine-to-machine communication and human machine interaction. Data acquisition and processing includes all the technologies used to collect data (sensor technology and the Internet of Things - IoT), and to store and protect it from intentional or unintentional damage (cloud and cyber security [27]), or to process it through simulation [28]. Machine-to-machine communication is about creating intelligent applications

30

M. Zanchi et al.

such as interconnected machines and equipment that can self-adapt to changes. This feature is achieved through excellent horizontal and vertical integration [29]. Finally, human-machine interaction encompasses two essential aspects, respectively the sharing of information through interfaces such as augmented reality (AR), virtual reality (VR) and smart devices, and direct collaboration between machines and employees, usually through cobots. Table 2 specifies the distinguishing features of I4.0. Table 2. I4.0 and its distinctive characteristics (based on [26]) I4.0 dimensions

Main techniques/approaches

Data acquisition and processing

Sensors and actuators; cloud computing; big data analytics; cybersecurity; simulation; Internet of Things (IoT)

Machine to machine communication

Vertical integration; Horizontal integration

Human-machine interaction

Virtual Reality (VR); Augmented Reality (AR); cobots

2.3 Job Satisfaction Job satisfaction constitutes one of the main qualitative criteria for determining whether a working environment is accommodating in respect of shopfloor personnel or not. Different interpretations of job satisfaction are available in the literature. Among others, [30] define job satisfaction as “…employees’ general feelings about their work and work elements such as work environment, working conditions, fair rewards, and communication with colleagues”. Many studies have attempted to identify the main determinants of job satisfaction [31–34]. On the basis of its different interpretations in literature, it is possible to outline common lines inherent to job satisfaction. These include: • • • • • • •

A healthier, safer and less stressful working environment The enlargement of tasks and the development of new skills The increased autonomy of workers Teamwork and respect for people The feedback from superiors An efficient and unambiguous communication An explicit operational performance evaluation and rewarding system.

3 Research Methodology The work was developed in four main stages, briefly outlined in the following section. 3.1 Model Development The model was built focusing the attention into the distinctive features and macrodimensions of LM and I4.0. In particular, the model of [19] was used to describe the peculiar dimensions of LM, while the model of [26] was adopted to categorize I4.0 technologies.

Effects of Lean and Industry 4.0 Technologies on Job Satisfaction

31

3.2 Model Populating Literature Analysis. A literature analysis was carried out to identify in which measure the intersection between LM and I4.0 technologies affects job satisfaction components. Combining a set of relevant keywords made it possible to identify a sample of 28 relevant articles concerning the integration of the three topics under observation. The search was limited to English-language reported studies, published in refereed scientific journals indexed in Scopus and related to the subject areas of engineering, management and social sciences. The implementation of a snowball approach combined with a forward search aimed at avoiding the limitations of retrospective research, thus making it possible to further expand the sample of analysis. The information contained in each article was then coded and mapped based on the built reference model. The categorization was carried out independently by three researchers. The map-patterns were then compared and aligned, in order to get at a unique and agreed interpretative description. The Study Case. The refinement of the model population was carried out by means of a business case, descriptive of a company that has been experiencing a Digital Lean implementation program. A first gemba walk session aimed to a better understanding of production operations organization. It was propaedeutic to the preparation of the research protocol and check-list used for the interviews, that were conducted face-to-face involving line operators, shopfloor supervisors and managers. In particular, the technological innovation manager, the quality manager, the TPM managers and the responsible of SMED projects were involved. The interview protocol was defined in relation to the interviewee’s job description and role of each respondent. The interviews were carried out adopting a semi-structured approach in order to avoid any influence on the respondents’ opinions. Furthermore, the interviews were recorded and then transcribed on an online platform in accordance with national privacy regulations. Subsequently, each researcher acted by coding the transcribed data with the help of an interpretative model [35]. Starting from the transcripts of the interviews, it was possible to subdivide the statements of the interviewees by similarity and to merge them into a first level of analysis. The research continued with the second level analysis, which involved extrapolating a theme from each interpretation identified at the previous level. The formulation of the second level was done by reconciling the themes that emerged from the case study and the theoretical aspects discussed in the literature review [35]. Finally, the last dimension of the analysis, i.e. the aggregate one, was obtained by combining the effects of two or more topics from the second level. The results were then compared to get a shared interpretation and to update the contents of the previous mapping model.

4 Model Populating In accordance with the definitions of LM and I4.0 presented in the background section, the model was built focusing around the dimensions of Just in Time (JIT), TQM, TPM and HRM for the former, and “Data acquisition and processing”, “Machine to machine communication” and “Human-machine interaction” for the latter (Fig. 1).

32

M. Zanchi et al.

Fig. 1. The research model (based on [19, 26])

4.1 Results from Literature The literature analysis made it possible to populate the model thus depicting relevant considerations on how and to what extent the integration of I4.0 technologies into LM methods can stimulate job satisfaction by acting on its main determinants. Firstly, the two paradigms appear positively correlated and their joint adoption enables better production performance, supporting each other in a virtuous cycle [36, 37]. Many of the LM tools can be enhanced by the use of one or more technologies, confirming the excellent relationship of the I4.0 paradigm on LM, in line with [38]. Regarding the effects of such integration on the determinants of job satisfaction, existing studies underline a strong interaction between data acquisition and data processing in the four main areas characterizing LM, namely JIT, TQM, TPM and HRM. Indeed, thanks to new technologies, it is possible to capture a lot of data and make more accurate decisions. However, this brings with it the need to acquire new skills, as pointed out by [39], especially in the area of problem solving. On the other hand, literature shows how the use of technologies that facilitate machine to machine interaction results in a reduction of the most repetitive operations carried out by workers, avoiding them having to manually control individual parts (in the case of TQM) or the direct control of machines in the case of stoppages or short interruptions (TPM). Operators are thus only called upon to intervene in case of serious problems [40] and can be allocated to new, higher-value roles, taking on new tasks and responsibilities. Thanks to the use of new technologies, the environment is safer, activities less physically demanding and work less alienating [41]. Moreover, the better management of personnel through the use of technology enables a more balanced distribution of tasks [42], thus resulting in a less stressful working environment. If, on the one hand, new I4.0 technologies that facilitate the human-machine interaction encourage operator’s autonomy (as in the case of AR technologies that act as supportive tools in the field of predictive maintenance), on the other hand such tools make communication among departments and with colleagues or managers more efficient. Indeed, the use of digital platforms implies the standardization of data input and data management logics.

Effects of Lean and Industry 4.0 Technologies on Job Satisfaction

33

This results in a better accessibility and utilization of information by people [43], thus leading to a greater decisional independency and more opportunities for teamwork and people collaboration, as in the case of augmented reality where a maintenance operator can be helped with his tasks by a technician located in another plant [44]. 4.2 The Study Case The selected company is a large Norwegian industrial company operating in the automotive sector leading manufacturer of suspension components and bumper support parts. The company has long applied LM principles. For the past few years, it has been introducing a very sophisticated set of I4.0 technologies in line with the idea of ‘lean first, then digital’ logic, convinced that effective digitization is only possible when the production process is already efficient [45]. In particular, a variety of technological devices (both hardware and software) are used to support the main Just In Time Lean techniques, aimed at improving the planning and control of setup and production activities and developing the skills necessary to manage new technologies. With reference to quality control, the company also makes use of a system for detecting errors and stopping production in the event of abnormal behavior, integrated with the operational performance reporting system and the factory scheduler. The use of an Andon 4.0 system, for example, allows operators to receive all the information on the status of the line (in terms of productivity, quality, availability) and at the same time to report any criticalities and anomalies (also in terms of safety) that can be used by management and/or other company functions to replan their activities and identify appropriate improvement actions. As far as maintenance is concerned, a number of 4.0 technologies are available to facilitate assistance activities, even remotely, as in the case of AR, used to allow expert personnel to assist the departmental operator. Finally, the company enjoys a very high level of sustainability strategies and practices, leveraging on availability of financial, organizational and skills resources. The visit at the company’s production plant and subsequent interviews highlighted 4 main effects, resulting from the integration of LM techniques with I4.0 digital technologies, that have an impact on the determinants of job satisfaction. They are briefly described in the following: 1. The enrichment of technical skills. The launch of I4.0 projects required the introduction of new technical skills, as in the case of a new project introduced to improve scheduling according to the Just in Time principle, which involved the activation of an ‘ad hoc’ training program for all shopfloor operators. 2. The reduction of repetitive operations and the allocation of operators to new tasks. The adoption of new technologies, such as automated weld control or the use of sensors to record and automatically signal stoppages, allowed operators to manage more lines or to devote their time to new tasks that involve taking on new responsibilities more related to decision-making. 3. A clearer and more transparent definition of short-term objectives. A pool of effective tools introduced in the company to ease the scheduling of activities, facilitated the operators in the prioritization of operational activities and SMED projects making clearer the monitoring of their operations via an interconnected cloud-based system. Thanks to this timely and unambiguous feedback system, operators have better control over the process, and are therefore more empowered and less stressed.

34

M. Zanchi et al.

4. Standardization and sharing of information. The information stored in the cloud provided a clear and standardized view of operational interventions, while an historical dataset, easily accessible by each operator through digital platform, made operators more autonomous in their decisions. Moreover, the use of augmented reality to connect different plants enabled knowledge sharing among technicians, so as to facilitate the development of a strong sense of cooperation between the different company departments in a teamwork perspective. The interviews also emphasized that the effect of the integration of LM and digital technologies on the determinants of job satisfaction also depends on the conditions of the implementation context. For instance, the absence of adequate support in performing the redefinition of scheduling processes, which is necessary upon implementation of the factory activity scheduler, increased the disappointment of some, who saw the time in which they would have experienced greater autonomy slipping away. Similarly, when the cost of technology limits or slows down its implementation, an imbalance is created within the organization that leads to discontent among those who cannot benefit from access to such technologies. For example, not everyone in the company has been equipped with hardware devices, such as tablets or smartphones, mainly due to the high cost of licenses and the reduced utilization of available technological resources. This situation is further aggravated when the company chooses to pursue a policy of training on the use of new technologies in accordance with an ‘e-learning’ approach to which staff who cannot make use of these technologies do not have access, but instead should ensure that all operators in the department are adequately trained through a specific course. Figure 2 shows the research model populated with the outcomes emerging from both the literature review process and the study case. The figure reports, in particular, some examples of the possible interactions found between LM and I4.0 technologies in the light of “Job satisfaction” factors.

Fig. 2. The populated research model

Effects of Lean and Industry 4.0 Technologies on Job Satisfaction

35

5 Discussion The integration of the information from the literature with the business case underlines that, through the use of technology, certain critical aspects of Lean management can be improved, and that through this integration certain factors characteristic of the creation of job satisfaction are stimulated. Specifically, up-skilling is fostered, by moving the operator from repetitive tasks, which are automated, to new and more complex tasks creates a gap between the necessary and available knowledge. At the same time, the operator is deprived of the obligation to perform unnecessary or lower-value tasks, thus paving the way for upskilling actions aimed at broadening his/her skills beyond the technical dimension to include problem solving, organization and management skills. Secondly, communication is improved and feedback is made clearer, as the communication of short-term goals and the sharing of information allows for an unambiguous flow of information within the company, thus confirming the results found from [34]. This fosters greater autonomy of the operators and at the same time improves the exchange of information, thus creating the prerequisite for the development of greater adherence in the work teams. The results gathered show how, ultimately, the implementation of Digital Lean processes inside the case study under review had a twofold effect. If, on the one hand, the application of such logics allowed to achieve a higher level of job satisfaction within the production context, as described above in terms of the reduction of repetitive tasks and greater information transparency, thereby fulfilling a “facilitator” role; on the other hand, achieving such a result necessitated taking the proper corrective measures at the initial production setting, such as developing a standard for the accurate transmission and sharing of information and designing an ad-hoc training program to encourage the expansion of staff members’ skill sets, thus acting as a ‘binding’ factor. This duality therefore shows how the implementation of Digital Lean logic does not so much alter the characteristic conditions of job satisfaction, described in Sect. 2.3, but rather accelerates its attainment process. Finally, the case study emphasizes that while the shift to the use of technology is always viewed positively by the operators, the management believes that the benefits that technologies inevitably provide in terms of employee well-being must be contrasted with the often excessively high costs of new technologies that significantly impact on economic performance. High costs, under-utilized technologies in daily work, unsuitable processes or reduced skills to be able to use them profitably lead management in some cases to select only a few technologies, thus slowing down the digital transformation process and instilling in some people a sense of frustration or disappointment, which acts in opposition to the benefits that the integration of LM and I4.0 can generate.

6 Conclusions, Limitations and Further Improvements This paper investigated the relationship between LM and I4.0 technologies with respect to social sustainability, exploring it through job satisfaction and its determinants. The use of a reference model built on the basis of the main determinants of LM and I40

36

M. Zanchi et al.

technologies made it possible to map how and to what extent the integration of the two paradigms influences job satisfaction, integrating information from the literature and from a case study carried out at a company that has been implementing Digital Lean for a few years. The results emphasize that the integrated adoption of LM and I4.0 promotes autonomy and upskilling of operators and improves feedback, thus contributing to the creation of an environment in which information is more easily usable, manageable and shareable, which stimulates job satisfaction. Nonetheless, the results of the case study point out that these benefits can be counteracted by opposing effects, in particular a growing disappointment due to the difficulty of disseminating technologies on the factory floor due to digital wastes, too high technology costs and a different sensitivity between management and shopfloor operators. Future studies also in other industrial sectors or in other cultural contexts would allow not only to consolidate and improve the considerations that emerged in this work, but also to analyze how and to what extent the organizational culture of a country mediates in the relationship between the implementation of Digital Lean logics and job satisfaction. Finally, since the work focused on how the integration of LM and I4.0 acts on the determinants of job satisfaction, rather than directly measuring its impact, future work could focus through some quantitative analysis on quantifying the adoption of Digital Lean on the satisfaction index.

References 1. Callan, S.J., Thomas, J.M.: Environmental Economics and Management: Theory, Policy, and Applications. 6th edn. South Western Cengage learning, Mason, OH (1999) 2. Grin, J., Rotmans, J., Schot, J.: Transitions to Sustainable Development: New Directions in the Study of Long Term Transformative Change, 1st edn. Routledge, New York, NY (2010) 3. Powell, D.J., Romero, D., Gaiardelli, P.: New and renewed manufacturing paradigms for sustainable production. Sustainability 14(1279) (2022) 4. Sundar, R., Balaji, A.N., Kumar, R.S.: A review on lean manufacturing implementation techniques. Procedia Eng. 97, 1875–1885 (2014) 5. Belekoukias, I., Garza-Reyes, J.A., Kumar, V.: The impact of lean methods and tools on the operational performance of manufacturing organisations. Int. J. Prod. Res. 52(18), 5346–5366 (2014) 6. Gaiardelli, P., Resta, B., Dotti, S.: Exploring the role of human factors in lean management. Int. J. Lean Six Sigma 10(1), 339–366 (2019) 7. Sanders, A., Elangeswaran, C., Wulfsberg, J.P.: Industry 4.0 implies lean manufacturing: research activities in industry 4.0 function as enablers for lean manufacturing. J. Indust. Eng. Manage. 9(3), 811–833 (2016) 8. Yilmaz, A., Dora, M., Hezarkhani, B., Kumar, M.: Lean and industry 4.0: mapping determinants and barriers from a social, environmental, and operational perspective. Technol. Forecast. Soc. Change 175, 121320 (2022) 9. Cherrafi, A., Elfezazi, S., Chiarini, A., Mokhlis, A., Benhida, K.: The integration of lean manufacturing, Six Sigma and sustainability: a literature review and future research directions for developing a specific model. J. Clean. Prod. 139, 828–846 (2016) 10. Kamble, S., Gunasekaran, A., Dhone, N.C.: Industry 4.0 and lean manufacturing practices for sustainable organisational performance in Indian manufacturing companies. Int. J. Product. Res. 58(5), 1319–1337 (2020)

Effects of Lean and Industry 4.0 Technologies on Job Satisfaction

37

11. Veleva, V., Ellenbecker, M.: Indicators of sustainable production: framework and methodology. J. Clean. Prod. 9(6), 519–549 (2001) 12. Bokhorst, J.A., Knol, W., Slomp, J., Bortolotti, T.: Assessing to what extent smart manufacturing builds on lean principles. Int. J. Prod. Econ. 253, 108599 (2022) 13. Dillinger, F., Bergermeier, J., Reinhart, G.: Implications of digital lean methods on relevant target dimensions: time, cost, quality, employee involvement, and flexibility. Procedia CIRP 107, 202–208 (2022) 14. Digalwar, A.K., Dambhare, S., Saraswat, S.: Social sustainability assessment framework for indian manufacturing industry. Mater. Today Proc. 28, 591–598 (2020) 15. Pang, Q., Fang, M., Wang, L., Mi, K., Su, M.: Increasing couriers’ job satisfaction through social-sustainability practices: perceived fairness and psychological-safety perspectives. Behav. Sci. 13(2), 125 (2023) 16. Kaya, E.: Job satisfaction of the librarians in the developing countries. In: 61st IFLA General Conference (1995) 17. Hancock, J.I., Allen, D.G., Bosco, F.A., McDaniel, K.R., Pierce, C.A.: Meta-analytic review of employee turnover as a predictor of firm performance. J. Manag. 39(3), 573–603 (2013) 18. Yang, M.G.M., Hong, P., Modi, S.B.: Impact of lean manufacturing and environmental management on business performance: an empirical study of manufacturing firms. Int. J. Prod. Econ. 129(2), 251–261 (2011) 19. Shah, R., Ward, P.T.: Lean manufacturing: context, practice bundles, and performance. J. Oper. Manag. 21, 129–149 (2003) 20. Womack, J.P., Jones, D.T.: Lean thinking–banish waste and create wealth in your corporation. J. Oper. Res. Soc. 48(11), 1148 (1997) 21. Sugimori, Y., Kusunoky, K., Cho, F., Uchikawa, S.: Toyota production system and Kanban system: materialisation of just-in-time and respect-for-human system. Int. J. Product. Res. 15(6), 553–564 (1977) 22. Dorota Rymaszewska, A.: The challenges of lean manufacturing implementation in SMEs. Benchmark. Int. J. 21(6), 987–1002 (2014) 23. Janji´c, V., Todorovi´c, M., Jovanovi´c, D.: Key success factors and benefits of Kaizen implementation. Eng. Manag. J. 32(2), 98–106 (2020) 24. Romero, D., Gaiardelli, P., Wuest, T., Powell, D., Thürer, M.: New forms of Gemba walks and their digital tools in the digital lean manufacturing world. In: Lalic, B., Majstorovic, V., Marjanovic, U., von Cieminski, G., Romero, D. (eds.) APMS 2020. IAICT, vol. 592, pp. 432–440. Springer, Cham (2020). https://doi.org/10.1007/978-3-030-57997-5_50 25. Cimini, C., Lagorio, A., Gaiardelli, P.: The evolution of operators’ role in production: how Lean Manufacturing and Industry 4.0 affect Job Enlargement and Job Enrichment. Int. J. Product. Res. 1–19 (2022) 26. Wagner, T., Herrmann, C., Thiede, S.: Industry 4.0 impacts on lean production systems. Procedia Cirp 63, 125–131 (2017) 27. Butt, J.: A strategic roadmap for the manufacturing industry to implement industry 4.0. Designs 4(2), 11 (2020) 28. Guala, F.: Models, simulations, and experiments. In: Magnani, L., Neressian, N.J. (eds.) Model-Based Reasoning, pp. 59–74. Springer, Boston, MA (2002) 29. Stock, T., Seliger, G.: Opportunities of sustainable manufacturing in industry 4.0. Procedia Cirp 40, 536–541 (2016) 30. Kim, W.G., Leong, J.K., Lee, Y.K.: Effect of service orientation on job satisfaction, organizational commitment, and intention of leaving in a casual dining chain restaurant. Int. J. Hosp. Manag. 24(2), 171–193 (2005) 31. Griffin, M.A., Patterson, M.G., West, M.A.: Job satisfaction and teamwork: the role of supervisor support. J. Organ. Behav. Int. J. Indust. Occup. Organiz. Psychol. Behav. 22(5), 537–550 (2001)

38

M. Zanchi et al.

32. Lam, T., Zhang, H., Baum, T.: An investigation of employees’ job satisfaction: the case of hotels in Hong Kong. Tour. Manage. 22(2), 157–165 (2001) 33. Linz, S.J.: Job satisfaction among Russian workers. Int. J. Manpow. 24(6), 626–652 (2003) 34. Giri, V.N., Pavan Kumar, B.: Assessing the impact of organizational communication on job satisfaction and job performance. Psychol. Stud. 55(2), 137–143 (2010) 35. Gioia, D.A., Corley, K.G., Hamilton, A.L.: Seeking qualitative rigor in inductive research: notes on the Gioia methodology. Organ. Res. Methods 16(1), 15–31 (2013) 36. Tortorella, G.L., Giglio, R., Van Dun, D.H.: Industry 4.0 adoption as a moderator of the impact of lean production practices on operational performance improvement. Int. J. Oper. Product. Manage. 39(6/7/8), 860–886 (2019) 37. Narula, S., et al.: Are Industry 4.0 technologies enablers of lean? Evidence from manufacturing industries. Int. J. Lean Six Sigma 14(1), 115–138 (2023) 38. Dombrowski, U., Richter, T., Krenkel, P.: Interdependencies of industry 4.0 and lean production systems: a use cases analysis. Procedia Manufac. 11, 1061–1068 (2017) 39. Fonda, E., Meneghetti, A.: The human-centric smed. Sustainability 14(1), 514 (2022) 40. Deuse, J., Dombrowski, U., Nöhring, F., Mazarov, J., Dix, Y.: Systematic combination of Lean Management with digitalization to improve production systems on the example of Jidoka 4.0. Int. J. Eng. Bus. Manage. 12 (2020) 41. Boudella, M.E.A., Sahin, E., Dallery, Y.: Kitting optimisation in Just-in-Time mixed-model assembly lines: assigning parts to pickers in a hybrid robot–operator kitting system. Int. J. Prod. Res. 56(16), 5475–5494 (2018) 42. Ejsmont, K., Gladysz, B., Corti, D., Castaño, F., Mohammed, W.M., Martinez Lastra, J.L.: Towards ‘Lean Industry 4.0  - Current trends and future perspectives. Cogent Bus. Manage. 7(1), 1781995 (2020) 43. Tortorella, G.L., Fogliatto, F.S., Cauchick-Miguel, P. A., Kurnia, S., Jurburg, D.: Integration of Industry 4.0 technologies into total productive maintenance practices. Int. J. Product. Econ. 240, 108224 (2021) 44. Valamede, L.S., Akkari, A.C.S.: Lean 4.0: digital technologies as strategies to reduce waste of lean manufacturing. In: Iano, Y., Saotome, O., Kemper Vásquez, G.L., Cotrim Pezzuto, C., Arthur, R., Gomes de Oliveira, G. (eds.) Proceedings of the 7th Brazilian Technology Symposium (BTSym 2021). BTSym 2021. SIST, vol. 295, Springer, Cham. https://doi.org/ 10.1007/978-3-031-08545-1_7 45. Nicoletti, B.: Lean and automate manufacturing and logistics. In: Prabhu, V., Taisch, M., Kiritsis, D. (eds.) Advances in Production Management Systems. Sustainable Production and Service Supply Chains, pp. 278–285. Springer Berlin Heidelberg, Berlin, Heidelberg (2013). https://doi.org/10.1007/978-3-642-41263-9_34

Lean Supply Chain and Industry 4.0: A Study of the Interaction Between Practices and Technologies Matteo Rossini(B) , Stefano Frecassetti , and Alberto Portioli-Staudacher Department of Management, Economics and Industrial Engineering, Politecnico di Milano, Milano, Italy [email protected]

Abstract. The ever-increasing competition due to the rapidly changing environment and businesses pushes to complement the adoption of Lean Supply Chain with different concepts or tools. Among these, since the Lean culture is extremely receptive to new technologies, Lean Supply Chain is being studied in relation to new trends such as Industry 4.0. In fact, Industry 4.0 has the same focus as Lean: the reduction of costs and wastes and the increase of efficiency along the whole supply chain. Therefore, this paper investigates which Lean Supply Chain practices are related to Industry 4.0 technologies and how they interact. Through the use of the DEMATEL methodology, this paper tries to answer this question with the final aim of providing a comprehensive overview of the interactions. This will provide both academicians and practitioners with thorough knowledge about the direction of their interrelation, representing a current literature gap. Keywords: Lean Supply Chain · I4.0 · Industry 4.0 · DEMATEL

1 Introduction For many years the market has been characterized by turbulent markets and a rapidly shifting competitive landscape; firms grapple with the challenge of providing superior value to demanding customers (Eisenhardt and Martin 2000). In this scenario, firms must provide an enhanced performance level characterized by a more effective cost structure and the ability to adapt quickly to the ever-changing market needs (Vanpoucke et al. 2014). Therefore, supply chains need to be more efficient and sustainable both in the short and the long term. One approach that can help supply chains overcome inefficiencies, reduce waste, and achieve sustainability is Lean Production, which originated from the Toyota Production System. In an environment characterized by increasing competitive pressure for shorter lead times, lower costs and better quality, Lean Production procedures, techniques, practices and tools are being implemented while managing the supply chain in the spirit of an integrative approach (Hines et al. 2004; Cudney and Elrod 2010). In this sense, the extension of the application of Lean principles and practices to the supply chain has been called Lean Supply Chain Management (Anand and Kodali © IFIP International Federation for Information Processing 2023 Published by Springer Nature Switzerland AG 2023 E. Alfnes et al. (Eds.): APMS 2023, IFIP AICT 689, pp. 39–53, 2023. https://doi.org/10.1007/978-3-031-43662-8_4

40

M. Rossini et al.

2008). This paradigm emphasizes using Lean Production practices synergistically to create high-quality production and logistics systems that produce and deliver according to customer demand (Shah and Ward 2007; Li et al. 2006). Therefore, due to the benefits provided to manufacturing environments, incorporating Lean Production principles and practices into the supply chain has culminated in differentiated results along the supply chain, surpassing those already achieved by the organizations individually (Arif-Uz-Zaman and Ahsan 2014). However, given the tough market condition of recent years, companies are continuously seeking new ways to optimize their supply chain management practices. The rise in digital technologies after the spread of the Industry 4.0 paradigm has brought significant changes in the manufacturing industry and opened up new opportunities to improve supply chain management. Thus, some researchers started investigating the relationship between Lean Supply Chain practices and Industry 4.0, showing that there could be a synergistic effect in several cases. However, the impact of Industry 4.0 technologies on Lean Supply Chain practices and vice-versa is still poorly understood. While Lean Supply Chain practices focus on reducing waste and enhancing efficiency, Industry 4.0 technologies can significantly improve supply chain management by providing real-time data, enabling autonomous decision-making, and improving traceability. In fact, some authors focused on the relationship between these two paradigms only at an operational level (Ciano et al. 2021) and not at a supply chain level. Others (Nunez Merino et al. 2020) pointed out how studies are scarce regarding how Lean Supply Chain practices can facilitate and boost Industry 4.0 technologies. Also, Rossini et al. 2023 showed the relationship between practices and technologies but only at a theoretical level. Therefore, this paper aims to empirically investigate the interplay between Lean Supply Chain practices and Industry 4.0 technologies. Specifically, the authors will explore how Lean Supply Chain practices can benefit from Industry 4.0 technologies and how Industry 4.0 technologies can be adopted to enhance the effectiveness of Lean Supply Chain practices. The research question of this paper is: “How do Lean Supply Chain practices and Industry 4.0 practices influence each other?”. This paper contributes to the growing literature on Lean Supply Chain and Industry 4.0. Giving insights into the interplay between Lean Supply Chain practices and Industry 4.0 technologies will provide practical implications for firms and managers. After this introductory section, this paper will present a literature review and the methodology used to answer the research question. Then, the results obtained will be presented and discussed in the existing body of literature, and the paper will end with the conclusion part.

2 Literature Review 2.1 Lean Supply Chain Lean Supply Chain has been defined as implementing Lean Production practices to the whole supply chain rather than a mere application at a company level. A Lean Supply Chain (LSC) can be defined as “a set of organizations directly linked by upstream and downstream flows of products, services, information, and funds that

Lean Supply Chain and Industry 4.0: A Study of the Interaction

41

collaboratively work to reduce cost and waste by efficiently pulling what is needed to meet the needs of individual customers” (Vitasek et al. 2005). A Lean Supply Chain aims to ensure that value is transferred downstream most efficiently. Within the paradigm of a Lean Supply Chain, there are several practices that companies can implement. For instance, Tortorella et al. (2017) with their study selected the 22 Lean Supply Chain practices most cited in the available literature and, starting from the eight pillars framework proposed by Jasti and Kodali (2016), have created four bundles of interrelated and internally consistent Lean Supply Chain practices. This instrument allows researchers to assess the state of LSCM implementation in companies and their supply chains and to test hypotheses about relationships between Lean Supply Chain Management and other characteristics that might affect performance (Tortorella et al. 2017). The four bundles identified by Tortorella et al. (2017) are: • • • •

Customer-supplier relationship management (CSRM) Logistics management (LOM) Elimination of waste and continuous improvement (EWCI) Top management commitment (TMC)

These practices can be beneficial for the performance of the whole supply chain. For instance, Green et al. (2014) investigated how using JIT purchasing, production, selling and information reduces logistics costs for suppliers and customers, suggesting that adopting these practices entails benefits for both of them. Other evidence of this approach is found in Jean et al. (2014) and Kohtamaki and Partanen (2016), which examined the influence of customer–supplier relationships for enhancing knowledge and learning along the supply chain for product innovation and services development. In general, these studies support the existence of a positive impact of certain Lean Supply Chain practices on performance, although they usually investigate them from a limited point of view. 2.2 Industry 4.0 Industry 4.0, which is currently taking place, sets a few challenges for manufacturing companies from the technological, organizational and management points of view. According to Lee et al. (2018), Industry 4.0 is “a result of the horizontal expansion of information technology”. In fact, information and communication technologies are used in a much more extensive way than before in all spheres, including business, government and everyday life (Kovács 2017). The term Industry 4.0 describes the “increasing digitization of the entire supply chain, which makes it possible to connect actors, objects and systems based on real-time data exchange” (Spath et al. 2013; Dorst et al. 2015). As a result of this interconnection, products, machines and processes with artificial intelligence will be able to adapt to changing environmental factors. The traditional approach to digitization defines it as “the use of computer and internet technology for a more efficient and effective economic value creation process” (Reddy and Reinartz 2017). Digitization is affecting all sectors, with traditional products that are either being replaced by digital counterparts or at least equipped with new digital features (Prem 2015). However, digital transformation or Industry 4.0 goes beyond product and process improvement. Indeed, it has created

42

M. Rossini et al.

new advanced production models and new technologies that allow the Digitalization of processes, products, services and business models, creating significant challenges for companies (Bleicher and Stanley 2016). In particular, the ultimate goal of Digitalization is to build connectivity and interaction between all parties involved in manufacturing. In the world of future production systems, some processes are expected to be simplified and others to become much more complex and embedded, leading to an increase in the number of higher-skilled jobs and a reduction in jobs requiring lower qualifications (Spath et al. 2013) that will require, in turn, changes in the education system. In line with the expected changes, companies are becoming more and more interested in the application of new technologies to ensure long-term competitiveness and to enable them to adapt to dynamically- changing environmental conditions, such as shortening product lifecycles, increasing diversity and changing consumer expectations (Spath et al. 2013). Therefore, Industry 4.0 will significantly impact both the labour market and society. In addition, Industry 4.0 gives rise to a new definition of the performance model of manufacturing processes, introducing the social aspect besides the economic, energy, environmental, and functional ones. 2.3 Lean Supply Chain and Industry 4.0 With the automation provided by Industry 4.0, questions about interoperability with the Lean approach have risen. According to Bittencourt et al. (2019), the Lean work environment nurtures a culture more receptive to new technologies, especially the ones that reduce waste. In this sense, Lean and Industry 4.0, despite with different approaches, can and should be complementary since they have the same goal of reducing costs and increasing productivity for companies. These authors conducted a systematic literature review to investigate this interaction, focusing on Lean’s role in this ongoing Industrial Revolution. They derived that Lean could become a facilitator in the implementation process of Industry 4.0. In fact, Lean concepts such as work standardization, organization and transparency are highlighted as support for implementing solutions linked to I4.0 by Kolberg and Zühlke (2015). Therefore, the level of maturity of the Lean system within a company is an important variable in the association process with Industry 4.0. A positive response about the support that Lean can provide to I4.0 was also achieved by Saxby et al. (2020). However, the same authors stressed the importance of understanding which Lean elements can support Industry 4.0 since implementing LSC practices that give little or no support to I4.0 would cost money but generate no return on that investment. Therefore, it is important to question the sustainability of Lean to support I4.0 effectively. In particular, Saxby et al. (2020) have suggested that what Lean needs is simply some updating to address I4.0 rather than its wholesale re-invention. Changing the perspective and investigating how Industry 4.0 can support the Lean Supply Chain, Haddud and Khare (2019) investigated where digitalizing supply chains may benefit businesses in a Lean environment. More specifically, the study examined the potential impacts of digitalizing supply chains on five selected lean operations practices that are Just-In-Time (JIT), Visual Management (VM), Total Productive Maintenance (TPM), Continuous Improvement (CI), and Autonomation (failure prevention) and PokaYoke (mistake-proofing). The authors confirmed the significant impact of digitalizing supply chains on all five examined lean operations practices.

Lean Supply Chain and Industry 4.0: A Study of the Interaction

43

Moreover, Nunez-Merino et al. (2020) found out that companies in the supply chain are using several information and digital technologies of I4.0 as an aid to implement Lean practices (LSCM). This effect coincides with the one that stands out in the internallevel IT-Lean relationship highlighted by prior studies (Pinho and Mendes 2017; Buer et al. 2018; Sony 2018). Nevertheless, when extended to the supply chain, a large group of companies uses many I4.0 instruments and Lean practices simultaneously, achieving a strong accumulated multiplying effect on efficiency and, ultimately, on supply chain results. More specifically, these authors stressed the tremendous impact on flexibility, agility, information sharing, synchronization and collaboration, integration, coordination, speed, reduced costs, deliveries, errors, reduced inventory levels, enabling a pull system, enabling traceability and inventory control, reduced risk, improved service quality and customer satisfaction.

3 Methodology To properly answer the research question, the authors have chosen to use the DEMATEL (Decision Making Trial and Evaluation Laboratory) methodology. This methodology has been used previously by other authors dealing with similar topics (Costa et al. 2019). This method is particularly useful when there is the need to demonstrate cause and effect relationship between different factors. Indeed, this technique has been implemented here first to highlight cause and effect relationships between Lean Supply Chain Practices and Industry 4.0 technologies. To do that, firstly, the authors decided to map both Lean Supply Chain Practices and Industry 4.0 from the literature to understand the most relevant ones. Afterwards, to implement the DEMATEL, it was necessary to develop a survey and to involve some experts to gather their responses. A data cleaning on the responses was performed as a last step, and then the DEMATEL methodology was applied. 3.1 Variables Selection Given the wide range of practices identified, especially regarding the Lean Supply Chain, a selection is necessary for these variables to run the analyses. Given the differences in the number and related characteristics, two approaches have been adopted for Lean Supply Chain and Industry 4.0 practices. A detailed explanation of the choices made is provided in the following paragraphs. Lean Supply Chain Practices Selection Since there are already robust and reliable studies regarding the LSC practices in the literature, the authors have decided to rely on pre-existing analysis. Specifically, the analysis performed in the study of Tortorella et al. (2017) is addressed to identify the most cited Lean Supply Chain practices. Relying on 26 articles, Tortorella et al. (2017) identified the 22 Lean Supply Chain practices most quoted in the literature. Considering the final aim of this paper and the characteristics of the four different bundles, the authors have obtained 13 Lean Supply Chain practices. Indeed, to be coherent with the orientation of Smart practices, the two operative bundles have been selected among the four. Logistics Management (LOM) and Elimination of Waste and Continuous Improvement

44

M. Rossini et al.

(EWCI) are more related to the operational application of the Lean paradigm. At the same time, Customer-Supplier Relationship Management (CSRM) and Top Management Commitment (TMC) are more related to the strategic one. The largest amount of Lean Supply Chain practices are contained in the Logistics Management (LOM) bundle: • P1 - Efficient and continuous replenishment: Anand and Kodali (2008). • P2 - Material handling systems: Anand and Kodali (2008). • P3 - Standardized work procedures to assure quality achievement: Anand and Kodali (2008), Vitasek et al. (2005). • P4 - Open-minded and in-depth market research conducted jointly: Vokurka (2000). • P5 - Inbound vehicle scheduling: Yusuf et al. (2004). • P6 - Outbound transportation: Anand and Kodali (2008). • P7 - Establishment of distribution centres: Arntzen et al. (1995). • P8 - Functional packaging design: Tortorella et al. (2017). The practices that belong to the second bundle selected, Elimination of Waste and Continuous Improvement (EWCI), are: • P9 - Kanban or pull system: Anand and Kodali (2008). • P10 - Levelled scheduling or Heijunka: Anand and Kodali (2008), Stratton and Warburton (2003). • P11 - Consignment stock: Cagliano et al. (2006). • P12 - Win-win problem-solving methodology: Stratton et al. (2003). • P13 - Value chain analysis or value stream mapping: Rother and Shook (1999). Industry 4.0 Practices Selection The Fourth Industrial Revolution was characterized by novel technologies that could be applied in the manufacturing field, leading to several benefits. The authors decided to include 9 Industry 4.0 technologies in this study. Therefore, the authors have decided to consider in further analyses the nine practices recommended by the expert consulted. In addition, considering all the nine practices enables a balanced analysis of Lean Supply Chain and Industry 4.0. Hence, for the Industry 4.0 environment, defining a criterion for the selection was not necessary. Below are represented the selected technologies: • • • • • • • • •

P14 - Big Data and Analytics P15 - Autonomous and Collaborative Robots P16 - Simulation P17 - Horizontal and Vertical Integration P18 - Industrial Internet Of Things (IoT) P19 - Cybersecurity P20 - Cloud P21 - Additive manufacturing P22 - Augmented reality

Lean Supply Chain and Industry 4.0: A Study of the Interaction

45

3.2 Survey Deployment Once all the variables belonging to both paradigms, namely Lean Supply Chain and Industry 4.0, were defined, a survey was created and distributed to different experts. They were required to fill the numerical survey with 1 – no influence, 2 – low influence, 3 – medium influence, 4 – high influence, 5 – extreme influence according to the level of influence between the practices under analysis. The surveys from the thirteen experts involved in this research have been collected and analyzed to check the quality and goodness of the responses. The review of the returned tables found that five responses had to be discarded from the study because of incompleteness or incorrect understanding of the request.

4 DEMATEL The approach used for the DEMATEL methodology is the same as that performed by several authors in the literature (Costa et al. 2019). In this section, the application of the DEMATEL technique to define the relationships between the selected LSC and Industry 4.0 practices is going to be discussed. Step 1: Generation of the Average Direct Relation Matrix “A” Starting from the eight eligible surveys, the experts’ responses were transformed from a 1–5 scale to a 0–4 scale, as required by the DEMATEL technique. Then, a 22 × 22 Average Direct Relation Matrix was built by averaging the responses obtained from the experts (Table 1). Table 1. Average Direct Relation Matrix “A”

Step 2: Computation of Normalized Initial Direct Relation Matrix “X” The second step has been to normalize the Average Direct Relation Matrix using the maximum value obtained among the sum of all the rows and columns of A to get S =

46

M. Rossini et al. Table 2. Normalized Initial Direct Relation Matrix “X”

0,01975. This value has allowed us to normalize the Average Direct Relation Matrix and to compute the Normalized Initial Direct-Relation Matrix X reported in Table 2. Step 3: Determination of the Total Relation Matrix “T” Then, using the Normalized Initial Direct Relation Matrix X, it has been possible to build the Total Relation Matrix T presented in Table 3 through T = 〖X(I−X)〗^(−1) where “I” is denoted as the identity matrix 22 x 22. Table 3. Normalized Initial Direct Relation Matrix “X”

Step 4: Compute Dispatcher and Receiver Groups Afterwards, with the Total Relation Matrix T values, the row and column sums for each row and column of the T matrix have been computed and subsequently used to determine each practice’s overall prominence and net effect (Table 4). From this, it was possible to distinguish the dispatcher practices with a positive net effect from those with receivers characterized by a negative net effect. In addition, from

Lean Supply Chain and Industry 4.0: A Study of the Interaction

47

the degree of prominence, it was possible to detect which practices develop stronger relationships and which ones have weaker relationships with others. Table 4. Prominence and Net Effect results

Step 5: Set a Threshold and Draw the Cause-Effect Diagram After consulting an expert, the threshold has been defined as the average value of the Total Relation Matrix T, resulting in being avg = 0.2125. In addition, a supplementary threshold obtained by adding one standard deviation to the average (avg + std dev = 0,2488) has been computed to understand the stronger relationships. The prominence and net effect values shown in Table 4 have allowed us to position each practice in the diagram. In addition, each relationship among the practices has been plotted in the diagram through an arrow if the value is higher than the threshold identified, as shown in Table 4. Finally, the Cause-Effect Diagram has been drawn and is presented in Fig. 1.

5 Results and Discussion This section aims to investigate the relationships between Lean Supply Chain and Industry 4.0 practices, given the results coming from the DEMATEL methodology (Fig. 1). In particular, the aim is to understand which practices are the most prominent, meaning that by properly implementing them, it is possible to achieve great benefits by adopting the other practices of the same and the other paradigm.

48

M. Rossini et al.

Fig. 1. Cause and Effect Diagram

The analysis of the results and the implications of applying the DEMATEL approach to investigate the relationships between the Lean Supply Chain and Industry 4.0 practices are discussed here. First, the rankings of the practices according to the net effect and the prominence are presented; then, the complex network of relationships will be analyzed. Dispatcher practices show positive values of net effect (R–C). Hence they exert a big influence on others and are the Causes of the system. The ranking of the dispatchers, according to descending score of the net effect, resulted in being: Win-win problemsolving methodology (P12) > Cloud (P20) > Cybersecurity (P19) > Value chain analysis or value stream mapping (P13) > Augmented reality (P22) > Industrial Internet of Things (IoT) (P18) > Big data and analytics (P14) > Additive manufacturing (P21) > Simulation (P16). Properly adopting the dispatcher practices brings tangible benefits to applying the remaining practices. On the other hand, receiver practices are characterized by a negative net effect, and therefore they can be said to be Effects, meaning that other practices influence them. The receivers have been ordered based on the descending net effect. The final ranking is Kanban or pull system (P9) > Open-minded and in-depth market research conducted jointly (P4) > Horizontal and vertical integration (P17) > Establishment of distribution centres (P7) > Outbound transportation (P6) > Autonomous and collaborative robots (P15) > Functional packaging design (P8) > Levelled scheduling or heijunka (P10) > Consignment stock (P11) > Standardized work procedures to assure quality achievement (P3) > Inbound vehicle scheduling (P5) > Efficient and continuous replenishment (P1) > Material handling systems (P2). From this first classification, it is already possible to notice a different trend in positioning the practices according to the paradigm and bundle. The predominant importance of Big data and analytics (P14), Industrial IoT (P18) and Cloud (P20) is confirmed by Haddud and Khare (2019), who classified them as the most relevant enabling technologies of digital supply chain given the high number of areas that they can impact.

Lean Supply Chain and Industry 4.0: A Study of the Interaction

49

Besides the large group of Industry 4.0 practices, in the dispatchers, there are also two LSC practices belonging to the EWCI (Elimination of waste and continuous improvement) bundle: Win-win problem-solving methodology (P12) and Value chain analysis or value stream mapping (P13). These practices in the top positions of the ranking indicate that a structured problem-solving and waste identification methodology is required to guide systemic improvements along the supply chain and influence most of the other practices, both LSC and Industry 4.0. Complementary to this vision, Haddud and Khare (2019) suggest that the usage of advanced and smart Industry 4.0 technologies within supply chains supports the implementation of continuous improvement practices. The practices have also been ordered according to the descending degree of prominence to get an insight into which practices develop more relationships in both senses. The resulting rank is Kanban or pull system (P9) > Efficient and continuous replenishment (P1) > Levelled scheduling or heijunka (P10) > Industrial Internet of Things (IoT) (P18) > Big data and analytics (P14) > Material handling systems (P2) > Value chain analysis or value stream mapping (P13) > Establishment of distribution centres (P7) > Consignment stock (P11) > Simulation (P16) > Outbound transportation (P6) > Inbound vehicle scheduling (P5) > Horizontal and vertical integration (P17) > Functional packaging design (P8) > Autonomous and collaborative robots (P15) > Standardized work procedures to assure quality achievement (P3) > Cloud (P20) > Cybersecurity (P19) > Win-win problem-solving methodology (P12) > Augmented reality (P22) > Open-minded and in-depth market research conducted jointly (P4) > Additive manufacturing (P21). Differently from what emerged from the analysis of the net effect, for what concerns the prominence, it was not possible to find a trend in the behaviour of the different types of practices. In fact, LSC and Industry 4.0 practices are spread across the ranking, some presenting a higher combined sum of the exerted and received effects and others a lower one. In order to conclude this DEMATEL analysis, it is important to focus on the distinction between the dispatcher factors and the receiver ones but also on the relationships undertaken between practices of the same paradigm. In particular, as already mentioned, most of the Industry 4.0 practices are dispatchers. However, some of them – Additive manufacturing (P21) and Augmented reality (P22) – show low prominence levels. In the Causes group, two Lean Supply Chain practices belong to the EWCI bundle – Win-win problem-solving methodology (P12) and Value chain analysis or value stream mapping (P13) –while all the LOM practices are receivers and can be easily influenced. Therefore, according to all the considerations made, the order of prioritization of implementation of the most influencing practices according to the positive externalities that they can produce on the other paradigm should be Industrial IoT (P18), Big data and analytics (P14), Simulation (P16) and then Value chain analysis or value stream mapping (P13). On the other hand, it is worth highlighting that the LSC practices that can be more easily influenced through the application of the top-priority I4.0 technologies are Efficient and continuous replenishment (P1), Material handling systems (P2), Kanban or pull system (P9) and Levelled scheduling or heijunka (P10).

50

M. Rossini et al.

Furthermore, besides the interactions between the two paradigms that have been extensively analyzed, it was also observed that both LSC and Industry 4.0 practices develop complex networks of relationships among each other that have not been specifically examined because they are not the focus of this analysis. From the analysis conducted to answer this Research Question, it has not been possible to detect a clear univocal direction in the relationship between Lean Supply Chain practices and Industry 4.0 technologies. Indeed, while it was immediately clear that the practices of each paradigm interact a lot with each other, in order to foster the positive externalities that exist and have been identified between Lean Supply Chain and Industry 4.0, it is necessary to address the implementation of some specific practices of both paradigms.

6 Conclusions and Limitations In this paper, the authors have studied the interaction between Lean Supply Chain and Industry 4.0 by analyzing the relationships between their practices and technologies using the DEMATEL methodology. From this analysis, it was discovered that the practices of both paradigms interact a lot inside the same model. However, to foster the positive externalities between the Lean Supply Chain and Industry 4.0, some specific practices must be addressed before others to achieve the most from both paradigms. This paper analyses the interrelation between Lean Supply Chain and Industry 4.0 and enriches the literature in this sense. Indeed, currently, in literature, several papers focused on this topic are not based on structured methodologies such as DEMATEL. Therefore, this study represents the fundamental starting point for academicians who want to develop further advanced research in the Lean Supply Chain and Industry 4.0 environments. Besides this, considering the four research questions, the authors’ findings can be properly located in the steam of literature based on the mutually beneficial relationship between Lean Supply Chain and Industry 4.0. Confirming the studies conducted by Davies et al. (2017), Mrugalska and Wyrwicka (2017) and Uriarte et al. (2018), there is a twofold positive effect between the application of these two paradigms, and this is evident through the detailed analysis developed in the previous paragraph. Therefore, it is worth highlighting that the outcomes obtained through this paper contribute to increasing the knowledge regarding the interrelations between Lean Supply Chain and Industry 4.0. Accordingly, the authors have provided academicians with an analytical framework based on a robust procedure that can answer the direction of the relation between these two paradigms. From a managerial point of view, this study investigates the interactions between Lean Supply Chain and Industry 4.0, providing a twofold and complete point of view. On the one hand, it offers suggestions on the action plan to follow if a firm is already implementing the Lean paradigm along the supply chain and wants to begin a new project involving Industry 4.0; on the other hand, it represents a guideline for companies already adopting Industry 4.0 and seeking to begin to implement also some Lean concepts. Nevertheless, this paper is not exempt from limitations, as any other research paper. Firstly, given the many practices considered in the analyses developed, the results may

Lean Supply Chain and Industry 4.0: A Study of the Interaction

51

suffer from biases due to the computational limitations inherent to DEMATEL. Also, the Lean Supply Chain practices used are based on and selected from the literature. Thus there is a slight risk of not considering practices that may be relevant. The absence of model validation through SEM could represent a further limitation of the work because DEMATEL methodology heavily relies on experts’ opinions; however, having consulted a variegate cluster of experts, the authors have ensured great content validity. Thus, given the topic’s novelty, this paper could be considered a preliminary analysis for further advanced research regarding the relationships between Lean Supply Chain and Industry 4.0, for instance, using different methodologies such as surveys, SEM or case studies. Another possible future step can involve the application of the framework of analysis developed here on specific situations, i.e. considering a specific industry, company size or geographical location. Indeed, considering an SME or a large enterprise could bring different outcomes in adopting Lean Supply Chain and Industry 4.0 simultaneously.

References Anand, G., Kodali, R.: A conceptual framework for lean supply chain and its implementation. Int. J. Value Chain Manage. 2(3), 313–357 (2008) Arif-Uz-Zaman, K., Ahsan, A.: Lean supply chain performance measurement. Int. J. Product. Perform. Manag. 63(5), 588–612 (2014) Bleicher, J., Stanley, H.: Digitization as a catalyst for business model innovation a three-step approach to facilitating economic success. J. Bus. Manag. 62–71 (2016) Buer, S.V., Strandhagen, J.O., Chan, F.T.S.: The Link between Industry 4.0 and lean manufacturing: mapping current research and establishing a research Agenda. Int. J. Prod. Res. 56(8), 2924– 2940 (2018) Cudney, E., Elrod, C.: Incorporating lean concepts into supply chain management. Int. J. Six Sigma Compet. Adv. 6(1/2), 12–30 (2010) Dorst, W., et al. Umsetzungsstrategie Industrie 4.0. Ergebnisbericht der Plattform Industrie 4.0 (2015) Eisenhardt, K.M., Martin, J.A.: Dynamic capabilities: what are they?”. Strat. Manag. J. 21(10), 1105–1121 (2000) Haddud, A., Khare, A.: Digitalizing supply chain benefits and impact on lean operations. Int. J. Lean Six Sigma 11(4), 731–765 (2019) Hines, P., Holweg, M., Rich, N.: Learning to evolve: a review of contemporary lean thinking. Int. J. Operations Prod. Manage. 24 (10), 994e1011 (2004) Jasti, N.V.K., Kodali, R.: An empirical study for implementation of lean principles in Indian manufacturing industry. Benchmark. Int. J. 23(1), 183–207 (2016) Jean, R., Sinkovics, R., Hiebaum, T.: The effects of supplier involvement and knowledge protection on product innovation in customer–supplier relationships: a study of global automotive suppliers in China. J. Prod. Innov. Manag. 31(1), 98–113 (2014) Kohtamaki, M., Partanen, J.: Co-creating value from knowledge-intensive business services in manufacturing firms: the moderating role of relationship learning in supplier–customer interactions. J. Bus. Res. 69(7), 2498–2506 (2016) Kolberg, D., Zuhlke, D.: Lean Automation enabled by Industry 4.0 Technologies. IFACPapersOnLine 28(3), 1870– 1875 (2015) Kovács, O.: Az ipar 4.0 komplexitása - II. Közgazdasági Szle, vol. 64 (2017)

52

M. Rossini et al.

Lee, M., et al.: How to respond to the fourth industrial revolution, or the second information technology revolution? Dynamic new combinations between technology, market, and society through open innova- tion. J. Open Innov. 4, 1–24 (2018) Li, S., Ragu-Nathan, B., Ragu-Nathan, T., Rao, S.: The impact of supply chain management practices on competitive advantage and organizational performance. Omega 34(2), 107–124 (2006) Pinho, C., Mendes, L.: IT in Lean-based manufacturing industries: systematic literature review and research issues. Int. J. Prod. Res. 55(24), 7524–7540 (2017) Prem, E.: A digital transformation business model for innovation. In: ISPIM Innovation Summit (Brisbane) (2015) Sanders, A., Elangeswaran, C., Wulfsberg, J.: Industry 4.0 implies lean manufacturing: research activities in industry 4.0 function as enablers for lean manufacturing. J. Ind. Eng. Manag. 9, 811 (2016). https://doi.org/10.3926/jiem.1940 Sony, M.: Industry 4.0 and lean management: a proposed integration model and research propositions. Product. Manufac. Res. 6(1), 416–432 (2018) Spath, D., Ganschar, O., Gerlach, S., Hämmerle, M., Krause, T., Schlund, S.: Produktionsarbeit der Zukunft-Industrie 4.0. (Stuttgart) (2013) Tortorella, G., Miorando, R., Marodin, G.: Lean supply chain management: empirical research on practices, contexts and performance. Int. J. Product. Econ. (2017) Vanpoucke, E., Vereecke, A., Wetzels, M.: Developing supplier integration capabilities for sustainable competitive advantage: a dynamic capabilities approach. J. Oper. Manag. 32(7–8), 446–461 (2014) Vitasek, K., Manrodt, K., Abbott, J.: What makes a lean supply chain. Supply Chain Manage. Rev. 9(7), 39–45 (2005) Rossini, M., Powell, D.J., Kundu, K.: Lean supply chain management and Industry 4.0: a systematic literature review. Int. J. Lean Six Sigma 14(2), 253–276 (2023) Costa, F., Lispi, L., Staudacher, A.P., Rossini, M., Kundu, K., Cifone, F.D.: How to foster Sustainable Continuous Improvement: a cause-effect relations map of Lean soft practices. Oper. Res. Perspect. 6, 100091 (2019). https://doi.org/10.1016/j.orp.2018.100091 Reddy, S.K., Reinartz, W.: Digital transformation and value creation: sea change ahead. GfK Market. Intell. Rev. 9(1), 10–17 (2017) Ciano, M.P., Dallasega, P., Orzes, G., Rossi, T.: One-to-one relationships between Industry 4.0 technologies and Lean Production techniques: a multiple case study, Int. J. Product. Res. 59, 5, 1386–1410 (2010) Núñez-Merino, M., Maqueira-Marín, J.M., Moyano-Fuentes, J., Martínez-Jurado, P.J.: Information and digital technologies of Industry 4.0 and Lean supply chain management: a systematic literature review. Int. J. Product. Res. 58(16), 5034–5061 (2020). https://doi.org/10.1080/002 07543.2020.1743896 Yusuf, Y., Gunasekaran, A., Adeleye, E., Sivayoganathan, K.: Agile supply chain capabilities: determinants of competitive objectives. Eur. J. Oper. Res. 159, 379–392 (2004) Vokurka, R.J.: Supplier relationships: a case study. J. Enterprise Resource Manage. 2, 2ndquarter, 78–83 (2000) Stratton, R., Warburton, R.D.H.: The strategic integration of agile and lean supply. Int. J. Product. Econ. 85(2003), 183–198 (2003) Rother, M., Shook, J.: Learning to See: Value Stream Mapping to Create Value and Eliminate Muda, Lean Enterprise Institute (1999) Davies, R., Coole, T., Smith, A.: Review of Socio-technical Considerations to Ensure Successful Implementation of Industry 4.0. Procedia Manufacturing (2017) Mrugalska, B., Wyrwicka, M.K.: Towards lean production in Industry 4.0. Procedia Eng. 182, 466–473 (2017)

Lean Supply Chain and Industry 4.0: A Study of the Interaction

53

Uriarte, A.G., Ng, A.H., Moris, M.U.: Supporting the lean journey with simulation and optimization in the context of Industry 4.0. Procedia Manufac. 25, 586–593 (2018) Saxby, R., Cano-Kourouklis, M., Viza, E.: An initial assessment of Lean Management methods for Industry 4.0. The TQM J. 32(4) 587–601 (2020) Bittencourt, V.L., Alves, A.C., Leão, C.P.: Lean Thinking contributions for Industry 4.0 : a systematic literature review. Int. Federat. Autom. Control (2019) Green, K., Inman, R., Birou, L., Whitten, D.: Total JIT (T-JIT) and its impact on supply chaincompetency and organizational performance. Int. J. Prod. Econ. 147, 125–135 (2014) Shah, R., Ward, P.T.: Defining and developing measures of lean production. J. Oper. Manage. 25(4), 785–805 (2007) Arntzen, B.C., Brown, G.G., Harrison, T.P., Trafton, L.L.: Global supply chainmanagement at Digital Equipment Corporation. Interfaces 21(5), 69–93 (1995) Cagliano, R., Caniato, F., Spina, G.: The linkage between supply chain integration and manufacturing improvement programmes. Int. J. Oper. Prod. Manag. 26(3), 282–299 (2006)

A Design Science – Informed Process for Lean Warehousing Implementation Anna Corinna Cagliano(B) , Giovanni Zenezini , Carlo Rafele , Sabrina Grimaldi , and Giulio Mangano Department of Management and Production Engineering, Politecnico di Torino, Corso Duca Degli Abruzzi. 24, 10129 Torino, Italy {anna.cagliano,giovanni.zenezini,carlo.rafele,sabrina.grimaldi, giulio.mangano}@polito.it

Abstract. Warehouses usually include a number of non-value-added activities that increase costs, so they are a suitable field for applying Lean Thinking. In fact, Lean Warehousing has recently established as a promising research topic. However, most of the available contributions lack a practical though generalizable and theoretically relevant approach. Additionally, Lean is currently not enough explored as a preliminary effort to pave the way for a smooth application of Industry 4.0, neither in manufacturing nor in logistics. Inspired by Desing Science Research, the present work develops a structured process for implementing Lean Warehousing aimed at fostering continuous improvement to facilitate the introduction of Industry 4.0 technologies. To this end, multiple Lean tools, such as Value Stream Mapping, Spaghetti Chart, 5W+1H, and 5S analysis, are integrated. The application to a warehouse in the food industry is illustrated. Future research will focus on carrying out a complete validation. Moreover, performance measurement through a set of warehouse Key Performance Indicators will be introduced to assess both wastes and the associated improvements after implementing the appropriate corrective actions. Keywords: Lean Warehousing · Design Science · Implementation Strategy · Industry 4.0

1 Introduction Nowadays the growing complexity of supply chains (SCs), mainly due to heterogeneous operational tasks and multiple partners sometimes with very distant geographical locations, introduces several non-value-added activities that hinder the related performances [1]. In this context, Lean Thinking principles have played a key role to streamline and optimize entire SCs in the last decade [2, 3]. A “Lean SC” can be defined as a set of suppliers and customers linked by flows of products, services, technologies, information, and money, with the aim of reducing costs and waste as well as minimizing the non-value added activities and thus increasing value to SC agents [4]. From a structural © IFIP International Federation for Information Processing 2023 Published by Springer Nature Switzerland AG 2023 E. Alfnes et al. (Eds.): APMS 2023, IFIP AICT 689, pp. 54–68, 2023. https://doi.org/10.1007/978-3-031-43662-8_5

A Design Science – Informed Process for Lean Warehousing

55

point of view, suppliers and customers are connected by a network whose nodes are represented by manufacturing plants and warehouses. In particular, warehouses constitute key SC nodes, representing the second expenditure item in overall logistics costs after transportation. Additionally, warehouse operations largely contribute to determine the SC customer service level, economic performance, and business competitiveness [5]. However, within SC processes, warehousing is often perceived as the “weak link” comprising several non-value-added activities that drain costs and efforts from companies. In fact, warehousing might include useless material movements or excessive stocks. For such a reason, warehouses are an interesting subject to be studied by applying the Lean SC concept, in order to optimize the associated costs and time [6]. These issues, together with the awareness that warehouses cannot be completely eliminated from SCs [7], being a vital connection among their echelons, paved the way to Lean Warehousing [8]. Lean Warehousing has emerged in the last ten years as a very relevant topic in Lean Thinking [9, 10]. It applies Lean concepts and tools to traditional warehouse management with the aim of enhancing the quality of the logistics service by optimizing the use of the available resources and activities through the reduction, when not the elimination, of wastes [11]. In such a way, Lean Warehousing helps achieving efficiency, which is a more and more compelling need to streamline processes in order to prepare them for the application of Industry 4.0 technologies, such as for instance Advanced Automation and Virtual and Augmented Reality, to warehouses [12]. In fact, it was proven that Lean and Industry 4.0 technologies can mutually benefit each other. The Lean culture helps strengthening the effects brought by Industry 4.0 technologies, which in turn increase the effectiveness of Lean approaches [13]. However, several organizations are having a difficult time adopting Lean SC practices due to a scarce awareness and lack of proper implementation strategies [14]. The present work is based on Design Science Research (DSR) and proposes a structured process for applying Lean Warehousing in order to address and remove all those criticalities that might hinder the success of Industry 4.0 technologies prior to their implementation in a warehouse environment. DS can be currently considered a frontier in Operations Management (OM) research, still dominated by the explanation approach, rather than the exploration, problem solving-oriented one [15, 16]. Grounded in Herbert Simon’s work [17], DSR develops a solution to a practical problem, named “artifact”, based on a specific practice setting. After that, the findings are generalized in order to prove the theoretical relevance of the solution. The goal of the contribution is twofold. First, it promotes a structured integration of several Lean tools to create the necessary logistics process settings and mindset to support a smooth implementation of Industry 4.0 in warehouses. Second, it aims at providing an application of DSR to warehouse management, one OM field where this approach is still pretty much unconsidered. In this way, the present work contributes to develop a culture of finding solutions to problems rather than only focusing on the explanation of them in OM. The developed Lean Warehousing process is discussed following the main DSR phases.

56

A. C. Cagliano et al.

2 Research Background 2.1 Warehouse Management and Lean Warehousing Warehouse management is a stream of research that has been long debated in literature from different perspectives, such as warehouse physical design, including material handling and storage systems, inventory management, storage allocation policies, product tracking, and performance measurement [18]. Most of the works are focused on meeting the main warehouse challenges in terms of enhanced productivity, a decreased number of defects, improved customer satisfaction, as well as stock and operational costs reduction. This can be summarized in just one phrase: “a need for efficiency” [19]. Therefore, in order to accomplish such a goal, it is quite straightforward to try to apply Lean Thinking to warehouse management, which have originated the Lean Warehousing topic [8]. A significant number of authors have developed literature contributions about Lean Warehousing so far. Among the most recent ones, a number of them provide classifications of warehouse wastes based on the Ohno’s taxonomy [20] and the main warehouse activities, namely receiving, put away, picking, and dispatching activities [2, 9, 11, 21]. Other authors implement specific Lean tools, either in a single or in a combined way. Value Stream Mapping (VSM) is by far the most commonly applied tool. [9] relies on VSM for developing solutions to reduce a number of warehouse wastes and assessing the associated benefits through lead time assessment. [22] combine VSM with Spaghetti Chart and Multi-Criteria ABC Analysis to address the material handling problems derived from a poor warehouse layout design. [23] use again VSM and MultiCriteria ABC Analysis, but together with 5S, Gemba, and Work Standardization tools, to improve inventory management so that to increase the delivery compliance indicator of a warehouse in the metalwork sector. Recently, [24] combine VSM with Discrete Event Simulation for the case of a warehouse in the pharmaceutical sector. Finally, another recent attempt to undertake a wide perspective on Lean Warehousing has been carried out by [25], who select the Lean Six Sigma approach to focus on all the warehouse processes of a third party logistics with the ultimate goal to increase productivity. Lean Warehousing works usually lack a perspective on the interconnection between Lean Thinking and Industry 4.0, although it has been the subject of a significant number of recent contributions in the manufacturing field (e.g. [26]). In particular, as highlighted by [27, 28], most of the research efforts have been spent on how digitalization, underpinning the Industry 4.0 notion, can support the application of Lean but how Lean principles can facilitate the introduction of Industry 4.0 is a topic that still deserve further attention. This is the debate the present research aims to contribute to through a lens which is still quite rare in OM research in general, and in warehouse management in particular, namely DSR [16]. The motivation for choosing such an approach is that companies need detailed and structured methods that can be implemented to streamline warehouse processes, where many wastes can be nested, and to be able to get full benefits from the application of Industry 4.0 technologies. 2.2 The Design Science Approach The DS approach lies at the intersection between problem solving and explanatory science to explore, create, and test solutions (the so called “artifacts”) to problems with

A Design Science – Informed Process for Lean Warehousing

57

a practical relevance [16]. It is not opposed to explanatory research but it complements it. In fact, its four steps are as follows, where the first two ones pertain to exploratory research while the last two ones are part of explanatory research [29]: • Solution incubation: the problem to be tackled is framed and a first potential solution is designed. • Solution refinement: the solution is subjected to empirical testing, usually in one real setting, and refined. • Explanation I – substantive theory: the theoretical relevance of the solution is established in order to understand to what research discourse it contributes to. It usually involves introducing the solution in multiple settings. A DSR effort may stop at this step. • Explanation II – formal theory: strengthening the theoretical and statistical generalizability of the solution to derive a formal theory from it. By following the present methodological approach, the next sections develop a Lean Warehousing process that can be implemented by organizations to support them to improve their logistics processes and eliminate wastes prior to the adoption of Industry 4.0 technologies. The Explanation I phase is performed by discussing the theoretical relevance of the proposed Lean process based on its comparison with literature evidence. The Explanation II phase has not been carried out at the moment.

3 Solution Incubation: The Lean Warehousing Process The first DSR step is understanding the problem to solve. This is here drawn from a schematic summary of the outcomes of the literature review discussed in Sect. 2.1: • Many authors have developed Lean Warehousing approaches to achieve efficiency. Several of them rely on an integrated application of multiple Lean tools. However, they are intended to meet several goals, such as for instance reducing congested warehouse routes or increasing the customer order fulfillment rate, but none of them is specifically aimed at preparing warehouse processes to Industry 4.0 implementation. • Practical approaches are needed providing companies with guidelines about how to apply Lean Thinking to pave the way for Industry 4.0. Thus, the problem can be posed as: how could a detailed Lean Warehousing process be structured in order to address and eliminate wastes in a systematic way as a preliminary step towards Industry 4.0? The Lean Warehousing process presented in this contribution, thus the “artifact” produced in the DSR effort, has been developed starting from the preliminary works of [30] and [31]. It consists of three interconnected steps that form a closed loop coherent with a continuous improvement effort (Fig. 1): 1. Process Mapping. 2. Waste Identification and Analysis with 5W + 1H. 3. Improvement Action Definition with 5S and Associated Assessment.

58

A. C. Cagliano et al.

Fig. 1. The proposed Lean Warehousing Process

3.1 Step 1: Process Mapping The first step of the Lean Warehousing Process is devoted to map the current situation of warehouse processes with the aim of identifying and assessing the existing wastes. This is performed through two selected Lean tools, namely the VSM and the Spaghetti Chart [32, 33]. Once the wastes have been identified, their current effects on warehouse performances are assessed by KPIs selected according to the nature of the affected activities and the related kinds of inefficiencies as highlighted by the VSM and the Spaghetti Chart. The values of these measures (i.e. time, distance KPIs, etc.) will be compared with the ones after the application of improvement actions (Sect. 3.3) to estimate the benefits provided by the process implementation. 3.2 Step 2: Waste Identification and Analysis with 5W+1H In the second step of the Lean Warehousing process the existing warehouse situation is analyzed, focusing on identifying the causes for wastes. Here the wastes in warehouse operations are investigated and classified and such an activity has to be carried out after an accurate process mapping. The causes for wastes are assessed via the 5W+1H tool [34]. This method allows to give an exhaustive overview of the warehouse wastes by answering the following questions: • • • •

What happens (What)? What is the source of the waste (Where)? Who is the person in charge (Who)? When in the process does the waste occur (When)?

A Design Science – Informed Process for Lean Warehousing

59

• What are the reasons for the waste occurrence (Why)? • What suggestions for improvements are needed to be done (How)? Inputs for the 5W+1H analysis stem from the VSM and the Spaghetti Chart. Hence, each warehouse waste is associated with its type according to the Ohno’s classification [20] in the “What” section of the 5W+1H tool, together with a quantification of the waste impact via a suitable KPI. Furthermore, the VSM tool supports the framing of the warehouse processes, which in turn highlights the locations where the wastes occur, both in terms of spatial location (“Where) and time-wise (“When”). Finally, the VSM identifies the responsible for each process, hence outlining the “Who” of the wastes. After process mapping and the analysis of the first 4Ws, there are enough informational data to gather insights and proceed onto framing the underlying reasons for waste occurrence (“Why”) and finally set up a response for waste reduction and elimination (“How”). This last step of the 5W+1H is often times left to the creativity of the warehouse management. The authors propose a more structured approach for this step in the following sub-section. 3.3 Step 3: Improvement Actions Definition with 5S and Associated Assessment The proposed Lean Warehousing process completes the “How” phase of the 5W+1H analysis by relying on a fourth Lean tool, namely the 5S analysis. It unfolds through five steps, Sort, Straitening, Shining, Standardizing, and Sustaining, that not necessarily are applied in a consecutive way [23]. Such steps suggest the main intervention areas to be addressed to reduce and control waste. Thus, they offer guidelines for formulating improvement actions and help defining standard processes and ensuring continuous improvement according to the Kaizen philosophy. Only selected “Ss” can be considered according to the nature of processes and wastes. The effectiveness of the improvement actions should then be monitored and, to this end, the present Lean Warehousing process suggests again undertaking the Process Mapping step and updating the VSM and the Spaghetti Chart after their implementation. On the one hand this will support assessing the benefits introduced by the process application. On the other end, a new analysis of warehouse processes will allow uncovering further waste that should be approached.

4 Lean Warehousing Process Implementation and Refinement The proposed Lean Warehousing process has been refined by applying it to a company operating a chain of grocery stores in North Eastern Italy and seeking to prepare its logistics processes to Industry 4.0 implementation. This company employs around 1,700 workers and sells products via a network of 69 shops. The authors took part in a 6-months process of Lean implementation in the largest warehouse of the company, working as part of a Project Team including the warehouse managers and workers and continuously receiving feedbacks from them in an iterative process. 4.1 AS IS Situation and Waste Analysis The authors have focused on non-perishable and fresh processed food products. Fresh groceries have not been considered because they are just cross-docked to the shop within

60

A. C. Cagliano et al.

few hours from the arrival at the docking bay. As part of Step 1 of the process, Fig. 2 shows the current VSM for non-perishable products including the main warehouse processes. In order to draw this state map the notation by [35] was adopted and several steps were undertaken, namely: • The stakeholders were identified and assigned to the different tasks and responsibilities. Three stakeholders take part in the warehousing process, namely the Supplier, the Buyer (i.e. the shop) and the Purchasing Department. • The ordered quantities and the daily shipments are retrieved from the company. Data showed that these two quantities are almost equal, as the percentage of backlog orders is very close to non-existing and can be neglected. • Warehouse processes are identified and are differentiated between processes entailing continuous product flows and processes which contain an interruption of the product flow. Interruption of the flows are associated with accumulation of stock, which in turn is one of the most relevant waste of a warehouse. • Data are compiled for each process. In particular, the cycle time and the number of employees are associated with each warehouse process. • Each warehouse process is associated with its guiding principle. Processes from orders receiving to stowing follow a “push” principle guided from the Purchasing Department, while processes from picking to shipping are driven by the customers’ orders and are thus associated with a “pull” principle. • Finally, each process is timed differentiating between value added and non valueadded processes. The overall warehouse lead time is calculated as the summation of all the warehouse processes. The time to perform the picking and packaging tasks is remarkably higher than the time needed to carry out the other warehouse activities because one roll pallet, delivered to a given shop, is constituted by heterogeneous products in very different quantities. Hence, three cycle times are calculated through the current state map analysis: i) time for value-added processes; ii) time for non-value-added processes; iii) total warehouse lead time. For non-perishable products they are equal to 66 min and 45 s, 255 min, and 321 min and 45 s respectively. For fresh processed products these time values are 66 min and 45 s, 235 min, and 301 min and 45 s. Such values have been computed by the Project Team through direct observation of activities and recording of the associated time. The non-value-added time does not include the average time products remain stocked in the warehouse. This comes from specific company policies and was out of the scope of the present application. A Spaghetti Chart of the general movements of warehouse workers has been drafted by observing daily operations across different days of the week. The Spaghetti Chart cannot be disclosed here due to the confidentiality of the complete warehouse layout. The Spaghetti Chart tool allows for calculating the minimum and maximum distance travelled by the warehouse personnel for the different warehouse activities, as shown in Table 1. Furthermore, the Spaghetti Chart analysis reveals that there are some areas of the warehouse where multiple paths of the workers overlap, creating potential wastes. In particular, flows overlap where goods are stored waiting to be shipped and at the entrance of the controlled temperature section of the warehouse where perishable products are

A Design Science – Informed Process for Lean Warehousing

61

Fig. 2. Current State VSM for non-perishables Table 1. Estimation of travelled distances via the AS IS Spaghetti Chart. Warehouse Activity

Minimum Distance [m]

Maximum Distance [m]

Unloading Non-Perishable Products

5

20

Unloading Fresh Products

17

45

Stocking Non-Perishable Products

35

180

Stocking Fresh Products

10

70

Picking Non-Perishable Products

–

607

Picking Fresh Products

–

595

Shipping Non-Perishable Products

40

83

Shipping Fresh Products

32

67

Truck Loading Of All Products

5

20

stored. The identification of overlapping flows enables to identify a first layer of warehouse wastes. This type of wastes can only be overcome through a reengineering of the warehouse layout. Both Spaghetti Chart and direct observation were used as methods to identify wastes and assess their causes in Step 2 of the process. The wastes are as follows: • • • • •

Long Waiting Time for Truck to Access Receiving Docks Long Waiting Time of Incoming Pallets before Storage Poor Attention to Storage Priority Long Waiting Time of Outgoing Pallets before Shipping Incorrect Positioning of Outgoing Pallets

62

A. C. Cagliano et al.

A sixth type of waste was identified in the damages suffered by the products being handled, and was found to be caused by the overlapping of workers’ movements in the controlled temperature area detected through the Spaghetti Chart. This waste was however deemed to not be relevant due to its negligible impact on the company’s performance. Table 2 shows an excerpt from the performed 5W+1H analysis focusing on the waste “Long Waiting Time for Truck to Access Receiving Docks”. The complete analysis is available from the authors upon request. It is worth reminding the reader that the HOW phase of the analysis will be performed in Step 3 of the process. Table 2. Excerpt from 5W analysis of warehouse wastes. Waste

5W step

Long Waiting Time for Truck to Access WHAT Receiving Docks

WHEN

Description Waste Type: Waiting Description: Trucks wait before connecting to an available unloading bay KPI: Average truck waiting time = 1.5 h Warehouse process: Receiving Warehouse sub-process: Prior to the unloading process

WHERE In the loading/unloading yard WHY

Insufficient number of unloading bays and workers; Lack of scheduling for truck arrival

WHO

Purchasing Department

4.2 Waste Reduction - Identification of Improvement Actions with 5S The aim of Step 3 of the Lean Warehousing process is defining improvement actions to reduce waste based on the outcomes of the analysis of the associated causes carried out through the 5W+1H methodology. To this end, the 5S approach was applied to the wastes identified for the case warehouse through several brainstorming sessions involving the Project Team. Table 3 presents an excerpt from the outcomes of such a process: only the relevant 5S steps are mentioned. Different integrated actions are suggested for each waste. The complete analysis is available from the authors upon request. An ex-ante assessment has been then carried out to evaluate the proposed improvement actions in terms of time and physical flow savings. As an example of the adopted approach, the following analysis is focused on the three wastes that most impact the nonvalue-added time, namely Long Waiting Time for Truck to Access Receiving Docks, Long Waiting Time of Incoming Pallets before Storage, and Long Waiting Time of Outgoing Pallets before Shipping. An online supplier delivery booking system has been proposed in order to reduce the waiting time for an available receiving dock. The Project Team estimated that up

A Design Science – Informed Process for Lean Warehousing

63

Table 3. Excerpt from 5S analysis of warehouse wastes. Waste

5S step

Improvement action

Long Waiting Time for Truck to Access Receiving Docks

Sort

Moving the outgoing pallets placed in the receiving area, so as not to occupy the receiving docks

Set in Order

Revising the schedule of the incoming trucks and making it coherent with the shipping schedule

Standardize

Implementing an online system where suppliers can book their delivery time within predefined time slots. This is completed by a GPS tracking system to detect any truck delay

Sustain

Disseminating the new procedures and checking their application periodically

to 5 truck arrivals can be scheduled every 20 min because the warehouse is equipped with 5 receiving docks and 20 min is the average unloading time computed based on a direct observation of this operation. Such a solution allows to reduce the average truck waiting time from 90 min to 20 min in case of delays in unloading the previous truck at a same dock. However, the waiting time might be even shorter thanks to a GPS tracking system that enables addressing delays by rescheduling the next arrivals. Coming to the goods waiting time before storage, increasing the number of workers assigned to storing activities is the identified solution. The associated time savings cannot be calculated in an ex-ante assessment, thus they were estimated through expert judgment. In particular, according to the Warehouse Manager, the waiting time after implementing this improvement action can be reduced from 90 to 30 min per each single pallet. Finally, the solutions to reduce the waiting time of pallets before being shipped to shops mainly rely on placing the goods close to the dock area shortly before the truck arrival. The time savings were estimated to be equal to 20 min and 30 min for fresh processed and non-perishable products respectively. It is very important to be able to reduce the waiting time of fresh products since, being perishable goods, they stop outside the cold room during this time and the cold chain is interrupted. The time savings introduced by the discussed improvement actions originate important reductions in the overall warehouse lead time and this can be highlighted by the Future State VSM. As already done for the current state analysis, two VSMs are developed for both fresh processed and non-perishable products. Table 4 summarizes the expected value-added time, non-value-added time, and total warehouse lead time for each single pallet, together with the percentage decrease compared with the current values. Future state VSMs are available from the authors upon request. In order to complete the analysis of the future warehouse situation, an ex-ante assessment of the staff movements after the potential application of the proposed improvement

64

A. C. Cagliano et al. Table 4. TO BE time and savings.

Fresh Processed food TO BE

DELTA

Value-added time

66 min and 45 s

0%

Non-value-added time

85 min

−41.4%

Warehouse lead time

151 min and 45 s

−28.4%

Non-perishables TO BE

DELTA

Value-added time

66 min and 45 s

0%

Non-value-added time

95 min

−42.4%

Warehouse lead time

161 min and 45 s

−30.3%

actions has been carried out and a Future State Spaghetti Chart drawn. In particular, moving outgoing pallets from the receiving area to the shipping one (Incorrect Positioning of Outgoing Pallets waste) allows to remove the overlaps in the staff flows that occurred in the area where goods are stored waiting to be shipped. Correctly placing outgoing pallets in the shipping area also originates relevant changes in the distances traveled by staff, as revealed by the quantitative evaluation of physical flows in the future warehouse situation. To be more precise, the minimum distance travelled to place outgoing fresh processed products increases from 40 m in the current situation to 60 m in the TO BE one, while the maximum distance to place non-perishable products decreases from 67 m to 45 m. However a significant flow streamline can be gained. All the other physical flows do not show any significant variations from the current state to future state configuration. Also the future state Spaghetti Chart cannot be disclosed for confidentiality reasons.

5 Relevance of the Proposed Lean Warehousing Process As the third step of a DSR approach, the findings should be generalized and the theoretical contribution demonstrated. Currently, such a phase has not been completed yet, thus the relevance of the proposed Lean Warehousing process is here discussed based on literature evidence. First, the relevance lies in the underlying research methodology. Through the development of the Lean Warehousing process the authors offer an application of DSR in OM and in particular in warehouse management. If DSR in OM is still in its infancy [36, 37], it is even more in the field of warehousing, where, to the best authors’ knowledge, no works exist. However, DSR could be highly beneficial in the logistics field, whose problems have practical roots and might find viable and effective solutions through a research approach that is centered around the real needs of professional practice. Adopting DSR in Lean Warehousing is also a way of introducing practical but also generalizable approaches for the application of Lean Thinking to warehouse management. In fact, the

A Design Science – Informed Process for Lean Warehousing

65

available literature either offers theoretical guidelines for the implementation of Lean in warehouses and the associated performance assessment [11] or focuses on case studies driven by the particular situations of specific organizations [7, 22]. Second, the authors think the relevance of the present research is also related to suggesting a detailed roadmap about how to streamline warehouse and logistics process to make them ready to get full benefits from Industry 4.0 technologies. Thus, it helps closing the research gap on the role of Lean as an endeavor that should start before the application of digital technologies [13, 28] to then continue after it to always ensure an appropriate operational environment.

6 Implications and Conclusions This study offers both theoretical and practical implications. From an academic point of view, as discussed in Sect. 5, it not only contributes to advance the state of the art about Lean Warehousing but also promotes research on integrating it with DSR. The presented process and its discussed application provide researchers with guidelines about structuring Lean Warehousing approaches. In this way, they foster deepening the knowledge of how Lean principles can positively affect not only production but also logistics flows, with the aim of preparing them to Industry 4.0 application. From a practical perspective, both manufacturing and logistics companies can benefit from the Lean Warehousing process. First of all, it might constitute for managers a ready-to-go approach to assess and improve their warehouse processes. The wastes discussed in the implementation help organizations to identify their own warehouse wastes by suggesting those activities that might be investigated. Then, the Lean improvement actions could offer companies a chance to leverage the implementation of Industry 4.0 technologies in logistics. Therefore, the Lean Warehousing process can become a means to approach the so-called Logistics 4.0 in an informed manner and not just by following a trend. In a nutshell, relying on structured methods to seek continuous improvement in warehouses is of paramount importance for their managers. In fact, as a front end towards the next SC echelon, warehouses need to always guarantee responsiveness by trying to “absorb” those delays and inefficiencies originating upstream in the chain. However, so far the proposed Lean Warehousing process has undergone just one single application. Moreover, the devised improvement actions were not implemented and the related benefits were estimated according to the company experts’ knowledge and experience. Future research efforts will be directed towards testing the process in multiple industries and warehouse settings in order to completely validate and refine it. Particular attention will be given to the actual implementation of the proposed actions to either limit or remove wastes, with the aim of fully capturing the effectiveness of the approach. The possible integration of additional Lean tools in the process will be explored as well, together with the introduction of a catalogue of warehouse KPIs to measure both AS IS wastes and their reduction after the process application.

66

A. C. Cagliano et al.

References 1. Jasti, N.V.K., Kurra, S.: An empirical investigation on lean supply chain management frameworks in Indian manufacturing industry. Int. J. Product. Perform. Manag. 66, 699–723 (2017) 2. Buonamico, N., Muller, L., Camargo, M.: A new fuzzy logic-based metric to measure lean warehousing performance. In: Supply Chain Forum: An International Journal. pp. 96–111. Taylor & Francis (2017) 3. Arif-Uz-Zaman, K., Nazmul Ahsan, A.M.M.: Lean supply chain performance measurement. Int. J. Product. Perform. Manag. 63, 588–612 (2014) 4. Librelato, T.P., Lacerda, D.P., Rodrigues, L.H., Veit, D.R.: A process improvement approach based on the value stream mapping and the theory of constraints thinking process. Bus. Process Manag. J. 20, 922–949 (2014). https://doi.org/10.1108/BPMJ-07-2013-0098 5. Kembro, J.H., Norrman, A., Eriksson, E.: Adapting warehouse operations and design to omnichannel logistics: a literature review and research agenda. Int. J. Phys. Distrib. Logist. Manag. 48, 890–912 (2018). https://doi.org/10.1108/IJPDLM-01-2017-0052 6. Dotoli, M., Epicoco, N., Falagario, M., Costantino, N., Turchiano, B.: An integrated approach for warehouse analysis and optimization: a case study. Comput. Ind. 70, 56–69 (2015) 7. Shah, B., Khanzode, V.: Designing a lean storage allocation policy for non-uniform unit loads in a forward-reserve model: an enterprise information management with an e-decision support system. J. Enterp. Inf. Manag. 31, 112–145 (2018). https://doi.org/10.1108/JEIM-01-20170018 8. Sharma, S., Shah, B.: Towards lean warehouse: transformation and assessment using RTD and ANP. Int. J. Product. Perform. Manag. 65, 571–599 (2016) 9. Abhishek, P.G., Pratap, M.: Achieving lean warehousing through value stream mapping. South Asian J. Bus. Manag. Cases. 9, 387–401 (2020). https://doi.org/10.1177/2277977920958551 10. Raghuram, P., Arjunan, M.K.: Design framework for a lean warehouse – a case study-based approach. Int. J. Product. Perform. Manag. 71, 2410–2431 (2022). https://doi.org/10.1108/ IJPPM-12-2020-0668 11. Abushaikha, I., Salhieh, L., Towers, N.: Improving distribution and business performance through lean warehousing. Int. J. Retail Distrib. Manag. 46, 780–800 (2018) 12. Mahdavisharif, M., Cagliano, A.C., Rafele, C.: Investigating the integration of Industry 4.0 and lean principles on supply chain: a multi-perspective systematic literature review. Appl. Sci. 12 (2022). https://doi.org/10.3390/app12020586 13. Frecassetti, S., Ferrazzi, M., Portioli-Staudacher, A.: Lean as a Facilitator for AGVs Implementation: A Case Study (2023). https://doi.org/10.1007/978-3-031-25741-4_10 14. Tortorella, G.L., Miorando, R., Tlapa, D.: Implementation of lean supply chain: an empirical research on the effect of context. TQM J. 29, 610–623 (2017). https://doi.org/10.1108/TQM11-2016-0102 15. van Aken, J., Chandrasekaran, A., Halman, J.: Conducting and publishing design science research: inaugural essay of the design science department of the Journal of Operations Management. J. Oper. Manag. 47–48, 1–8 (2016). https://doi.org/10.1016/j.jom.2016.06.004 16. Romme, A.G.L., Holmström, J.: From theories to tools: calling for research on technological innovation informed by design science. Technovation. 121 (2023). https://doi.org/10.1016/j. technovation.2023.102692 17. Simon, H.A.: The Sciences of the Artificial. 3rd (1996) 18. Ramirez, C., Neuenfeldt, A.L., Siluk, J.C., de Ataíde, L.: A literature overview about warehouse management. Int. J. Logist. Syst. Manag. 42, 153–175 (2022). https://doi.org/10.1504/ IJLSM.2022.124184

A Design Science – Informed Process for Lean Warehousing

67

19. Kolinski, A., Sliwczynski, B.: Evaluation problem and assessment method of warehouse process efficiency. In: Proceedings of the 15th International Scientific Conference Business Logistics in Modern Management, October 15, 2015, pp. 175–188. Osijek, Croatia (2015) 20. Ohno, T.: Toyota Production System: Beyond Large-Scale Production. Productivity Press, New York (1988) 21. Salhieh, L., Altarazi, S., Abushaikha, I.: Quantifying and ranking the “7-Deadly” Wastes in a warehouse environment. TQM J. 31, 94–115 (2019) 22. Aksüto˘glu, A., Bıkmaz, E., Sırkıntı, N.E., Pala, M.Ö., Ekinci, E.: Lean Warehouse Application in a Meat Producer Company (2021). https://doi.org/10.1007/978-3-030-62784-3_76 23. Anton, A., Roque, M., Ramos, E.: Lean warehousing model to improve the order fulfillment in a metalworking company. In: ACM International Conference Proceeding Series, pp. 90–95 (2022). https://doi.org/10.1145/3568834.3568881 24. Abideen, A., Mohamad, F.B.: Improving the performance of a Malaysian pharmaceutical warehouse supply chain by integrating value stream mapping and discrete event simulation. J. Model. Manag. 16, 70–102 (2021). https://doi.org/10.1108/JM2-07-2019-0159 25. Adeodu, A., Maladzhi, R., Kana-Kana Katumba, M.G., Daniyan, I.: Development of an improvement framework for warehouse processes using lean six sigma (DMAIC) approach. A case of third party logistics (3PL) services. Heliyon. 9 (2023). https://doi.org/10.1016/j.hel iyon.2023.e14915 26. Rossini, M., Costa, F., Tortorella, G.L., Portioli-Staudacher, A.: The interrelation between Industry 4.0 and lean production: an empirical study on European manufacturers. Int. J. Adv. Manuf. Technol. 102, 3963–3976 (2019). https://doi.org/10.1007/s00170-019-03441-7 27. Buer, S.-V., Strandhagen, J.O., Chan, F.T.S.: The link between industry 4.0 and lean manufacturing: mapping current research and establishing a research agenda. Int. J. Prod. Res. 56, 2924–2940 (2018). https://doi.org/10.1080/00207543.2018.1442945 28. Núñez-Merino, M., Maqueira-Marín, J.M., Moyano-Fuentes, J., Martínez-Jurado, P.J.: Information and digital technologies of Industry 4.0 and Lean supply chain management: a systematic literature review. Int. J. Prod. Res. 58, 5034–5061 (2020). https://doi.org/10.1080/002 07543.2020.1743896 29. Holmström, J., Ketokivi, M., Hameri, A.-P.: Bridging practice and theory: a design science approach. Decis. Sci. 40, 65–87 (2009). https://doi.org/10.1111/j.1540-5915.2008.00221.x 30. Mustafa, M.S., Cagliano, A.C., Rafele, C.: A proposed framework for lean warehousing. Pioneer. Solut. Supply Chain Perform. Manag. Concepts, Technol. Appl. 137–149 (2013) 31. Cagliano, A.C., Grimaldi, S., Schenone, M.: Proposing a new framework for lean warehousing: first experimental validations. In: Proceedings of the Summer School Francesco Turco (2018) 32. Bait, S., Di Pietro, A., Schiraldi, M.M.: Waste reduction in production processes through simulation and VSM. Sustainability. 12, 3291 (2020) 33. Al-Zubaidi, S.Q.D., Coli, E., Fantoni, G.: Automating production process data acquisition towards spaghetti chart 4.0. Int. J. Ind. Eng. Manag. 13, 145–157 (2022). https://doi.org/10. 24867/IJIEM-2022-3-308 34. Petrillo, A., De Felice, F., Zomparelli, F.: Performance measurement for world-class manufacturing: a model for the Italian automotive industry. Total Qual. Manag. Bus. Excell. 30, 908–935 (2019) 35. Chen, J.C., Cheng, C.-H., Huang, P.B., Wang, K.-J., Huang, C.-J., Ting, T.-C.: Warehouse management with lean and RFID application: a case study. Int. J. Adv. Manuf. Technol. 69, 531–542 (2013)

68

A. C. Cagliano et al.

36. Dresch, A., Lacerda, D.P., Cauchick-Miguel, P.A.: Design science in operations management: conceptual foundations and literature analysis. Brazilian J. Oper. Prod. Manag. 16, 333–346 (2019). https://doi.org/10.14488/BJOPM.2019.v16.n2.a13 37. Su, S.-L.I., Fan, X., Shou, Y.: A design science-based case study of retail chain delivery operations and its implications. Int. J. Phys. Distrib. Logist. Manag. 51, 212–235 (2021). https://doi.org/10.1108/IJPDLM-11-2019-0352

Digitally Enhancing Kanban Lean Practice in Support of Just-in-Time Reconfigurable Supply: A Case Study Christina Papadimitropoulou1 , Anne Zouggar Amrani1(B) Helena Macedo3 , and David Romero4

, Daryl Powell2

,

1 University of Bordeaux, CNRS, IMS, UMR 5218, Talence, France

[email protected], [email protected] 2 SINTEF Manufacturing, Trondheim, Norway [email protected] 3 University of Minho, Braga, Portugal [email protected] 4 Tecnológico de Monterrey, Mexico City, Mexico

Abstract. This paper explores the benefits of digitally enhancing Kanban Lean practice in support of Just-in-Time (JIT) reconfigurable supply processes for increased efficiency and reliability. Ensuring JIT-supply of raw materials and parts is essential to achieve Lean production and reduce inventory storage and capital tied up in unprocessed inventory. After a literature positioning, this research work highlights the importance for the industry, more than ever before, to build reliable and reconfigurable supply chains to easier the absorption of demand diversification while using different raw materials and parts at the beginning of their production flows. Kanban, as a solid JIT approach with visual signaling, is still a suitable Lean lever to activate given demand diversification, however, some limits are observed when using physical Kanban cards in a mass-customization context and the Industry 4.0 era. Hence, this work aims to analyze the usefulness of Digital Kanban Systems (DKSs) in support of demand diversification absorption that will increase the reliability of production lines and supply chains. Keywords: Digitalization · Lean · Digital Lean Manufacturing · Kanban Systems · Supply Chain · BLE · RFID · IoT · Case Study

1 Introduction The successful adoption of the Just-in-Time (JIT ) philosophy [1] in modern manufacturing systems and supply chains still presents many obstacles and is requiring better frameworks and step-by-step methods to help Lean production and supply chain managers in its fruitful implementation at each activity of a process value chain; from raw materials sourcing to the manufacturing of goods to their quality assurance and final © IFIP International Federation for Information Processing 2023 Published by Springer Nature Switzerland AG 2023 E. Alfnes et al. (Eds.): APMS 2023, IFIP AICT 689, pp. 69–83, 2023. https://doi.org/10.1007/978-3-031-43662-8_6

70

C. Papadimitropoulou et al.

delivery logistics to the customer. Furthermore, with the arrival of (new) smart manufacturing and logistics technologies [2], these have made a step up in the continuous effort of optimizing production and supply chain efficiency. The possibilities offered by the ongoing digitalization of key business processes such as production and supply chain management for complying with the Lean principles have been reinforced over the last decade and will continue to increase in the foreseeable future as highlighted by the emerging Digital Lean Manufacturing paradigm [3]. This research work analyzes the critical points that ought to be considered for the successful adoption of the JIT philosophy [1] leveraged by new Digital Lean tools [3] in the Industry 4.0 era addressing possible constraints and opportunities. With this aim in mind, Digital Kanban Systems (DKSs) are the new Internet of Things (IoT) solutions supporting the JIT supply of raw materials and parts in production lines and supply chains for more reliable and efficient operations [3]. Jidoka and JIT are the two main pillars of Lean Manufacturing principles. The objective of this research work is to propose a “Kanban-based digital solution” that enhances supply processes and increases the reliability of production lines and supply chains. Additionally, the digitalization of JIT philosophy enables the effective integration of Industry 4.0 technologies in Lean Manufacturing systems. Even though there is a large number of research papers in the scientific literature that focus on JIT, Pull, and Kanban systems, there is a noticeable gap in the digitalization of Kanban systems and their corresponding case studies. Through this applied research work, it was possible to provide an overview of DKS solutions by analyzing two illustrative examples from the automotive industry to introduce the reader to a decision-aided approach to selecting the most appropriate ones in each case. This paper is organized as follows: Sect. 2 presents a brief literature review on the Digital Lean Manufacturing paradigm and the historical evolution of the Digital Enhancements of Kanban systems; Sect. 3 presents the methodology followed for this research work; Sect. 4 is dedicated to the “Supply Reliability” challenge in supply chains and the valuable role of DKSs to address it; and Sect. 5 browses the modelling selection of DKS solutions. In Sect. 6, two different case studies will be analyzed along with the benefits obtained from a DKS adoption, and Sect. 7 will present conclusions and further work.

2 Literature Review Lean Manufacturing (LM) refers to a set of management techniques and principles designed to eliminate waste and simplify activities that add value to products from a customer perspective [4]. LM has been implemented by a high number of enterprises around the world since it is proven to provide many benefits in support of operational excellence. By eliminating waste in production processes, LM speeds up production flows and emphasizes visual and transparent control, making it easier to identify failures towards continuous improvement [5]. Similarly to LM, Industry 4.0 offers many new opportunities as its main characteristic is the “digitalization” and “networking” of business processes by using Information Technology (IT) and the Internet. Industry 4.0 users see enormous potential in improving the productivity, quality, and flexibility of

Digitally Enhancing Kanban Lean Practice

71

their production systems [6]. Another means to achieve Digitalization is the application of modern Information and Communication Technologies (ICTs) as well as Data Analytics [7]. Researchers have shown a high interest in the outcomes of digitalizing traditional Lean tools [3]. For example, [8] et al. analyzed five new technology-enhanced Gemba Walk forms and their effects on communication and overall performance. Similarly, [9] et al. performed a study on how digital technologies can be applied to enhance PokaYokes as error-proofing tools and their digital benefits and drawbacks. Additionally, [10] et al. analyzed the advantages and disadvantages of implementing different digital technologies to the Lean philosophy of Visual Management or Mieruka (as it is called in Japanese). However, more “applied research” is needed on this trend to find the most suitable technologies to “digitally enhance” the traditional Lean tools and provide detailed guidance to Lean managers on how to adopt these successfully. This paper focuses specifically on the “Digital Enhancement of Kanban systems”. The traditional Kanban system was initially developed by Toyota and was one of the pillars of the Toyota Production System (TPS). It offers several benefits such as waste detection, quality improvement, inventory and breakdown reduction. The traditional Kanban system is a physical Kanban cards-based method that transmits information between workstations, providing what must be produced, when to produce it, and in which quantity. However, the transition from its physical cards to digital cards had to be done due to many drawbacks such as the loss and mishandling of the physical Kanban cards, and the limited real-time information sharing [11, 12]. Many DKSs case studies can be found in the scientific literature, using digital technologies (tools) such as the Internet of Things (IoT), Radio Frequency Identification (RFID), Bluetooth Low Energy (BLE), and Web applications. 2.1 Brief Review of the Digital Lean Manufacturing Paradigm By reviewing the available scientific literature on the Digital Lean Manufacturing paradigm (in Scopus and Web of Science databases), many examples of ways of digitalizing LM tools and their positive impact on a manufacturing enterprise’s operational excellence can be found. Each literature piece, being a journal article, conference paper, or book chapter, contains many Lean practices and Digital tools cited along with their known combinations and their impact on a company’s operational performance (e.g., reduced distance to be covered, increased efficiency, reduced lead times in material allocation, to mention a few). Based on information gathered from multiple scientific publications (viz. [3–5, 13], and [14]), it is commonly agreed that “Lean is a prerequisite for successful Digital Transformations” [15]. Furthermore, Kanban was one of the most frequent Lean practices cited in scientific literature and its known relations with Digital tools include: Sensors, RFID, Big Data Analytics, Automated Guided Vehicles (AGVs), and Simulations. Based on [13] et al.’s quantitative research and case study, nine Lean practices and nine Digital tools are strongly related to each other in their aim to contribute to “operational excellence”, making possible the emergence of Digital Lean tools [3] such as “Digital Kanban Systems”. Each known “Digital tool(s)-Lean practice(s)” combination can contribute to the increase of the efficiency, productivity, and traceability of a company’s performance.

72

C. Papadimitropoulou et al.

Nevertheless, since this paper focuses on DKSs, it will look more deeply into these in the following subsection. 2.2 Historical Evolution of the Digital Enhancement of Kanban Systems and the Role of DKSs After a brief review of the Digital Lean Manufacturing paradigm [16], this paper will focus on the “Digital Enhancement of Kanban Systems” history as narrated by the scientific literature. Since its conception, the Kanban system has undergone several significant improvements with the emergence of new technologies that have increased its effectiveness and precision [17]. In the scientific literature, various terms regarding Kanban systems evolution can be found due to different advances in technology such as: “e-Kanban Systems”, “Digital Kanban Systems”, and “Smart Kanban Systems”, however, there is no clear understanding of the subtleties and differences between these. For this reason, a Kanban Systems Taxonomy has been created (see Fig. 1) based on information extracted from multiple scientific publications to differentiate the existing generations in the Kanban systems evolution according to their digital (and analytical) transformations. As stated by Liker [18], the first generation of Kanban systems, called here Kanban Systems 1.0, was introduced by the Toyota Production System (TPS) around the 1940s and was based on analogue information systems where physical Kanban cards or containers are/were used to signal the need for more (raw) materials and parts between the warehouse and production and across the production line [19, 20]. The entire production was controlled by the physical Kanban cards circulating the workstations [21] and the number of cards needed to be dynamically adjusted to not cause excess inventory [22]. However, this first-generation system was limited by its physical nature, and as industries became more digitized, there was a need for more efficient and automated systems. Hence, Kanban Systems 2.0 are based on electronic information systems supported by computers and communication networks [23], referred to as “e-Kanban Systems”, which allow their e-Kanban cards to be electronically transmitted using an Internet connection to call for the required quantity of (raw) materials and parts from a supplier, warehouse, or workstation [24]. Kanban Systems 2.0 are more effective than their first generation since they allow for real-time tracking and performance monitoring along with data analysis [25]. According to [26], an e-Kanban System infrastructure consists of a service engine, a database, and a user interface, and was extremely helpful when the supplier and the customer were separated by long distances on the shop floor [27, 28]. Moreover, [11] et al. presented the importance of e-Kanban cards and their many benefits (e.g., ↓ human errors, ↓ work-in-progress, ↓ stock, real-time inventory, etc.). With the introduction of barcode technology in the 1990s, the third generation of Kanban systems, or Kanban Systems 3.0, was born allowing the automation of tracking and signaling, the reduction of errors, and the improvement of efficiency [29]. [30] et al. supported the third generation of Kanban systems by attaching BLE (Bluetooth Low Energy) beacons/buttons to the now digital Kanban cards to create a “Digital Kanban System”. In addition, this third generation differentiates from the previous ones by utilizing technical information such as barcodes and digital messages. These barcodes are scanned in various stages of production, monitoring the inventory and giving information for the replenishment procedure [31]. Lastly, Kanban Systems 4.0 were introduced

Digitally Enhancing Kanban Lean Practice

73

after 2011 along with the introduction of different Industry 4.0 technologies. [32, 33] and [34] researched how DKSs could be fully automated by minimizing human intervention and by implementing Industry 4.0 technologies such as active RFID tags, the Internet of Things (IoT), and Artificial Intelligence (AI) towards the creation of “Smart Kanban Systems” to enable greater automation and optimization of supply processes [35, 36].

Fig. 1. A Taxonomy of Kanban Systems Evolution based on their Digital Transformations

In Table 1, some extracts of the scientific literature review conducted along with their outputs and some critical points were summarized. Based on [30], the combination of BLE technology with a DKS was implemented for production management purposes in small and medium-sized manufacturing enterprises (SMEs). BLE beacons were affixed to Kanban cards to facilitate the transmission of information within a network, regardless of whether they were placed inside or outside a container. The primary feedback obtained from this case study indicated that this system is particularly suitable for SMMEs due to its cost-effectiveness and the elimination of additional manual scanning tasks for workers. Other proposed or implemented Smart Kanban systems that were identified in the scientific literature included various Industry 4.0 technologies such as IoT devices, smart sensors, active RFID, smart bins, and Webbased software. In regards to the current role of Kanban systems, traditional Kanban systems (based on physical Kanban cards) are still used in many shop floors and warehouses and can remain useful in some cases but in many others, this first-generation system has reached its limits in the context of diversified demand and high production cadency [37]. Indeed, the risk of losing physical Kanban cards exists when the plants are physically distant, and the lack of knowing the production points that require sourcing of (raw) materials and parts induces the logistician to visit and check all the production points. As a consequence, the work in non-value-added activities increases and makes the transportation of (raw) materials and parts perform with a high rate of waste, making the process of supply not agile. To make a supply process more reliable and reconfigurable with eco-friendly transportation, organized Kanban systems have to be intelligent enough to better inform the procurement organization or department about the required needs for the specific

74

C. Papadimitropoulou et al. Table 1. Insights for the Scientific Literature on Advanced Digital Kanban Systems

Authors

Kanban System Scope

Supporting Technology

Deployment Feedback

Critical Points

Sector Implemented

Shima et al. (2021) [30]

Production

BLE

• No additional operations by frontline workers are needed • Secured system

• Ideal for SMEs • Low cost of implementation • No specialized reader is needed • Possible beacon malfunction • Annual need for button battery replacement

Manufacturing SMEs

Xiang et al. (2017) [31]

Procurement

RFID

• Inventory • Hardware accuracy requirements • Decision-making • Efficiency skills needed by • Automation replenishment • ERP responsible software • Not applicable in case of theft or missing parts

Industrial backroom warehouse and shop floor

Thurer et al. (2016) [33]

Logistics

IoT

• Better time • Conceptual management suggestions • Real-time without information implementation • Visibility

Public solid waste collection services

Ghelichi et al. (2014) [34]

Inventory Management

RFID

• Increased efficiency and speed of production • Automated orders Kanban inventory reduced by 50%

Industrial production line

• Inventory items need to be properly categorized • Privacy and Security challenges • Data storage challenges • Ownership transfer challenges • Expensive technology and risk of obsolescence

Digitally Enhancing Kanban Lean Practice

75

nodes (i.e., production plants, or workstations). The concept of transportation holds a significant role within production and supply chain logistics as it can ensure the timely delivery of (raw) materials and parts, and its reconfiguration is deemed necessary. The possibility of reconfiguring the vehicle paths according to the digitalized Kanban cards information is an interesting support for increasing the reconfigurability of a supply process as can be depicted in Fig. 2.

Fig. 2. Digital Enhancements to Kanban Systems for Reconfigurable Supply Processes

It is suggested to analyze two possibilities/types of “Digital Kanban Systems”, the first one based on a (Digital) Kanban system enhancement using IoT/RFID technologies, and the second one based on a (Digital) Kanban system enhancement using IoT/BLE technologies. The first one is based on passive technologies, while the second one is based on active emission technologies.

3 Research Methodology This research work was developed based on a scientific literature review (in Scopus and Web of Science databases) that focused on “Kanban systems” and their available digital and analytical capabilities, and two illustrative “Digital Kanban Systems” case studies from the automotive industry. During the study of Kanban systems scientific literature, particularly of their digitalization trend, notable gaps regarding technology selection and the availability of case studies were identified. The main outcomes of the literature analysis were then organized and presented in Table 1. Moreover, to assist Lean production and supply chain managers in the successful adoption of DKSs, a taxonomy of Kanban systems evolution based on their digital transformations was introduced in Fig. 1 along with the development of a decision-support tool (framework) utilizing IoT/BLE and IoT/RFID technologies for Kanban systems digitalization as can be depicted in Fig. 4.

4 The Digitalization Challenge of Supply Chain Supply Chain Management (SCM) presents many local and global challenges. As such, the SCOR model presents the main macro processes that have to be reinforced to increase supply chain efficiency: Plan, Source, Make, Deliver, Return [38]. Other Supply Chain

76

C. Papadimitropoulou et al.

(SC) performance measures that can be assessed by the SCOR system are a company’s reliability, responsiveness, and agility [39]. The SC can be extended to the whole network and can be referenced to the internal one. For instance, in the automotive industry, the production plants are organized in such a way that they can be assimilated into the internal supply chain gathering plants in charge of sourcing raw materials and procurement parts, plants in charge of production, and the services in charge of shipment and organizing the transportation. In this paper, the focus is on the “supply process” to catch the right parts when required, namely “just-in-time”. Increasing supply reliability is a challenge today because it is the triggering point of any production flow. The more a supply process is “reliable”, the easier and faster a production flow can start in the internal supply chain. Nowadays, increasing the reliability of “supply” in global supply chains has become of the utmost importance in the automotive and aeronautic industries. [40] et al. remind the importance of avoiding shortages to make a reliable supply chain and sustain Lean (supply chain) projects. The importance of supporting supply processes is a key activity in SCM. Ensuring the arrival to the production line of the right (amount) raw materials and parts, to the right placement, and at the right time without consuming a high duration in their searching, excluding errors of picking, and bringing to the production lines has become today the specific objectives that collectively production and supply chain managers should think about under the light of the digital and smart capabilities. The target is to support the JIT philosophy in their manufacturing systems and supply chains based on Digital and Smart Kanban Systems. 4.1 A (Digital) Kanban System Enhanced by IoT/RFID Technologies Radio Frequency Identification (RFID) technology consists of a reader, an antenna, and a tag [41]. Its presence is becoming more and more popular in inventory management, and in its combination with the Kanban Lean practice [42]. This is because it does not require line-of-sight scanning and can read multiple tags simultaneously at a distance [43]. Moreover, it can collect real-time production data to provide timely and effective solutions through a real-time multi-objective decision-making model [44]. An IoT system allows devices to be accessible to the user in a remote way. Combining RFID and IoT technologies has some advantages like information transparency, product traceability, compatibility, scalability, and flexibility. In SCM, the RFID-IoT combination can automatically identify the status of an object, stock, equipment, machine, and even workers by capturing real-time data. In addition, it can detect the start or finish of a process from the signal of arriving and leaving the RFID-IoT covered area [41]. However, there are also some drawbacks to RFID technology use. First, the cost of using it can be expensive in the long run due to problems of continuous improvement and the risk of obsolescence. In addition, due to the huge amount of data generated by RFID tags, it is necessary to have economical computing resources to appropriately use and manage the data produced [34]. Lastly, external radio signals, metals, and liquids can interfere with RFID signals and technology [43]. In this paper, passive RFID tags are considered because of the opportunity that arises to study this in a practical case study in the automotive industry.

Digitally Enhancing Kanban Lean Practice

77

4.2 A (Digital) Kanban System Enhanced by IoT/BLE Technologies A BLE beacon is a small and low-cost component that transmits Bluetooth waves at predetermined intervals [30]. These devices have a low power consumption and can operate for years due to low power idle mode [45]. According to [46] et al., extensive research on BLE technology indicated the feasibility and suitability of IoT infrastructure due to its low-cost hardware, ease of deployment, and the freedom given to developers. BLE is known to be compatible with IoT devices like smartphones, watches, sensors, laptops, and keyboards [45]. IoT is a system of systems that allows for adjustments and flexibility [47]. IoT/BLE devices work as a twoway communication system by exchanging data through a gateway. The BLE gateway transmits the data to online cloud storage and then the data are transferred to electronic devices and/or analytics data storage. In addition, commands can be communicated from electronic devices straight to IoT devices. A proposed architecture for an IOT/BLE system based on [48] (see Fig. 3), includes a data source, a data collector, and a data gateway. A smart object can fulfil one or more of these roles and BLE technology is used for forwarding the data from the collector to the gateways.

Fig. 3. IoT/BLE Technology (Adapted from [47])

In sectors such as supply chain and logistics, BLE beacons can also transmit simple state data instead of big data packages. The information of whether a supply is needed or not is transmitted easily without the need for big infrastructure costs and components.

5 Modelling Selection of Digital Kanban System Solutions Figure 4 is the depiction of a DKS selection process model that was developed and is applied to two scenarios (Sect. 6) for supporting “Reconfigurable Lean Supply Processes”. The development of this decision-support tool arose out of the necessity to provide essential assistance to Lean production managers aiming to implement a DKS in their manufacturing facility. It can help them gain a comprehensive understanding of the characteristics and benefits associated with each scenario (described in Subsects. 6.1 and 6.2) as they strive to define their Kanban system constraints and the objectives of their company. Initially, a company will define what are its current objectives, and what

78

C. Papadimitropoulou et al.

it aims to improve in its current Kanban system (see O1 to O6 in Fig. 4). Afterwards, it is important to specify the potential constraints of their current Kanban system (see C1 to C6 in Fig. 4), as they can serve as decisive factors in determining the choice of the DKS solution. Each DKS solution’s characteristics and advantages can also drive the company’s decision. Thus, Fig. 4 provides a comprehensive summary of the aforementioned information (e.g. advantages, constraints, and scenario summary).

Fig. 4. Digital Kanban System Selection Process Model

The industrial feedback, received by the collaborating companies of IMS Laboratory, encompassed the expectations and the benefits recorded thanks to the design of a new DKS solution. The Digitalization of the Lean Kanban practice was grasped accurately by the field requirements (viz. safety, maintainability, cost, …) and applied accordingly to the context constraints that may augment the supply efficiency by capitalizing on the learnings from this process and by making human interventions simpler and more effective in future projects. In that way, a mutually beneficial and synergistic relationship emerges between humans and technology.

6 A Case Study on Selecting a Digital Kanban System In the subsequent subsections, a detailed analysis will be conducted on two distinct implemented, real-case scenarios originating from French automotive companies. These case studies, denoted as “Scenario 1” and “Scenario 2”, have been designated according to the DKS selection process model presented in Fig. 4. 6.1 Scenario 1: An IoT/RFID-Enabled Digital Kanban System As the example shown in Fig. 5, before implementing a DKS solution, the shop floor in Scenario 1 was using the traditional Kanban system with physical Kanban cards.

Digitally Enhancing Kanban Lean Practice

79

A logistician had to make a specific route multiple times a day to collect the physical Kanban cards (if available) and deliver them to the raw materials handling point (D10). After the raw material boxes were re-filled, they were then returned to the specific shops along with the physical Kanban cards. Depending on the size of the company and the production rate, the logistician could be doing multiple routes every day around all the workshops without having to collect or deliver anything; this situation was harmful to fuel consumption and the carbon print effect. So, a decision was made to implement IoT/RFID-enabled Digital Kanban System where the ‘WIP Area & Conditioning’ and ‘Tribofinishing’ stations were equipped with RFID reader surfaces. Once the shops (‘Tribofinishing & Reconditioning’) have a raw material shortage, the empty boxes equipped with RFID tags on them will be placed near the reader and the logistician will be electronically notified. The logistician will benefit from a reconfigurable route and will only pass through specific workshops to collect and deliver the materials.

Fig. 5. Scenario 1: Implementation of IoT/RFID Digital Kanban System on the Shopfloor

Some of the main benefits of the IoT/RFID-enabled Digital Kanban System are the decrease of the carbon print produced by the logistician’s vehicle, efficiency, productivity, and transparency increase due to the lack of errors or physical Kanban-cards loss. Now, the logistician will be notified on a digital display screen for the real-time shortages in reconditioning and tribofinishing shops and will follow a reconfigured routing in real-time. This DKS offers various tangible and quantifiable improvements, including a reduction in material pick-up time and a decrease in the distance covered during operations. In the example (see Fig. 5), we can observe how this DKS can be implemented on a Shopfloor. NFC terminals have been placed next to the workshops, therefore when there is a material shortage, the logistician is directly informed on his digital display screen and can follow the green loop show in green. The time and distance that will be required for the logistician to provide raw materials to the workshops that are equipped

80

C. Papadimitropoulou et al.

with NFC terminals, will be reduced by up to 50% since the use of Kanban cards will no longer be required. 6.2 Scenario 2: An IoT/BLE-Enabled Digital Kanban System The second use case considers an IoT/BLE-enabled Digital Kanban System. Before implementing a DKS solution, the warehouse used the traditional Kanban system with physical Kanban cards. When the workshops require new items to be replenished by the warehouse stock, the logistician responsible needs to browse through the warehouse aisles to find the correct reference and amount and deliver it to the corresponding workshop along with the physical Kanban card. When a manufacturing facility implemented an IoT/BLE-enabled Digital Kanban System in its warehouse management, the architecture shown in Fig. 6 was followed. The grey rectangles are the storage aisles where the BLE tags have been placed. These tags transmit small packages of data to a gateway, and these are then uploaded to a cloud computing environment. These data are then communicated through an IoT platform or smart devices such as smartphones, tablets, digital displays, or computers. In this scenario, the logistician will receive an e-notification to his/her smart device about material shortages in the lines along with their exact position (aisle, shelf) in the warehouse. Two additional options that could be integrated into the architecture are the “Pick-byLight” & “Put-to-Light”. Both options use a light, usually LED, mounted on a shelf or a beacon. The light indicates which part or product needs to be removed.

Fig. 6. Scenario 2: Implementation of IoT/BLE Digital Kanban System.

This method provides a higher picking performance, transparency, and reduces errors, thus efficiency [40]. The measurable improvements offered by this system are time and effort reduction in material picking and allocation and a decrease in the distance covered during operations. The “Put-to-Light” is essentially the inverse procedure, and it is used for replenishment or re-stocking picking areas. By utilizing “Pick-to-Light”/“Putto-Light” BLE beacons, productivity and working rate can be increased and errors in picking or re-supply will be avoided [40].

7 Conclusions and Further Work The Digitalization of traditional Lean tools has been proven by many to provide multiple advantages to an industrial environment [3, 16]. Since the traditional Kanban systems with physical Kanban cards have reached some limits in the context of diversified

Digitally Enhancing Kanban Lean Practice

81

demand and high production cadence, new technologies like IoT/RFID and IoT/BLE are being put into practice with expected positive results like improvement of efficiency and productivity. The two case study scenarios presented in this paper can show how these technologies and architectures are applied in real-life situations. However, for the moment, there is still an R&D gap and no detailed methodology exists for the selection of the right “Digitally Enhanced Kanban System” to help Lean production and supply chain managers guide the digitalization of their traditional Kanban Systems. As the future work of this research, it is aimed to build an Analytic Hierarchy Process (AHP) tool based on the constraints of a company and its objectives when aiming to digitally transform/enhance its Kanban system. Moreover, in addition to the aimed AHP tool development, it is targeted the creation of an algorithm that can be used by Lean production and supply chain managers to achieve correct and appropriate technological selection for their industry and company in regards to DKS technologies. A DKS solution selection process supported by an algorithm-based AHP tool might be of great help to guide the Lean managers in their digitalization strategy gaining managerial insights from this research work. Lastly, though this paper presents several useful reflections on the “DigitallyEnhancement of Kanban Systems”, it is suggested that further research could benefit from additional case studies. Acknowledgements. This research work was supported by the project BEST of the University of Bordeaux, France. The funding has been granted by Work Package (WP) 2 dealing with the “Design of Autonomous, Mobile, and Reconfigurable Industrial Systems”. The authors would like to thank the leader and co-leader of the WP2.

References 1. Pinto, J.L.Q., Matias, J.C.O., Pimentel, C., Azevedo, S.G., Govindan, K.: Just in time. In: Pinto, J.L.Q., Matias, J.C.O., Pimentel, C., Azevedo, S.G., Govindan, K. (eds.) Just in Time Factory. Management for Professionals, pp. 25–38. Springer, Cham (2018). https://doi.org/ 10.1007/978-3-319-77016-1_3 2. Sameer, M., Muztoba, K., Romero, D., Wuest, T.: Smart manufacturing: characteristics, technologies and enabling factors. Eng. Manuf. 233(5), 1342–1361 (2017) 3. Romero, D., Gaiardelli, P., Powell, D., Wuest, T., Thürer, M.: Digital lean cyber-physical production systems: the emergence of digital lean manufacturing and the significance of digital waste. In: Moon, I., Lee, G., Park, J., Kiritsis, D., von Cieminski, G. (eds.) APMS 2018. IFIPAICT, vol. 535, pp. 11–20. Springer, Cham (2018). https://doi.org/10.1007/978-3319-99704-9_2 4. Rahardjo, B., Wang, F.-K., Yeh, R.-H., Chen, Y.-P.: Lean manufacturing in industry 4.0: a smart and sustainable manufacturing system. Machines 11(1), 72 (2023) 5. Valamede, L., Akkari, A.: Lean 4.0: a new holistic approach for the integration of lean manufacturing tools and digital technologies. Math. Eng. Manag. Sci. 5(5), 851–868 (2020) 6. Jäger, J., Schöllhammera, O., Lickefetta, M., Bauernhansla, T.: Advanced complexity management strategic recommendations of handling the “Industrie 4.0” complexity for SMEs. In: 49th CIRP Conference on Manufacturing Systems, pp. 116–121. Elsevier (2016) 7. Thoben, K., Wuest, T., Wiesner, S.: “Industrie 4.0” and smart manufacturing – a review of research issues and application examples, automation technology 11(1), 4–16 (2017)

82

C. Papadimitropoulou et al.

8. Romero, D., Gaiardelli, P., Wuest, T., Powell, D., Thürer, M.: New forms of gemba walks and their digital tools in the digital lean manufacturing world. In: Lalic, B., Majstorovic, V., Marjanovic, U., von Cieminski, G., Romero, D. (eds.) APMS 2020. IFIPAICT, vol. 592, pp. 432–440. Springer, Cham (2020). https://doi.org/10.1007/978-3-030-57997-5_50 9. Romero, D., Gaiardelli, P., Powell, D.J., Zanchi, M.: Intelligent Poka-Yokes: error-proofing and continuous improvement in the digital lean manufacturing world. In: Kim, D.Y., von Cieminski, G., Romero, D. (eds.) APMS 2022. IFIPAICT, vol. 664, pp. 595–603. Springer, Cham (2022). https://doi.org/10.1007/978-3-031-16411-8_68 10. Romero, D., Zanchi, M., Powell, D.J., Gaiardelli, P.: Cyber-physical visual management systems in the digital lean manufacturing world. In: Kim, D.Y., von Cieminski, G., Romero, D. (eds.) APMS 2022. IFIPAICT, vol. 664, pp. 575–585. Springer, Cham (2022). https://doi. org/10.1007/978-3-031-16411-8_66 11. Romeira, B., Cunha, F., Moura, A.: Development and application of an e-Kanban system in the automotive industry. IOEM Soc. Int., 613–624 (2021) 12. Zouggar Amrani, A., Sailliol, J.-L.: Dassault using 4.0 technologies, showcase France. Internal document – industry for Europe project (I4EU) (2020) 13. Naciri, L., Mouhib, Z., Gallab, M., Nali, M., Abbou, R., Kebe, A.: Lean and industry 4.0: a leading harmony. Procedia Comput. Sci. 200, 394–406 (2022) 14. Mrugalska, B., Wyrwicka, M.: Towards lean production in industry 4.0. Procedia Eng. 182, 466–473 (2017) 15. Powell, D., Morgan, R., Howe, G.: Lean first … then digitalize: a standard approach for industry 4.0 implementation in SMEs. In: Dolgui, A., Bernard, A., Lemoine, D., von Cieminski, G., Romero, D. (eds.) APMS 2021. IFIPAICT, vol. 631, pp. 31–39. Springer, Cham (2021). https://doi.org/10.1007/978-3-030-85902-2_4 16. Powell, D.J., Romero, D.: Digital lean manufacturing: a literature review. In: IEEE International Conference on Industrial Engineering and Engineering Management, pp. 659–662 (2021) 17. Lage, M., Godinho, M.: Variations of the Kanban system: literature review and classification. Prod. Econ. 125(1), 13–21 (2010) 18. Liker, J.K.: The Toyota Way: 14 Management Principles from the World’s Greatest Manufacturer. McGraw-Hill, New York (2004) 19. Sugimori, Y., Kusunoki, K., Cho, F., Uchikawa, S.: Toyota production system and Kanban system materialization of just-in-time and respect-for-human system. Prod. Res. 15(6), 553– 564 (1977) 20. Ohno, T.: Toyota Production System: Beyond Large-Scale Production (1988) 21. Gupta, S.M., Al-Turki, Y.A.Y.: An algorithm to dynamically adjust the number of Kanbans in stochastic processing times and variable demand environment. Prod. Plan. Control 8(2), 133–141 (1997) 22. Tardif, V., Maaseidvaag, L.: An Adaptive approach to controlling Kanban systems. Eur. J. Oper. Res. 132(2), 411–424 (2001) 23. Kotani, S.: Optimal method for changing the number of Kanbans in the e-Kanban system & its applications. Prod. Res. 45(24), 5789–5809 (2007) 24. Razafuad, R., Ridwan, A.Y., Santosa, B.: Development of e-Kanban application using stock-needs rule prioritizing policy to reduce 0-pick for pharmaceutical warehousing. In: International Conference on Information and Communication Technology, pp. 310–318 (2018) 25. Wan, H., Chen, F.F.: A Web-based Kanban system for job dispatching, tracking, and performance monitoring. Adv. Manuf. Technol. 38(9–10), 995–1005 (2008) 26. Menanno, M., et al.: Implementing industry 4.0 technologies in lean production through eKanban automotive production. In: XXIV Summer School Francesco Turco, pp. 458–463 (2019)

Digitally Enhancing Kanban Lean Practice 27. 28. 29. 30.

31. 32. 33. 34.

35. 36. 37.

38.

39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49.

83

Ansari, A., Modarress, B.: Wireless Kanban. Prod. Invent. Manag. 36(1), 60–64 (1995) Cullen, T.J.: Toyota speeds parts delivery with e-Kanban. Automot. News Eur. (2002) Harmon, R.L., Peterson, L.D.: Reinventing the factory (1990) Shima, K., et al.: Status estimation and in-process connection of kanbans using BLE beacons and LPWA network to implement intra-traceability for the Kanban system. Sensors 21, 5038 (2021) Xiang, C., Savino, M., Palmieri, A.: Shelf Replenishment with RFID-ERP-Kanban System: A Case Study in Large Distribution (2017). https://hdl.handle.net/20.500.12070/42984 Sapry, H.R.M., Sabli, S.F.H.M., Ahmad, A.R.: Exploring e-Kanban application in the inventory management process. Crit. Rev. 7(8), 6–10 (2020) Thürer, M., et al.: Internet of Things (IoT) driven Kanban system for reverse logistics: solid waste collection. Intell. Manuf. 30, 2621–2630 (2019) Ghelichi, A., Abdelgawad, A.: A study on RFID-based Kanban System in inventory management. In: IEEE International Conference on Industrial Engineering and Engineering Management, pp. 1357–1361 (2014) Pekarcikova, M., et al.: Material flow optimization through e-Kanban system simulation. Simul. Model. 19(2), 243–254 (2020) Jaipriya, S., et al.: Development of smart Kanban system for stores. In: 7th International Conference on Advanced Computing and Communication Systems, pp. 945–948 (2021) Amrani, A., et al.: Lean manufacturing challenges to sustain flexible aeronautic industry. In: International Conference on Information Systems, Logistics & Supply Chain, Bordeaux (2016) El Kihel, Y., Zouggar A.A., Ducq, Y., Ameguouz, D., Lfakir, A.: Methodology combining industry 4.0 technologies and KPI’s reliability for supply chain performance. Comput. Integr. Manuf. (2023) Zulkarnaen, I., et al.: Modeling of beef supply chain performance measurement information system based SCOR. In: IOP Conference Series: Earth & Environmental Science (2022) Powell, D.J.: Integration of MRP logic and Kanban shopfloor control. In: Handbook of Research on Design and Management of Lean Production Systems. IGI Global (2014) Tan, W., Sidhu, M.: Review of RFID and IoT integration in supply chain management. Oper. Res. Perspect. 9, 100229 (2022) Powell, D.J., Skjelstad, L.: RFID for the extended lean enterprise. Int. J. Lean Six Sigma 3(3), 172–186 (2012) Barlow, R.: Stripes, tags can chip away at supply chain inefficiencies, tracking-and-tracing troubles. Healthcare Purchasing News (2022) Gong, G., Yang, Y., Wang, S.: Information and decision-making delays in MRP, KANBAN, and CONWIP. Prod. Econ. 156, 208–213 (2014) Rault, T., Bouabdallah, A., Challal, Y.: Energy efficiency in wireless sensor networks: a top-down survey. Comput. Netw. 67(4), 104–122 (2014) Jeon, K., She, J., Soonsawad, P., Ng, P.: BLE beacons for Internet of Things applications: survey, challenges and opportunities. IEEE IoT A(B) (2018) Benghozi, P., Bureau, S., Massit-Folléa, F.: The Internet of Things | What Challenges for Europe? Éditions de la Maison des Sciences de l’Homme, Paris (2009) Boualouache, A.E., et al.: A BLE-based data collection system for IoT. In: International Conference on New Technologies of Information and Communication, Algeria, pp. 1–5 (2015) Battini, D., Calzavara, M., Persona, A., Sgarbossa, F.: A comparative analysis of different paperless picking systems. Ind. Manag. Data Syst. 115(3), 483–503 (2015)

Sociotechnical Approach to Self-reporting in PMM Systems for HSE and Digital Security Jarle Nyberg

and Sverre Sørbye Larsen(B)

Department of Industrial Economics and Technology Management, Norwegian University of Science and Technology, Gjøvik, Norway {jarlenyb,Sverrsla}@stud.ntnu.no

Abstract. The purpose of this paper is to address self-reporting as performance measurements. The PMM system in the case organisation works well with automated performance measurements, but as self-reporting is implemented, the implications are not known. The authors collected qualitative data in a single unit of a Norwegian light-metal production company. The authors show that although the employees are highly engaged in their job, successful implementation of selfreporting on HSE and digital security is challenging with existing cultural barriers. The research performed is limited to a single-case study. Further research in other organisations implementation of self-reporting is required to confirm our findings. This paper contributes to research on PMM with a behavioural approach and further explores the interplay and balancing of social and technical controls, as called for by Smith & Bititci (2017). This paper integrates organisational and sociotechnical theory into their framework. The integrated framework is used to identify barriers for self-reporting at the case organisation. Keywords: PMM · Job Crafting · Organisational Culture · Self-reporting

1 Introduction Ongoing changes in society caused by digitisation and sustainability affects the reporting demand in organisations and how performance measurements and performance management (PMM) are performed. This is discussed in parts of the PMM literature. For organisations, Health, Safety and Environment (HSE), as well as Digital Security (DS) are under scrutiny of legal demands, social responsibility and ethics that imposes expectations to organisations from the surroundings. The efficiency of self-reporting is a known issue regarding how reliability can be achieved for reporting in PMM systems. To create a participating and empowering environment which creates engagement for improvement work for HSE and DS, employees’ self-reporting on the performance of both them and their co-workers is utilised. This approach was implemented in the unit of our analysis in 2022. The case organisation went through a major cyberattack only few years ago as well as an increase in injury statistics. This paper aims to explore to explore the relationship between self-reporting and performance measurements, and is centred around the following research questions: © IFIP International Federation for Information Processing 2023 Published by Springer Nature Switzerland AG 2023 E. Alfnes et al. (Eds.): APMS 2023, IFIP AICT 689, pp. 84–96, 2023. https://doi.org/10.1007/978-3-031-43662-8_7

Sociotechnical Approach to Self-reporting in PMM Systems

85

RQ 1) How does the self-reporting on HSE and DS function as a performance measurement in the department? RQ 2) What are the most important factors creating enablers and barriers for self-reporting as performance measurement in a PMM system?

2 Theoretical Background 2.1 Organisation Theory An organisation’s goals, strategies, and structure are formal organisational factors influencing and aligning employee’s behaviour with strategic choices [1]. Organisational culture is an informal organisational factor which can greatly influence behaviour and strengthen or weaken the formal organisational factors effect on behaviour [1]. Job design containing work tasks further structures, constrains and influence employee’s behaviour [2]. Research on job design also focuses on employee’s individual and collective job crafting and customisation of their own work [2]. Job design can emphasise employee motivation [2, 3] by the perceived intrinsic and extrinsic motivational factors in the content of the work itself [3]. Employee’s motivation can affect their attitudes with different denominations for job- and organisational satisfaction, engagement, and commitment [3]. Organisation theory utilises socialisation, control and formal governance as means for influencing behaviour [1, 3]. 2.2 Sociotechnical Systems Theory (STS) In sociotechnical work systems the social subsystem consists of work roles [4], individuals who socially interact in teams with a need for coordination, control, and boundaries [5]. The technical subsystem consists of technology and associated work organisation [4] [5]. Sociotechnical principles seek to balance the subsystems in relation to each other and the environment in which they exist [4]. The purpose is joint optimisation of the subsystems to achieve productivity and human psychosocial well-being in the design of the technical subsystems [4]. STS focuses on work systems and teams rather than individual work roles [4]. 2.3 Performance Measurement and Performance Management (PMM) Organisational control theory concerns how organisations can influence employee’s behaviour to achieve goals [6]. Rooted in organisational control theory Smith & Bititci proposes a theoretical framework for PMM which provides a basis for further research [7]. The framework shown in Fig. 1, integrates performance management/social controls and performance measurement/technical controls [7]. Systematisation of the PMM literature’s description of performance management/social controls can be categorised as: A) communication/evaluation/feedback on performances [8, 9] B) influencing organisational culture [7, 9, 10, 12, 13] C) job design [7, 14] D) the purpose/use of performance measurements [8, 13, 14]. Approaches to performance management/social control spans from authoritarian to empowering orientation [7]. The PMM literature describes performance measurements/technical controls as

86

J. Nyberg and S. S. Larsen

Fig. 1. Smith & Bititci theoretical framework

results of strategic goals, aiding deployment of strategies [7], implemented in the formal structure [7] of the organisation as measurements [8], rules and guidelines [9], procedures [10], standards and solving of tasks [11] and organisation of individuals/groups [12]. According to Smith & Bititci [7] improvement interventions in performance measurements can be configured as social controls and/or technical controls with positive or negative effect on intended outcome. Further Smith & Bititci indicates that an appropriate balance between the controls can lead to increased employee engagement and performance. Authors in the field of PMM argues that PMM interacts with organisational goals, strategies, structure, and culture [20]. Parts of the PMM literature references the sociotechnical nature of PMM [13, 14].

3 Research Method The case study analyses the integration of security in an already existing performance measurement system, called Team Performance (TP). Security has been integrated both in the digital tool for tracking performance, and in an individual meeting structure like the one in the original Team Performance, but with security as the sole subject. The study was conducted in a single department of a Norwegian electrochemical plant which we in this paper call Department. This department was chosen to explore the problem statement both because the Department had recently introduced Security as part of TP, as well as because of its success in the use of TP to support continuous improvement of production parameters. During two separate visits to the same plant in November 2022 and March 2023, qualitative data was collected. The authors observed two days of production in November and conducted 17 semi-structured interviews over a three-day visit in March. The informants represent several levels of the organisation. The organisational levels

Sociotechnical Approach to Self-reporting in PMM Systems

87

covered were operators, safety representatives, team leaders, shift managers and department manager. An overview of the informants is presented in Table 1. Some informants cover several roles, e.g., Team Leader and Safety Representative. Table 1. Data collection overview Role

No. of informants

Avg. Length of interviews

Avg. Number of coded segments

Department Manager

1

52 min

110

Shift Manager

3

57 min

103

Team Leaders

6

43 min

106

Safety Rep

3

41 min

71

Operator

5

27 min

59

TOTAL

17

11 h 18 min

1350

Additionally, the authors gathered secondary data about TP through the research project’s longstanding collaboration with the case organisation, which involved visiting multiple plants, participating in workshops, and engaging in other relevant activities. The data analysis started by splitting the task in two - one researcher categorised the material by relying on theoretical propositions [18] and the other worked the data from the “ground up” [18] by coding initially in a total of 130 categories. The results were then discussed to develop common interpretations and agree on the most important findings. These findings are presented as individual subsections in Sect. 5, Results. Comparing the findings with theory clarified the need for an integration of several fields of theory into one model – one field of theory is not enough to answer the research questions. The authors therefore suggest a revised model, presented in our discussions in Sect. 6.

4 Case Description In 2007, the company created a production system tailor made for their organisation. The system was implemented in the different plants and departments over a few years following. The production system incorporates traditional lean principles for process control and continuous improvement, as well as concepts of self-managing teams and reduced hierarchy. Standardisation of repetitive tasks is implemented through Standard Operating Procedures (SOPs), although operators are also encouraged to have the competence to reply to deviations in chemical processes. In 2016, the department was chosen for piloting the implementation of Team Performance (TP) a measurement system for collecting and displaying data from various machines and systems, presented to employees as Key Performance Indicators (KPIs). The KPIs can be tracked by employees to see developments in performance over time, and to compare to other teams and departments. Team Performance Security (TPS) was implemented as an extension of TP in the summer of 2022. The KPIs implemented were

88

J. Nyberg and S. S. Larsen

“Amount of employee reported non-conformities» and “Average pass rate in control observations (WOC)”. In addition to the KPIs added, a dedicated TPS meeting is held every five weeks, where the team, supported by the Shift Manager and Safety Representative, agrees on security focus areas for the following five weeks. An example of a focus area can be to focus on the security KPIs in TP. For the reporting of non-conformities, an established digital system is used, functional with both smart phones and computers. Anyone in the company can access written reports, and most employees have access to the system. The reports are compiled in the scorecard in TP, as well as read out loud on every change of shift by the Shift Manager. For the average pass rate in control observations, manual observations done in production is the source of data. These observations are called Walk-Observe-Communicate (WOC). Anyone can perform a WOC on anything, although most WOCs are performed by the Shift Managers. On a regular basis WOCs of the SOPs are performed.

5 Results 5.1 Job Satisfaction, Engagement, and Commitment The general impression is that employees enjoy working at Department, the degree of job satisfaction is high. The most notable reasons are working conditions, pride in working for an important part of the local community, and identification with company values. Employees at Department regard their everyday tasks, although repetitive, as well adapted to their skill sets and competencies. The employees are committed to the company’s strategic goals. Most respondents also feel valued, and a critical, integral part of the Departments function. The respondents are engaged in Cyber Security, rating it on an equal level of importance as HSE. Employees have completed Cyber Security training in the company’s safety training app, although many respondents feel the need for a higher degree of understanding of Cyber Security risks (Table 2). Table 2. Testimonies regarding job satisfaction, engagement, and commitment Role

Testimony

Team Leader “It’s pretty good, I like it here. The company values, I can find myself in them in a way. This place has things I can work for” Operator

“Honestly, I don’t feel sufficiently competent. Considering we had an impactful cyber incident some years ago, that is not good. (…) You can see the paralysis in the organisation when an incident happens. After the incident we had a training module in the app, but it was around 7 min of training and long forgotten. So no, I don’t think I’m sufficiently competent, and I think there is a high chance one of our many employees makes the same mistake again. (…) I would absolutely like to have more training in Cyber Security, I think it would be a huge advantage”

Sociotechnical Approach to Self-reporting in PMM Systems

89

5.2 Job Design, Job Crafting and Impact on Self-reporting Overall, the job design at the department can be described as attending employees’ psychosocial needs. Findings are that even though work can be repetitive with strict procedural impositions, a considerable number of interviewees responded positively regarding; variation in tasks, task identity, importance of work, received feedback, social recognition/support, and performance achievement. Job design is influenced by continuously quality finetuning of production processes, requiring adherence to standard operating procedures (SOPs) to ensure production of high-quality aluminium. The workplace by nature contains risk with potential for harm which influences the job design. Operators are potentially exposed to high voltage electricity, high heat, large vehicles, gases, complexity of different job tasks/processes co-existing and simultaneously performed. Risks are sought to be controlled through strict SOPs maintaining HSE. Job design is influenced by DS requirements and there is consensus among the interviewees that DS is strongly emphasised with policies and guidelines to follow. Job design regarding TPS; self-reporting is regarded as an important part of TPS encouraged to be performed by all employees. In general, the plant has a low threshold for reporting deviances. Data collected from self-reports are to be used in TPS meetings with the purpose of learning and feedback. These meetings delegates responsibility and decisionmaking to create autonomy in the operators HSE work. A weakness is the lack of access and mastery of the digital system for reporting in parts of the work force. Job crafting in production processes can occur mainly due to time pressure and shortcuts taken. A general impression is that job crafting regarding some of the SOPs for HSE occurs. Job crafting is frequent for self-reporting on HSE incidents committed by employees. It seems to be a threshold for perceived severity of which incidents to report, with a preference to give direct feedback after incidents. Regarding DS the interviewees had faith in the corporate functions maintaining security, leading to minimal job crafting and general adherence to the policies/guidelines. In general job crafting is reduced due to the possibility to suggest improvements and give feedback on existing SOPs, either to leaders or in the digital system where SOPs are described (Table 3). 5.3 The Culture and Its Impact on Self-reporting The safety/security focus at the department is strong. All interviewees are conscious of HSE and its importance. HSE has been increasingly improved and focused on since the 1980’s. The remnants from the old days are a reason for why the culture can be described as fragmented, with significant differences between and within shifts. Culturally the value of self-reported incidents can be perceived as something negative in parts of the work force. A general perception is a natural discomfort in reporting on other people. There are uncertainties of how co-workers react when reported on. There is a tendency that younger employees to a larger extent perceives self-reporting of incidents in a more positive manner. There are certain norms related to self-reporting incidents. The severe incidents are reported, still there is a norm for the threshold for which incidents to self-report, and for

90

J. Nyberg and S. S. Larsen Table 3. Testimonies regarding job design, job crafting and its impact on self-reporting.

Role

Testimony

Department Manager

“TPS is totally dependent of the collection of information”

Shift Manager “I am not the most knowledgeable on DS but follow the rules, which are so clear that I feel confident that I will not make any mistakes” “I trust the self-reported HSE incidents to be real, but do not have confidence in the numbers. It could and should be much higher. A lot of incidents are not reported, but the severe incidents are reported, it is the incidents that seems less severe I question the numbers for” Team Leader

“It is the team that chooses focus areas to work on, they discuss internally, and it is the operators themselves who should choose the focus areas” “I trust the corporation DS functions and regard our systems as being safe, I expect that DS in the corporation are in order” “If observing an incident, I message/talk to the person involved to notify him about the mistake and inform him that it should not happen again. I do not report it” “I am confident that severe incidents are reported in my team, less severe incidents should be filtered in regard to what is and is not important, I am the only person in my team self-reporting incidents”

Operator

“The signals from management are to self-report everything that is not in order, the smallest thing” “I do not self-report myself, I notify the team leader who does it, I would probably self-report if I had access” “In essence, the SOP should be followed to the letter. But sometimes you must use discretionary judgement” «You sometimes see the rule being broken, I don’t think it’s done on purpose, or it can be done on purpose” “The safety rules govern the work processes, one should not really break them, but most people do, not all rules are broken only some of them”

which not to report. There is also a norm for giving direct feedback to persons involved in incidents, rather than reporting the incident. There are assumptions in the culture impacting self-reporting of incidents (Table 4).

6 Discussion of Findings 6.1 Department’s Experiences of Self-reporting and TPS (RQ 1) TPS performance management can be described as empowering, in contrast to the authoritarian nature of HSE/DS systems. Performance measurements in TPS has formative purposes to continuously improve HSE and DS. This is reflected in the TPS methods for evaluation and feedback of performances. Job design in TPS attends psychosocial needs and focuses on autonomy and participation. The department has a strong focus and a cultural disposition for safety and improvement work. The employees can be said

Sociotechnical Approach to Self-reporting in PMM Systems

91

Table 4. Testimonies regarding the culture and its impact on self-reporting Role

Testimony

Department Manager “Changing culture and behaviour is not easy regarding safety/security. There are variations between the teams, some are doing well, and others are struggling to achieve a good process for TPS. It has something do with age; the younger employees react more positively. Those that started working in the 1980’s, when almost everything was permitted, they are a bit characterised by it. Such a person can influence the mood and tone within a team” Shift Manager

“There are differences in the team’s motivation for TPS. Some are very motivated, whilst other teams have a mix of members that are less motivated, which can be challenging” “The operators reacts differently when they are reported in an incident” “If a mistake has happened, reporting is dependent on which shift that replaces own shift. Self-reporting on HSE is related to results of actions and not why they have occurred. Still, it can be perceived as a pillory between shifts. My impression is that incidents occurring on a shift you have a good relationship with are not reported. It is more likely to report incidents occurring on shifts that never replaces you in the shift schedule”

Safety Rep

“There is a significant difference in culture between the shifts, and between teams on the same shift”

Team Leader

“If working overtime at other shifts you notice that they do things differently compared to the shift I work on. Things are done differently with differences between the shifts” “I do not know for sure about other shifts, some has a reputation for being cowboys (take more risk)” “I can’t find any negative consequences of self-reporting incidents, except the feeling of putting your buddy in a nasty situation” “There is a clear threshold for which incidents to report and which that is communicated directly to the person(s) involved” “There is reluctancy to report HSE incidents in general and especially on other shifts, there is fear of theme revenging and reporting on own shift” (continued)

92

J. Nyberg and S. S. Larsen Table 4. (continued)

Role

Testimony

Operator

“The co-workers focus, and attitude are that everyone wants to perform well and do their jobs according to the rules. I feel the safety culture is good” “There is a difference between the older and younger ages. It lingers from the old days” “Self-reporting of incidents is beneficial in the long run. It is positive that the involved persons are anonymised since it prevents it from becoming personal” “The colleague wouldn’t report an incident when considering the possible reactions. It was not a severe incident, and some people thought it would be bad to reported for less severe incidents. Many incidents can be regarded as insignificant” “I believe there is widespread reluctancy to report incidents, since they can have a comradely relationship”

to have good job satisfaction, engagement and commitment also indicated by the TP results. Still the department has not experienced or achieved sufficient motivation and numbers for TPS and self-reporting. Job crafting is not aligned to the Department’s goal and strategies for TPS. Job crafting for self-reporting in TPS are more misaligned and frequent compared to other work tasks. The norm of employees giving each other direct feedback when involved in incidents reflects their strong HSE focus, it is consequently done, and undesirable behaviour are reacted to and dealt with by the employees. There have been very few self-reports regarding DS. This is because of clear policies and central DS functions efforts to create well-secured digital systems. Also, the interviewees experiences are concentrated towards HSE as they to a larger degree relate to this in their daily work. The papers contribution to self-reporting on DS are affected by this matter. The answer to the first research question is that self-reporting does not function adequately as a performance measurement at the department. There is a lower frequency of self-reports compared to the frequency of actual incidents occurring. 6.2 Barriers and Enablers to Self-reporting (RQ 2) Job design regarding self-reporting with autonomy and decreased direct control of employees performing it leads to opportunities for job crafting. Self-reporting is dependant of employees’ attitudes and values. If these are not in accordance with the demands that self-reporting poses job crafting will occur. The culture in the Department is fragmented between and within shifts. This is reflected in how self-reporting HSE/DS incidents is valued and denominated differently between individuals, teams, and shifts. This causes both reluctancy to self-report on others and negative reactions when being reported on. Identified norms arising from

Sociotechnical Approach to Self-reporting in PMM Systems

93

these value denominations are: 1) Preference for giving direct feedback to co-workers involved in incidents, alternatively do nothing 2) There is a common threshold for the perceived severity of incidents to self-report, where incidents regarded as severe are reported, incidents regarded as not severe are not reported. The cultural value and assumptions regarding self-reporting can lead to gaming between shifts. Tactical reporting occurs based on 1) Which shifts replaces each other in the schedule 2) their comradely relationships 3) pillory attitude and fear of revenge and 4) inequal standards for reporting between shifts can lead to perceived injustice. Gaming prevents effective self-reporting and can lead to decreased lack of trust and perceived procedural injustice for self-reporting in TPS. The digital system utilised for self-reporting in TPS is not accessible and mastered by all employees. A considerable number of the interviewees described it as being unnecessary cumbersome and complicated, especially at the initial use of the system. The lack of simplification and optimisation of the system in relation to the users doesn’t ensure widespread use, negatively affecting numbers for self-reporting. Sociotechnical joint optimisation relates to how use of technology and work organisation should reflect the needs of the work roles. For the department the needs of the employees are not aligned with self-reporting. The cultural value of self-reporting needs to be altered to enable self-reporting. Also, the digital system used for self-reporting needs to be simplified for use, or alternatively parts of the employees’ skills of utilising it should be increased. Increased focus on TPS as a work system regarding employees as components in this system prone to making mistakes, where the department will learn from incidents and improve the work on HSE and DS, can reduce the lack of self-reporting with preference of giving direct feedback which instead leads to learning between individuals and teams. According to Smith & Bititci [7] interventions in performance measurements can be configured as either technical or social controls with unknown effect on outcome. For the department the intervention has been technical with a change in the methods used for performance measurement and the outcome has been negative. Employees avoid self-reporting except for severe incidents. Smith & Bititci [7] argues that an appropriate balance between technical and social can lead to improved employee engagement. The social controls created by the culture needed for self-reporting are not in place at the department. Intervention in the social controls regarding the cultural value of self-reporting affecting norms and assumptions are demanded. Summarised the main barrier for self-reporting are the cultural value affecting cultural norms and assumptions with a negative impact of self-reporting. Lack of access and mastery of the digital system used for reporting also poses a barrier. The main enabler for self-reporting is the alignment between the culture and demands posed by self-reporting. This can create the social controls needed for self-reporting at the department. Mastery and access to the digital system for reporting is also an enabler for self-reporting. 6.3 Integration of Theory into Smith and Bititci’s Theoretical Framework The papers research question relates to what employees’ choses to do when observing an incident that should be reported, where the sum of their actions influences the PMM system using self-reporting. All three fields of theory help to explain the barriers and

94

J. Nyberg and S. S. Larsen

enabler of self-reporting. Organisation theory focuses on human behaviour in organisations [1] and strengthens the framework ‘s explanatory properties. Both organisation theory and STS consists of technical/formal and social/informal components that can be integrated into Smith & Bititci [7] technical and social axes. Formal organisational factors highlight how employee’s behaviour is sought to be influenced and aligned with goals through chosen strategies and structure. They are antecedents of employee’s task solving and performance that PMM measures and utilises with various technical controls. Informal factors of organisational culture as well as attitudes, motivations and job crafting influences task solving, performances and goal alignment. In this social sphere performance management and social controls of PMM are implemented. They consist of job design, purpose of performance measurements, socially influencing the organisation culture as well as various methods for evaluating and giving employees feedback on their performances. The rationale for integrating STS is system thinking and the highlighting of interaction between social and technical subsystems/structures [4, 5]. Design of the technical subsystem in relationship to the needs of the work roles in the social system can lead to productivity. In this paper productivity for the work system TPS is improved selfreporting compared to actual incidents that occurs. Sociotechnical control where the work roles have autonomy and control of their own work is suited as social control for self-reporting. PMM’s social and technical controls are respectively implemented in the social and technical subsystems, as controls for employee’s behavioural alignment with organisational goals and strategies.

Fig. 2. Integration of theory into Smith & Bititci’s [7] theoretical framework

A behavioural approach to PMM rooted in organisational control theory, focusing on how organisations can influence behaviour puts the sociotechnical work roles at the centre of the integrated framework. Similar to the sociotechnical principle of joint optimisation of subsystems [4], Smith & Bititci discusses how intervention in either social or technical

Sociotechnical Approach to Self-reporting in PMM Systems

95

controls could lead to the need of intervention in the opposite control [7]. The behaviour and working conditions for the work roles are affected by the technical controls in the technical subsystem which consists of organisations formal factors, technology used and work organisation equivalent to structures in organisation theory. The behaviour of the work roles is also affected by performance management with social controls as well as the conditions in the social subsystem. The different fields of theory can interact and be complementary for influencing and facilitating behaviour that leads to desired performances in accordance with organisational goals. Integration of theory is illustrated in Fig. 2.

7 Concluding Remarks The discussed barriers lead to frequent job crafting related to self-reporting in TPS. Job crafting can be positive or negatively associated with organisational goals, strategies, and values [19]. At the Department job crafting related to self-reporting in TPS has an undesirable effect. The sociotechnical principle of joint optimisation of both subsystems [4] is actualised regarding the effectivity of TPS, utilising this principle and considering the employees needs could prevent barriers for self-reporting at the department. Smith & Bititci discusses the need to balance technical and social controls and their study concluded with interventions in the technical controls could lead to need for interventions in the social controls [7]. This paper confirms their conclusion. Intervention in technical controls by self-reporting in TPS, leads to a need for performance management practices influencing and altering cultural values and norms. Smith & Bititci [7] comments the gap of research on the design of organisational controls in context of the process managed. This paper contributes with research on the use of self-reporting as performance measurement in PMM system regarding HSE and DS. Self-reporting is dependent on suitable social controls and proposes emphasis on them when self-reporting is utilised. Smith & Bititci proposes their framework as basis for further research [7]. This paper has contributed to this call by proposing integration of other fields of theory to create a more holistic behavioural approach to PMM. The paper proposes further research on the suggested framework with integrated theory.

References 1. Jacobsen, D., Thorsvik, J.: Hvordan Organisasjoner Fungerer, 4th edn. Fagbokforlaget, Bergen (2013) 2. Oldham, G.R., Fried, Y.: Job design research and theory: past, present and future. Organ. Behav. Hum. Decis. Processes, 20–35 (2016). https://doi.org/10.1016/j.obhdp.2016.05.002 3. Kaufmann, A., Kaufmann, G.: Psykologi i organisasjon og ledelse, 5th edn. Fagbokforlaget, Bergen (2015) 4. Mumford, E.: The story of socio-technical design: reflections on its success, failures and potential. Inf. Syst. J. 16, 317–342 (2006). https://doi.org/10.1111/j.1365-2575.2006.00221.x

96

J. Nyberg and S. S. Larsen

5. Carayon, P., Hancock, P., Leveson, N., Noy, I., Sznelwar, L., Van Hootegem, G.: Advancing a sociotechnical systems approach to workplace safety - developing the conceptual framework. Ergonomics 58, 548–564 (2015). https://doi.org/10.1080/00140139.2015.1015623 6. Liu, L., Borman, M., Gao, J.: Delivering complex engineering projects: reexamining organizational control theory. Int. J. Proj. Manag. 32(5), 791–802 (2014). https://doi.org/10.1016/ J.IJPROMAN.2013.10.006 7. Smith, M., Bititci, U.S.: Interplay between performance measurement and management, employee engagement and performance. Int. J. Oper. Prod. Manag. 37(9), 1207–1228 (2017). https://doi.org/10.1108/IJOPM-06-2015-0313 8. Bourne, M., Pavlov, A., Franco-Santos, M., Lucianette, L., Mura, M.: Generating organisational performance; the contributing effects of performance measurement and human resource management practices. Int. J. Oper. Prod. Manag. 33(11), 1599–1622 (2013). https://doi.org/ 10.1108/IJOPM-07-2010-0200 9. Schleicher, D.J., Baumann, H.M., Sullivan, D.W., Levy, P.E., Hargrove, D.C., Barros-Rivera, B.A.: Putting the system into performance management systems: a review and agenda for performance management research. J. Manag. 44(6), 2209–2245 (2018). https://doi.org/10. 1177/0149206318755303 10. Garengo, P., Betto, F.: The role of organisational culture and leadership style in performance measurement and management: a longitudinal case study. Prod. Plan. Control (Ahead-ofprint), 1–19 (2022). https://doi.org/10.1080/09537287.2022.2058431 11. Simons, R.: How new top managers use control systems as levers of strategic renewal. Strateg. Manag. J. 15(3), 169–189 (1994). https://doi.org/10.1002/smj.4250150301 12. Tessier, S., Otley, D.: A conceptual development of Simons Levers of control framework. Manag. Account. Res. 23(3), 171–185 (2012). https://doi.org/10.1016/j.mar.2012.04.003 13. Malmi, T., Brown, D.A.: Management control systems as a package - opportunities, challenges and research directions. Manag. Account. Res. 19(4), 287–300 (2008). https://doi.org/10. 1016/j.mar.2008.09.003 14. Nudurupati, S.S., Garengo, P., Bititci, U.S.: Impact of the changing business environment on performance measurement and management practices. Int. J. Prod. Econ. 232 (2021). https:// doi.org/10.1016/j.ijpe.2020.107942 15. Bititci, U.S., Bourne, M., Cross, J.A., Nudurupati, S.S., Sang, K.: Editorial: towards a theoretical foundation for performance measurement and management. Int. J. Manag. Rev. 20(3) (2018). https://doi.org/10.1111/ijmr.12185 16. Franco-Santos, M., Otley, D.: Reviewing and theorizing the unintended consequences of performance management systems. Int. J. Manag. Rev. 20(3), 696–730 (2018). https://doi. org/10.1111/ijmr.12183 17. Neely, A., Gregory, M., Platts, K.: Performance measurement system design: a literature review and research agenda. Int. J. Oper. Prod. Manag. 15(4), 80–116 (1995). https://doi.org/ 10.1108/01443579510083622 18. Yin, R.K.: Case Study Research: Design and Methods, 5th edn. SAGE Publications, Inc., Thousand Oaks (2014) 19. Wrzesniewski, A., Dutton, J.E.: Crafting a job: revisioning employees as active crafters of their work. Acad. Manag. Rev. 25, 179–201 (2001) 20. Melnyk, S.A., Bititci, U.S., Platts, K., Tobias, J., Andersen, B.: Is performance measurement and management fit for the future? Manag. Account. Res. 25(2), 173–186 (2014)

The Productivity Leap: Effects of an Industry Program for Norwegian SMEs Eivind Reke(B) , Natalia Iakymenko, and Mette Holmriis Bugger SINTEF Manufacturing, Raufoss, Trondheim, Norway [email protected]

Abstract. Industry programs have been a method of lifting the competence of especially Small and Medium Sized Enterprises (SMEs) that otherwise would not be interested or could not afford such efforts. In this paper we present a case study of four Norwegian SMEs that participated in a Lean Production based industry program for Small and Medium sized Businesses in Norway. The aim of the program is to lift the productivity level of the participating companies through teaching and coaching lean production tools, methods, and leadership ideals. The data was collected through semi-structured interviews and documents obtained from the companies with the intent of discovering how and if the companies successfully applied the content of the program and the effects of the program on company performance. The companies that participated in this study was part of the first round of the program and had all gone through the whole program, as such the study is retrospective. It was found that companies experienced some positive results from the program but struggled to spread the learning from the program to other parts of the business, and to continue its improvement journey once the program had ended. Keywords: Lean Production · Industry Programs · Case Study

1 Introduction How companies can increase their productivity has always been of interest to both owners, employees, managers, and researchers alike. In a high-cost Scandinavian country like Norway where Small and Medium sized Enterprises (SMEs) are seen as cornerstone companies, key to the survival of local communities, efforts to improve productivity are considered paramount for the survival of these companies. As such, there is a need to spread information and knowledge about how productivity can be increased to these companies. One such way of spreading knowledge is through industry programs. “Produktivitetsspranget”, “The Productivity Leap (PL)” is a Norwegian industry program developed by Lean Forum Norway (LFN). The PL was designed to support Small and Medium Sized Enterprises (SMEs) to develop knowledge and capabilities for increased productivity. The program was developed based on the experiences of a similar program in Sweden; “Produktivitetslyftet”, that commenced in 2007. In fact, in © IFIP International Federation for Information Processing 2023 Published by Springer Nature Switzerland AG 2023 E. Alfnes et al. (Eds.): APMS 2023, IFIP AICT 689, pp. 97–108, 2023. https://doi.org/10.1007/978-3-031-43662-8_8

98

E. Reke et al.

an evaluation of the effects of the first part of that program (the period from 2007–2010), researchers in Sweden found that their PL program had led to a significant increase in turn-over when compared to similar companies over the same period that had not participated in the program [1]. This paper presents the initial findings from the first effort on studying the effects of the sibling Norwegian program. The pilot version of PL in Norway commenced in 2018 with the aim of developing and adopting the concept for companies, recruiting and qualifying companies, as well as applying the program in six companies. Currently 16 companies have finished or are in the process of finishing the program. The program was designed to increase the productivity of SMEs based on central elements of the Norwegian model, lean production, and leadership (including long-term thinking and consensus-based decisionmaking), and the application of enabling technology which will be presented more indepth in the next section followed by methodology used, findings from the cases, and finally discussion and conclusion. This paper aims to unfold the topic PL implementation in Norway by answering following questions: (1) What is the status of PL program implementation in Norwegian companies? (2) Does the PL program affect the companies’ operational and financial performance? Following hereon we present a literature study on lean production and the content and intent of the program. We then present our methodology and findings. Finally, we present our analysis and discussion of the data and in the end a conclusion with suggestions for further research and improvements for the industry program.

2 Theoretical Foundation 2.1 The Norwegian Model The Norwegian model share many similarities with the Nordic model and refers to the collaborative relationship between employers (confederations of enterprises), employees (trade unions), and the state (laws and legislations) with regards to industrial development and workers’ rights. Nordic countries have been more successful than others in combining economic growth with a well-regulated labor market, social cohesion and a fairer distribution of wealth [2]. [3] discuss three ways in which this model can be understood. 1) As a socially constructed discourse or dogma, 2) a set of institutionalized procedures and understandings, and 3) as a result of social power relations and strategy in the class struggle. According to [4], the idea of the model was to bring resources and measures from employers, unions and the government together to reduce conflict in working life. For the program design, the Norwegian model manifests itself in the fact that both union leaders and senior management are part of the steering committee of each company. LO, the largest union of employees and NHO the largest employer organization are both part of the overall steering committee of the program itself and have seats on the board of Lean Forum Norway.

The Productivity Leap: Effects of an Industry Program

99

2.2 Lean Production Originating in Toyota, Japan as an alternative to Ford and GMs Mass Production principles lean production was introduced as a complete business model in [5]. After discovering the superior performance of Japanese automakers, and in particular Toyota, Jones and Womack set out to find other companies that had adopted lean production. Their findings were presented in the seminal book “Lean Thinking” [6]. The catchy nature of the term “lean” and the success of its early adopters led to a new manufacturing paradigm that has become one of the, if not the most influential of our time [7]. In Norway, the diffusion of activities under the term lean gathered pace in 2009 and the terminology of lean has since been adopted outside of manufacturing [8]. In 2014 the book “Lean blir Norsk” (Lean becomes Norwegian) found that four basic sets of tools had made its way to Norway. Standardization, Flow, Visualization and Continuous Improvement [9]. Re-ordering the different tools and methods from the original Toyota Production System into these four main categories and in the process losing some tools. Tools and methods related to these four categories are also the main part of the programs tools box with its focus on 5S, Value Stream Mapping (VSM), Standardization, and Single Minute Exchange of Die (SMED). In addition, Total Productive Maintenance (TPM) tools are also used when applicable. Measuring the impact of lean on firm performance has been of interest to academics in multiple countries [10–12] and the first article that set out how to measure impact on firm performance was [13]. Since, the potential positive effect of lean on organizational performance and productivity, especially the concept of Just-in-Time has been explored in several different studies [14–20] that do show a positive effect on both operational and financial performance. 2.3 Enabling Technology Since the term was first launched as a catch-all by the German governments effort to protect the competitiveness of its manufacturing industry, Industry 4.0 has been an umbrella term that encompass new and old technology that enables automated and connected manufacturing processes [21]. A recent studies of adoption of Industry 4.0 in SMEs found that companies who emphasized the development of social capital first, stood a much higher chance of succeeding with new technology adaption [22], and that lean organizational structures support the effective adoption of Industry 4.0 technologies. The latter also found that the introduction of new technologies leads to a need for higher levels of non-technical competences, as well as new types of jobs. However, the potential of Industry 4.0 technologies in SMEs are not fully exploited, especially in utilizing the potential for more flexible and nimble information systems [23]. Finally, [24] suggest that both financing and knowledge are key constraints that hinder the adoption of enabling technology.

3 The Productivity Leap Program The PL program is built around the three theoretical cores presented above and so far sixteen companies have started the program in Total. The case studies presented here are four of the five companies that completed the pilot round of the program that started

100

E. Reke et al.

in 2019. The program is supported financially by a private foundation called “Stiftelsen Teknologiformidling” and the outside expertise is supplied by consultants from SINTEF Manufacturing, C2U, Kiwa, as well as lecturers from Chalmars and NTNU. The project organization sits in the NCE Raufoss cluster and reports to the board of LEAN Forum Norway. The participating companies commit to a payment of approximately 25.000 euros and at least 4000 h of internal time spread over 18 months. In exchange they receive 390 h of consulting as well as formal training (University certified) for at least two employees. During the program the participating companies receive visits and carry out development work in the form of workshops every fortnight the first year and then every month for the last 6 months. Workshops are formed around three main themes each with its own goals, activities, methods, and tools. Phase one of the program is designed to help the different levels of the company get an overview of as-is and develop a common understanding of the company across the different levels of the organization. It consists of a stakeholder analysis, value stream mapping, a lean self-assessment, as well as developing a future vision of the company based on guiding principles such as pull-production, built-in-quality, customer focus, standardization, and stable processes. Finally this first phase sets the mandate for the second phase pilot. The second phase is the pilot phase where the participating company develops a “model-line” to show for themselves and to the rest of the organization what is possible to achieve through improvements. This phase is supported by coaches from the program. The third and final phase of the program centers around leadership and is designed to help the participating company share and spread the learning from the pilot to the rest of the organization. The first and third phase of the program are the most structured and standardized parts of the productivity leap. The middle pilot phase is more open and flexible to allow for the uniqueness and specificities of each company and of the different pilots.

Fig. 1. The three phases of the productivity leap program

4 Methodology Since the aim of the study is to obtain an in-depth understanding of PL program implementation status and its effect on companies, as well as to answer the “how” research question using contextual data, a case study research method was chosen. In order to strengthen external validity and draw more robust results, a multiple case study approach was chosen [25, 26]. In total 16 companies have been through or are in the process of finishing the PL program. Of those, 8 companies that had finished the program were contacted and four

The Productivity Leap: Effects of an Industry Program

101

companies agreed to participate in this study. Generally, between 4 and 10 cases work well for balancing the depth of the study and the generalizability of the results [26] The unit of analysis in this study is a single company where PL program has been implemented. The primary source of qualitative data was interviews with company representatives. Eight semi-structured interviews (see Table 1) were carried out online. The interviews were recorded, then transcribed, and finally coded. Coded data from each case were combined in a spreadsheet and compared by looking at the commonalities and differences in case dimensions with regards to stages of PL program: consensus, pilot, and learning [26, 27]. In addition, quantitative operational data was obtained from the four companies, as well as publicly available financial data. The financial data was then analyzed for indications of improved performance. Especially increased productivity of customers (increase sales), assets (return on assets), materials (gross margin and inventory holdings) and people (salary as part of turnover). Table 1. Semi-structured interview questions Theme

Consensus

Pilot

Performance

Question 1

What did you do to achieve consensus

Describe the pilot

What have you done to spread the learning from the pilot

Question 2

What did you achieve in this phase?

What was the biggest take-away from the pilot?

Have you achieved any results from the program (operational or financial)

Question 3

What did you learn from the consensus process

What did you learn from the pilot?

5 Findings Case A is the largest company in this study with a turnover of 265 million Norwegian kroners (MNOK) in 2018, the year before the program commenced. Employing approximately 130 employees. It casts and machines products from iron and their biggest market is the Norwegian public sector. Its processes are machine intensive with several stages of machining applied to form the finished product. Case B is the second largest company in this study with a turnover of 251MNOK in 2018. It employs approximately 110 people, and its main markets are oil and gas, wind power, cranes, and other heavy engineering equipment. It sells both standard parts as well as parts engineered to order. The company is part of a larger multinational corporation. Its products are manufactured from steel and come in a multitude of different variations depending on the needs of the customers. Its processes are highly automated and use machining to form its products. Case C is the third largest company of the study with a turnover of 208MNOK in 2018. It employs approximately 110 people. The company builds bespoke add-ons to

102

E. Reke et al.

trucks, mainly in the Norwegian market, and its production processes are mostly based on manual labour. Case D is the smallest company in this study with a turnover of just 20MNOK in 2018. In 2021 it employeed 39 people. This company also makes products from machining processes. However, the main material used is aluminum. Their main manufacturing processes are casting, and machining. Most products are built to order but they can also engineer bespoke parts and products based on customer specifications. The companies are located in different parts of Norway, from the west coast to the eastern inlands. As part of the PL program the companies received the same amount, and the same type of coaching. Three of the case companies started the PL program in Q42018 and finished Q12020 and the last case company started in Q42019 and finished in Q12021 The program has three main stages: 1) creating consensus, 2) developing a pilot, 3) spreading the learning from the pilot to the rest of the company. First Stage: The data from the consensus stage suggests that this was helpful for the companies. All cases found that different parts of the organization had a different understanding of the company, the challenges they faced and its current state. The informants reported that this was an eye-opener to them. In addition, they reported that, inside the company, they learned and discovered new things about themselves. The consensus stage also helped improve communication and information flow. Pilots were then selected based on input from the first stage and from discussions with the coaches. Case B and D chose pilots dealing with multiple processes, while case A and C had more focused pilots studying a specific operation (A) or machine (C). Pilot Stage: All the companies reported on an improved work environment from the pilot such as a cleaner workplace, better planning, better understanding of work. They also reported learning lean tools or practices such as VSM, 5S, TPM, SMED, A3 Problem-solving, and so forth. In terms of improved productivity, two companies reported improved safety metrics, one company could document improved quality, two reported reduced lead-times, and finally all the three machining companies reported increased OEE. However, we were not able to verify these numbers independently. In terms of people engagement, a clear target for the program, two of the companies found it hard to involve everyone from operators to managers in the pilot, while the two others reported the involvement of operators as a success factor. Third Stage: Based on the data analysis from the interviews, the last stage proved to be the toughest for the case companies. Two cases reported that the Covid pandemic hindered them in continuing after PL. One case found it difficult to continue once the coaches left and one found it difficult in general to continue with improving productivity. The application of enabling technology was stated as part of the program goals, however we found no evidence of coaches or companies working with or benefiting from enabling technologies. Since the focus of the interventions have been basic lean tools, it was not possible to draw any conclusions on the effects of implementing enabling technologies. Be that as it may, our findings show that both the case companies successfully implemented both the consensus and the pilot stage of the program in accordance with the Norwegian model of worker participation and involvement and by applying lean tools. While at the same time, all four cases found it hard to roll-out the tools and

The Productivity Leap: Effects of an Industry Program

103

methods along with the productivity improvements across the organization in phase 3 of the program. See Fig. 1 on the next page for a more detailed overview of the main effects reported in each stage of the program for each case (Table 2). Table 2. Summary of the main findings related to the effect of each stage of the program Theme

Case A

Case B

Case C

Case D

Phase 1 Consensus

Recognized different points of view, importance of broad involvement

Improved communication and information flow. Gave a better understanding of where the challenges are in the company

Different functions had different ideas on how to improve the company. Recognized the different points of view

Better understanding of the overall process, Improved communication, and information flow

Phase 2 Pilot

Learnt different lean tools and saw the effect of involving everybody in improving the process, increased productivity of the pilot process

Managed to develop a foundation for further work on improving processes. Increased cross-functional collaboration. Understood the importance of being on the shop-floor

Chose a department with a poor work environment, in hindsight not the best place to start

Increased machine efficiency due to improvements in flow and 5S. Gave a better overview of processes. Also better planning

Phase 3 Roll-out

Stopped due to Corona, afterwards have been able to spread some of the tools (such as 5S and stand-up meetings) to other parts of the company

Difficult to get continuity as many people who were part of the pilot have since left the company. Have started to develop a “X-way” for lean, and have continued to work on creating an improvement culture

Struggled with roll-out due to corona and loss of income in that period which lead to lay-offs. Continues with daily management for senior management team

Face unique challenges when implementing different types of processes. Different degrees of difficulty. Difficult to keep up continuity without outside support

(continued)

104

E. Reke et al. Table 2. (continued)

Theme

Case A

Case B

Case C

Case D

Overall effects

Reduced lead-time and increased sales, reduced work-in-process stock

Not much data on overall effects from the interviewees. However document obtained suggest decrease in sick-leave, reduced scrap and reduced unplanned stops

Improved meeting structure and daily management meetings, increased focus on flow in the production

Better flow between departments. Better communication between departments. More systematic in how work is done

Operational Results: In addition to the interviews covering the three stages of the program, data was collected on the case companies’ operational performance and which lean tools were applied in the different cases. For the lean tools, 5S, Value Stream Mapping (VSM) and A3 were applied in all four cases. In addition, SMED was applied in three of the cases. Two cases reported some improvements in absenteeism, and one reported a reduction in noise levels in the factory, which could indicate a safer work environment. The companies did not report any data for quality except for Case B which reported a decrease in scrap. In terms of lead-time data, two of the companies had measured the improvements reported in the interviews and one of those reported a dramatic decrease in lead-time, from eleven to two days. Finally, the three companies with machining equipment reported an increase in OEE, although the numbers could not be verified by the researchers. These findings are summarized in Table 3. The operational data is a comparison of before and after with a baseline year of 2017 for all the three companies compared with data from the year the program ended for the cases (2020 and 2021). Financial Performance: In terms of financial improvements, all four cases improved their revenue over the period from a 2017 baseline, when compared with the overall growth of revenue in Norwegian industry1 (23% from December 2017 to December 2021), three of the four companies grew their turnover faster than the average. Three out of four increased their return on assets, all four reduced their inventory holding days, and three of four reduced their salary to sales ratio. Indicating that some of the operational effects did translate into financial performance (See Table 3 below for a summary of the numbers.) for the companies involved. However, the quantitative data set is not large enough for any conclusions to be drawn. As such further data collection and analysis are required to fully answer RQ2 in terms of the effects the program had on financial performance (Table 4). 1 https://www.ssb.no/energi-og-industri/industri-og-bergverksdrift/statistikk/omsetning-I-olje-

og-gass-industri-bergverk-og-kraftforsyning.

The Productivity Leap: Effects of an Industry Program

105

Table 3. Operational results Theme

Case A

Case B

Case C

Case D

Lean-tools

5S, VSM, A3, SMED, Work Standards

TPM, VSM, 5S, SMED, A3, Fishbone

5S, SMED, TPM

5S, VSM, A3

Safety

Reduced noise-levels

Reduced sick-leave from 9% to 5,5%

No data

Sick-leave from 4,5% to 4%

Quality

Not reported

Reduced scrap from 12,9% to 5,7%

No data

No data

Lead-time

From 11 days to No data 2 days

No data

From 10 days to 8 days

OEE (Overall Equipment Efficiency)

From 37% to 67%

Reduced From 43,47% to unplanned stop 74% time from 30% to 15%

Not Applicable

Table 4. Financial Results Theme

Case A

Case B

Case C

Case D

Revenue (increased customer productivity)

From 224MNOK to 330MNOK (47% improvement)

From 221MNOK to 281MNOK (28% improvement)

From 172MNOK to 208MNOK (14% improvement)

From 19MNOK to 31MNOK (63% improvement)

Return on Assets From 11% to (increased assets 13% (18% productivity) improvement)

From −16% to − From 2% to −2% From −10% to 8% (50% (200% decrease) −2% (80% improvement) improvement)

Inventory From 223 days to From 192 days to From 77 days to holding (material 162 days (27% 122 days (36% 59 days (23% productivity) improvement) improvement) improvement)

From 294 days to 147 days (50% improvement)

People productivity (salary as part of revenue)

From 82% to 67% (18% improvement)

From 44% to 44% (no improvements)

32% to 24% (25% improvement)

From 32% to 33% (3% down)

6 Discussion The results from the interviews show that the most successful parts of PL program are consensus creation and pilot development. Pilot development included finding bottlenecks in the production processes and process improvement. All companies emphasized

106

E. Reke et al.

that they came to a common understanding of company’s processes and goals (consensus creation), which lead to better communication and information sharing. Work on pilots showed positive operational results, even though not all the results were possible to quantify. Additionally, the direct impact of the PL program on the overall financial results of the company is challenging to prove, but the timing of PL program implementation coincides with timing of financial improvements. Additionally, the pandemic period was turbulent in economic terms and has probably affected the case companies differently. Therefore, it was difficult to compare the financial improvements in the case companies. It is however interesting to note that the one company that mainly uses manual labor processes is the company that did not experience any positive operational, nor financial effects from participating in the program. There could of course be many reasons for this, however it could be worth exploring further and for example compare with coaches’ industry experience and competence. The success of consensus creation and pilot development could also be explained by the fact that both processes were controlled by external coaches with a lot of experience. After the program ended and coaches left, the work with lean slowed down considerably. Some of the companies stated that this is due to the pandemic or layoffs in the company, but it could be also because companies were not ready to work on their own without coaches present to help them. PL program promises “long-termism and utilization of enabling technologies” [28]. The results show that long-term results were limited since companies struggled to spread the learning to the rest of the company. All the companies stated that they wanted to pick up speed and continue working with Lean since they saw positive results from the pilots. Utilization of enabling technologies never came up as part of our data collection. None of the companies mentioned implementation or utilization of enabling technologies because of the program, although one did mention investing in a new ERP system.

7 Conclusion, Implications and Limitations This paper reported on the results of a multiple case study investigating PL program implementation. Through the combination of qualitative and quantitative analysis, the study showed that results of PL implementation across companies are similar. All companies were successful with the first two steps of PL program – consensus building and pilot development. All companies struggled with continuity of lean use and spreading of the knowledge across the company in the third phase, and especially when there was no one to guide them. None of the companies reported the use of enabling technologies which then questions why it is mentioned as part of the program. For the design of industry programs like PL, we therefore suggest that they take into consideration how companies can become more self-guided, and also suggest that this could be of interest to researchers to explore further why companies struggle when the coaches leave. Even though the qualitative data analysis points towards some effects on operational and financial performance for the four case companies, it is difficult to determine if the cause of these improvements are due to the improvements on the shop floor, or if there is a relationship between improved operational performance and improvement of financial performance. Therefore it is not possible to answer RQ2 with the dataset

The Productivity Leap: Effects of an Industry Program

107

derived and analysis done so far. However, this is an interesting area of research worth further exploration. There are some limitations with this study especially related to the analysis of quantitative data. This analysis was done to give an indication of what effects one might expect and the findings are in line with previous findings on operational performance [13] and financial performance [14]. As such it should be of interest to researchers to further explore this line of research in the future.

References 1. Olsson, M.: Effektutvärdering av Produktionslyftet - Fas 1: 2007-2010, p. 40 2. Andersen, T.M.: The Nordic Model: Embracing Globalization and Sharing Risks. Taloustieto Oy, Helsinki (2007) 3. Bergene, A.C., Hansen, P.B.: A historical legacy untouched by time and space? The hollowingout of the Norwegian model of industrial relations. NJWLS 6, 5 (2016). https://doi.org/10. 19154/njwls.v6i1.4907 4. Gustavsen, B.: The Nordic model of work organization. J. Knowl. Econ. 2, 463–480 (2011). https://doi.org/10.1007/s13132-011-0064-5 5. Womack, J.P., Jones, D.T., Roos, D.: Machine That Changed The World. Harper Perennial, New York (1990) 6. Womack, J.P., Jones, D.T.: Lean Thinking. Simon & Schuster UK Ltd., London (2003) 7. Holweg, M.: The genealogy of lean production. J. Oper. Manag. 25, 420–437 (2007). https:// doi.org/10.1016/j.jom.2006.04.001 8. Madsen, D.Ø., Storsveen, M., Klethagen, P., Stenheim, T.: The diffusion and popularity of lean in Norway: an exploratory survey. Cogent Bus. Manag. 3, 1–22 (2016). https://doi.org/ 10.1080/23311975.2016.1258132 9. Rolfsen, M.: Lean blir Norsk. Fagbokforlaget (2014) 10. Rahman, S., Laosirihongthong, T., Sohal, A.S.: Impact of lean strategy on operational performance: a study of Thai manufacturing companies. J. Manuf. Technol. Manag. 21, 839–852 (2010). https://doi.org/10.1108/17410381011077946 11. Taj, S., Morosan, C.: The impact of lean operations on the Chinese manufacturing performance. J. Manuf. Technol. Manag. 22, 223–240 (2011). https://doi.org/10.1108/174103811 11102234 12. Camuffo, A., Gerli, F.: The complex determinants of financial results in a lean transformation process: the case of Italian SMEs. In: Berger, E., Kuckertz, A. (eds.) Complexity in Entrepreneurship, Innovation and Technology Research. FGF Studies in Small Business and Entrepreneurship, pp. 309–330. Springer, Cham. (2016). https://doi.org/10.1007/978-3-31927108-8_15 13. Karlsson, C., Åhlström, P.: Assessing changes towards lean production. Int. J. Oper. Prod. Manag. 16, 24–41 (1996) 14. Fullerton, R.R., McWatters, C.S., Fawson, C.: An examination of the relationships between JIT and financial performance. J. Oper. Manag. 21, 383–404 (2003). https://doi.org/10.1016/ S0272-6963(03)00002-0 15. Fullerton, R.R., Kennedy, F.A., Widener, S.K.: Lean manufacturing and firm performance: the incremental contribution of lean management accounting practices. J. Oper. Manag. 32, 414–428 (2014). https://doi.org/10.1016/j.jom.2014.09.002 16. Fullerton, R.R., Wempe, W.F.: Lean manufacturing, non-financial performance measures, and financial performance. Int. J. Oper. Prod. Manag. 29, 214–240 (2009). https://doi.org/10. 1108/01443570910938970

108

E. Reke et al.

17. Balakrishnan, R., Linsmeier, T.J., Venkatachalam, M.: Financial benefits from JIT adoption: effects of customer concentration and cost structure. Account. Rev. 71, 183–205 (1996) 18. Shah, R., Ward, P.T.: Lean manufacturing: context, practice bundles, and performance. J. Oper. Manag. 21, 129–149 (2003). https://doi.org/10.7166/26-1-644 19. Ahmad, A., Mehra, S., Pletcher, M.: The perceived impact of JIT implementation on firms’ financial/growth performance. J. Manuf. Technol. Manag. 15, 118–130 (2004). https://doi. org/10.1108/09576060410513715 20. Modarress, B., Ansari, A., Lockwood, D.L.: Kaizen costing for lean manufacturing: a case study. Int. J. Prod. Res. 43, 1751–1760 (2005). https://doi.org/10.1080/00207540500034174 21. Müller, J.M., Buliga, O., Voigt, K.-I.: Fortune favors the prepared_ how SMEs approach business model innovations in Industry 4.0. Technol. Forecast. Soc. Change, 2–17 (2018). https://doi.org/10.1016/j.techfore.2017.12.019 22. Agostini, L., Nosella, A.: The adoption of Industry 4.0 technologies in SMEs: results of an international study. MD 58, 625–643 (2019). https://doi.org/10.1108/MD-09-2018-0973 23. Moeuf, A., Pellerin, R., Lamouri, S., Tamayo-Giraldo, S., Barbaray, R.: The industrial management of SMEs in the era of Industry 4.0. Int. J. Prod. Res. 56, 1118–1136 (2018). https:// doi.org/10.1080/00207543.2017.1372647 24. Masood, T., Sonntag, P.: Industry 4.0: adoption challenges and benefits for SMEs. Comput. Ind., 103–261 (2020). https://doi.org/10.1016/j.compind.2020.103261 25. Yin, R.K.: Case Study Research Design and Methods. SAGE, London (2003) 26. Eisenhardt, K.M.: Building theories from case study research. Acad. Manag. Rev. 14, 532–550 (1989). https://doi.org/10.5465/amr.1989.4308385 27. Miles, M.B., Huberman, A.M., Saldaña, J.: Qualitative Data Analysis: A Methods Sourcebook. Sage Publications (2018) 28. Om ProduktivitetsSpranget. leanforum. http://www.leanforumnorge.no/produktivitetsspra nget/produktivitetsspranget/om-oss/om-produktivitetsspranget. Accessed 1 Dec 2022

Integrating Smart Manufacturing to Lean: A Multiple-Case Study of the Impact on Shop-Floor Employees’ Autonomy and Empowerment Thomas Bortolotti1(B) , Stefania Boscari1 , Etta Morton1 , and Daryl Powell2 1 University of Groningen, PO Box 800, 9700 AV Groningen, The Netherlands

[email protected] 2 SINTEF Manufacturing, Horten, Norway

Abstract. This study investigates the effects of Smart manufacturing implementation on shop-floor employees’ autonomy and empowerment in Lean companies, an area that has largely remained unexplored in the existing literature. A theoretical framework based on the Job Characteristics Model is used to assess the impact of ‘soft’ Lean practices, such as teamwork and group problem-solving, on employees’ job perceptions. The framework is then used to assess the effects of Smart manufacturing adoption, in terms of changes in shop floor employees’ perceptions. To this aim, a multiple case-study methodology is employed, collecting data from four large manufacturing companies with established Lean systems and varying degrees of Smart manufacturing adoption, as well as a consultant company specializing in their integration. This study’s findings show that Smart manufacturing supports job rotation in the shop-floor. In relation to employees’ autonomy, it supports faster and more informed decisions, as well as (Lean) practices related to employee controls, with no significant alteration of problem-solving. The speed of process feedback has increased with the use of Smart Manufacturing technologies, while other forms of feedback appear to be unaffected at this time, such as in-person coaching and person-to-person feedback. Overall, these findings support that many of the “soft” Lean practices are maintained when adopting Smart manufacturing, which the latter being adapted to support the existing Lean systems. By showcasing the effects of Smart manufacturing on shop-floor employees’ work, our insights provide valuable information for practitioners who are considering implementing Smart manufacturing in their organizations. Keywords: Lean · Smart manufacturing · Industry 4.0

1 Introduction Lean is a well-established managerial approach aimed at reducing costs and increasing flexibility and responsiveness by eliminating non value-adding activities [1, 2]. However, companies need to create an engaged and flexible workforce by utilizing ‘soft’ © IFIP International Federation for Information Processing 2023 Published by Springer Nature Switzerland AG 2023 E. Alfnes et al. (Eds.): APMS 2023, IFIP AICT 689, pp. 109–124, 2023. https://doi.org/10.1007/978-3-031-43662-8_9

110

T. Bortolotti et al.

Lean practices to achieve these benefits, such as small-group problem solving, continuous improvement and employees’ training [3]. Soft Lean practices offer autonomy and responsibility for work, leading to better individual performance, and consequently, higher organizational performance [4, 5]. Smart manufacturing, the underlying technology of the concept of Industry 4.0 [6], has gained prevalence [7]. It heavily focuses on using technology to automate processes, such as MES systems, product tracking and flexible automation [6], which could ultimately leads to a factory that runs without workers [8]. Previous research has examined the combination of Smart manufacturing and Lean practices [7, 8], demonstrating their potential to enhance performance across a wider spectrum of operational outcomes [7]. However, as Smart manufacturing emphasizes the use of autonomous technologies, questions arise about how companies can implement soft Lean practices that empower workers in this context. Therefore, it is crucial to examine the compatibility of these two management strategies and how they can be integrated to create a sustainable advantage for companies. An example of how Smart manufacturing technologies impact Lean is the use of real-time planning and self-optimization in production plants, allowing for faster analytics and problem-solving, thereby creating substantial changes in plant operations [9]. However, autonomy is a critical aspect of an engaged and motivated workforce [4], and higher autonomy levels lead to higher Lean commitment and better overall corporate performance [10]. Continuous improvement on the shop floor cannot occur without the involvement of employees [11]. As new technologies are implemented to increase autonomy, it raises concerns about how the essential autonomy of shop floor employees will be affected by the increased use of Smart manufacturing technologies. Although there is evidence that Smart manufacturing and Lean can offer benefits when used together [7], very little has been discussed about the effects of increased use of Smart manufacturing technology on the implementation of soft Lean practices, especially concerning the autonomy and empowerment of employees on the shop floor. This paper aims to investigate these effects by examining the research question is: In Lean companies, how does the implementation of Smart manufacturing affect the autonomy and empowerment of shop floor employees? As shop floor processes become more automated and decisions more autonomous, questions arise regarding, for example, whether shop floor employees become more integral to the process and management of these machines, or if the tighter control over the machines translates into tighter controls over the workers. This paper uses a multiple case-study methodology to address the research question, collecting data from 4 large manufacturing companies with an established Lean system and varying degrees of Smart manufacturing adoption, as well as a consultant company specializing in their integration. This allows for a deep and novel understanding in the existing literature of the implications of the advent of Smart manufacturing for shop floor employees in Lean companies. We also derive interesting implications for practitioners.

Integrating Smart Manufacturing to Lean

111

2 Theoretical Background 2.1 Lean Lean management, originating from the Toyota Production System, focuses on waste reduction and resource utilization [1]. Lean is an integrated socio-technical system [5]. That is, the hard, technical and analytical practices alone could not provide a competitive advantage; success hinged on also adopting soft Lean practices such as continuous improvement and small group problem-solving [3]. Benefits from Lean are achieved by combining multiple elements, such as decentralized decision-making through teamwork and group problem-solving fostering more ownership and better improvement suggestions from employees [12]. Lean practices can be classified into four bundles: Human Resource Management (HRM), Total Quality Management (TQM), Just-in-Time (JIT), and Total Preventative Maintenance (TPM) [13]. HRM is the main focus of this paper. It focuses on employee management, with cross-trained flexible employees and self-directed teams as key concepts [13]. HRM practices form the foundation for sustainable competitive advantages offered by Lean [3]. 2.2 Smart Manufacturing and Its Integration with Lean Smart manufacturing emerged in 2011 and encompasses advancements in Cyber Physical systems, Internet of Things, Internet of Service, Robotics, Big Data, Cloud Manufacturing, and Augmented Reality [14]. The primary goal of Smart manufacturing is to drive strategic innovation within the manufacturing industry and increase automation to minimize human intervention in factories [8]. Prior studies have analyzed the integration of Smart manufacturing with Lean practices [7, 8], showing that together, they can lead to better performance across a broader range of operational outcomes [7]. However, the implications for shop floor employees are largely unclear. While some envision fully autonomous factories in the future thanks to Smart manufacturing technologies, it is currently more practical for automation to assist and enable shop floor workers, with innovations like touch-sensitive robots ensuring worker safety [15]. Moreover, workers in a Lean production system not only assemble products but also play a pivotal role in identifying issues and improving processes [15]. This role is hardly replaceable by machines. Accordingly, some scholars suggest that the role of shop floor employees will evolve rather than be replaced by technology. Nevertheless, some new practices of Smart manufacturing might conflict with old Lean ones [16]. Organizations may adapt the new practices to fit the existing system. This can result in the preservation of the original system, despite the integration of new technologies [7]. Ultimately, the harmonization of Lean practices and Smart manufacturing will likely redefine the roles of shop floor workers and their interactions with technology, leading to more efficient and effective production processes.

112

T. Bortolotti et al.

2.3 A Theoretical Framework: The Job Characteristics Model Hackman and Oldham’s Job Characteristics Model (JCM) [17, 18] explains how core job dimensions impact the critical psychological states of workers on the shop floor, thereby influencing their motivation and job performance. According to the JCM, objective changes in a job affect workers’ perceptions of three critical psychological states: experienced meaningfulness of work, responsibility for outcomes of work, and knowledge of results of the work [19]. These psychological states are broken down into various measurable dimensions, including [17, 19]: • skill variety: the degree to which a job allows an employee to use different skills by carrying out a variety of tasks in their work; • task identity: the extent to which a job can be identified as a “whole” task or piece of work; • autonomy: how much a job allows the employee freedom, independence, and choice in how and when the work is done; and • feedback: the information employees receive about the effectiveness of their performance and can come from the process itself, supervisors, or other employees. The JCM can be utilized to gauge the impact of Lean on employees’ job perceptions, which can then be compared to the effects of Smart manufacturing to analyze the changes in perceptions. When reviewing the Lean literature, various studies contribute to the understanding of how Lean practices influence various job dimensions, showing that these dimensions are interrelated and work together to create the overall perception of the job [4]. Table 1 summarizes key contributions in the Lean literature on the relevant four core job dimensions. As Smart manufacturing technologies are implemented, job characteristics are expected to change, with higher complexity jobs gaining more autonomy and skill variety. Information and communication systems can automate routine tasks, freeing up employees to focus on diversified work and learning opportunities. This shift is expected to result in a decrease in low-skilled work and an increase in high-skilled work. Even as lower-level tasks are automated, workers remain essential for addressing out-of-control systems and driving continuous improvement. Automation of monotonous tasks allows employees to engage in higher-level, more creative work, leveraging Smart manufacturing to enhance their skills and abilities. This shift positions workers as strategic decisionmakers and flexible problem-solvers in increasingly complex environments. Finally, Smart manufacturing technologies also facilitate better communication between operators and machines, enabling real-time notifications and immediate feedback through smartphones or tablets. This improved communication allows for more efficient feedback, empowering employees on the shop floor and fostering greater involvement in the production process. Overall, these statements are largely speculative, thus further research is needed.

Integrating Smart Manufacturing to Lean

113

Table 1. Effects of Lean on Job Characteristics Model Dimensions Article

Skill Variety

[20] Flynn, Sakakibara & Schroeder 1995

Problem-solving and workers moving to help bottlenecks

Take Identity

[21] Powell 1995

Autonomy

Feedback

Decentralized management of processes to find solutions

Information Feedback is needed

Employee empowerment correlated with performance

[22] Delbridge, Lowe & Oliver 2000

Handled day-to-day and quality tasks only. No upskilling. Involved in problem-solving

Control stays with team leader. However, employees involved in improvement activities

Leader sets pace for work and handles any disputes or disagreements

[23] Jackson & Mullarkey 2000

Higher skill variety due to changes in work structure

Job enlargement

Higher production responsibility, more process autonomy, higher

More interaction with team members leading to more feedback. More pressure from managers to perform

[13] Shah and Ward 2003

Cross-trained employees, Job rotation

Job enlargement

Self-directed teams, problem-solving groups

Quality management programs, measurement capability, inventory management

[10] Angelis et al. 2011

Workers should be Individual involved in both feedback to lessen “on the line” work ambiguity and in improvement projects, multi-skilled workers, job rotation

Higher autonomy means higher worker commitment. Involvement in improvement programs improves commitment

Supervisor training to teach them not to blame workers but help the workers. Individualized feedback on displays. Task support from both leaders and coworkers increase Lean commitment

114

T. Bortolotti et al.

3 Methodology While Lean is well-established, Smart manufacturing is still an emergent topic that had yet to be thoroughly investigated, especially in relation to how it affects the autonomy and empowerment of shop floor employees. In light of the limited existing research, an inductive case methodology is deemed the most suitable to address the research question [24]. Case studies are particularly effective for exploring “how” and “why” questions, and they offer valuable insights for generating new theories and ideas [24, 25]. They allow for the development of an in-depth understanding of a phenomenon, including the contextual conditions under which it occurs [24], here different levels of Lean and Smart manufacturing technology implementation. To strengthen the validity and generalizability of the findings, multiple cases have been selected for analysis [25]. The unit of analysis is the firm, as Smart manufacturing and Lean practices can differ significantly between organizations. 3.1 Case Selection The research setting focused on large manufacturing companies. Due to the capitalintensive nature of Smart manufacturing, which can be a barrier for smaller companies [26], the study targeted large multinational companies with the financial capacity to invest in these technologies. All of the selected companies are multinational corporations employing over 1,500 individuals. While they exhibit high levels of Lean implementation, their levels of Smart manufacturing implementation vary. By including varying levels of Smart manufacturing implementation, we address theoretical replication. General information about the selected companies is included in Table 2. Table 2. Cases and interviewees. Case

Industry

Country

Interviewee

Interviewee position

1

Automotive

Netherlands

1A

Quality Manager

2

Aerospace

Netherlands

2A

Lean Expert

2B

Process Engineer

2C

Smart manufacturing Project Leader

3A

Lean Expert

3B

Digitalization Team Member

3C

Digitalization Team Member

3

Manufacturing

Germany

3D

Lean Expert

4

Automotive

Germany

4A

Smart manufacturing Project Manager

Consultant

–

–

C1

Operation Consultant

Integrating Smart Manufacturing to Lean

115

3.2 Data Collection To ensure diverse perspectives and minimize bias [25], managers from both Lean and Smart manufacturing departments were chosen whenever possible. The data was primarily collected through semi-structured interviews, utilizing an interview protocol to ensure the reliability of the case studies (available upon request). These interviews gathered information about the companies’ Lean implementation, the types of Smart technologies used, and the impact of these technologies on employees’ decision-making abilities and other job characteristics related to autonomy. Moreover, observations of the facilities were also used to enhance the understanding gained through the interviews. To further support data triangulation, an additional interview was conducted with an operations consultant who is an expert in integrating Lean and Smart manufacturing. The interviews lasted between 1 and 4 h. The interviewees and their positions within the companies are listed in Table 2. Each interview was recorded with the interviewee’s permission. 3.3 Data Analysis After the interviews were conducted, they were transcribed and combined with observations to be analyzed using an inductive approach. Since the inductive approach does not have first-order codes already identified [24], initial codes were discovered by reading through interview transcripts. Codes were developed by examining the main concepts, such as employee autonomy, skill variety, task identity, employee feedback (see Sect. 2), and any changes in the employee’s environment due to Smart manufacturing. Cases were first analyzed individually to identify any internal patterns. An emphasis was put on shop floor employee autonomy (e.g., what decisions employees were able to make, what responsibilities they had, etc.) in Lean, and how shop floor employees have been affected by the changes resulting from Smart manufacturing implementation. This analysis contributes to the internal validity of the study. Afterwards, a cross-case analysis was conducted to look for patterns in the overall data and further the external validity of the findings. To enhance internal validity and reliability, a codebook was used (available upon request).

4 Analysis and Discussion This section reports the findings based on the cross-case analysis and comparison of the four analyzed companies. Although the job characteristics varied in each company for many reasons, here we focus on some areas that all interviewees agreed had been affected by Smart manufacturing. Table 3 summarizes the case comparison. Below, findings are discussed in light of existing literature.

116

T. Bortolotti et al. Table 3. Case comparison.

Case

1

View of Smart manufacturing related to Lean Job Rotation

3

4

As a tool for Lean Enabler for Lean

Next step after Lean

Next step after Lean

Yes

No

Yes

Yes

Process autonomy Some, teams set standards

Low

Low

Low

Increased controls No due to Smart manufacturing

Yes

No

No

Structure for problem-solving

Individual and Team

Individual

Individual - high level only; Lower automated

Team

2

4.1 Skill Variety and Task Identity All the interviews revealed that the integration of Smart technologies significantly reduces task complexity, as information accessibility is amplified. This enhancement enables swifter, more reliable decision-making for shop floor employees. Moreover, the implementation of these cutting-edge technologies streamlines processes, consequently reducing job complexity and stress among employees. Although job roles evolve and necessitate increased technical acumen, they will not become more complex, partly due to the information consolidation facilitated by Smart technologies. Additionally, real-time data provision ensures that only pertinent details are displayed to shop-floor employees. Three of the four companies highlighted the continued importance of job rotation in the era of Smart manufacturing. As stated by Interviewee 1A: “The tasks will not become more and more complex, and they [employees] will still have to rotate. I think that’s a precondition. Also, from a health point of view.” And Interviewee 2C: “…in the beginning it was a lot of information. Now it’s reduced to the specific set that you need for the activities. It’s not a matter of overload. It’s more focused.” The literature on Smart Manufacturing corroborates the idea that technology should be employed for process simplification and enhanced information availability [26, 27]. Smart Manufacturing leverages real-time data to automate lower-level decisions and simplify information for workers, thereby allowing them to focus on other tasks [26]. These advanced technologies enable shop floor employees to access information anytime, anywhere, and receive relevant data in real time [28]. This leads to the notion that job rotation will be facilitated by the increased and concurrently simplified information. Interviewee 3A emphasized this point: “Because what you actually do with helping people and complexity, giving the right information at the right point, actually you make, at least I would say, the work easier. Let’s say you instantly get the information on what to do, it’s not like that you have to be trained for weeks and months for a certain task, because the system will help you and tell you what to do.” This supports the idea that

Integrating Smart Manufacturing to Lean

117

information is not only more accessible and understandable, but also aids employees in learning additional tasks while automating or eliminating monotonous tasks, allowing them to concentrate on higher-level tasks. 4.2 Autonomy Autonomy refers to the degree of freedom, independence, and choice granted to employees [19]. Some scholars distinguish between two primary types of autonomy: choice and responsibility [4]. In Lean companies, employees are not granted greater choice autonomy, but rather increased responsibility through problem-solving and process improvements [22]. This perspective was echoed by the interviewees, who emphasized the importance of autonomy but found the autonomy primary types insufficient to describe their experiences. Interviewees indicated that autonomy was mostly granted through problem-solving but also, in some cases, by involving employees in the Smart Manufacturing change process. Consequently, this study proposes redefining employee autonomy to better align with Lean and Smart Manufacturing contexts, encompassing employee empowerment, problem-solving, and involvement in the Smart Manufacturing change process. The discussion will first address employee empowerment and its impact on basic decision-making, followed by an exploration of problem-solving aspects and effects. Finally, the analysis will consider employee involvement in the Smart Manufacturing change process. 4.2.1 Employee Empowerment Empowerment can be defined as enabling employee decision-making and increasing their involvement in processes [21]. All companies agreed on the importance of allowing employees to make decisions. The interviewed consultant shared the view that Smart technologies and the Lean principle of decentalized decision making complement each other: “Decentralized decisions are more likely to be better, because the people have more information or can evaluate the situation better than a centralized person who just has portions of the information. So that’s why I would say yes, the Lean principles and Smart manufacturing are matching…” (C1). In Case 2 and Case 3, employees were expected to make the same number of decisions, but with the introduction of Smart manufacturing, these decisions were made at greater speed and accuracy. However, Case 4 suggested that decision-making on the shop floor was constrained by Smart manufacturing. Despite this limitation, the interviewee concurred that Smart Manufacturing facilitates faster decision-making “…there is a very strict process to follow. I don’t see that they have to [make] that many decisions. What they have to do is to perform the right tasks and also perform the tasks right.… They have to look … at the screen and then immediately know what to do. So that is kind of a decision, which part to pick then and which task to perform. This is definitely supported by digital information.” (4A). These case findings align with the literature on Smart Manufacturing, which highlights the increased information availability and accessibility offered by such technologies [14, 26]. Furthermore, employees are poised to become “knowledge workers” with enhanced involvement and understanding of the external environment [29].

118

T. Bortolotti et al.

The cases revealed varying degrees of employee control implementation on the shopfloor. Case 2 and Case 3 displayed contrasting positions on using employee data. Case 2 shows that Smart technology is used to implement stricter controls on employees: “It’s helping me add more hard-controls to ensure people with the right certifications are doing the job. Currently, when I ship my products, I need to have all the papers… there is a lot of administration.” (2C). Smart manufacturing enables Case 2 to more easily enforce the legal requirements and to manage the information they must provide when selling their products. In contrast, Case 3 expressed disinterest in individual worker data, focusing instead on higher-level performance data and KPIs. The divergent focuses of these cases can be partially explained by the legal requirements and restrictions of the companies. Case 2 operates in a heavily regulated industry, necessitating the demonstration that certified professionals perform the work in an approved manner. Case 3, however, is prohibited from using individually identifiable data. Despite these differences, both companies continue to collect data, and the introduction of Smart Manufacturing has not altered this practice. 4.2.2 Problem-Solving Problem-solving can be regarded as an indicator of an employee’s independence. Most interviewees emphasized problem-solving as an area where employees exhibit autonomy within their own domain. However, the extent to which employees were able to address problems and the resources provided varied significantly across companies. In the first three cases, employees were encouraged to resolve issues independently before seeking guidance or assistance. In Case 1, the emphasis is on team-based problem-solving before requesting help from an immediate supervisor. Workstations are organized into small teams or pairs. Interviewee 1A: “[the employee] gets the support of the team leader, … Then the team leader should solve the problem. If the team leader can’t solve the problem or it will take too much time, they have to decide what they do.” In Case 2, employees are encouraged to address their own problems. The expected behavior is described as self-empowered teams eager to resolve issues independently first. While the initial problem-solving approach was similar to Case 1, the escalation process differed. Employees could choose the person they believed is best suited to resolve the issue rather than turning to their immediate supervisor. In Case 3, individuals were expected to tackle issues independently before seeking guidance from a leader through a coaching process. Case 4 is the only instance where technology significantly impacted employee problem-solving. In this case, machines automatically detect issues and alert an expert to address them. Workers solely operate the machines without engaging in problemsolving. Machines were also able to determine required maintenance, suggesting that Smart Manufacturing may limit problem-solving autonomy. From a Lean standpoint, shop floor employees should participate in problem-solving processes. Lean production necessitates substantial human support for effective operation, as employees are expected to anticipate and prevent issues that may disrupt output [1]. Smart Manufacturing research aligns with this view, emphasizing that employees are needed to tackle abnormalities and improve systems in an increasingly complex environment [15, 26, 27].

Integrating Smart Manufacturing to Lean

119

Despite the differing opinion in Case 4, the other interviews corroborate the literature on employees’ problem-solving roles, supporting the notion that even with limited decision-making autonomy, employees are still expected to participate in problemsolving [22] within a Smart Manufacturing context. Interviewee C1 reinforced this idea: “those people who can solve problems and overlook the complexity are required, and not those who perform the simple tasks, because this will be replaced with AI and machines. So, there will be a shift more towards IT and away from the simple workers.” 4.2.3 Involvement in the Smart Manufacturing Change Process All companies recognized the value of employee feedback during Smart Manufacturing implementation, but only two companies, Case 3 and Case 4, stressed the importance of involving employees in the development phase. These companies aimed to engage employees in the Smart manufacturing transformation. In Case 4, employees were encouraged to participate in the development of Smart technologies, despite not being involved in problem-solving or decision-making on the shop floor. Interviewee 4A emphasized this saying “Where you need employee involvement is while developing the system […] looking at what individual workers need and develop software along the needs […] That is where you need the involvement of the employees, you need their feedback, you need their ideas” (4A). Similarly, Case 3 involved employees in the change process, with a focus on providing resources and understanding their requirements for optimal performance. Interviewee 3B: “… [we should] focus more on enabling and providing people with the resources they need to do the best job they can. That also involves asking them what kind of tool or solution or what is it that they need to do the best job they can do.” (3B). Both cases accentuated the importance of drawing inspiration and direction for innovation from shop floor employees. Conversely, Case 1 and Case 2 developed their technologies without direct involvement from shop floor employees. Case 2 initially attempted to engage employees but later abandoned this approach when it failed to yield the desired results. Nevertheless, employees were actively engaged in continuous improvement processes and encouraged to suggest ideas. In Case 1, the Smart Manufacturing development department operated in a separate factory, expressing a cautious approach to technology implementation to ensure employee safety. Both cases acknowledged the necessity of communication with shop floor employees but did not involve them directly in the development process. The differing approaches could be partly attributed to how the Smart Manufacturing departments perceived themselves. Case 1 and Case 2 viewed Smart Manufacturing as a tool to support or enable Lean practices: “It’s providing you with new tools, new solutions, that will solve problems that you are facing, and it will absolutely help you, but applied in a meaningful way.” (1A), while Case 3 and Case 4 regarded it as an evolution of Lean, aiming to achieve higher organizational performance. Case 4 considered Lean to be limited, suggesting Smart Manufacturing as the next level of efficiency and complexity management. In Case 3, the Smart Manufacturing department believed Lean thinking to be limited, while Smart Manufacturing encouraged rethinking entire processes and systems. Although the Smart Manufacturing departments in Case 3 and Case 4 saw themselves as separate from Lean, they still operated under Lean thinking principles.

120

T. Bortolotti et al.

Literature suggests that employee involvement leads to higher commitment and shop floor-driven changes [10, 29]. These findings confirm that this thinking persists during Smart Manufacturing implementation. Case 3 and Case 4 support the extant research [28], emphasizing the necessity of employees in Smart Manufacturing, particularly for creative tasks such as implementing and developing tools. Buer et al. [7] identified research streams representing both Smart Manufacturing supporting Lean and Lean supporting Smart Manufacturing, and this study further corroborates the presence of both in managerial practice. 4.3 Feedback All case studies indicated that employee feedback methods in Smart Manufacturing environments resembled those in Lean-only settings. Companies employed digitized shop floor boards to broadcast real-time process data to shop floor employees. Coaching from leaders served as a significant source of feedback, with decision-making autonomy granted to employees in most cases, except for Case 4. Cases 1, 2, and 3 highlighted that employees were encouraged to solve issues independently, provided with tools and support to do so. Smart Manufacturing implementation has not significantly altered this process. In Case 3, new and old technologies were used interchangeably for communication. Similarly, in Case 2, the introduction of technology had not substantially impacted feedback methods, with in-person coaching still emphasized. Although technology has been integrated into processes, overall person-to-person feedback techniques remain largely unchanged. This contrasts with literature predictions that Smart Manufacturing would enable real-time data feedback from both processes and supervisors [26, 30]. The absence of internal social networks or other informal communication channels could explain the lack of change in feedback forms. Literature anticipations may not yet be fully realized. Sanders et al. [26] expected employees to have handheld smart devices for system and technology usage, but many interviewees reported limited technology on the shop floor. This discrepancy suggests that literature predictions might be ahead of current shop floor realities.

5 Conclusion Lean is a well-established approach, while Smart Manufacturing remains in its nascent stages in both academia and practice. This study seeks to understand how Smart Manufacturing influences shop-floor employees’ autonomy in a Lean environment through managers’ interviews and shop-floor observations. Table 4 summarizes the literature and the empirical evidence from the cases, divided by job characteristics, and provides a summary of the findings of this study. Currently, technology is employed to facilitate decision-making and expedite general information dissemination [31]. However, other dimensions of autonomy vary based on the perception of Smart Manufacturing and potentially industry requirements. As more technologies are implemented, these findings may evolve.

Integrating Smart Manufacturing to Lean

121

This research contributes to the growing body of literature exploring the intersection between Smart Manufacturing and Lean [7, 9, 15, 26] by examining the impact of changes due to Smart Manufacturing implementation in Lean companies with a focus on shop-floor employees’ jobs, through the lens of the job characteristics model. The study offers a novel approach to dividing autonomy into distinct dimensions and helps managers comprehend Smart Manufacturing’s interaction with established "soft" Lean practices, particularly concerning employee autonomy. Our findings provide valuable insights not only for researchers but also for practitioners who are considering implementing smart manufacturing in their organizations. By understanding the potential effects of Smart manufacturing on shop-floor employees’ work, practitioners can make informed decisions and anticipate changes that might occur in their work processes and employee motivation. Additionally, the study confirms that many advantages of "soft" Lean practices can be preserved in a Smart Manufacturing setting. Table 4. Summary of the findings. Literature

Cases

Findings

Skill variety and task identity: Job rotation is important and Higher as simple tasks will made possible by Smart become automated manufacturing, thanks to the availability of information and simplification

Smart manufacturing supports job rotation by making jobs less complex and information more readily available

Employee empowerment: in Lean, decision-making may be limited. Together with smart, employees are expected to make decisions that are both greater in scope and executed with greater speed

Employees are expected to make more informed decisions, both better and faster, thanks to the increased availability of information

Smart Manufacturing supports faster and more informed decisions on the shop floor Smart Manufacturing further enables existing (Lean) practices related to employee controls

Problem solving: Employees must still be involved until machines become self-learning

In most cases, employees were encouraged to independently solve issues and were provided with varying levels of resources and options. However, there was one particular case where all issues were handled autonomously

Smart Manufacturing does not have a significant impact on the autonomy of problem-solving among shop floor employees

(continued)

122

T. Bortolotti et al. Table 4. (continued)

Literature

Cases

Findings

Involvement in Smart manufacturing change process: Employees should remain actively engaged in the development of new ideas, particularly by harnessing their creativity

In firms that considered Smart manufacturing as the next step, employees were actively involved in the development process. However, in firms where Smart manufacturing was seen as a tool for lean practices, the changes were communicated to employees, but they were not directly involved

When Smart Manufacturing is viewed as the next evolution of Lean, employees are more likely to participate in the development process. If Smart Manufacturing is perceived as merely a tool for Lean, employees are less likely to be involved in the development process

Feedback: feedback processes can become faster as the overall speed of communication improves

For the majority of employees, there was a strong encouragement to independently address issues. They were provided with resources and support to facilitate problem-solving. However, if they encountered challenges beyond their ability to solve, they were encouraged to seek additional support

The speed of process feedback has increased with the use of Smart Manufacturing technologies, while other forms of feedback appear to be unaffected at this time

This study has limitations. Firstly, no firms examined exhibit high levels of Smart Manufacturing implementation. As Smart Manufacturing is an evolutionary process, results may change as more technologies are developed and adopted. Secondly, the study’s scope is restricted to Western Europe, limiting the applicability of its findings. Furthermore, focusing on employee autonomy changes from a managerial perspective may introduce bias. As an exploratory study, all findings require empirical testing.

References 1. Womack, J., Jones, D., Roos, D.: The Machine That Changed The World : How Japan’s Secret Weapon in The Global Auto Wars Will Revolutionize Western Industry, 1st Harper. Harper Perennial, New York (1991) 2. Danese, P., Manfè, V., Romano, P.: A systematic literature review on recent lean research: state-of-the-art and future directions. Int. J. Manag. Rev. 20(2), 579–605 (2018) 3. Bortolotti, T., Boscari, S., Danese, P.: Successful lean implementation: organizational culture and soft lean practices. Int. J. Prod. Econ. 160, 182–201 (2015) 4. De Treville, S., Antonakis, J.: Could lean production job design be intrinsically motivating? Contextual, configurational, and levels-of-analysis issues. J. Oper. Manag. 24(2), 99–123 (2006)

Integrating Smart Manufacturing to Lean

123

5. Shah, R., Ward, P.T.: Defining and developing measures of lean production. J. Oper. Manag. 25(4), 785–805 (2007) 6. Bokhorst, J.A., Knol, W., Slomp, J., Bortolotti, T.: Assessing to what extent smart manufacturing builds on lean principles. Int. J. Prod. Econ. 253, 108599 (2022) 7. Buer, S.-V., Strandhagen, J.O., Chan, F.T.S.: The link between industry 4.0 and lean manufacturing: mapping current research and establishing a research agenda. Int. J. Prod. Res. 56(8), 2924–2940 (2018) 8. Sony, M.: Industry 4.0 and lean management: a proposed integration model and research propositions. Prod. Manuf. Res. 6(1), 416–432 (2018) 9. Arcidiacono, G., Pieroni, A.: The revolution lean six sigma 4.0. Int. J. Adv. Sci. Eng. Inf. Technol. 8(1), 141 (2018) 10. Angelis, J., Conti, R., Cooper, C., Gill, C.: Building a high-commitment lean culture. J. Manuf. Technol. Manag. 22(5), 56 (2011) 11. de Menezes, L.M., Wood, S., Gelade, G.: The integration of human resource and operation management practices and its link with performance: a longitudinal latent class study. J. Oper. Manag. 28(6), 455–471 (2010) 12. Boscari, S., Danese, P., Romano, P.: Implementation of lean production in multinational corporations: a case study of the transfer process from headquarters to subsidiaries. Int. J. Prod. Econ. 176, 53–68 (2016) 13. Shah, R., Ward, P.T.: Lean manufacturing: context, practice bundles, and performance. J. Oper. Manag. 21(2), 129–149 (2003) 14. Pereira, A.C., Romero, F.: A review of the meanings and the implications of the industry 4.0 concept. Procedia Manuf. 13, 1206–1214 (2017) 15. Rüttimann, B.G., Stöckli, M.T.: Lean and industry 4.0—twins, partners, or contenders? A due clarification regarding the supposed clash of two production systems. J. Serv. Sci. Manag. 9(6), 485–500 (2016) 16. Maguire, K.: Lean and IT—working together? An exploratory study of the potential conflicts between lean thinking and the use of information technology in organisations today. In: Chiarini, A., Found, P., Rich, N. (eds.) Understanding the Lean Enterprise. Measuring Operations Performance, pp. 31–60. Springer, Cham (2016). https://doi.org/10.1007/978-3319-19995-5_2 17. Hackman, J.R., Lawler, E.E.: Employee reactions to job characteristics. J. Appl. Psychol. 55(3), 259–286 (1971) 18. Hackman, J.R., Oldham, G.R.: Motivation through the design of work: test of a theory. Organ. Behav. Hum. Perform. 16(2), 250–279 (1976) 19. Hackman, J.R., Oldham, G., Janson, R., Purdy, K.: A new strategy for job enrichment. Calif. Manag. Rev. 17(4), 57–71 (1975) 20. Flynn, B.B., Sakakibara, S., Schroeder, R.G.: Relationship between JIT and TQM: practices and performance. Acad. Manag. J. 38(5), 1325–1360 (1995) 21. Powell, T.C.: Total quality management as competitive advantage. Strateg. Manag. J. 16(1), 15–37 (1995) 22. Delbridge, R., Lowe, J., Oliver, N.: Shopfloor responsibilities under lean teamworking. Hum. Relat. 53(11), 1459–1479 (2000) 23. Jackson, P.R., Mullarkey, S.: Lean production teams and health in garment manufacture. J. Occup. Health Psychol. 5(2), 231–245 (2000) 24. Karlsson, C.: Research Methods for Operations Management (2016) 25. Yin, R.: Applications of Case Study Research - University of Bath. SAGE, London (1993) 26. Sanders, A., Elangeswaran, C., Wulfsberg, J.: Industry 4.0 implies lean manufacturing: research activities in industry 4.0 function as enablers for lean manufacturing. J. Ind. Eng. Manag. 9(3), 811–833 (2016)

124

T. Bortolotti et al.

27. Mrugalska, B., Wyrwicka, M.K.: Towards lean production in industry 4.0. Procedia Eng. 182, 466–473 (2017) 28. Zolotová, I., Papcun, P., Kajáti, E., Miškuf, M., Mocnej, J.: Smart and cognitive solutions for operator 4.0: laboratory H-CPPS case studies. Comput. Ind. Eng. (2018) 29. Davies, R., Coole, T., Smith, A.: Review of socio-technical considerations to ensure successful implementation of industry 4.0. Procedia Manuf. 11, 1288–1295 (2017) 30. Kühl, A., Franke, J., Erll, A., Weigelt, M., Mayr, A., Grimm, S., Potzel, M.: Lean 4.0 - a conceptual conjunction of lean management and Industry 4.0. Procedia CIRP 72, 622–628 (2018) 31. Waschull, S., Bokhorst, J.A.C., Molleman, E., Wortmann, J.C.: Work design in future industrial production: transforming towards cyber-physical systems. Comput. Ind. Eng. (2019)

Applying the Value Stream Map to Streamline Energy Consumption: Analysis of an Italian Company Matteo Ferrazzi(B)

and Alberto Portioli-Staudacher

Department of Management, Economics and Industrial Engineering, Politecnico di Milano, Via Lambruschini 4, b, Milano, Italy [email protected]

Abstract. In recent years, sustainability has become an increasingly important issue for businesses. Growing awareness of the environmental impacts of business activities has prompted many organizations to adopt more sustainable strategies. Lean production techniques are increasingly being evaluated and used as a catalyst for developing better strategies for sustainable production. The literature showed that the extent of Value Stream Mapping (sus-VSM), an important technique used in lean manufacturing to identify environmental impacts and waste, was present in most studies. Its application is illustrated through a case study, conducted in a plant of an Italian company involved in the collection, treatment and final destination of end-of-life tires. The results indicate that the proposed PDCA-based approach for sus-VSM can be an effective alternative for improving the environmental performance of operations. In conclusion, this research served to provide further evidence of the application of lean methods and tools to address critical environmental issues and promote sustainability within organizations. Keywords: Value Stream Mapping · Lean manufacturing · Environmental Sustainability · Energy consumption

1 Introduction Industrial activities are one of the most significant impacts on the environment. The impact of manufacturing industries can include emissions of hazardous chemicals into the air, toxic discharges to water, and hazardous wastes deposited in landfills. These activities can have a negative impact on human health and the environment. The environmental concerns have contributed to organizations taking an active role in developing more sustainable services and production processes. Some fundamental prerequisites for achieving this goal are the use of effective strategies and operations, enabled by an energy and resource-saving approach, adequate technology, human capabilities, and an effective organizational context. Lean manufacturing is recognized for its persistent goal of eliminating waste from industrial workshops and service providers. The Lean organizational model is based on the Toyota Production System of Toyota Motor Company, which was © IFIP International Federation for Information Processing 2023 Published by Springer Nature Switzerland AG 2023 E. Alfnes et al. (Eds.): APMS 2023, IFIP AICT 689, pp. 125–136, 2023. https://doi.org/10.1007/978-3-031-43662-8_10

126

M. Ferrazzi and A. Portioli-Staudacher

able to provide an approach that allowed the company to remain competitive and even thrive in the highly competitive market [1]. The lean manufacturing has driven the provision of solutions that return more value in more effective ways using less time, inventory, energy and resource. However, current challenges require not only highly productive and responsive production systems, but also eco-efficient systems, systems that provide more value with a lower environmental impact [2]. Lean Manufacturing can be considered as a set of practices that if applied with proper criteria can help manufacturing companies achieve better production efficiency. In the article published by Bai et al. 2019 [21] Lean Manufacturing practices are divided into 4 macro Bundles: Supplier, Production planning and control, Process technology, Human resources and Customer. Within these Bundles are several Lean Manufacturing practices. There are several studies in the literature that show how the use of Lean practices can improve the operational performance of manufacturing companies [19, 20]. On the other hand, in today’s literature it has not yet been studied in depth how Lean practices can have a positive impact on performance, not only operational but also environmental. In order to have a deeper understanding of how Lean practices can foster improved eco-efficiency performance, this paper analyzes a real case study of a Lean practice application. Specifically, the aims of this paper is to analyze how the VSM can be useful in reducing energy consumption in an Italian manufacturing company. In the following section this article will exploit the existing knowledge regarding this topic. In Sect. 2 provides the background of the study, namely the concepts of Lean Manufacturing, Eco-Efficiency, and their interaction and the results obtained from the analysis of the literature review. Section 3 presents the case study in which show how the Value Stream Map was use to improve the energy efficiency in a manufacturing company. Finally, Sect. 4 presents the final results, and the Sect. 5 managerial and academic implications and possible future studies are analyzed.

2 Literature Review The following session explores what is meant by eco-efficiency and how this concept is related to the lean manufacturing approach. 2.1 Interaction Between Lean manufacturing and Eco-Efficiency Over the decades, the lean manufacturing approach has been able to adapt to different contexts and be used for different problems in companies worldwide [3]. In recent years, many researchers [4] have investigated whether, in addition to productivity goals, the lean manufacturing approach was also useful for environmental sustainability goals in manufacturing companies. The managing operations following a logic aimed at environmental sustainability has become the managers’ mission in many manufacturing companies. Effects such as rising energy prices, increasing environmental pollution and market demand for green practices have prompted companies to rethink their strategies from a green perspective [5]. Environmental sustainability initiatives consider business improvements, which aim to transform an input into an output using as few resources

Applying the Value Stream Map to Streamline Energy Consumption

127

as possible [6]. The first and most important positive outcome for companies that apply green thinking is reducing their environmental impact. Eco-efficiency can be considered one of the possible solutions to the need to reduce the gap between economic development and environmental protection. Eco-efficiency is achieved by the delivery of competitively priced goods and services that satisfy human needs and bring quality of life, while progressively reducing ecological impacts and resource intensity throughout the life-cycle to a level at least in line with the earth’s estimated carrying capacity [7]. The notion of eco-efficiency, along with other related and unrelated concepts, represents an active set of resources that can be utilized to solve the most significant effects of a number of urgent challenges facing modern civilizations. The eco-efficiency concept is grounded on the safeguarding of nature by providing conservation on its resources, namely with [8]: a) b) c) d)

reduction of material intensity; reduction of energy intensity; reduction of the quantity and level of toxicity of substances; promotion of closed cycles and the use of significant end-of-life strategies; e) promotion of renewable, abundant, and local resources; e) improvement of product durability; f) intensification of service use.

Innovation efforts at the product and production process levels as well as a unique viewpoint on the assessment of a product’s environmental performance are necessary to deliver more value with less effect. The eco-efficiency represents a possible solution to the need to reconcile economic development with environmental protection. It is a holistic approach that involves all business functions and requires constant commitment from companies. However, the adoption of eco-efficient practices can represent an opportunity for companies to increase their value and improve their reputation, in the face of a market increasingly sensitive to environmental sustainability issues [17]. To achieve the goal of reconciling economic development and environmental sustainability, the integration of eco-efficiency and Lean Production represents a winning combination. In fact, Lean Production focuses on eliminating waste, which also represents sources of environmental impact [18]. 2.2 Investigating the Use of VSM for Eco-Efficiency Performance In order to have complete overview of the application of Lean manufacturing practices for eco-efficiency performance, the authors conducted a literature review. The literature search was conducted in the databases: SCOPUS, Google Scholar, Elsevier, Emerald, Taylor & Francis, IEEE, Springer, Wiley, MDPI, ASTM, and Research Gate. In addition, the reviewed articles year a publication date between 2010 and 2023. The keywords used for the literature search were: on the one hand words related to Lean manufacturing (“lean manufacturing” or “lean tools” or “lean practices”) while on the other hand words related to environmental sustainability performance (“eco-efficiency” or “environmental performance”). Through the literature review, the Lean practice most cited by scholars is VSM [9]. Some studies showing how VSM is applied in relation to issues related to eco-efficiency

128

M. Ferrazzi and A. Portioli-Staudacher

performance will be presented below. The Value Stream Map (VSM) is the most widely used lean practice. However, it can be observed from the literature that not all scholars use this lean practice in the same way. W. Faulkner and F. Badurdeen 2014 [10] present the sustainable VSM (Sus-VSM). The sus-VSM begins with the identification and measurement of seven environmental or green wastes: (1) energy, (2) water, (3) materials, (4) waste, (5) transportation, (6) emissions, and (7) biodiversity. Several researchers, as Faulkner & Badurdeen, 2014 [10], have modified and adapted and adapted the VSM tool to make it more effective in solving problems related to environmental sustainability (Sus-VSM). Thanks to this were able to visually capture the sustainability performance of the production line. It emerged that the metrics and visual symbols used captured the most important criteria related to environmental and social sustainability performance for a subsequent in-depth analysis with other more comprehensive techniques. The SusVSM helped the company to have an initial evaluation of the production line status and to understand which processes were critical in terms of sustainability by calculating the different cycle times and the number of stocks. This information was then used to calculate the overall lead time. In addition to the value-added ratio, which can also be identified through the traditional VSM, the use of Sus-VSM allows the evaluation of energy consumption, water usage, and raw material usage. Articles as Garza Reyes et al. [11] aligned the VSM approach with the PDCA cycle, an effective method for enabling the implementation of E-VSM studies in a systematic, repeatable, and continuous improvement cycle. Regarding E-VSM, the strategic objective for its application in the case organization was to effectively identify and measure environmental waste in value stream processes and formulate and undertake appropriate strategies to eliminate/minimize them towards the future state. On the other hand, Rezeian et al. [12] used an energy-oriented VSM to map the entire production process and identify sources of energy waste. They subsequently analyzed the collected information and identified opportunities for improvement, suggesting solutions to reduce energy consumption and improve the overall efficiency of the production system. The energyoriented VSM used in this study helped to better understand the flow of energy within the production process and identify critical points for improving energy efficiency. In this way, it was possible to develop a targeted improvement strategy, based on concrete data, which led to an increase in the energy efficiency of the production system. Another widely discussed approach in the literature is the integration of VSM and Lean Six Sigma practice. For example, in Vieira et al. [13] article, researchers applied the Lean Six Sigma methodology and Value Stream Mapping (VSM) to improve the efficiency of energy assessments. They used VSM to map the workflow of energy assessments and identify any unnecessary or delay-causing activities. Later, they used the Lean Six Sigma DMAIC (Define, Measure, Analyze, Improve, Control) framework to analyze the collected data, identify the root causes of problems, and propose solutions to improve the energy assessment process. The combined approach of VSM and Lean Six Sigma reduced the time required for energy assessments and improved the quality of the analyses. Additionally, researchers also used Lean Six Sigma visual management tools to monitor progress and ensure process control over time. In conclusion, the Value Stream Map has proven to be an effective practice for identifying and removing waste

Applying the Value Stream Map to Streamline Energy Consumption

129

and negative environmental impacts, thanks to its ease of application and its ability to visualize the entire process on a single sheet. In particular, what emerges from the literature is the use of an extension of VSM, called sus-VSM. The sus-VSM is able to identifying value-added and non-value-added activities, also identifies activities that generate waste, negative environmental impacts, or inefficiencies in resource use [14]. Despite this, the academic literature on the utilization of this lean manufacturing practice can still be widely studied. In fact, the following study aims to understand how Value Stream Map can be suitable for coping with eco-efficiency problems. Specifically through the analysis of a case study of an Italian manufacturing company, it will analyze how the Value Stream Map has been used to reduce energy consumption. The case study will analyze a project guided by the PDCA approach and with the support of a sus-VSM.

3 Methodology The company under examination is one of the main Italian companies engaged in the management of end-of-life tires (ELT), a topic of great relevance and importance for environmental protection and the promotion of a circular economy. Its main goal is to recover, recycle, and valorize ELTs in order to reduce the environmental impact of waste and promote a circular economy. To achieve these goals, the company manages a series of plants in Italy, including a series of ELT processing and recovery centers. These plants are designed to ensure maximum efficiency and maximum environmental sustainability through careful design and the use of the most modern technologies. In particular, the company uses innovative processes for ELT processing that allow the separation of the different materials that make them up (rubber, metal, fabrics, etc.) and the recovery of those that can be reused. The current energy crisis has led the company to need to decrease energy consumption in its plants. Since the company has plants throughout Italy, a questionnaire was filled out for each one. The questionnaire submitted to all the company’s plants under analysis investigated annual consumption (KWh consumption during 2022) and tons of finished product in the same year. By going to analyze the ratio of energy used in the production process to total annual production, the company was able to identify which plants had the most problematic eco-efficiency indices (KWh consumed in 2022/ Tons produced in 2022). After obtaining the results, it was decided to intervene in a plant that had high energy consumption. The different phases of the PDCA approach that led to the implementation of the sus-VSM in this specific plant will be reported below. 3.1 PDCA Approach for sus-VSM An sus-VSM, like the traditional VSM, should be considered a continuous improvement process in which, based on the definition of a current state map and after achieving the proposed future state map, successive future state maps can be traced to allow for a cycle of continuous improvement. Therefore, in order to continuously eliminate/minimize waste, the VSM approach has been aligned with the PDCA cycle. Based on this logic, the PDCA-based approach to sus-VSM implementation presented has been proposed

130

M. Ferrazzi and A. Portioli-Staudacher

to provide an effective method for enabling sus-VSM studies to be implemented in a systematic, repeatable, and continuous improvement cycle [15]. The PDCA approach is divided into 4 phases: • Plan: During this phase, the organization must define the objectives of the E sus-VSM implementation, establish the project team, identify the process to be mapped, define success criteria, and plan implementation activities. Additionally, the team should establish a plan for data collection, map creation, and presentation of results. • Do: During the execution phase, the team collects the necessary data to create the value stream map, using the tools of the sus-VSM. This data must be accurate, and representative of the process being mapped. Once the data has been collected, the team creates the value stream map. • Check: The verification phase is when the project team evaluates the value stream map, verifies if the collected data is accurate, and if the map correctly represents the process being mapped. The team should also compare the map with the objectives established in the planning phase. • Act: The final phase of the PDCA approach is action. During this phase, the team must establish an action plan to address any problems that emerged during the verification phase. The team should also implement the necessary changes to the process and create a plan for continuous process monitoring. 3.2 Plan Stage In the planning phase, formulating objectives is important to provide a focal point for the implementation of sus-VSM. The strategic objective for its application in the organization’s case was to effectively identify and measure energy waste in the plant’s value stream, and then formulate and undertake appropriate strategies to eliminate/minimize them in order to move towards the future state. Since employees’ commitment to the project is essential, and their participation and feedback will provide a link to continuous improvement and the creation and achievement of current and future state maps [16]. It was decided to map the entire plant process, as it was comprised 6 processing stages. This choice was also made because the company’s staff wanted to have an overall idea of the energy consumption of the entire plant to understand which processing stages consumed more energy. In a VSM study, it is necessary to define a rigorous and systematic data collection methodology to identify waste sources and find ways to eliminate them. The same applies to sus-VSM implementation. For this project, the data collection strategy consisted of formulating and following a pre-established data collection plan that included collecting data on energy consumption at different characteristics of the plant’s value stream. In particular, the collected data included characteristics such as (1) energy consumption; (2) the utilization rate of machinery operating in each stage of the value stream to identify the major consumers and reduce their usage; (3) the number of employees involved in each stages; (4) the number of hours each machinery operates; (5) the cycle and maintenance times of each machinery. The data were collected through observations and direct measurements of the value stream.

Applying the Value Stream Map to Streamline Energy Consumption

131

3.3 Do Stage The organization in this case produces two different types of products, namely rubber granules and “cobblers”. Rubber granule was chosen for this study: since it is the product that requires the most processes to be produced and its energy consumption for its production is the greatest. After identifying the chosen product and defining the production process in detail, the procedure for constructing the actual sus-VSM begins with the identification and measurement of seven environmental or green wastes. Considering the seven environmental wastes of W. Faulkner and F. Badurdeen (2014) (presented in Sect. 2), the company under analysis decided to target only energy consumption. Specifically because the plant did not have data on other environmental wastes. Consequently, without accurate and timely quantification of environmental wastes, using VSM would be useless. Moreover, this was the first time for the company to tackle an improvement project related to environmental sustainability, and consequently it wanted to proceed with a progressive approach, attacking one environmental wastes at a time. Of these seven wastes, only energy waste was considered. The VSM created thus focuses exclusively on reducing energy consumption that occurs when more energy is consumed than necessary.

Fig. 1. Current state map

This step preceding the creation of the current state map was carried out in conjunction with data collection, as described in the Plan Stage section. After identifying and measuring green waste, the last step of the Do Stage is the creation of the current state map. For this purpose, the traditional VSM was adapted, as previously stated, to incorporate the green wastes previously discussed. In fact, in the “data box” in addition to the classic data used in the traditional VSM (cycle time, changeover time, machine availability, number of shifts and number of operators for each process) energy consumption

132

M. Ferrazzi and A. Portioli-Staudacher

metrics have also been included (the annual consumption in kWh of each stage and the KWh/ton efficiency index) (Fig. 1). 3.4 Check Stage In the Check Stage, the current state map was analyzed, the different processes with higher energy consumption within the plant and the non-value-added activities were identified. The process that showed the highest energy consumption was the 7–4 mm rubber granulation with a consumption of about 520 MWh per year. Another process shown by this analysis to have a highly energy consumption is the 2 mm granule refining process. In this process in fact only a part of the product is processed about 1600 tons per year (20%) and its energy consumption is about 333 MWh. Although this is not the highest, it is very high for the amount of granulate it actually produces. This aspect was further emphasized through the calculation of the efficiency coefficient for each process. In fact, in this process the value of the coefficient is much higher than in the others, with a KWh/ton ratio of about 200, compared with an average for the other processes of about 36 KWh/ton. These data therefore show that an intervention on these two processes is necessary since their high consumption is directly related to their inefficiency. As for non-value-added activities, there were large amounts of inventory among the plant processes, resulting in a processing time of about 13 days. The next step was to analyze these processes to define the nature of their inefficiencies. In the 4–7 mm granulation process, it was observed that the components of the two machines used for processing were very worn, particularly the rotors and bearings. Also in the 2 mm refining process, it was observed that the two single rotor machines were the main causes of process inefficiency. In fact, these two machines are very dated and were adapted to carry out the rubber refining activity because they were not initially designed for this practice. The last phase of the Check Stage concerns the definition of strategies to minimize and reduce the identified waste, resulting in the creation of the Future State Map. The solution implemented for the 4–7 mm granulation process was bearing replacement and a complete rotor overhaul of the two machines. This replacement is expected to bring a 5% reduction in energy consumption, a cost of 33000e. For the 2 mm refining process, on the other hand, it was chosen to replace both refining machines, with a single new machine. This machinery has a lower power output than the previous two machines of 110 KW, it was estimated that it can bring 30% reduction in energy consumption to the process. In addition, the new machinery has a production capacity of about 1.5 ton/h, almost three times that of the previous one (Fig. 2). 3.5 Act Stage Once the future state map has been created, the last stage of the proposed method is the Act Stage. This stage consists of developing a plan for the implementation of strategies designed to eliminate/reduce waste and, in this way, “transform” the studied value stream into its desired future state. Goals and objectives will also serve as the basis for monitoring and controlling the progress and success of sus-VSM implementation. Therefore, to develop an implementation plan, the model includes setting individual goals and targets to address green waste. In addition, there is provision for the use of KPIs for

Applying the Value Stream Map to Streamline Energy Consumption

133

Fig. 2. Future state map

each individual goal or target to measure and track the progress of the implementation of a future state of sus-VSM. The implementation of the plan formulated in the previous stage is the last phase of the proposed approach to sus-VSM. Therefore, this phase corresponds to the execution phase that will bring about and enable the “transformation” from the current to the future state of sus-VSM.

4 Findings In terms of quantitative benefits through the implementation of sus-VSM, it has been possible to identify the most energy consuming activities and implement strategies to reduce them. In particular, the 22-mm rubber granulation process was improved by replacing the bearings and rotors of the two granulators. This replacement resulted in reducing the energy consumption of the entire process by about 5%. In addition, thanks to the replacement of material components that were worn out, the productivity of the process was partially increased. Table 1 summarizes the improvements brought to the plant. Regarding the 2 mm rubber refining process as mentioned above, it was determined to replace the refining machines, due to inefficiencies of technological origin. The machine selected for the replacement of the two refiner machines, will have a power of about 110 KW, 30% lower than the two previous machines. Moreover its production capacity will be about 1.1 ton/h, almost three times higher than before. Thus, this replacement would allow a decrease of about 30% in energy consumption of that particular process, and also the reduction of processing stock since it will be able to work at a higher production capacity. Table 2 summarizes the improvements brought to the plant.

134

M. Ferrazzi and A. Portioli-Staudacher Table 1. Planned energy improvements for granulation process 4–7 mm.

Granulation stage 4–7 mm (Current state)

Granulation stage 4–7 mm (Future state)

Annual Energy consumption 518 MWh

Annual Energy consumption 497 MWh

Energy efficiency index 46 KWh/ton

Energy efficiency index 45 KWh/ton

Production capacity 2,5 ton/h

Production capacity 2,7 ton/h

Table 2. Planned energy improvements for granulation process 4–7 mm Refining stage 2 mm (Current state)

Refining stage 2 mm (Future state)

Annual Energy consumption 333 MWh

Annual Energy consumption 236 MWh

Energy efficiency index 208 KWh/ton

Energy efficiency index 78 KWh/ton

Production capacity 400 kg/h

Production capacity 1,1 ton/h

Overall, at a general level through these improvements the plant’s production process had a reduction in total energy consumption of about 5%. This reduction would allow the plant to save about 43000e annually, against an investment 183000e, with a payback time from ‘investment of 4 years. Finally, regarding the environmental benefits brought through this work, using data for sustainability, awareness of this issue within the company has grown. The company is now more supportive in promoting new sustainable practices and policies. Sustainability has thus become of great importance to the company, and the main goal for the future is to implement it in all areas of the company, ensuring that the company moves toward a sustainable business model. Inclusion of sustainability at all levels of the company requires a strong and ongoing commitment, but as has been shown it can lead to great benefits. Through the development of a sustainable model, the company will be able to meet the coming environmental challenges of the future.

5 Conclusion In conclusion, this case study was able to show how the application of sus-VSMs was able to reduce the company’s energy consumption without going to change its operational performance. This is a confirmatory case study on how Lean manufacturing can have an impact on eco-efficiency performance. The contribution of this paper for managers of manufacturing companies is to be able to show how the VSM tool is effective

Applying the Value Stream Map to Streamline Energy Consumption

135

in addressing problems related to environmental sustainability. Moreover, improvement in environmental performance does not coincide with a deterioration in manufacturing performance. In fact, environmental and operational performance under the right conditions can improve simultaneously. For the academic and research side, the contribution of this paper is further evidence of the effectiveness of the lean manufacturing approach for problems related to environmental sustainability. Future studies on this area of research are related to the possibility of expanding the research using a multiple case study analysis. The selection of case studies will be guided to have a heterogeneity of companies with different degrees of lean maturity and different degrees of environmental awareness. This will help to understand how and whether these factors can play a moderating role between lean manufacturing practices and eco-efficiency performance.

References 1. Womack, J.P., Jones, D.T., Roos, D.: The Machine That Changed The World (1991) 2. Garza-Reyes, J.A.: Lean and green-a systematic review of the state of the art literature. J. Clean. Prod. 102, 18–29 (2015) 3. Shah, R., Ward, P.T.: Lean manufacturing: context, practice bundles, and performance. J. Oper. Manag. 21(1), 129–149 (2003) 4. Dhingra, R., Kress, R., Upreti, G.: Does lean mean green? J. Clean. Prod. (2014) 5. Gaikwad, L., Sunnapwar, V.: Integrated lean-green-six sigma practices to improve the performance of the manufacturing industry. Concepts Appl. Emerg. Oppor. Ind. Eng. (2021) 6. Benabdellah, A.C., Zekhnini, K., Cherrafi, A., Garza-Reyes, J.A., Kumar, A.: Design for the environment: an ontology-based knowledge management model for green product development. Bus. Strateg. Environ. 30(8), 4037–4053 (2021) 7. Maxime, D., Marcotte, M., Arcand, Y.: Development of eco-efficiency indicators for the Canadian food and beverages. J. Clean. Prod. 14, 636–648 (2006) 8. Hendrickson, C.T., Lave, L.B., Matthews, H.S.: Environmental Life Cycle Assessment of Goods and Services: An Input-Output Approach. Resources for the Future (2006) 9. Abreu, M.F., Alves, A.C., Moreira, F.: Lean-Green models for eco-efficient and sustainable production. Energy 137, 846–853 (2017) 10. Faulkner, W., Badurdeen, F.: Sustainable value stream mapping (Sus-VSM): methodology to visualize and assess manufacturing sustainability performance. J. Clean. Prod. 85, 8–18 (2014) 11. Garza-Reyes, J.A., Torres Romero, J., Govindan, K., Cherrafi, A., Ramanathan, U.: A PDCAbased approach to environmental value stream mapping (EVSM). J. Clean. Prod. 180, 335–348 (2018) 12. Rezaeian, J., Parviziomran, I., Mahdavi, I.: Increasing energy productivity in lean production system with energy oriented value-stream mapping. Int. J. Prod. Qual. Manag. 24(4), 495–506 (2018) 13. Viera, R.J., Sardoueinasab, Z., Lee, J.: Improving the efficiency of energy assessments with application of lean tools—a case study. Energy Effic. 12(7), 1717–1728 (2019) 14. Schillig, R., Stock, T., Müller, E.: Energy value-stream mapping a method to visualize waste of time and energy. In: Umeda, S., Nakano, M., Mizuyama, H., Hibino, N., Kiritsis, D., von Cieminski, G. (eds) APMS 2015. IFIPAICT, vol. 459, pp. 609–616. Springer, Cham (2015). https://doi.org/10.1007/978-3-319-22756-6_74 15. Milosevic, M., Djapan, M., D’Amato, R., Ungureanu, N., Ruggiero, A.: Sustainability of the production process by applying lean manufacturing through the PDCA cycle – a case study

136

16. 17. 18.

19. 20.

21.

M. Ferrazzi and A. Portioli-Staudacher in the machinery industry. In: Hloch, S., Klichová, D., Pude, F., Krolczyk, G.M., Chattopadhyaya, S. (eds.) ICMEM 2021. LNME, pp. 199–211. Springer, Cham (2021). https://doi.org/ 10.1007/978-3-030-71956-2_16 Abdulmalek, F.A., Rajgopal, J.: Analyzing the benefits of lean manufacturing and value stream mapping via simulation: a process sector case study. Int. J. Prod. Econ. 107(1), 223–236 (2007) Marques, T., et al.: Monitoring and evaluating eco-efficiency by three different ways in a beverage company: a lean-green approach. Smart Sustain. Manuf. Syst. (2022) Gouveia, J., Gonçalves, M., Rocha, R., Schwab, D.J., Monteiro, H.: Efficiency framework to assess aeronautic composite panel production: tracking environmental and process performance. Sustain. Prod. Consum. (2022) Torri, M., Kundu, K., Frecassetti, S., Rossini, M.: Implementation of lean in IT SME company: an Italian case. Int. J. Lean Six Sigma 12(5), 944–972 (2021) Rossini, M., Audino, F., Costa, F., Cifone, F.D., Kundu, K., Portioli-Staudacher, A.: Extending lean frontiers: a Kaizen case study in an Italian MTO manufacturing company. Int. J. Adv. Manuf. Technol. 104(5–8), 1869–1888 (2019) Bai, C., Satir, A., Sarkis, J.: Investing in lean manufacturing practices: an environmental and operational perspective. Int. J. Prod. Res. 57(4), 1037–1051 (2019)

Crossroads and Paradoxes in the Digital Lean Manufacturing World

A Systematic Literature Review on Combinations of Industry 4.0 and Lean Production Kristian Ericsson(B)

and Antonio Maffei

Department of Production Engineering, KTH Royal Institute of Technology, 151 81 Södertälje, Sweden [email protected]

Abstract. Prior literature reviews on combining Industry 4.0 (I4.0) and Lean Production (LP) in production has often described the paradigms as “supportive”, where either “I4.0 supports LP” or “LP supports I4.0”. In this systematic review of 50 studies from this growing area of research, we find evidence of cases where combinations have not been “supportive”. We also find evidence of causal interactions that cannot be subordinated “I4.0 supports LP” nor “LP supports I4.0”, and that plainly go beyond those two categories. Additionally, we find that several studies evaluate the merits of I4.0- and LP combinations without looking at their effects on results in production, which substantially reduces the use of these evaluations to production managers. We encourage future studies to use nomenclature that does not unnecessarily limit the overall perception of the properties of I4.0- and LP combinations, to evaluate the merits of such combinations more in line with the requirements of production managers, and to be more cautious when concluding on causal interactions between I4.0 and LP. Keywords: Industry 4.0 · Lean Production · Lean Manufacturing · operations performance · I4.0/LP combinations

1 Introduction Industrie 4.0/Industry 4.0 (I4.0) has been described as a future-oriented paradigm for how firms should operate their production [1]. The paradigm consists of elements in the form of both digital technologies – like artificial intelligence, big data, cloud computing, cyber-physical systems, the Internet of Things, and advanced simulations – and principles for how to combine these digital technologies [2]. While I4.0 is the future-oriented paradigm, Lean Production (LP) can be regarded as the legacy paradigm for how firms should operate their production [3]. Just like I4.0, the LP paradigm also consists of elements in the form of principles and practices, like continuous flow, kanban, pull production, setup time reduction (SMED), standardization, Total Productive Maintenance, and Value Stream Mapping [4]. © IFIP International Federation for Information Processing 2023 Published by Springer Nature Switzerland AG 2023 E. Alfnes et al. (Eds.): APMS 2023, IFIP AICT 689, pp. 139–156, 2023. https://doi.org/10.1007/978-3-031-43662-8_11

140

K. Ericsson and A. Maffei

Producing firms have already begun combining I4.0 and LP, and authors predict that this will become increasingly common in the future [5]. Prior studies have presented great potential benefits from the use of both paradigms, concerning for example cost, productivity, quality, and customer satisfaction [6]. However, scholars have also underlined the ease with which such an endeavor can fail [7]. The combined use of both I4.0 and LP in production has been referred to by several different terms, for example “the integration of lean manufacturing and industry 4.0” [5], “the relationship between LP practices and the implementation of I4.0” [3], “leandigitized manufacturing” [8], and “the impacts of Industry 4.0 technologies on the different Lean principles” [9]. Our assumption is that most of these terms might inadvertently limit the perception of what the concept of ‘combined use of I4.0 and LP in production’ can entail. For example, terms with positive or negative connotations – like “synergy between I4.0 and LP” or “synthesis of I4.0 and LP” – risk creating value-laden biases around the concept. Terms assuming far-reaching, intended interaction between the paradigms – like “integration of I4.0 and LP” or “conjunction of I4.0 and LP” – can exclude uncomplicated forms of combined use, like an I4.0 element and a LP element used in sequence, and combined use emerging through evolution rather than intent. Contrariwise, terms that describe I4.0 and LP as separate entities – like “the link between I4.0 and LP”, “the relationship between I4.0 and LP”, “impact of I4.0 on LP”, or “effect of I4.0 on LP” – risk excluding cases where I4.0 and LP are so integrated that they form a third, eliminating the “between” (like in the case of “lean-digitized manufacturing”). For the purposes of this study, we therefore use the broad term “combination of I4.0/LP” to refer to the use of at least one I4.0 element in cooperation with at least one LP element in production. Hence, “I4.0/LP combinations” can refer to everything from simple setups of one I4.0 element delivering input to one LP element, to heavily integrated, complex systems with many elements from both paradigms. The divergence in terms used to refer to I4.0/LP combinations is perhaps a symptom of the current state of the literature on the subject. The research area has experienced a rapid rise in interest during the last years (as evidenced by [10] and [11]) but is still labelled as under-researched and full of opportunities for future research [10]. We speculate that the rapid rise of studies has led to a divergence in the research area. There is a need for consolidation and synthesis so that the area can proceed towards offering guidance to managers in how to use I4.0/LP combinations to their advantage. Consistent use of key terms and assumptions seems to be of particular need.

2 Background There have been several structured literature reviews on I4.0/LP combinations. For example, [1] is a well-cited, relatively early study intended to map out the literature in the (by then) new research area. Two influential contributions from this study were (1) the conclusion that prior studies generally perceive the relationship between I4.0 and LP as “supportive”, and (2) a clustering of prior studies based on whether they perceived that “I4.0 supports LP” (18 studies) or “LP support I4.0” (1 study). The study also presented manufacturing performance as a “key area of interest”.

A Systematic Literature Review on Combinations of Industry 4.0

141

Despite a rapidly growing number of studies in the field, the early study’s conclusions concerning (1) the “supportive” nature of I4.0/LP combinations, and of a (2) bidirectional relationship of support, have remained in later structured literature reviews. The review [10], for example, confirms the two-way support, the review [12] only investigates literature on how I4.0 supports LP, and review [13] only concerns literature on how LP supports I4.0. This study intends to revise these central contributions from study [1] now that the research area has grown and (assumingly) matured. The first aim of our study is to investigate if more recent studies also perceive “supportive” I4.0/LP combinations in general, and if these resemble either “I4.0 supports LP” or “LP supports I4.0”. For the second aim of this study, we wish to investigate how authors evaluate the merits of I4.0/LP combinations. While [1] labelled results in production as a “key area of interest”, we assume that the potential to reach improved results in production is the main driver behind implementing I4.0 by itself, LP by itself, as well as any combinations between the two. Production managers should consider implementing I4.0/LP combinations as means to achieve improved results in production (i.e. not ends in themselves) and thereby as one alternative among other means to achieve those same ends. Consequently, the merit of any I4.0/LP combination lies primarily in its ability to improve results in production. To be of practical relevance, statements like “I4.0 supports LP” should imply that the I4.0/LP combination enables the achievement of improved results in production, or at least assists it. Labelling an I4.0/LP combination as “supportive” should be backed up with evidence of (a) improved results in production, or at least evidence of (b) positive effects on the “causal chains” through which I4.0 elements, LP elements, or their combinations, affect results in production. Therefore, the second aim of this study is to investigate to what degree prior studies have reached conclusions like “supportive” or “I4.0 supports LP” through evaluating I4.0/LP combinations’ effects on results in production.

3 Review Method To reach the aims above, a structured literature review was conducted of documents describing I4.0/LP combinations. The method used can be described in four steps. 3.1 Finding Keywords for I4.0 and LP The keywords used to find relevant literature were designed to represent I4.0 and LP. However, no keywords for results in production were used in the searches, since documents about I4.0/LP combinations do not necessarily discuss results in production. Researchers have used a wide variety of keywords to refer to I4.0, LP, and the combinations of the two. To find relevant literature, lists of relevant keywords were compiled by using iterative searches and snowball sampling. The lists were designed to include studies coming from different backgrounds and thus using (slightly) different wording. They included keywords synonymous with I4.0 or LP (e.g. “lean manufacturing”, “lean program”, “lean transformation”…). They also

142

K. Ericsson and A. Maffei

included keywords where either I4.0 or LP likely makes up a substantial share (e.g. “Lean Six Sigma”, “continuous improvement initiative”, “Operational Excellence”…). They also included several ways to refer to the “production” context (“manufacturing”, “factory”, “operations”…). However, the lists did not include elements of I4.0 or LP, i.e. individual digital technologies or principles for digital transformations, or individual LP principles or practices. This was to maintain the study on an aggregate level, foremost investigating what is true for a I4.0/LP combinations in general rather than in single cases. A few exceptions were made for elements that are closely associated with their entire paradigms, like “value stream (mapping)” or “cyber physical production systems”. The lists also included slight variants of the words in the terms, like singular and plural nouns, forms with and without hyphen, etc. This was to not be reliant on the search engine’s ability to include both “factory” and “factories”, “manufacturing” and “manufacturer”, “lean-digitized” and “lean-digitised”, etc. However, the lists did not include single-word terms, and abbreviations of only a few letters, like “lean” and “ICT”. This was to avoid instances where the same spelling could appear by chance in irrelevant contexts. Instead, keywords with two or more words (in quotation marks) were promoted. The resulting lists were extensive. They contained 33 keywords that represented “Industry 4.0 in production”, 94 keywords for “Industry 4.0”, 21 keywords for “production”, 63 keywords for “Lean Production”, and 12 keywords for “combinations of Industry 4.0 and Lean Production”. 3.2 Designing a Search Query The keywords were put together to find documents that describe I4.0/LP combinations. The query sought to generate studies that (a1) mentioned both I4.0 as well as LP, or (a2) used a known keyword for “combinations of Industry 4.0 and Lean Production”, and (b) concerned production. In search engine language, the query was: Title, abstract, keywords (((([keywords for “Industry 4.0 in production”] OR ([keywords for “Industry 4.0”]) AND [keywords for “Lean Production”]) OR [keywords for “combinations of Industry 4.0 and Lean Production”]) AND [keywords for “production”]). 3.3 Searching with Known Limitations The search query was used middle of April 2023 in the Scopus database, and included all documents published until that date. All document types were considered (i.e. mostly conference papers and articles) but only those published in English. The search generated a set of 1721 documents. 3.4 Selecting Documents to Include in the Review Due to the large set of generated documents, the analysis of document contents was carried out on a subset of 50 documents. The total 1721 documents were arranged in a row according to their level of influence. The processing of the documents began with the

A Systematic Literature Review on Combinations of Industry 4.0

143

documents with the highest influence and descended along the row, until 50 documents had been found. We used a citation quota to represent level of influence for each document, calculated as total number of citations divided by the expected number of years since publication. For example, documents published in 2022 were expected to be 0,75 years since publication, so a document from 2022 with 6 citations would receive a citation quota of 8. We avoided arranging the studies by total citations, as this might have given too much weight to older studies, with many total citations but less attention seen over their lifespans. It would also have given too little weight to newer studies, with fewer total citations but on average more attention. A limitation to our method of arranging the document is that it only concerns confirmed influence and not potential influence (e.g. very new documents will be overlooked even if from high-ranked journals). The most influential documents were reviewed until 50 documents had been found that (a) studied I4.0 or some concept (remotely) similar to it, (b) studied LP or some concept (remotely) similar to it, (c) gave importance to both paradigms (e.g. not merely mentioned them), and (d) discussed the properties of I4.0/LP combinations. If in doubt whether the criteria were met, studies were included rather than excluded. Documents that were excluded typically (1) mentioned a synonym for either I4.0 or LP, but did not refer to a concept similar to the paradigms discussed here (examples of excluded studies: [14, 15]), (2) gave little importance to one of the paradigms (examples: [15, 16]), (3) bundled I4.0 and LP together to contrast them to a third (paradigm), but did not discuss combining I4.0 and LP (example: [17]), (4) discussed at least one I4.0/LP combination, but only concerned a single element from one of the paradigms (example: [18]), or (5) discussed at least one I4.0/LP combination, but only concerned e.g. its application and not its properties (example: [19]). Many studies were excluded; 86 documents were reviewed before 50 had been found (i.e. 58% were retained). The high number of excluded documents rules out any meta-data investigations of the 1721 documents, since many are probably excludable. 3.5 Analysing the Included Documents The 50 included documents were reviewed for (1) to what extent they presented conclusions similar to “supportive”, “I4.0 supports LP”, or “LP supports I4.0”, (2) to what extent they reached these conclusions by linking them to results in production. Each document’s results concerning (1) and (2) were gathered in a table (similar to Table 1), which allowed an aggregated synthesis of the documents’ contributions.

4 Results During the review, two concepts emerged that describe important characteristics of I4.0/LP combinations. These correspond to two trends in the reviewed literature. The first trend is the description of other types of internal compatibilities in I4.0/LP combinations than “supportive”. The studies imply that the use of the element(s) from one paradigm can both support (positive), disregard (neutral), or disturb (negative) the

144

K. Ericsson and A. Maffei

performance of the element(s) from the other paradigm. This property is treated as inherent to each I4.0/LP combination and we label it “combination compatibility”. The second trend is that the studies often describe an interplay within I4.0/LP combinations that go beyond and deeper than “I4.0 supports LP” or “LP supports I4.0”. Examples are (a) I4.0 is capable of implementing LP, (b) LP makes it easier to reach higher levels of I4.0, and (c) I4.0 and LP can create a synergistic combination (all examples retrieved from Table 1 below). The authors seem to refer to a wide range of causal relationships between I4.0 and LP. We label the causal interplay within an I4.0/LP combination as its “combination mechanisms”. With these two concepts in place the literature can be summarised as in Table 1. Table 1. Summary of the studies reviewed. #

Study

Citations/year

Concludes on combination compatibility

General statement on combination compatibility (and combination mechanisms, if available)

Describes own perception of results in production

1

[1]

90,7

Yes

I4.0 supports LP, LP supports I4.0

Yes

2

[2]

86,5

Yes

I4.0 is an enabler of LP

Yes

3

[3]

76,6

Yes

The concurrent implementation of I4.0 and LP leads to larger improvements of results in production

Yes

4

[4]

74,3

Yes

A synergistic relationship, LP is a necessary basis for I4.0 execution

NO

5

[5]

73,3

Yes

I4.0 is capable of implementing LP

NO

6

[8]

56,7

Yes

A mutually supporting relationship

Yes

7

[9]

55,3

Yes

A strongly supporting NO relationship

8

[20]

53,4

Yes

A supporting relationship

NO (continued)

A Systematic Literature Review on Combinations of Industry 4.0

145

Table 1. (continued) #

Study

Citations/year

Concludes on combination compatibility

General statement on combination compatibility (and combination mechanisms, if available)

Describes own perception of results in production

9

[21]

50,1

Yes

I4.0 applied to LP gives combined benefits

NO

10

[12]

49,1

Yes

A synergistic relationship, a positive interaction

NO

11

[22]

45,7

Yes

A synergistic relationship

Yes

12

[23]

42,9

Yes

A very strong NO synergistic relationship, I4.0 is used to aid the implementation of LP

13

[24]

42,8

Yes

A stabilising and supporting relationship

NO

14

[25]

42,7

Yes

Rooted LP practices guides I4.0 efforts to more success

Yes

15

[26]

42,3

Yes

A mutually supporting, enabling, and empowering relationship

NO

16

[6]

40,0

Yes

Having an LP basis allows an effective outcome from I4.0

NO

17

[27]

35,8

Yes

LP helps to implement I4.0

NO

18

[28]

35,3

NO

(Does however refer to I4.0 and LP as a “union”)

NO

19

[7]

34,7

Yes

A supplementary, supporting relationship

NO

(continued)

146

K. Ericsson and A. Maffei Table 1. (continued)

#

Study

Citations/year

Concludes on combination compatibility

General statement on combination compatibility (and combination mechanisms, if available)

Describes own perception of results in production

20

[29]

34,3

Yes

LP is essential for and Yes positively influences I4.0

21

[30]

33,3

Yes

LP makes it easier to Yes reach higher levels of I4.0

22

[13]

28,6

Yes

A complementary relationship, LP enabling and triggering I4.0

NO

23

[31]

27,8

Yes

I4.0 applied to LP methods combines their benefits

NO

24

[32]

27,4

NO

-

Yes

25

[33]

26,7

Yes

LP can eliminate waste in I4.0 implementations

NO

26

[34]

26,7

Yes

I4.0 and LP are relatively easy to combine due to shared objectives

NO

27

[35]

26,4

NO

-

Yes

28

[36]

26,3

Yes

Most LP and I4.0 “features” (aims) are either mildly or strongly positively correlated

Yes

29

[37]

26,2

Yes

I4.0 enable enhanced LP practices

NO

30

[38]

25,1

Yes

A synergistic combination

NO

31

[39]

24,0

NO

-

NO (continued)

A Systematic Literature Review on Combinations of Industry 4.0

147

Table 1. (continued) #

Study

Citations/year

Concludes on combination compatibility

General statement on combination compatibility (and combination mechanisms, if available)

Describes own perception of results in production

32

[40]

22,7

Yes

I4.0 technologies can create synergistic effects between LP (efficiency) and resilience

NO

33

[41]

21,2

Yes

I4.0 technologies can be incorporated into LP, to bring LP to higher levels

NO

34

[42]

21,1

Yes

Synergistic, I4.0 technologies and LP practices can be integrated

NO

35

[43]

21,0

Yes

LP can enable a successful and sustainable implementation of I4.0

NO

36

[44]

20,2

Yes

Mutually supportive, LP enables I4.0 implementation and vice versa

NO

37

[45]

20,0

Yes

The combined Yes application of LP and I4.0 can lead to a wide range of benefits

38

[46]

19,2

Yes

I4.0 technologies highly facilitate LP, and are crucial to achieve higher levels of LP

39

[47]

18,5

Yes

I4.0 technologies can NO improve design and implementation of LP value streams

Yes

(continued)

148

K. Ericsson and A. Maffei Table 1. (continued)

#

Study

Citations/year

Concludes on combination compatibility

General statement on combination compatibility (and combination mechanisms, if available)

Describes own perception of results in production

40

[48]

17,7

NO

(Does, however, assume higher levels of “Lean Automation” to be beneficial)

NO

41

[49]

17,3

Yes

The integration of I4.0 and LP - “Lean 4.0” - makes firms more competitive

NO

42

[50]

17,3

Yes

LP is a prerequisite to NO implement I4.0, LP practices can support I4.0 initiatives

43

[10]

17,3

Yes

A synergistic NO relationship, LP support implementing I4.0, I4.0 enhance LP

44

[51]

17,1

Yes

A supportive relationship

NO

45

[11]

17,1

Yes

Combining I4.0 and LP is necessary to reach higher performance in the future

NO

46

[52]

16,4

Yes

I4.0 technologies significantly impact LP practices

NO

47

[53]

16,0

NO

(Does, however, Yes assume higher levels of “Lean Automation” to lead to improved results in production) (continued)

A Systematic Literature Review on Combinations of Industry 4.0

149

Table 1. (continued) #

Study

Citations/year

Concludes on combination compatibility

General statement on combination compatibility (and combination mechanisms, if available)

Describes own perception of results in production

48

[54]

15,6

Yes

I4.0 and LP used together improve results in production substantially

Yes

49

[55]

15,3

Yes

LP can guide successful implementation of I4.0

NO

50

[56]

15,2

Yes

I4.0 technologies can bring LP practices to higher levels

NO

As is apparent, the level of influence, i.e. the average number of citations per year since publication, decreases considerably from high up to far down in the list. Additionally, 44 out of 50 studies contain some statement on combination compatibility. The general statements on combination compatibility are positive. Although some of the studies mention neutral or even negative I4.0/LP combinations, or aspects of such combinations, all of the 44 generally report of supportive compatibilities. The combination mechanisms, however, vary a lot more between the different studies than the combination compatibilities. Several studies do not mention any causal relationships between I4.0 and LP in the combinations. Some studies describe how I4.0 can enhance LP, while others describe the reverse causality. Yet other studies conclude that I4.0 elements and LP elements can come together to form a third, that does not resemble the initial elements. Some studies merely conclude that one paradigm is capable of implementing the other paradigm but do not describe the product after implementation as either I4.0, LP, nor a combination of the two. Finally, most studies do not describe their perception of results in production, but 15 out of the 50 studies do so to some degree. Among these, the most common terms for results in production are “operational performance” and terms referring to the three pillars of sustainability, like “operational-, environmental and economic performance”, “economic, environmental, and social sustainability”, among others.

150

K. Ericsson and A. Maffei

5 Analysis and Discussion As can be seen in the results, many studies have provided valuable insights concerning the characteristics of I4.0/LP combinations. Our literature review adds to this body of knowledge by identifying and labelling two key properties of I4.0/LP combinations, and by identifying issues in how these properties have been determined. 5.1 Extending the Nomenclature There has so far been no systematic way of referring to important concepts surrounding I4.0/LP combinations. This study is an attempt to establish more congruence. In the introduction chapter we adopted the term “combination of” to refer to the combined use of both I4.0 element(s) and LP element(s) in production, with the motivation that our choice of term should not narrow our understanding of the concept. The literature review has supported this choice by describing I4.0/LP combinations as diverse: positive and negative, interactive and separate, independent and merged, etc. In the results chapter we presented and defined the term “combination compatibility” as a label for a property that prior authors have repeatedly identified in I4.0/LP combinations, and that go beyond the “supportive” nature often found in prior literature reviews. Our study also shows that combination compatibilities can have different magnitudes, going from “low positive” [24] to “very strong synergistic” [23]. Furthermore, the review overall concluded on positive combination compatibility, but there are several critical evaluations (e.g. [9, 25, 54]). The results chapter also introduced “combination mechanisms” to refer to the causal interplays that prior authors have identified within I4.0/LP combinations, that go beyond and deeper than “I4.0 supports LP” and “LP supports I4.0”. The literature review shows that authors use combination mechanisms to explain how combination compatibilities are achieved, e.g. if what makes them “supportive” is the (a) parallel vs. sequential use, or (b) separate vs. merged use of the paradigms, or if it comes from the paradigms (c) influencing the contents vs. implementation of the other paradigm. Since combination compatibility can be both positive, neutral, and negative, it’s natural that combination mechanisms describe a variety of causal paths. Although our proposed terms, and their definitions are mere suggestions at this stage, we hope they will help bring the view on I4.0/LP combinations beyond “supportive”, “I4.0 supports LP”, and “LP supports I4.0”. At the same time, we hope they will make future discussion on I4.0/LP combinations easier to comprehend. Furthermore, our literature review has identified two common issues among prior studies, that puts their ability to offer valid statements on the merits of I4.0/LP combinations and their mechanisms into question. The following two sections will outline and discuss these issues. 5.2 Issue #1: Studies that Indicate Merits of I4.0/LP Combinations in Ways that are of Limited Use to Production Managers A majority of the studies investigated in this review present conclusions on the compatibility of I4.0/LP combinations. This is most often done without simultaneously

A Systematic Literature Review on Combinations of Industry 4.0

151

presenting a link to results in production. As can be seen in our results, most studies do not include a description of results in production. Also, few studies contain descriptions of how I4.0/LP combinations are causally connected to such results (“causal chains”). Referring back to the discussion in the introduction chapter, this would mean that most studies’ conclusions have limited practical use as indicators of the merits of I4.0/LP combinations. Production managers, who wish to use I4.0/LP combinations intentionally as means to achieve improved results in production, receive limited guidance from statements on combination compatibility, simply since these statements have been made without a clear link to results in production. If positive combination compatibility would carry promises of improved results in production, production managers would probably be more willing to invest time and money in establishing positive compatibility I4.0/LP combinations. However, as the research has been conducted so far, statements on whether a combination is “supportive”, “synergistic”, etc. does not carry any such promises. We also see a risk that omitting the link to results in production might have removed an important, independent scale against which to evaluate combination merits, which might have contributed to the positive bias permeating prior studies. We therefore recommend future authors, who aim to indicate the merits of a I4.0/LP combinations, to always equip their indicators with an evaluation of the combination’s effects on results in production. This would give production managers indications of which I4.0/LP combinations are suitable means in order to reach their aims, while combination mechanisms would subsequentially direct them how to combine the I4.0- and LP elements to reach the same aims. We also speculate that future studies would benefit from reducing their efforts in evaluating combination compatibility. Instead, they could use these efforts to evaluate the overall suitability of I4.0/LP combinations compared to other means to achieve the same ends; a kind of “combination competitiveness”-indicator. “Combination competitiveness” would necessarily include an evaluation of the combination’s effects on results in production, but also include, for example, indicators of the costs and risks caused by running the combination, the costs and risks from implementing the combination, etc. This would give production managers an improved capacity to decide between I4.0/LP combinations and other means to improve their results in production. We also encourage more studies to put efforts into identifying and describing “causal chains” between (a1) I4.0- and LP elements, (a2) I4.0/LP combinations, and (b) results in production – both in theory and in practice. Merits of I4.0/LP combinations could then be evaluated by investigating their effects on these “causal chains”. 5.3 Issue #2: Studies that Present Combination Mechanisms for I4.0/LP Combinations Despite only Observing Combination Compatibility One clearly distinguishable group of studies evade issue #1 and evaluate the merits of I4.0/LP combinations more in line with the requirements of production managers (e.g. [2, 3, 22], and [32]). These are quantitative studies, that describe results in production, and relate (a1) I4.0 levels, (a2) LP levels, and (b) levels of results in production. They are some of the few truly empirical studies in this review, and they add valuable insights – in particular by offering a more critical view on I4.0/LP combinations among studies that are otherwise overwhelmingly positive (see for example [25] and [54]).

152

K. Ericsson and A. Maffei

However, these studies also tend to make statements on combination mechanisms, although the studies have not observed the actual combinations. This renders the statement speculative. We assume it could be negative for future studies to base their perception of I4.0/LP combinations on speculative statements on their mechanisms. Therefore, future quantitative studies relating (a1) I4.0 levels, (a2) LP levels, and (b) levels of results in production should be encouraged, as they have proven to give valid, critical descriptions on compatibility. However, they should be careful to make speculative statements on combination mechanisms. Instead, we encourage future studies that are intentionally tailored to investigate such combination mechanisms – for example in the form of case studies.

6 Conclusions Prior literature on I4.0/LP combinations includes a rich variation of studies, and as the research area expands its diversity will probably also increase. Our aim with this systematic literature review has been to guide the future development of this body of knowledge towards increased consistency, but also increased diversity. Concerning theoretical implications, this study introduces concepts to further characterise important phenomena in the research area, like “I4.0/LP combinations”, “combination compatibility”, and “combination mechanisms”. The use of more neutral terms when addressing these phenomena might enable both increased consistency but also more diverse perspectives on combinations of I4.0 and LP in production in the future. We also highlight limitations of prior studies that may have led to biases. Concerning practical implications, we have argued for consistency in how studies evaluate the merits of combinations – in particular the necessity to link such indicators to results in production – and in how they investigate combination mechanisms. We hope that this will eventually give production managers a diverse range of options for what I4.0/LP combinations they can use to their advantage, and how to use them effectively and efficiently.

References 1. Buer, S.-V., Strandhagen, J.O., Chan, F.T.S.: The link between industry 4.0 and lean manufacturing: mapping current research and establishing a research agenda. Int. J. Prod. Res. 56(8), 2924–2940 (2018). https://doi.org/10.1080/00207543.2018.1442945 2. Kamble, S., Gunasekaran, A., Dhone, N.C.: Industry 4.0 and lean manufacturing practices for sustainable organisational performance in Indian manufacturing companies. Int J Prod Res 58(5), 1319–1337 (2020). https://doi.org/10.1080/00207543.2019.1630772 3. Tortorella, G.L., Fettermann, D.: Implementation of industry 4.0 and lean production in Brazilian manufacturing companies. Int. J. Prod. Res. 56(8), 2975–2987 (2018). https://doi.org/10. 1080/00207543.2017.1391420 4. Tseng, M.L., Tran, T.P.T., Ha, H.M., Bui, T.D., Lim, M.K.: Sustainable industrial and operation engineering trends and challenges toward Industry 4.0: a data driven analysis. J. Ind. Prod. Eng. 38(8), 581–598 (2021). https://doi.org/10.1080/21681015.2021.1950227 5. Sanders, A., Elangeswaran, C., Wulfsberg, J.: Industry 4.0 implies lean manufacturing: research activities in industry 4.0 function as enablers for lean manufacturing. J. Industrial Eng. Manage. 9(3), 811–833 (2016). https://doi.org/10.3926/JIEM.1940

A Systematic Literature Review on Combinations of Industry 4.0

153

6. Chiarini, A., Kumar, M.: Lean six sigma and industry 4.0 integration for operational excellence: evidence from italian manufacturing companies. Production Planning and Control 32(13), 1084–1101 (2021). https://doi.org/10.1080/09537287.2020.1784485 7. Mayr, A., et al.: Lean 4.0-A conceptual conjunction of lean management and Industry 4.0. Procedia CIRP 72, 622–628 (2018). https://doi.org/10.1016/J.PROCIR.2018.03.292 8. Ghobakhloo, M., Fathi, M.: Corporate survival in Industry 4.0 era: the enabling role of leandigitized manufacturing. J. Manuf. Technol. Manag. 31(1), 1–30 (2020). https://doi.org/10. 1108/JMTM-11-2018-0417 9. Rosin, F., Forget, P., Lamouri, S., Pellerin, R.: Impacts of industry 4.0 technologies on lean principles. Int. J. Prod. Res. 58(6), 1644–1661 (2020). https://doi.org/10.1080/00207543. 2019.1672902 10. Antony, J., McDermott, O., Powell, D., Sony, M.: The evolution and future of lean six sigma 4.0. TQM Journal (2022). https://doi.org/10.1108/TQM-04-2022-0135 11. Ejsmont, K., Gladysz, B., Corti, D., Castaño, F., Mohammed, W.M., Martinez Lastra, J.L.: Towards ‘lean industry 4.0’–current trends and future perspectives. Cogent Business and Manage. 7(1) (2020). https://doi.org/10.1080/23311975.2020.1781995 12. Pagliosa, M., Tortorella, G.L., Carlos, J., Ferreira, E., Ferreira, J.C.E.: Industry 4.0 and lean manufacturing a systematic literature review and future research directions. J. Manuf. Technol. Manag. 32(3), 543–569 (2019). https://doi.org/10.1108/JMTM-12-2018-0446 13. Bittencourt, V.L., Alves, A.C., Leão, C.P.: Industry 4.0 triggered by lean thinking: insights from a systematic literature review. Int. J. Prod. Res. 59(5), 1496–1510 (2021). https://doi. org/10.1080/00207543.2020.1832274 14. Rai, A. Patnayakuni, A., Seth, N.: Firm performance impacts of digitally enabled supply chain integration capabilities. MIS Q 30(2), 225–246 (2006). https://doi.org/10.2307/25148729 15. Mumford, E.: The story of socio-technical design: reflections on its successes, failures and potential. Inf. Syst. J. 16(4), 317–342 (2006). https://doi.org/10.1111/J.1365-2575.2006.002 21.X 16. Pacchini, P.T., Lucato, W.C., Facchini, F., Mummolo, G.: The degree of readiness for the implementation of Industry 4.0. Comput. Ind. 113 (2019). https://doi.org/10.1016/J.COM PIND.2019.103125 17. Dev, N.K., Shankar, R., Qaiser, F.H.: Industry 4.0 and circular economy: operational excellence for sustainable reverse supply chain performance. Resour. Conserv. Recycl. 153 (2020) 18. Tortorella, G.L., Fogliatto, F.S., Cauchick-Miguel, P.A., Kurnia, S., Jurburg, D.: Integration of industry 4.0 technologies into Total Productive maintenance practices. Int. J. Prod. Econ. 240 (2021). https://doi.org/10.1016/J.IJPE.2021.108224 19. Pérez, L., Rodríguez-Jiménez, S., Rodríguez, N., Usamentiaga, R., García, D.F.: Digital twin and virtual reality based methodolo-gy for multi-robot manufacturing cell commissioning. Applied Sciences (Switzerland) 10(10) (2020). https://doi.org/10.3390/APP10103633 20. Mrugalska, B., Wyrwicka, M.K.: Towards lean production in industry 4.0. Procedia Eng. 182, 466–473 (2017). https://doi.org/10.1016/J.PROENG.2017.03.135 21. Kolberg, D., Zühlke, D.: Lean automation enabled by indus-try 4.0 technologies. IFACPapersOnLine 28(3), 1870–1875 (2015). https://doi.org/10.1016/J.IFACOL.2015.06.359 22. Buer, S.-V.V., Semini, M., Strandhagen, J.O., Sgarbossa, F.: The complementary effect of lean manufacturing and digitalisation on operational performance. Int. J. Prod. Res. 59(7), 1976–1992 (2021). https://doi.org/10.1080/00207543.2020.1790684 23. Núñez-Merino, M., Maqueira-Marín, J.M., Moyano-Fuentes, J., Martínez-Jurado, P.J.: Information and digital technologies of Industry 4.0 and Lean supply chain management: a systematic literature review. Int. J. Prod. Res. 58(16), 5034–5061 (2020). https://doi.org/10.1080/ 00207543.2020.1743896

154

K. Ericsson and A. Maffei

24. Wagner, T., Herrmann, C., Thiede, S.: Industry 4.0 impacts on lean production systems. Procedia CIRP 63, 125–131 (2017). https://doi.org/10.1016/J.PROCIR.2017.02.041 25. Tortorella, G.L., Giglio, R., van Dun, D.H.: Industry 4.0 adoption as a moderator of the impact of lean production practices on operational performance improvement. Int. J. Oper. Prod. Manag. 39, 860–886 (2019). https://doi.org/10.1108/IJOPM-01-2019-0005 26. Ciano, M.P., Dallasega, P., Orzes, G., Rossi, T.: One-to-one relationships between Industry 4.0 technologies and lean production techniques: a multiple case study. Int. J. Prod. Res. 59(5), 1386–1410 (2021). https://doi.org/10.1080/00207543.2020.1821119 27. Sony, M.: Industry 4.0 and lean management: a proposed integration model and research propositions. Prod. Manuf. Res. 6(1), 416–432 (2018). https://doi.org/10.1080/21693277. 2018.1540949 28. Kipper, L.M., Furstenau, L.B., Hoppe, D., Frozza, R., Iepsen, S.: Scopus scientific mapping production in industry 4.0 (2011–2018): a bibliometric analysis. Int. J. Prod. Res. 58(6), 1605–1627 (2020). https://doi.org/10.1080/00207543.2019.1671625 29. Tortorella, G., Miorando, R., Caiado, R., Nascimento, D., Portioli Staudacher, A.: The mediating effect of employees’ involvement on the relationship between Industry 4.0 and operational performance improvement. Total Quality Management and Business Excellence 32(1–2), 119–133 (2021). https://doi.org/10.1080/14783363.2018.1532789 30. Rossini, M., Costa, F., Tortorella, G.L., Portioli-Staudacher, A.: The interrelation between Industry 4.0 and lean production: an empirical study on European manufacturers. Int. J. Adv. Manuf. Technol. 102(9–12), 3963–3976 (2019). https://doi.org/10.1007/S00170-01903441-7 31. Kolberg, D., Knobloch, J., Zühlke, D.: Towards a lean automa-tion interface for workstations. Int. J. Prod. Res. 55(10), 2845–2856 (2017). https://doi.org/10.1080/00207543.2016.1223384 32. De Giovanni, P., Cariola, A.: Process innovation through industry 4.0 technologies, lean practices and green supply chains. Research in Transportation Economics 90 (2021). https:// doi.org/10.1016/J.RETREC.2020.100869 33. Tripathi, V., Chattopadhyaya, S., Mukhopadhyay, A.K., Sharma, S., Li, C., Di Bona, G.: A sustainable methodology using lean and smart manufacturing for the cleaner production of shop floor management in industry 4.0. Mathematics 10(3) (2022). https://doi.org/10.3390/ MATH10030347 34. Dixit, A., Jakhar, S.K., Kumar, P.: Does lean and sustainable manufacturing lead to industry 4.0 adoption: the mediating role of ambidextrous innovation capabilities. Technol Forecast Soc Change, 175 (2022). https://doi.org/10.1016/J.TECHFORE.2021.121328 35. Varela, L., Araújo, A., Ávila, P., Castro, H., Putnik, G.: Evaluation of the relation between lean manufacturing, industry 4.0, and sustainability. Sustainability (Switzerland) 11(5) (2019). https://doi.org/10.3390/SU11051439 36. Ciliberto, C., Szopik-Depczy´nska, K., Tarczy´nska-Łuniewska, M., Ruggieri, A., Ioppolo, G.: Enabling the circular economy transition: a sustainable lean manufacturing recipe for Industry 4.0. Bus Strategy Environ 30(7), 3255–3272 (2021) 37. Shahin, M., Chen, F.F., Bouzary, H., Krishnaiyer, K.: Integration of lean practices and industry 4.0 technologies: smart manufacturing for next-generation enterprises. Int. J. Adv. Manuf. Technol. 107(5–6), 2927–2936 (2020). https://doi.org/10.1007/s00170-020-05124-0 38. Amjad, M.S., Rafique, M.Z., Khan, M.A.: Leveraging optimized and cleaner production through industry 4.0. Sustain Prod Consum 26, 859–871 (2021). https://doi.org/10.1016/J. SPC.2021.01.001 39. Luz Tortorella, G., Cauchick-Miguel, P.A., Li, W., Staines, J., McFarlane, D.: What does operational excellence mean in the fourth industrial revolution era?. Int. J. Prod. Res. 60(9), 2901–2917 (2022). https://doi.org/10.1080/00207543.2021.1905903

A Systematic Literature Review on Combinations of Industry 4.0

155

40. Ivanov, D., Dolgui, A., Das, A., Sokolov, B.: Digital supply chain twins: managing the ripple effect, resilience, and disruption risks by data-driven optimization, simulation, and visibility. International Series in Operations Research and Management Sci. 276, 309–332 (2019). https://doi.org/10.1007/978-3-030-14302-2_15 41. Meudt, T., Metternich, J., Abele, E.: Value stream mapping 4.0: Holistic examination of value stream and information logistics in production. CIRP Ann Manuf. Technol. 66(1), 413–416 (2017). https://doi.org/10.1016/J.CIRP.2017.04.005 42. Tortorella, G.L., Sawhney, R., Jurburg, D., de Paula, I.C., Tlapa, D., Thurer, M.: Towards the proposition of a lean automation framework: integrating industry 4.0 into lean production. J. Manuf. Technol. Manag. 32(3), 593–620 (2021). https://doi.org/10.1108/JMTM-01-20190032 43. Dombrowski, U., Richter, T., Krenkel, P.: Interdependencies of industrie 4.0 & lean production systems: a use cases analysis. Procedia Manuf. 11, 1061–1068 (2017). https://doi.org/10.1016/ J.PROMFG.2017.07.217 44. Davies, R., Coole, T., Smith, A.: Review of socio-technical considerations to ensure successful implementation of industry 4.0. Procedia Manuf. 11, 1288–1295 (2017). https://doi.org/10. 1016/J.PROMFG.2017.07.256 45. Yilmaz, A., Dora, M., Hezarkhani, B., Kumar, M.: Lean and industry 4.0: Mapping determinants and barriers from a social, envi-ronmental, and operational perspective. Technol. Forecast Soc. Change 175 (2022). https://doi.org/10.1016/J.TECHFORE.2021.121320 46. Ghobakhloo, M., Azar, A.: Business excellence via advanced manufacturing technology and lean-agile manufacturing. J. Manuf. Technol. Manag. 29(1), 2–24 (2018). https://doi.org/10. 1108/JMTM-03-2017-0049 47. Tortorella, G.L., Pradhan, N., Macias de Anda, E., Trevino Martinez, S., Sawhney, R., Kumar, M.: Designing lean value streams in the fourth industrial revolution era: proposition of technology-integrated guidelines. Int. J. Prod. Res. 58(16), 5020–5033(2020). https://doi. org/10.1080/00207543.2020.1743893 48. Tortorella, G.L., Rossini, M., Costa, F., Portioli Staudacher, A., Sawhney, R.: A comparison on industry 4.0 and Lean Production between manufacturers from emerging and developed economies. Total Quality Management and Business Excellence 32(11–12), 1249–1270 (2021). https://doi.org/10.1080/14783363.2019.1696184 49. Javaid, M., Haleem, A., Singh, R.P., Rab, S., Suman, R., Khan, S.: Exploring relationships between lean 4.0 and manufacturing industry. Ind. Robot. 49(3), 402–414 (2022). https://doi. org/10.1108/IR-08-2021-0184 50. Ferreira, W.P., Armellini, F., de Santa-Eulalia, L.A., Thomasset-Laperrière, V.: Extending the lean value stream mapping to the context of industry 4.0: an agent-based technology approach. J. Manuf. Syst. 63, 1–14 (2022). https://doi.org/10.1016/J.JMSY.2022.02.002 51. Bin Abd Rahman, M.S., Mohamad, E., Bin Abdul Rahman, A.A.: Development of IoT— enabled data analytics enhance decision support system for lean manufacturing process improvement. Concurr. Eng. Res. Appl. 29(3), 208–220 (2021). https://doi.org/10.1177/106 3293X20987911 52. Haddud, A., Khare, A.: Digitalizing supply chains potential benefits and impact on lean operations. Int. J. Lean Six Sigma 11(4), 731–765 (2020). https://doi.org/10.1108/IJLSS-032019-0026 53. Tortorella, G.L., Narayanamurthy, G., Thurer, M.: Identifying pathways to a high-performing lean automation implementation: An empirical study in the manufacturing industry. Int. J. Prod. Econ. 231 (2021). https://doi.org/10.1016/J.IJPE.2020.107918

156

K. Ericsson and A. Maffei

54. Yadav, N., Shankar, R., Singh, S.P.: Impact of industry4.0/icts, lean six sigma and quality management systems on organisational performance. TQM J. 32(4), 815–835 (2020). https:// doi.org/10.1108/TQM-10-2019-0251 55. Butt, J.: A strategic roadmap for the manufacturing industry to implement industry 4.0. Designs (Basel) 4(2), 1–31 (2020). https://doi.org/10.3390/DESIGNS4020011 56. Huang, Z., Kim, J., Sadri, A., Dowey, S., Dargusch, M.S.: Industry 4.0: development of a multi-agent system for dynamic value stream mapping in SMEs. J. Manuf. Syst. 52, 1–12 (2019). https://doi.org/10.1016/J.JMSY.2019.05.001

Lean and Digital Strategy Role in Achieving a Successful Digital Transformation Stefano Frecassetti(B) , Anna Presciuttini , Matteo Rossini, and Alberto Portioli-Staudacher Department of Management, Economics and Industrial Engineering, Politecnico Di Milano, Milan, Italy [email protected]

Abstract. Nowadays, digital technologies are completely transforming the way in which companies are managed. After the massive spread of digital technologies thanks to the Industry 4.0 phenomenon, companies needed to find a way to integrate them into their processes, particularly in the manufacturing industry. Thus, in the last years, several researchers investigated ways to integrate them in the best way to exploit the benefits they can give. It has been discovered that one fundamental pillar that can facilitate their implementation is Lean Production. This study wants to focus even more on this relationship by showing how companies with a higher degree of Lean Production adoption tend to have a more structured strategy in their digital transformation path and how this has an effect on the implementation performance. By using a survey, this study shows how Lean Production is associated with the company’s digitalisation strategy and how leveraging this strategy is fundamental to achieving outstanding implementation performance. Keywords: Lean · Strategy · I4.0 · Performance · Survey

1 Introduction Recently the global economy and all industries have been subjected to many strong drawbacks and challenges, starting with the COVID-19 outbreak and damage from geopolitical events all around the world and being overwhelmed with the new environmental conditions and regulations. The COVID-19 breakout put some industries out of business and overwhelmed others with demand. A major spinoff of this pandemic effect on manufacturing has been disruption to the production pillars and supply chains. The Supply chain disruption resulted in many problems, such as delivery delays, cost increases, and uncertainty. On the demand side, the pandemic impact on manufacturing has resulted in two distinct outcomes: decreased demand in some areas and increased demand in others. Due to current world crises and inflation, almost all industries stumbled again post-COVID-19. In addition to the former consequences, manufacturers nowadays suffer from increased energy prices, raw materials prices, and higher production costs. The consequences are evident and drastic for manufacturers. They still must find new ways to power their production facilities, manage their supply network, and deliver their © IFIP International Federation for Information Processing 2023 Published by Springer Nature Switzerland AG 2023 E. Alfnes et al. (Eds.): APMS 2023, IFIP AICT 689, pp. 157–170, 2023. https://doi.org/10.1007/978-3-031-43662-8_12

158

S. Frecassetti et al.

products. Furthermore, they must do so at a time of unprecedented customer expectations around price, personalisation, and service. Manufacturers who are familiar with the vital role Lean Production (LP) can play in overcoming the presented challenges have been able to cope with them and keep competing [1]. In fact, for all industries, the keywords were efficiency and waste reduction, two main pillars of the Lean philosophy. The LP consists of several practices that save energy and material, hence, a less vulnerable production process when it comes to constantly disrupted supply chains [2]. One of the LP practices’ greatest byproducts is lower prices, which is a definite competitive edge to any industry, especially amidst all the above-mentioned price-increasing situations. However, the LP practices can support even more companies in facing these vigorous challenges through its integration with Industry 4.0 (I4.0) technologies. In fact, thanks to the use of I4.0 digital technologies, the company can obtain even more advantages from LP, having more accurate and time-sensitive data and tools. Lean and I4.0 are a powerful combination with common goals of extreme efficiency, productivity, and performance [3]. Data-driven systems provide actionable insights that save time, reduce error, and maximise opportunities far beyond the capabilities of humans. According to the recent literature, LP practices are likely to have a pivotal role in the fourth industrial revolution. Some literature shed light on how LP is essential for making I4.0 a real solid and growing technological breakthrough [4]. Other studies revealed that companies adopting LP before and during I4.0 implementation always showed better results than companies relying solely on I4.0 ([3–5]). Given all of these facts, it is clear how LP positively influences the companies’ digital transformation and greatly impacts how companies make investments. Particularly, [4] highlighted how companies with a higher LP maturity tend to follow a different and more structured path toward I4.0 implementation. Nevertheless, there is a scarcity of contributions underlining the impact of having a structured strategy in successful I4.0 technologies implementation and how much LP influence the company strategy. Thus, this paper will try to answer the following research questions: – “Are the LP practices associated with the firms’ strategy for I4.0 implementation?” – “Which impact do LP and strategy have on I4.0 implementation performance?” This article is organised in the following way: Sect. 2 presents the literature review related to the topics under analysis. Section 3 describes the methodology used, while Sect. 4 highlights the results from the data analysis. Section 5 links the results to what is already known in the literature, while Sect. 6 presents the conclusion and limitations of this study.

2 Literature Review 2.1 Lean Production (LP) LP is based on the principle that waste is the enemy of productivity and efficiency and offers a structured methodology that maximises value for the customer by eliminating waste from an organisation’s activities. It emphasises removing wasteful steps in a process and taking only value-added steps. As a result, the Lean method ensures high quality and customer satisfaction [2].

Lean and Digital Strategy Role in Achieving a Successful

159

The concept of Lean has gained significant popularity over the past few decades, especially in the manufacturing industry. It has become a widely accepted management philosophy, strongly emphasising continuous improvement and waste reduction [6]. However, the concept of Lean has also been applied to other industries such as healthcare, services, and construction [7]. Recently, a particular interest has been given to integrating LP with the innovative digital technologies from I4.0 [3], positively impacting firms’ operational performance [4]. 2.2 Industry 4.0 (I4.0) I4.0 is a recent concept representing the fourth industrial revolution, characterised by the integration of advanced digital technologies and the Internet of Things (IoT) in the manufacturing industry [8]. This new paradigm shift promises to transform the manufacturing industry by improving production efficiency and flexibility and reducing costs [9]. The convergence of various technologies, including artificial intelligence (AI), robotics, machine learning, cloud computing, and big data analytics, enables the realisation of I4.0 [10]. The concept of I4.0 is based on the interconnection of cyber-physical systems, which enable the seamless exchange of information between machines, products, and people [11]. This interconnectedness creates a “smart factory,” where machines, equipment, and products can communicate in real-time and make decentralised decisions based on the data collected [12]. The technologies related to the concept of I4.0 are several, as depicted in Table 1.

Additive Manufacturing (3D printing)/Rapid Prototyping

Bai et al., 2020

Ciano et al., 2020

Dalenogare et al., 2018

Bai et al., 2017

Wang et al., 2015

Muhuri et al., 2019

Frank et al., 2019

Table 1. I4.0 Technologies

X X

Artificial Intelligence/Machine Learning

X

X

Augmented Reality

X

X X

Autonomous Robots (Robotics)

X

Big Data and Analytics

X

Blockchain Technologies Cloud Computing Cyber Security

X X X X X X

X

Automatic Guided Vehicles (AGV)/Drones Internet of Things/Industrial Internet of Things (IoT)

X X

X

RFID Tags

X X

Sensors and Actuators

X X

Simulation

X X

160

S. Frecassetti et al.

The benefits of I4.0 are vast, with increased productivity, efficiency, and quality as the primary advantages [9]. Moreover, I4.0 can help companies reduce costs, optimise their supply chain, and improve their response time to market demands [10, 28]. A list of the main performance indicators firms is willing to improve after I4.0 technologies implementation is displayed in Table 2.

Profitability

X

X

Return on Investment (ROI)

X

X

Costs

X

X

X

X

X

Quality Productivity/Production Efficiency

X X

X

X

X

X

X

X

Flexibility

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

Lin et al., 2019

Bravi & Murmura, 2021

Pansare et al., 2022

Nara et al., 2021 X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X X

X X

Frederico et al., 2020

Acioli et al., 2021

X

X

X X

Antony et al., 2023

Bai et al., 2020

Marcucci et al., 2022

Szász et al., 2020 X

X

Process Time/Cycle Time Delivery/Time-To-Market

X

Calış Duman et al., 2021

Dalenogare et al., 2018

Kamble et al., 2020

Li et al., 2020

Table 2. Performance Indicators Impacted by I4.0 technologies implementation

X

X

X

X

X

X

X

X

Adopting I4.0 can also help companies achieve sustainability goals by reducing energy consumption and waste generation [1, 13]. In addition, the use of data analytics and AI can provide insights for better decision-making and enhance the overall performance of the manufacturing process [11]. 2.3 Lean Production and Industry 4.0 I4.0 and LP are popular paradigms that have gained significant attention recently. While some authors argued that I4.0 would reduce the need for LP, others believe the two paradigms can coexist and complement each other. In fact, several authors [3, 14] argued that I4.0 and LP can complement each other. They suggest LP practices can be applied to I4.0 technologies to ensure they are used effectively. For instance, LP can help identify and eliminate waste in implementing I4.0 technologies. Similarly, LP can help ensure that I4.0 technologies are integrated seamlessly into the organisation’s processes, reducing the likelihood of errors and improving overall efficiency. [15] stated that I4.0 will not become an industrial revolution if it is not integrated into LP. There are three types of relationships between I4.0 and LP [27] stated that the first relationship is one of integration, where I4.0 and LP are perfectly integrated and exploit the benefits of both paradigms. The second one is the positive effect that can be obtained through the digitalisation of LP practices through I4.0 technologies. While the third one is the facilitating effect that LP has on I4.0 technologies implementation.

Lean and Digital Strategy Role in Achieving a Successful

161

2.4 Lean Production Facilitating Industry 4.0 Some authors pointed out how implementing I4.0 requires a structured and systematic approach that involves several steps. Before implementation, there is a need for a current state assessment to spot possible improvements to be made and identify the most appropriate technologies, and also plan their integration with existing processes [16]. After doing this, the company should develop an implementation plan [17]. In doing this, allocating resources and planning other contingent activities is important, such as training the operators on using the new technologies [16, 30] and planning regular monitoring to assess whether the planned outcomes are reached or not [17]. In this sense, continuous improvement is essential to ensure that the implementation of Industry 4.0 remains current with the latest technological advancements and business needs [16]. 2.5 Strategic Elements for the Implementation of I4.0 Many authors have pointed out that to have a successful digital transformation, firms need to have a structured and well-designed strategy to cope with barriers and thus implement digital technologies successfully [31]. Some of them considered in their transformation framework factors, such as strategic partnerships, as a fundamental part of embarking on the digital path [29]. Instead, others pointed out how having a “human-centred” strategy can help have a better and easier digital journey [33]. Nonetheless, when strategies are applied to a company, they must be aligned with the embedded firm’s strategy and cultural aspects. In fact, according to some authors, “digital transformation strategies have a cross-functional character and need to be aligned with other functional and operational strategies” [32]. Thus, being aligned with companies’ operational strategies or, even more, leveraging them can exploit the benefits deriving from the technologies implemented. Among the others, LP has been considered by some authors as a strategic element to leverage for achieving a successful digital transformation. In fact, having a structured implementation of I4.0 technologies, coupled with adopting LP practices, can contribute positively towards the success of technology implementation. By having a clear plan and expertise, companies can leverage the technology and data to achieve their goals more efficiently, significantly improving their performance [18]. In fact, LP is mentioned by several authors as a requirement for the successful implementation of I4.0 solutions. This is supported by the hypothesis that automating inefficient processes will magnify their inefficiency. LP aims to eliminate all forms of waste in manufacturing by recognising and removing not-value-adding activities, streamlining the process, and creating standardised routines. Automation and digitalisation of the manufacturing process are much better when workstations of low levels of complexity [19]. According to [9], a production process that has previously incorporated LP practices seems to be more likely to be shaped and governed. Therefore, they argue that an LP-based environment is a better foundation for constructing a smart manufacturing platform. As a result, I4.0 necessitates a certain degree of process configuration, with defined processes, suppliers and customers, activities, and timelines. The development and implementation of efficient processes with defined standards and a high customer focus is a major expectation and goal of LP systems. In addition, according to the study conducted by [5], companies with

162

S. Frecassetti et al.

higher levels of performance improvement in the last few years that have been implementing LP extensively are more likely to adopt I4.0 technologies concurrently. Also, [20] proved with their case study that the absence of lean before technology implementation could highly affect the outcome. Lastly, others demonstrated how LP is a foundation and the training act as a driver for firms’ digital transformation [30]. In conclusion, the main strategical drivers which can facilitate I4.0 technologies implementation can be summarised as follows: • Having a structured process in I4.0 technologies implementation. • Assessment of the current state before the implementation of I4.0 technologies. • Definition of the goals and objectives to identify the specific areas to apply I4.0 technologies. • Plan development for the technical integration to identify specific technologies to be used and their integration in the process. • Implementation plan development to integrate I4.0 technologies. • Continuous evaluation and monitoring of the results coming from the implemented I4.0 technologies. • Training and workshops to involve workers in the use of the new technologies • Use of LP practices to standardise the processes before I4.0 technologies implementation. • Use of LP practices to remove waste before I4.0 technologies implementation. • Use of LP as a base structure for I4.0 technologies implementation. • Lean was implemented before I4.0 technologies. Some authors identified how companies with high LP maturity usually undertake a different digital transformation strategy [21]. According to them, Lean companies usually go for a more standardised and planned digital transformation, while others use a more disruptive approach. The first is known to be more successful in the long term, which is highly attributed to having a strong base of LP. Nevertheless, it is still difficult to understand quantitatively how frequent this is and how much the strategy differs, particularly regarding implementation success. In this sense, this study will try to overcome this gap by focusing on how LP and digital strategy are associated and how they impact the final implementation outcome.

3 Methodology 3.1 Survey Creation and Data Collection Questionnaire Development. The authors used the survey methodology to answer the research questions. The questionnaire was developed in six different sections: after the introductory section, in which questions related to general information about the company (e.g., size, years of lean implementations, etc.), the questionnaire provided a section completely dedicated to the LP assessment in which the 41 questions related to the [2] model were asked. Each practice is described in a statement that was evaluated according to a Likert scale that ranged from 1 (fully disagree) to 5 (fully agree) [22]. Afterwards, the second section

Lean and Digital Strategy Role in Achieving a Successful

163

presented questions related to I4.0 technologies (as in Table 1), while the third section comprised questions related to the company’s implementation strategy (Sect. 2.4). A 5-point Likert scale ranging from 1 (not used) to 5 (fully adopted) was applied to each question of these two sections. The last section discussed performance improvement after implementing I4.0 technologies (Table 2). A 5-point scale ranging from 1 (not being met) to 5 (completely met) is used in this last part. Data Collection. Responses to the questionnaire were collected through Microsoft Forms during February and March 2023. The applied criteria for respondents’ selection followed the ones proposed by [5]. The authors have decided to have no control in terms of the sector of the respondents. The final sample for the analysis comprised 73 valid responses (22.1% response rate). As a first step before analysing the data, we evaluated the differences in means between the early (respondents to the first e-mail (n1 = 46) and the late respondents (after follow-ups; n2 = 27) respondents using Levene’s test and a T-test [23]; both had significant results (p-value > 0.05) showing that the means between the two groups were not different. Then, we computed Cronbach’s alpha values for all the different sections of the study (LP, I4.0, Strategy, Performance) to ensure the reliability of the responses. All the sections were considered reliable since all the alpha values were higher than the threshold set of 0.6. ([4, 24]). In particular, the values were 0.947 (LP), 0.888 (I4.0), 0.937 (Strategy), and 0.911 (Performance). 3.2 Data Analysis Descriptive Analysis. The analysis of the responses started with a descriptive analysis of the respondents by using the data gathered in the demographic section. The majority (66%) of the respondents were from large enterprises (>250 employees), and the companies had a corporateowned type of ownership (66%). Also, 71% of the respondents declared to have B2B (business-to-business) as a business operating model. In contrast, for the sector, 63% belonged to the manufacturing sector, while the others were distributed as in Fig. 1.

Fig.1. Number of respondents by sector

164

S. Frecassetti et al.

Also, the majority of respondents (62%) showed a high expertise in LP (more than five years of experience), while only 23% had more than five years of experience in I4.0 (Fig. 2 and Fig. 3). As the last step of the descriptive analysis, the average level of adoption of I4.0 technologies was plotted in Fig. 4, where the majority of the companies responded to employ in substantial way technologies such as Cybersecurity, Cloud Computing, and Sensors, while other technologies such as Blockchain or Augmented Reality were not commonly used in their processes.

Fig.2. Number of respondents by years of LP expertise

Fig.3. Number of respondents by years of I4.0 expertise

Statistical Analysis. As a first step for the statistical analysis, we followed an approach similar to previous studies ([4, 5]). Thus, one cluster for each section of the questionnaire was created based on the LP practices (Sect. 1), I4.0 technologies adoption (Sect. 2), strategy (Sect. 3), and implementation performance (Sect. 4). The first step used the hierarchical method to define the suitable number of clusters (k). Subsequently, the observations were reorganised in the k-clusters using the k-means clustering methodology [25]. The number of clusters (k) was equal to two for all four different sections. Regarding LP, the respondents were divided into two clusters: High Lean Implementation (HLI) and Low Lean Implementation (LLI). The High Lean Implementation

Lean and Digital Strategy Role in Achieving a Successful

165

Fig.4. I4.0 technologies’ level of adoption

cluster comprised 36 respondents with a higher average LP level, while Low Lean Implementation consisted of 37 responses. The same approach identified two clusters for I4.0 technologies adoption: Low Digital Adopters (LDA) and High Digital Adopters (HDA). The first one included 46 respondents (with a lower average level of adoption), while the second one included the remaining 27. For the strategy level, the identified clusters were High Strategy Adoption (HSA) which comprised 41 respondents, and Low Strategy Adopters (LSA), which had 32 respondents. The last clustering analysis was performed on the performance; two main clusters were identified as Low Performance (LP) and High Performance (HP), with 28 respondents and 45 respondents. After this clustering phase, we obtained different categorical variables for each respondent, which were further analysed to answer the research questions. Thus, the authors selected the chi-square test with contingency tables as a technique that was deemed to be the most appropriate methodology to address the scope of this paper ([4, 26]). Thus, the authors proceeded with a confirmatory study to assess whether the LP’s level was associated with I4.0 and operational performance [22]. Then, the association between LP and strategy was tested together with the implementation performance.

4 Findings This section will present the survey results and their interpretation to answer the research question. The first step was confirming what was obtained by previous studies, particularly the one of [4], thus that LP implementation level is associated with I4.0 technologies adoption, and both impact firms’ performance, the chi-squared tests related to this were carried out (Tables 3 and 4). As can be seen, the presence of LP practices affects the firms’ implementation performance. Companies with higher levels of LP are the ones who get the most benefits in terms of performance, while companies with low levels are more likely to obtain less. Instead, I4.0 technologies adoption certainly has an impact on the performance but seems to be more restrained. In fact, on the one hand, companies with high levels of adoption almost certainly increase their performance. In contrast, companies with few technology adoptions can still have great results in terms of performance.

166

S. Frecassetti et al. Table 3. Contingency table for LP and Performance Lean Production HLI

LLI

HP

30**

15**

LP

6*

22*

Total

36

37

Performance

*significant at 1% **significant at 5%

Table 4. Contingency table for I4.0 and Performance Industry 4.0 HDA

LDA

HP

22***

23

LP

5**

23***

Total

27

46

Performance

*significant at 1% **significant at 5% ***significant at 10%

Afterwards, the chi-squared tests for computing the association between LP and Strategy (Table 5) were computed, and second between Strategy and Performance (Table 6). Table 5 shows that companies with a high degree of LP practices implementation are mostly oriented to have a structured strategy in their digital technologies implementation. They are more likely to leverage LP practices before introducing I4.0 technologies in their processes. On the other hand, companies with a low level of LP practices tend to have a less structured strategy. Looking at Table 6, it is also clear how a structured strategy impacts the final outcome of the I4.0 technologies implementation. In fact, in most cases, companies adopting a structured strategy have a higher chance of getting greater benefits from technology implementation and, thus, as an outcome, better implementation performance. On the other hand, companies with low performance tend to be the ones who do not have a structured strategy while embarking on the I4.0 technologies implementation path.

5 Discussion At first, the results presented in the Findings section confirm what was found in previous studies; thus, companies with a higher degree of LP adoption are more likely to perform better. Also, the presence of LP and I4.0 practices is associated. Thus, companies likely to embark on the digital transformation journey must first move toward LP [4]. Besides this, this study also confirms the strategies differences between LP and nonLP companies highlighted by [21]. In fact, the results showed how companies with high

Lean and Digital Strategy Role in Achieving a Successful

167

Table 5. Contingency table for LP and Strategy Lean Production HLI

LLI

HSA

30*

11*

LSA

6*

26*

Total

36

37

Strategy

*significant at 1%

Table 6. Contingency table for Strategy and Performance Performance HP

LP

HSA

33**

8**

LSA

12**

20*

Total

45

28

Strategy

*significant at 1% **significant at 5%

levels of LP tend to have a structured strategy, thus meaning that they have a more structured and planned approach to digitalisation. Instead, the other companies tend to have a less structured approach without an extensive assessment and planning phase or leveraging LP practices to introduce digital technologies. This approach is not beneficial, especially compared to a structured strategy. In most cases, the final performances will be affected negatively. Thus, this study focused even more on the importance of having a strategy before introducing I4.0 technologies. In fact, the results showed how having a structured strategy significantly impacts the implementation performance and thus has a huge impact on successful implementation. A structured strategy has been recognised as a key success factor in successful digitalisation. Furthermore, LP has a fundamental role in the strategy since many of the strategic actions that companies do, such as training the employees [30], supplier relationships [29], and people involvement [33], are referred to and achievable through LP practices. From this, it is possible to assume that companies with high LP maturity tend to align more with most of the best practices and driving forces highlighted in the literature. Firstly, it is possible to hypothesise that firms having high expertise in LP tend to be more eager to change and cope with the challenges that digital transformation is bringing, given their dedication to continuous improvement. Thus, it could be possible that these companies tend to leverage the LP practices, on the one hand, to create a more standardised and straightforward environment, which can ease the implementation of I4.0 technologies, but on the other hand, to create a synergy between humans and technologies to achieve an improved digital transformation process. In fact, among the others, one of the key pillars of the LP principles is the involvement of workers with training,

168

S. Frecassetti et al.

decision making and problem-solving. Thus it can be hypothesised that companies with high LP maturity grades will incorporate these factors into their digital transformation strategy. In light of the latest advances in terms of Industry 5.0, where the human-centric perspective is the master, the results from this study shed light on the importance of integrating LP with the latest digital technologies to achieve improved implementation outcomes and, at the same time to keep the focus on workers centricity and wellbeing. Hence, this study stresses the importance of implementing LP before digitalisation and how much leveraging its practices is key to a successful digital transformation path.

6 Conclusions and Limitations This study explored the association between LP and the companies’ digitalisation strategies and their impact on the implementation performance. The findings indicated that companies with higher levels of LP are more willing to adopt a structured strategy for I4.0 technologies. Furtherly, adopting a structured strategy positively affects the successful implementation of I4.0 technologies in terms of performance. Furtherly, this study confirmed what was obtained from previous studies confirming a prevailing effect of LP over I4.0 to achieve performance improvement. Nevertheless, being Lean is not enough since the presence of LP practices is not the only factor to consider to achieve a successful digital transformation. Companies need to be aware of using and designing a correct strategy. In this sense, leveraging the existing LP practices is crucial and useful to achieve this goal. Thus, the contributions of this study are many and can be divided into theoretical and practical. Theoretically, this paper enriches the literature regarding LP and I4.0 by providing novel insights regarding strategy and confirming previous studies. From a practical perspective, this piece will be useful for managers and companies willing to embark on a digital transformation journey by providing a study depicting the importance of LP practices and creating a structured strategy. In that way, the implementation process of I4.0 technologies can be done in a more conscious, robust, and effective way. This study also presents its own limitation, which has to be underlined. Firstly, the sample size is limited and could be enhanced to consider other contextual factors or develop deeper statistical analyses. Also, by extending the sample size, it will be possible to generalise the findings from this study more reasonably. Secondly, simplifying the clustering analysis phase could be considered a weakness. In fact, all the LP-related parameters were considered in one single variable, and the same approach has been used for all the other themes (I4.0, Strategy, Performance). Nonetheless, this could be too simplistic because some specific relation between the different themes could be neglected. Further studies with a bigger sample size can focus on this part by adding a deeper level of analysis and using other statistical tools to analyse and interpret the results. Lastly, another possible weakness of this study is the absence of consideration regarding contextual factors regarding the implementation of LP and I4.0. According to previous studies, these factors have been recognised as influential to both paradigms. However, even though they were asked in the questionnaire, they have not been included in the analysis. Thus, further studies can include these factors in a more complex and nuanced analysis.

Lean and Digital Strategy Role in Achieving a Successful

169

References 1. Ghobakhloo, M., Fathi, M.: Corporate survival in Industry 4.0 era: the enabling role of leandigitised manufacturing. J. Manuf. Technol. Manag. 31(1), 1–30 (2020). https://doi.org/10. 1108/JMTM-11-2018-0417 2. Shah, R., Ward, P.: Defining and developing measures of lean production. J. Oper. Manag. 25(4), 785–805 (2007) 3. Buer, S.V., Semini, M., Strandhagen, J.O., Sgarbossa, F.: The complementary effect of lean manufacturing and digitalisation on operational performance. Int. J. Prod. Res. 59(7), 19761992 (2020). https://doi.org/10.1080/00207543.2020.1790684 4. Rossini, M., Costa, F., Tortorella, G.L., Portioli-Staudacher, A.: International Journal of Advanced Manufacturing Technology (2019). https://doi.org/10.1007/s00170-019-03441-7 5. Tortorella, G., Fettermann, D.: Implementation of Industry 4.0 and lean production in Brazilian manufacturing companies. Int. J. Prod. Res. 56(8), 2975–2987 (2018) 6. Holweg, M.: The genealogy of lean production. J. Oper. Manag. 25(2), 420–437 (2007) 7. Tortorella, G.L., Fogliatto, F.S., Kurnia, S., Thürer, M., Capurro, D.: Healthcare 4.0 digital applications: an empirical study on measures, bundles and patient-centered performance. Technological Forecasting and Social Change, 181, 121780 (2022) 8. Schwab, K.: The Fourth Industrial Revolution. World Economic Forum (2016) 9. Wang, C., von Tunzelmann, N.: Industry 4.0: a perspective of lean production. J. Manuf. Syst. 39, 176–181 (2016) 10. Chen, Y., Preston, D.S., Swink, M.: The impact of industry 4.0 on lean principles. J. Manufacturing Technol. Manage. 30(8), 1181–1197 (2019) 11. Lee, J., Bagheri, B., Kao, H.A.: A cyber-physical systems architecture for industry 4.0-based manufacturing systems. Manufacturing Letters 3, 18–23 (2015) 12. Kagermann, H., Wahlster, W., Helbig, J.: Recommendations for implementing the strategic initiative INDUSTRIE 4.0: Final report of the Industrie 4.0 working group. For-schungsunion (2013) 13. Ghobakhloo, M., Azar, A., Fathi, M.: Industry 4.0 and sustainability implications: a two-stage fuzzy approach. Resour. Conserv. Recycl. 144, 38–51 (2019) 14. Rossini, M., Costa, F., Tortorella, G.L., Valvo, A., Portio-li-Staudacher, A.: Lean production and industry 4.0 integration: how lean automation is emerging in manufacturing industry. Int. J. Prod. Res. 60(21), 6430–6450 (2022) 15. Rüttimann, B.G., Stöckli, M.T.: Lean and industry 4.0-twins, partners, or con-tenders? a due clarification regarding the supposed clash of two production systems. J. Serv. Sci. Manag. 9, 485–500 (2016) 16. Schumacher, A., Erol, S., Sihn, W.: A maturity model for assessing Industry 4.0 readiness and maturity of manufacturing enterprises. Procedia CIRP, 52 (2016) 17. Zhao, Y., Cheng, H., Wang, X., Zhang, Z., Ji, Z.: An industry 4.0-oriented manufacturing process planning framework for smart manufacturing. Int. J. Advanced Manufacturing Technol. 107(9–10), 3255–3270 (2020). https://doi.org/10.1007/s00170-020-05263-2 18. Bascavusoglu-Moreau, E., Schoormans, M.: Industry 4.0: a structured implementation framework. IEEE Eng. Manage. Rev. 47(4), 202–216 (2019) 19. Kolberg, D., Zühlke, D.: Lean automation enabled by industry 4.0 technologies. IFAC PapersOnLine 48(3), 1870–1875 (2015) 20. Frecassetti, S., Ferrazzi, M., Portioli-Staudacher, A.: Lean as a facilitator for AGVs implementation: a case study. In: McDermott, O., Rosa, A., Sá, J.C., Toner, A. (eds) Lean, Green and Sustainability. ELEC 2022. IFIP Advances in Information and Communi-cation Technology, 668. Springer, Cham (2023). https://doi.org/10.1007/978-3-031-25741-4_10

170

S. Frecassetti et al.

21. Rossini, M., Cifone, F.D., Kassem, B., Costa, F., Portioli-Staudacher, A.: Being lean: how to shape digital transformation in the manufacturing sector. J. Manufacturing Technol. Manage. 32(9), 239–259 (2021). https://doi.org/10.1108/JMTM-12-2020-0467 22. Rossini, M., Costa, F., Staudacher, A.P., Tortorella, G.: Industry 4.0 and lean production: an empirical study. IFAC-PapersOnLine, Volume 52, Issue 13, pp. 42–47, ISSN 2405–8963 (2019) 23. Armstrong, J., Overton, S.: Estimating nonresponse bias in mail surveys. J. Mark. Res. 14(3), 396–402 (1977) 24. Meyers, L., Gamst, G., Guarino, A.: Applied Multivariate Research. Sage Publica-tions, Thousand Oaks (2006) 25. Rencher, A.: Methods of Multivariate Analysis. Wiley-Interscience, New Jersey (2002) 26. Tabachnick, B., Fidell, L.: Using Multivariate Statistics. Pearson, Upper Saddle River, NJ (2013) 27. Buer, S.-V., Strandhagen, J.O., Chan, F.T.S.: The link between Industry 4.0 and lean manufacturing: mapping current research and establishing a research agenda. Int. J. Prod. Res. 56(8), 2924–2940 (2018) 28. Frank, A.G., Dalenogare, L.S., Ayala, N.F.: Industry 4.0 technologies: implementation patterns in manufacturing companies. Int. J. Prod. Econ. 210, 15–26 (2019), ISSN 0925–5273. https://doi.org/10.1016/j.ijpe.2019.01.004 29. Bai, C., Kusi-Sarpong, S., Sarkis, J.: An implementation path for green infor-mation technology systems in the Ghanaian mining industry. J. Clean. Prod. 164, 1105–1123 (2017). https:// doi.org/10.1016/j.jclepro.2017.05.151 30. Singh, P., Shirish, A., Kumar, A., O’Shanahan, J.: Lean training as a driver for microbusinesses’ digital transformation. In: McDermott, O., Rosa, A., Sá, J.C., Toner, A. (eds) Lean, Green and Sustainability. ELEC 2022. IFIP Advances in Information and Communication Technology, 668. Springer, Cham (2023). https://doi.org/10.1007/978-3-031-25741-4_12 31. Horváth, D., Szabó, R.Z.: Driving forces and barriers of industry 4.0: do multinational and small and medium-sized companies have equal opportunities?. Technological Forecasting and Social Change, 146, 119–132 (2019), ISSN 0040–1625. https://doi.org/10.1016/j.techfore. 2019.05.021 32. Matt, C., Hess, T., Benlian, A.: Digital transformation strategies. Bus. Inf. Syst. Eng. 57, 339–343 (2015). https://doi.org/10.1007/s12599-015-0401-5 33. Patricio, R., Gomide, P., Rocha, L.: Taking the digital innovation journey beyond technology: a human-centered design approach. J. Innov. Manage. 10(4), 26–46 (2022). https://doi.org/ 10.24840/2183-0606_010.004_00021

Tying Digitalization to the Lean Mindset: A Strategic Digitalization Perspective Victor Eriksson1(B) , Sourav Sengupta1 , Ann-Charlott Pedersen1 , Elsebeth Holmen1 , Heidi Carin Dreyer1 , Marte Daae-Qvale Holmemo1 Signe Sagli1 , Sigrid Eliassen Sand1 , Sunniva Økland1 , Daryl Powell2 , Natalia Iakymenko3 , Serkan Eren3 , and Eirin Lodgaard4

,

1 Department of Industrial Economics and Technology Management, Norwegian University of

Science and Technology, Alfred Getz vei 3, 7034 Trondheim, Norway {victor.eriksson,sourav.sengupta}@ntnu.no 2 SINTEF Manufacturing, Vestfold Innovation Park, 3184 Borre, Norway [email protected] 3 SINTEF Manufacturing, S.P. Andersens vei 3, 7031 Trondheim, Norway 4 SINTEF Manufacturing, Grøndalsvegen 2, 2830 Raufoss, Norway

Abstract. Through qualitative case research involving two cases of digital lean implementation projects, we identify and explicate how grit, vision, and pragmatism (as enablers) and technology maturity, implementability, and idiosyncrasy (as barriers) interact with and affect strategic digitalization initiatives. Seeminglyparadoxical strategic approaches to digitalization emerge from the associations between the enablers and barriers with micro-foundations in continuous learning and purposeful implementation—in other words, a lean mindset. Lean thinking changes how firms work with production processes and practices, approach management culture, envision the larger operational landscape, and manage strategy. In today’s business environment, firms need to consider how to make digital technologies fit their own strategic and operational priorities. The contribution of this study is its presentation of strategic digitalization as a lean approach to competitive differentiation. Keywords: Lean-digital · Paradox · Strategic enabler · Strategic barrier

1 Introduction Amid the hype surrounding digital technologies in the industrial sector, a crucial consideration is how these technologies can assist companies in furthering their efficiencies. The challenge lies in determining how to integrate digital technologies so that they align with the company’s strategic and operational priorities. Management research on lean and digitalization is on the rise [1, 2], and the combination of lean thinking or principles with information and digital technologies is gaining traction among practitioners and academics [3, 4]. However, understanding of which © IFIP International Federation for Information Processing 2023 Published by Springer Nature Switzerland AG 2023 E. Alfnes et al. (Eds.): APMS 2023, IFIP AICT 689, pp. 171–183, 2023. https://doi.org/10.1007/978-3-031-43662-8_13

172

V. Eriksson et al.

organization models and principles strengthen and reap the best of the new digital technologies is inadequate [5]. Nevertheless, previous studies have shown that putting digital technology first contradicts the fundamental lean principle that technology should serve people and processes [6]. How lean is implemented has always been essential, but is perhaps more so now that digital technologies may change the framing of lean principles. Previous studies argue for establishing lean as the foundation and then adding the digital, as lean lays the groundwork for stable processes that can then be digitalized [7]. Firms should avoid creating waste and other digital messes and instead integrate digital solutions into their lean processes [7, 8]. Firm decisions are, to a large extent, becoming more data-driven [9], and there is a need to understand how local data [10] can be fused with external data to make even better decisions [11, 12], as well as to use planning to coordinate supply chain activities with the business strategy [13]. Digital technologies allow for higher visibility and information sharing but also simultaneously lead to increased variability and uncertainty [14, 15]. Digital technologies support organizations in several ways, and they should be formulated to support continuous improvement activities and operational decision-making [10]. However, today’s working environment tends to separate these two streams, which creates a larger distance within the organization and more specialization—instead of integration—of the two. Digital technologies’ potential to transform performance is now widely recognized, and some link efforts in digital transformation with lean principles [3, 8, 11]. One question is how manufacturing firms can use digital technologies and lean principles to simultaneously develop and create value in relationships with their strategically important suppliers. In a recent study, Stark et al. [16, p. 1] state that “without an appropriate operations strategy to capture the value of digitalization, manufacturing companies will be unable to focus on technological investments and operational changes [and] to address this concern, operations management academics must develop new theory through active engagement in the practice of digitalization in manufacturing.” Against this background, this study aims to identify and explicate the barriers to and enablers of the integration of lean principles and digital technologies in manufacturing companies.

2 Theoretical Background 2.1 Lean Management Lean management is arguably the most successful approach to business improvement of our generation, outlasting many other approaches and being adopted by organizations in all kinds of industries worldwide [17]. Ultimately, in aiming to solve more customer problems with fewer resources, lean thinking and practice present managers and executives with a strategy for sustainable (and profitable) business growth. Lean first emerged in the 1990s as a superior business system for better understanding the customer, designing better products, running the factory better, better coordinating the supply chain, and

Tying Digitalization to the Lean Mindset: A Strategic Digitalization Perspective

173

generally presenting managers with a better way of managing the (manufacturing) enterprise [18]; in many instances, lean quickly became a set of best practices to implement in order to extract more value from production processes. However, as we have learned, this is not the case: more recently, researchers have linked lean success and organizational learning [19], in particular the complementarity of lean and action learning [20]. Developing an integrative theory of lean and action learning, Saabye et al. [20, p. 134] suggest that “as a learning capability, lean is about learning in and from action.” This significant reframing shifts the focus from simply implementing lean tools and techniques to using them as learning frameworks to develop people. 2.2 Digitalization The fourth industrial revolution—i.e., Industry 4.0—was a German government initiative presented to the world at the Hannover fair in Germany in 2011. A predominately technology-driven approach, it was set to revolutionize the way in which manufacturing organizations operate, not just in Germany but across the globe. Its vision was clear: the seamless connectivity of smart products and processes in and across smart factories, supply chains, and ecosystems [21, 22]. In reality, the vision of the smart factory remains just that—a vision—although companies have reported successful implementations of so-called key enabling technologies, such as the industrial internet of things (IIoT), artificial intelligence (AI), digital twins, and cloud computing solutions. However, research has shown that 90% of digitalization projects are failing and companies will, on average, waste $4.5 million on failed digitalization projects in the coming years. That said, research has also demonstrated that digitalization can improve firms’ operation performance and financial performance [23, 24]. 2.3 Lean and Digitalization: Intersections and Crossroads The crossroads of lean and digitalization is of growing interest to both industry and academia [25]. Even though many studies assert that digital technologies and lean are complementary [8], the dilemmas and paradoxes emerging from the convergence of strategic or top-down digitalization programs and more grounded, people-centric, bottom-up lean practices, along with issues such as digital waste, remain largely unexplored yet highly relevant [26]. Recent papers in the Journal of Operations Management [27–29] and the International Journal of Operations & Production Management [30] indicate the relevance and timeliness of lean and digitalization research. A distinct research void exists around the associations between the micro-foundations of lean (such as considering continuous learning and purposeful implementation) and digitalization. We address this void in our current study by explicating how these micro-foundations influence the barriers to and enablers of digitalization so that companies can realize the potential of digitalization for efficiency and revenue growth.

174

V. Eriksson et al.

3 Research Methodology This study is part of a four-year multi-work-package research project called “The lean-digitalization paradox: Toward strategic digitalization,” funded by the Norwegian Research Council and aimed at developing theories and managerial guidelines for the digitalization of Norwegian industry based on lean principles. The study herein comprises two cases: the first case (Alpha) includes a single manufacturing research and training center for lean digitalization practices; the second case (Zeta) is a consolidated case compiled from 19 interviews with Norwegian companies about their lean and digitalization practices. We primarily collected data from in-depth interviews (Zeta and Alpha) and participant observations (Alpha), but we also collected secondary data including project reports, websites, and white papers [31, 32]. We analyzed the cases by following recommendations from both Miles et al. [33] and Creswell [34], in that we organized the data (i.e., process), coded the data (i.e., process), and then wrote a representation of the data (product). We followed the Gioia approach for rigor of the data analysis. First, we grouped the first-order empirical findings, after which we identified second-order themes through abstraction. Finally, we aggregated these abstractions into the enablers of and barriers to strategic digitalization. In the next chapter, we present our findings from an overall observational perspective and at a deeper level, with explanations of the underlying structures (i.e., barriers and enablers) related to lean and digitalization.

4 Case Background 4.1 Alpha Case The Alpha case focuses on augmented reality (AR) glasses in a lean manufacturing simulator for training purposes. Alpha is a lab in Raufoss at one of Norway’s first catapult centers and was first opened in August 2009 as a full-scale analog lean manufacturing simulator. The lab offers a lean environment: the workplaces along with parts and materials are organized based on 5S principles, the operations are standardized and described on standard operation paper sheets, the production is leveled and based on pull principles, and continuous improvement work—based on manually-gathered data analyzed by humans—is manually performed on analog boards. Furthermore, the current level of digitization is at a basic actuator and monitoring level, meaning no experience with digital technology for the operators. In addition, practical training is demonstrated in a training arena that consists of a small-scale model treehouse assembly and disassembly line comprising various components such as windows, doors, air-conditioning units, shades, chimneys, and the main body, which are attached to each other with bolts and nuts using a drilling machine. Through multiple production rounds, different lean techniques are applied and compared so as to evaluate their effectiveness. The Alpha case aims to test AR technology in a generic environment to document technology usage. For example, to support quality control and disassembly of the houses, operators in the lab were provided with AR glasses and were required to determine the

Tying Digitalization to the Lean Mindset: A Strategic Digitalization Perspective

175

optimal method for component assembly and disassembly. Two operators simultaneously worked to tighten 48 nuts gauging the right torque (thus ensuring the correct assembly of 14 different parts) as well as disassemble all of the 14 parts in the correct order, place them in the correct box, and check that all parts were positioned correctly in the boxes (not upside down or turned 180 degrees). As a digital lean tool, the AR glasses helped guide the operators through the various steps. 4.2 Zeta Case The Zeta case focuses on how lean management and digital technologies are used in Norwegian industrial companies and how these are intertwined or complementary, thereby demonstrating the plurality of simultaneous lean and digital practices. The digital approach is implemented across many firms in Norway, not merely in leading industrial firms. Interviewees’ experiences working with lean and digitalization ranged from a few months to 20+ years. The interview study comprised various firms and sections of Norwegian industry, including firms working with electronic components, industrial lighting, windows and doors, furniture, construction, defense, steel components, and dairy products. The involved firms produce standardized and customized products with varying repetitiveness. The firms’ sizes range widely, from small and medium-sized enterprises (SMEs) with 50 employees and a turnover of NOK 50 million to large established firms with 3500 employees and NOK 4.5 billion. Moreover, the firms have worked with lean and digitalization to varying degrees by focusing on different aspects. For example, the majority of the firms see lean as a toolbox and not as a philosophy, and many have tried to find their own ways to work with the lean “toolbox,” whereby some firms have hired consultants to increase their lean capabilities. Moreover, the firms’ lean maturity levels (in terms of how they use lean to measure lean’s influence on operational results) differ, as some firms cannot account for it, others can exemplify, and still others can quantity the results.

5 Enablers and Barriers to Strategic Digitalization From the two cases, we identified three high-level enablers—namely, grit, vision, and pragmatism—and three barriers—namely, technology maturity, implementability, and idiosyncrasy—that we discuss in the following subsections (see Fig. 1 for enablers and Fig. 2 for barriers). 5.1 “Grit” as an Enabler It has been argued that the concept of “grit,” defined as the “passion and perseverance for long-term goals” [35, p. 269], plays an essential part in the successful adoption of new technologies [36]. Our research shows that grit as a personal quality is required at all levels, from executive management to factory operators, to make the leap to digital systems. Resistance to change, a lack of skills and knowledge, and insufficient resources or capabilities can all hinder digitalization initiatives. Grit drives individuals and organizations to continuously learn and grow from their failures and challenges and to persevere for

176

V. Eriksson et al.

the greater good in the long term despite initial setbacks. Leaders who are passionate and perseverant about digitalization are crucial to a top-down strategy, but that enthusiasm must filter down through the organization so that even the lowest-level employees understand the long-term benefits of getting involved. Upskilling the workforce or engaging in bottom-up innovation in exploiting digital opportunities would be extremely challenging without grit. 5.2 “Vision” as an Enabler Vision helps organizations have a purpose and a clear direction toward digitalization for problem-solving [37]. Vision not only provides a sense of direction or a road map but also helps spot anomalies and potential disruptions along the way. This is why a strategic vision for digitalization involves not only foreseeing the potential opportunities and gains that may result from this pursuit but also anticipating the challenges that may arise [38]. For instance, if the leadership cultivates trust and inclusivity among the workforce and subtly conveys the message that future digital initiatives are not meant to replace anyone but rather to support their daily activities, they could avert disruptions and boost participation and ownership. Furthermore, an organization’s digitalization efforts are more likely to be focused, effective, and aligned with its overall lean objectives if a compelling vision for digitalization has been developed and communicated. A clear “vision” emerges as an essential enabler of strategic digitalization – vision structures goals, and guides organizational alignment, motivation, and decisions. 5.3 “Pragmatism” as an Enabler Pragmatism is a way of thinking about and solving problems that emphasizes practicality and an action-oriented approach, forming a connection between means and ends, and focusing on long-term goals [39]. When seen through a pragmatist’s lens, strategic digitalization is an action whose consequences are considered from an empirical perspective so as to render them self-evident or understandable without the need for elaborate rationalization. For example, as seen in the Alpha case, the high customization needs of AR technology do not prevent its use, but it is not used to help in every operation. Instead, low-complexity operations are chosen for AR technology use, because they are better suited to adaptation. This aligns with the continuous learning mindset: implementing new technologies to improve processes leads to a better understanding of those processes and exposes other underlying problems—i.e., lean transformation. The pragmatist worldview centers on the values of curiosity and open-mindedness, emphasizes experimentation and innovation through improved collaboration and communication, and requires an iterative, case-by-case review of choices and their implications [40]. Its role in fostering an outcome-driven environment that rewards experimentation and iteration, embraces change, and promotes collaboration and communication makes pragmatism a macro enabler of strategic digitalization that is in line with the organization’s larger lean objectives.

Tying Digitalization to the Lean Mindset: A Strategic Digitalization Perspective

177

5.4 “Technology Maturity” as a Barrier There are benefits to adopting new technologies, but managers often remain skeptical because they are still in the “emerging” stage of maturity [41]. Hype-induced perception [42] toward emerging technologies can be optimistic or pessimistic, but from a realistic standpoint, the technology must also attain a level of maturity whereby it can be relied on as a solution and in a manner that is superior to the other more mature but less-promising or modern alternatives that are available. Given that we recognize the existence of more-mature-but-less-promising alternatives, technological maturity in terms of merely accumulating years of existence and being tried and tested does not necessarily equate to effectiveness. Maturity is a continuum: what matters is that the solution is suitable to the needs at hand, ideally outperforms alternatives in terms of both costs and benefits, indicates its ability to do so even in the future, and has the capacity to grow and adapt to changing situations. With this interpretation, new technologies’ lack of maturity is not a deal-breaker, and it is also possible that more mature alternatives are not the most viable alternatives. Overemphasis of maturity in terms of years or usage is a barrier that impedes innovation and radical problem-solving, and maturity becomes a barrier when a technology solution is implemented without the above-mentioned micro evaluations and does not have interoperability with the legacy systems. For example, in the Alpha case, we observed that for the AR technology to function, the treehouses needed to be adjusted to the ground and not in their usual positions; to fix this issue, the AR team developed an add-in software solution, but it eventually failed. The failure was not a result of the low maturity of the AR technology but rather the associated capabilities and expertise involved in implementing the AR technology. 5.5 “Implementability” as a Barrier Implementability is not a one-size-fits-all concept, and many organizations face technological and cultural barriers, among others [43, 44]. In the Alpha case, they needed customization for each application, which sacrificed efficiency for effectiveness. Implementability necessitates the development of supportive capabilities and infrastructure for effectiveness; as Firm 15 from the Zeta case observes, this may require an update of work processes and a thorough understanding of how the transition is to be approached and carried out. Digitalization can potentially solve a problem, but only if in the process it does not negatively impact the many other implicitly and explicitly interconnected activities. The most promising digital solutions may not be feasible in a specific business setting, due to the other interdependencies that the solutions affect; a holistic or systems view may reveal such interdependencies as underlying micro barriers to strategic digitalization. 5.6 “Idiosyncrasy” as a Barrier Idiosyncrasy is related to the unique business setting in which the technology is used rather than the technology itself [38, 44]. A solution that works well in one business context may not work well in another, and when the microenvironment is ignored, a

178

V. Eriksson et al.

promising or “hyped” technology may fail to meet expectations. For example, while AR headsets appear to be effective in problem-solving in general, they are not appropriate for workers with impaired vision or for use over extended periods of time during the day. Furthermore, as Firm 11 of the Zeta case described, a single digital platform is needed that supports all of the organization’s activities, and each organization has a unique set of activities that a platform should support. Thus, effectiveness requires more than just a collection of individually-advanced technologies; said technologies should work well together or complement each other as part of a larger system. Digitalization is more of a dynamic capability-building process that seeks to leverage the combination that best suits the business context, resulting in competitive differentiation rather than equifinality. Representative empirical illustrations

Having basic technology knowledge/skills for the operators is crucial to implementing the tools efficiently – Alpha Case We have a fleet of machines that have built-in sensors that gather data; we have had these for 10–12 years but have not used the data for anything – Zeta Case, Firm 19 Now we will run some courses together with the operators. Because we want all operators to move up to a higher digital level. So that we get everyone together to use it more actively – Zeta Case, Firm 16

Operators should be informed that technology is not a substitute for them but a tool that makes their daily tasks easier – Alpha Case Yes, digital processes support lean or our improvement work. That is the point. For example, it better supports solving or understanding a problem if you have data. If you have data about a problem that you have, then you understand more easily – Zeta Case, Firm 1 In the future, the operational focus will have more digitization in it, because that is where the gains can be made – Zeta Case, Firm 14

Due to the high customization requirements of the technology, operations with low complexity are chosen – Alpha Case I hear many contradictions between lean and digital lean, considering the philosophy and the way of thinking. I do not see the intention; we have to change and look at how we can go from analog work to digital – Zeta Case, Firm 19 Lean has always been part of us, but I think what is meant by lean is going to change for quite a few companies going forward, because it is going to be more about digitalization – Zeta Case, Firm 14

Abstraction

Aggregation - Enablers

Need to acquire new skills at all hierarchical levels of the organization This would require passion and perseverance at all levels Therefore, all must be able to associate the effort needed with their personal set of goals to put in the effort; that altogether adds up to the organizational goal of enabling topdown and bottom-up engagement simultaneously

Operators need assurance. Need to specify upfront the need and implications of the technologies on their job and continuous learning Clarity is needed on how and why the technology is to be used, what is the value added in the short and long term, and how it contributes to continuous improvement

All technologies are not equally applicable as they are in all contexts Need to realistically and purposefully plan to embrace technologies by identifying the right set of problems that these could address

Fig. 1. Data structure for the enablers of strategic digitalization.

Tying Digitalization to the Lean Mindset: A Strategic Digitalization Perspective Abstraction

Representative empirical illustrations

Not a 100% off-the-shelf product yet – Alpha Case Another challenge encountered was the need to fix the tree houses to the ground for certain operations in order to ensure the correct functioning of the AR. The AR provider attempted to address this issue through the development of a software solution, but it was ultimately unsuccessful – Alpha Case

Aggregation - Barriers

Maturity of the technology is not always the cause for failure in implementation; often, it is the implementation capabilities and expertise involved to adjust to the onthe-ground needs

It required customization for each application – Alpha Case

Implementation is not a onesize-fits-all approach

Technological development runs faster than people’s ability to absorb it. Setting up a new system is straightforward, but the work process must be updated. How exactly are we going to attack these things then? – Zeta Case, Firm 15

Technology-based solutions are only as good as the human abilities to skillfully adjust and use them

The solution is not suitable for people with impaired vision – Alpha Case

User of the technologies and the context in which they are used must be in synergy

What I have seen and thought about the most is the challenge of getting digitization to support lean work. It is difficult to choose good platforms that support everything we do, because the company does so many different things – Zeta Case, Firm 11

179

Technology is a tool that must be adapted while keeping people in mind, not the other way around

Fig. 2. Data structure for the barriers to strategic digitalization.

6 Uncovering the Paradoxical Strategic Approaches Following our analysis of the enablers and the barriers, we observed associations between them that translated into four seemingly paradoxical strategic approaches to digitalization (Fig. 3). We elaborate on these in the following subsections.

ar in Pr ep

ity

n

tio ica

n

fo rG

en

er

al

Barriers

tio up ty

Sp ec

li bi

ifi

ci

ty

ta rS

fo

M

if pl

g

to

m Si

Enablers

sr Di

y

rit

u at

sm

ati

gm

Pra

d

be

it

te

Un

Gr

ica

pr

pl

ep

m

ar

Co

ed

Vision

Implementability

Fig. 3. Framework for strategic digitalization.

Id io

sy

nc

ra

sy

180

V. Eriksson et al.

6.1 Preparing to Be Unprepared The demands of the environment are constantly shifting, as is technology. Therefore, focusing on developing the micro-abilities and -structures to embrace continuous uncertainties and to accommodate ever-“maturing” technologies—with “grit” catalyzing the transitions and facilitating a longer-term orientation—is more important than being distracted by the shorter-term “gain” of the potential technology use, i.e., finding a match between what is known as a problem and what is available as a mature technical solution today. We call this paradoxical strategic approach to problem-solving “preparing to be unprepared”, because it emphasizes planning for the unknown over quick fixes of the known. 6.2 Complicated Simplification The process of digitalization is intricate; it may not be “pragmatic” in every situation, and it may necessitate a more thorough examination and comprehension of its “implementability.” Moreover, a longer-term futuristic “vision” would seem to complicate otherwise routine activities and even require additional upskilling, restructuring, and other preparations without producing immediate results, which may not be well received by the workforce. As a result, implementation may face significant resistance unless leadership steps in to explain how what appears complex at first will actually streamline processes over time. We call this paradoxical strategic approach a “complicated simplification”, due to the fact that the road to digital transformation in efficiency or automation is difficult in the short term but could pay off in the long term. 6.3 Specificity for Generality The “pragmatic” approach requires tailoring one’s level of openness and innovation to a given situation. Due to their non-generalizability and lack of standardization, digital technologies favor differentiation over uniformity. Furthermore, the effectiveness of an “implementation” calls for a set of capabilities to be developed that are considered to be unique to the problem at hand. However, when these capabilities are combined with various technologies, solutions may emerge that are not problem-specific but rather generalizable. By leveraging an assortment of technologies and capabilities to deal with “idiosyncrasies”, managers can create solutions that can be applied to a wide range of problems in the whole system. 6.4 Disruption for Stability Because “grit” is an enabler of strategic digitalization, it implies that the outcomes of digitalization may be disruptive, involve failures, and possibly destabilize the status quo in the short term. However, initial frictions and difficulties should not deter the process; continuous learning through disruptions and perseverance should be promoted, effectively balancing them in order to develop a new and more robust form of stability over the long term. In the short term, this would also require the development of the necessary capabilities for “implementability”, and disruptions can also reveal hidden interdependencies that could be exploited to design for “maturity” and effectiveness.

Tying Digitalization to the Lean Mindset: A Strategic Digitalization Perspective

181

7 Contributions and Conclusions A firm’s strategy is guided by its goal; lean is the measure to achieve that goal, while technology is the lever to adjust those measures. Thus, strategic digitalization involves a mix of top-down and bottom-up approaches. Identifying opportunities for achieving goals is a bottom-up approach through defined measures (lean) that are closer to the shop floor, but it has to be approached via levers designed at the top (digitalization). The notion is that these flow down and back up again—the ebb and flow of lean and digitalization. In other words, we conceive of strategic digitalization as a lean approach to competitive differentiation. The contributions of this research are twofold. First, we identify six high-level enablers and barriers that interact with and affect strategic digitalization initiatives. These enablers and barriers highlight what firms need to consider in their endeavors to integrate lean with digitalization, to ensure that they proceed cautiously. Second, based on the identified enablers and barriers, we present a framework for strategic digitalization, highlighting four seemingly paradoxical associations between the enablers and the barriers, which expose their micro-foundations in continuous learning and purposeful implementation. In summary, lean thinking changes how firms work with production processes and practices, approach management culture, envision the larger operational landscape, and manage strategy. In today’s business environment, firms need to have a lean mindset in order to fit digital technologies to the firm’s strategic and operational priorities.

References 1. Sanders, A., Elangeswaran, C., Wulfsberg, J.: Industry 4.0 implies lean manufacturing: research activities in industry 4.0 function as enablers for lean manufacturing. J. Industrial Eng. Manage. 9(3), 811–833 (2016) 2. Cattaneo, L., Rossi, M., Negri, E., Powell, D., Terzi, S.: Lean thinking in the digital era. In: Ríos, J., Bernard, A., Bouras, A., Foufou, S. (eds.) Product Lifecycle Management and the Industry of the Future. PLM 2017. IFIP Advances in Information and Communication Technology, 517. Springer, Cham (2017). https://doi.org/10.1007/978-3-319-72905-3_33 3. Pinho, C., Mendes, L.: IT in lean-based manufacturing industries: systematic literature review and research issues. Int. J. Prod. Res. 55(24), 7524–7540 (2017) 4. Santhiapillai, F.P., Chandima Ratnayake, R.M.: A lean digital approach to after-delivery services: a case study from the multi-supplier retail industry. In: Kim, K. Y., Monplaisir, L., Rickli, J. (eds.) Flexible Automation and Intelligent Manufacturing: The Human-Data-Technology Nexus. FAIM 2022. Lecture Notes in Mechanical Engineering. Springer, Cham (2023). https:// doi.org/10.1007/978-3-031-17629-6_27 5. Bodroži´c, Z., Adler, P.S.: The evolution of management models: a neo-schumpeterian theory. Adm. Sci. Q. 63(1), 85–129 (2018) 6. Liker, J.K.: The Toyota Way: 14 Management Principles from the World’s Greatest Manufacturer. McGraw-Hill, New York (2004) 7. Hänggi, R., Fimpel, A., Siegenthaler, R.: LEAN Production–Easy and Comprehensive: A Practical Guide to Lean Processes Explained With Pictures. Springer Nature, Berlin (2022) 8. Buer, S.-V., Semini, M., Strandhagen, J.O., Sgarbossa, F.: The complementary effect of lean manufacturing and digitalisation on operational performance. Int. J. Prod. Res. 59(7), 1976– 1992 (2021)

182

V. Eriksson et al.

9. McAfee, A., Brynjolfsson, E.: Big data: the management revolution. Harv. Bus. Rev. 90(10), 60–68 (2012) 10. Ingvaldsen, J.A.: Organizational learning: bringing the forces of production back in. Organ. Stud. 36(4), 423–444 (2015) 11. Haddud, A., Khare, A.: Digitalizing supply chains potential benefits and impact on lean operations. Int. J. Lean Six Sigma 11(4), 731–765 (2020) 12. Sestino, A., Prete, M.I., Piper, L., Guido, G.: Internet of things and big data as enablers for business digitalization strategies. Technovation 98, 102173 (2020) 13. Jonsson, P., Holmström, J.: Future of supply chain planning: closing the gaps between practice and promise. Int. J. Phys. Distrib. Logist. Manag. 46(1), 62–81 (2016) 14. Büyüközkan, G., Göçer, F.: Digital supply chain: literature review and a proposed framework for future research. Comput. Ind. 97, 157–177 (2018) 15. Hofmann, E., Sternberg, H., Chen, H., Pflaum, A., Prockl, G.: Supply chain management and Industry 4.0: conducting research in the digital age. Int. J. Physical Distribution & Logistics Manage. 49(10), 945–955 (2019) 16. Stark, A., et al.: Hybrid digital manufacturing: capturing the value of digitalization. J. Operations Manage., 1–21 (2022) 17. Jones, D.T., Womack, J.P.: The evolution of lean thinking and practice. In: Netland, T., Powell, D.J. (eds.) The Routledge Companion to Lean Management, Routledge, New York, pp. 3–8 (2016) 18. Womack, J.P., Jones, D.T., Roos, D.: The Machine that Changed the World. Free Press, New York (1990) 19. Powell, D.J., Coughlan, P.: Rethinking lean supplier development as a learning system. Int. J. Oper. Prod. Manag. 40(7/8), 921–943 (2020) 20. Saabye, H., Powell, D.J., Coughlan, P.: Lean and action learning: towards an integrated theory? Int. J. Oper. Prod. Manag. 43(13), 128–151 (2023) 21. Westerman, G., Bonnet, D., McAfee, A.: Leading Digital: Turning Technology into Business Transformation. Harvard Business Press, Boston (2014) 22. Björkdahl, J.: Strategies for digitalization in manufacturing firms. Calif. Manage. Rev. 62(4), 17–36 (2020) 23. Abou-Foul, M., Ruiz-Alba, J.L., Soares, A.: The impact of digitalization and servitization on the financial performance of a firm: an empirical analysis. Prod. Planning & Control 32(12), 975–989 (2021) 24. Hautala-Kankaanpää, T.: The impact of digitalization on firm performance: examining the role of digital culture and the effect of supply chain capability. Bus. Process. Manag. J. 28(8), 90–109 (2022) 25. Ashrafian, A., et al.: Sketching the landscape for lean digital transformation. In: Ameri, F., Stecke, K., von Cieminski, G., Kiritsis, D. (eds) Advances in Production Management Systems. Production Management for the Factory of the Future. APMS 2019. IFIP Advances in Information and Communication Technology, vol. 566. Springer, Cham (2019). https://doi. org/10.1007/978-3-030-30000-5_4 26. Alieva, J., Powell, D.J.: The significance of employee behaviours and soft management practices to avoid digital waste during a digital transformation. Int. J. Lean Six Sigma 14(1), 1–32 (2022) 27. Holmström, J., Holweg, M., Lawson, B., Pil, F.K., Wagner, S.M.: The digitalization of operations and supply chain management: theoretical and methodological implications. J. Oper. Manag. 65, 728–734 (2019) 28. Cusumano, M.A., et al.: Commentaries on “the lenses of lean.” J. Oper. Manag. 67(5), 627–639 (2021) 29. Hopp, W.J., Spearman, M.S.: The lenses of lean: visioning the science and practice of efficiency. J. Oper. Manag. 67, 610–626 (2021)

Tying Digitalization to the Lean Mindset: A Strategic Digitalization Perspective

183

30. Frank, A.G., Marodin, G.A., Godinho Filho, M., Thürer, M.: Call for papers – Beyond Industry 4.0 – integrating lean, digital technologies and people. International Journal of Operations & Production Management (2023) 31. Flick, U.: An Introduction to Qualitative Research. SAGE Publications, London (2014) 32. Mills, J., Birks, M.: Qualitative Methodology. SAGE Publications, London (2014) 33. Miles, M.B., Huberman, A.M., Saldana, J.: Qualitative Data Analysis: A Methods Sourcebook. SAGE Publications, Thousand Oaks (2014) 34. Creswell, J.: Qualitative Inquiry and Research Design: Choosing Among Five Approaches. SAGE Publications, Thousand Oaks (2013) 35. Duckworth, A.: Grit: The Power of Passion and Perseverance. Scribner, New York (2016) 36. Panicker, P.: Embedding Culture and Grit in the Technology Acceptance Model (TAM) for Higher Education. arXiv preprint arXiv:2005.11973 (2020) 37. Albukhitan, S.: Developing digital transformation strategy for manufacturing. Procedia Computer Sci. 170, 664–671 (2020) 38. Bughin, J., Kretschmer, T., van Zeebroeck, N.: Digital technology adoption drives strategic renewal for successful digital transformation. IEEE Eng. Manage. Rev. 49(3), 103–108 (2021) 39. York, J.G.: Pragmatic sustainability: translating environmental ethics into competitive advantage. J. Bus. Ethics 85, 97–109 (2009) 40. Jacobs, D.C.: A pragmatist approach to integrity in business ethics. J. Manag. Inq. 13(3), 215–223 (2004) 41. Kretschmer, T., Khashabi, P.: Digital transformation and organization design: an integrated approach. Calif. Manage. Rev. 62(4), 86–104 (2020) 42. Korotkova, N., Benders, J., Mikalef, P., Cameron, D.: Maneuvering between skepticism and optimism about hyped technologies: building trust in digital twins. Information & Manage. 60(4), 103787 (2023) 43. Tabesh, P., Mousavidin, E., Hasani, S.: Implementing big data strategies: a managerial perspective. Bus. Horiz. 62(3), 347–358 (2019) 44. Wimelius, H., Mathiassen, L., Holmström, J., Keil, M.: A paradoxical perspective on technology renewal in digital transformation. Inf. Syst. J. 31(1), 198–225 (2021)

Characterization of Digitally-Advanced Methods in Lean Production Systems 4.0 Simon Schumacher1(B) , Roland Hall1 , Michael Hautzinger1 Jan Schöllmann1 , and Thomas Bauernhansl1,2

,

1 Fraunhofer Institute for Manufacturing Engineering and Automation IPA, Stuttgart, Germany

[email protected] 2 Institute of Industrial Manufacturing and Management, University of Stuttgart, Stuttgart,

Germany

Abstract. Lean Production Systems are enterprise-specific, methodical frameworks for the continuous orientation of all enterprise processes to the customer in order to achieve overall objectives. Due to the increasing complexity of digital transformation, the management of methods in Lean Production Systems 4.0 (LPS 4.0) is a challenging task for industrial engineering practice. Starting with a set of 35 Lean methods, this paper aim to review the existing literature regarding digitally-advanced LPS 4.0 methods. Academic contributions were analyzed in a systematic literature review (SLR). The SLR screening of 706 scientific papers led to the identification of 54 approaches for LPS 4.0 methods. All 54 approaches were assessed regarding their maturity level for detailed documentation. As a result, 18 LPS 4.0 methods are described in a uniform standardized LPS 4.0 method template. The new LPS 4.0 methods are highly relevant for practical application in industrial engineering practice, which has already been tested in a first implementation of methods in an industrial engineering toolbox (Toolbox Lean 4.0). Keywords: Lean Production System · Lean Production · Industrie 4.0 · Industry 5.0 · Knowledge Management · Production Management

1 Introduction Lean Production Systems (LPS) are enterprise-specific, methodical frameworks for the continuous orientation of all enterprise processes to the customer in order to achieve overall objectives [1]. While LPS were essentially shaped by the Toyota Production System and Lean Production approaches in the past decades, a change of LPS is taking place due to the influence of Industrie 4.0 to more advanced Lean Production Systems 4.0 (LPS 4.0) [2–6]. These changes address all levels of LPS [2]. LPS 4.0 can increase process efficiency when implemented in line with Industrie 4.0 requirements, building on an LPS and thus increasing the entrepreneurial target dimensions of quality, costs, and time, among others [4]. © IFIP International Federation for Information Processing 2023 Published by Springer Nature Switzerland AG 2023 E. Alfnes et al. (Eds.): APMS 2023, IFIP AICT 689, pp. 184–199, 2023. https://doi.org/10.1007/978-3-031-43662-8_14

Characterization of Digitally-Advanced Methods in Lean Production Systems 4.0

185

In particular, changes in the operative part of LPS can be seen [6]. The interactions of lean methods [1] and Industrie 4.0, which are becoming apparent as a result of the digital transformation, are of significant operational importance. The quantity and complexity of methods and tools are increasing due to digital transformation [7]. Some of these methods and tools are already part of research projects and are being systematized and scientifically processed [3]. However, often there are unsystematic catalogs, collections, wikis, and database approaches for the application of methods and tools in companies [8]. This presents industrial engineering with the challenge of implementing innovative approaches created by modern information and communication technologies on existing methods and tools [9]. This described problem thus justifies the need for a systematic description of LPS 4.0 methods and tools for industrial engineering [6]. This work aims to conduct a systematic literature review for methods published in the literature in LPS 4.0 and, based on this, to develop individual fact sheets for the target group Industrial engineering. Figure 1 illustrates the research process followed. The methodological approach is a forward-looking, step-by-step literature search using the systematic literature review (SLR) method based on selected databases, search terms, and search criteria [10]. The underlying state of research is relevant literature from standard works and standard literature, as well as current research contributions, which are named individually in the research process. Starting with a set of 35 Lean methods from an established methods catalog [11], this paper aims to review the existing literature regarding digitally-advanced LPS 4.0 methods. For this, the study focuses on the following research question: How can digitally-advanced methods for Lean Production Systems 4.0 be systematically identified, described, and analyzed?

Fig. 1. Overview about research process

This paper is structured as follows. Section 2 gives the necessary definitions and terminology from previous works related to this research. Section 3 lays out the research methodology of this study, focusing on the systematic literature review method and the derivation of characteristics for a standardized template for LPS 4.0 methods. Section 4 comprises the operationalization and results of the SLR, while Sect. 5 provides the reader with the results of the LPS 4.0 method fact sheets. Section 6 describes the first practical implementation results of the analysis and Sect. 7 provides the concluding remarks and an academic outlook.

186

S. Schumacher et al.

2 Related Works and Terminology This section presents relevant definitions and research contributions on managing Lean Production Systems 4.0 (LPS 4.0). A production system is a socio-technical system that transforms inputs into outputs through value-adding processes as part of the transformation process [12]. Production systems have developed via Taylorism, the Toyota Production System, and Lean Production in the twentieth century. Through Taylorism, industrial work was transferred to economical mass production with methodical standardization and rationalization measures. Ford developed the automobile industry’s first comprehensive production system for mass production [5, 13, 14]. The Toyota Production System (TPS) is a holistic, overall system of methods to shorten lead times, reduce costs, and increase quality. The TPS is founded on eliminating waste and the continuous improvement process. The achievement of goals in TPS is made possible by the two pillars of Just in Time and Autonomous Automation (Jidoka) [5, 15]. Lean Production, or Lean Manufacturing, has introduced methodical components of TPS to management practices of production systems and other business areas in the Western world [5, 16–19]. The German guideline for Lean Production Systems (LPS) by [1] consolidates the developments of TPS and Lean Production and, at the same time, expands the definition of production systems by integrating the concept of holism in a company-wide understanding. According to [1], a Lean Production System is defined as ‘an enterprisespecific, methodical system of rules for the continuous orientation of all enterprise processes to the customer in order to achieve the targets set by the enterprise management’ [1]. The general structure of Lean Production Systems is composed of the five levels objectives, processes, principles, methods, and tools, while the content within these levels is company-specific [1]. Today, numerous companies have developed and customized an individual LPS to their specific requirements [1, 5]. Industrie 4.0 is the German vision of a fourth industrial revolution and refers to the intelligent networking of machines and processes for industry based on information and communication technology (ICT) [20–23]. Industrie 4.0 holds various potentials to change the design of LPS, which is being researched by several academics [2, 24–33]. Building on [1, 4] defines Lean Production Systems 4.0 as follows: ‘A Lean Production System 4.0 represents a synergetic approach based on LPS and Industrie 4.0 to improve enterprise processes, which enables the implementation of Industrie 4.0 on the basis of an LPS in order to improve individual enterprise objectives’ [4]. Other relevant concepts and terms comparable to or complementing this definition of LPS 4.0 are Digital Lean Cyber-Physical Production Systems [25], Digital Lean Manufacturing [25], Lean Industry 4.0 [34], Smart Manufacturing [35]. This paper specifically deals with the industrial engineers’ management of methods in LPS 4.0. Industrial engineering is the application of methods and findings for the holistic analysis, evaluation, and design of complex systems, structures, and processes of producing organizations [7, 36]. According to [1], a method is defined as a specific standardized procedure assigned to a principle and used to achieve enterprise targets within LPS [1]. As an evolution of this, a LPS 4.0 method is a specific standardized procedure that is assigned to a principle and is used to achieve enterprise targets within LPS 4.0 [4].

Characterization of Digitally-Advanced Methods in Lean Production Systems 4.0

187

Industrial engineering practice offers several approaches to manage methods in LPS systematically. Typical approaches comprise attribute-based template descriptions with a user-oriented presentation of necessary information per method. See [4, 9, 11, 36, 37] for examples of LPS method templates and catalogs.

3 Research Methodology The analysis performed throughout this paper is based on a sound methodological research procedure. To analyze existing literature, the systematic literature review method is used. To provide users with a standardized as well as systematic description of methods, a standardized template for the description of LPS 4.0 methods is introduced. 3.1 Systematic Literature Review on LPS 4.0 Methods The systematic literature review (SLR) method is described as a conclusive, systematic, and reproducible method for conducting a literature search to identify, evaluate, and synthesize existing literature [10]. Such an SLR always addresses a specific topic and serves to determine the current state of research. Thus, it is a secondary analysis that contributes to the evaluation of already researched content [38]. Based on [10], the methodological approach for this SLR is divided into the following six phases: (1) Definition of a research question; (2) identification of relevant and suitable databases; (3) formulation of appropriate search terms for each database; (4) definition of criteria for the selection process of the searched results; (5) performance and documentation of the literature review; (6) synthesis of the results of the SLR. The corresponding operationalization of the SLR process is explained in detail in Sect. 4. 3.2 Standardized Template for LPS 4.0 Methods Technology fact sheets are an essential tool for innovation and technology planning. For example, they provide a decision-making basis for the selection of technologies in order to identify and provide suitable resources for developing and applying technologies [39]. The collected knowledge can be made available in a structured form for transparency documentation, and information [40]. There are various approaches to the design of the content of technology fact sheets, but there is little agreement on which elements are relevant to the standard of such a fact sheet. The preparation of the information must meet the requirements of the respective target group, here industrial engineering. At the same time, the structured, compact preparation should filter out relevant information [39]. To fulfill this and in order to design comparable fact sheets, templates are used. A basic structure that forms the building blocks of a template is the division into text or image areas. Furthermore, information on name, organizational information, a general description, and references to further information carriers are often presented [39].

188

S. Schumacher et al.

In order to derive a standardized template for LPS 4.0 methods, four established Lean methods catalogs were analyzed and compared for similarities and differences [11, 36, 37, 39]. Attributes of each approach were listed in a table and analyzed in a workshop within the project team. The synthesis of this process led to the following attributes combining the applied standardized template for LPS 4.0 methods used in this paper. – Name. The name field on each template indicates the method’s name by which it is known based on the literature source studied. – Abbreviation. Common abbreviations of the method name exist for some LPS methods. This block is provided for any abbreviations of LPS 4.0 methods. – Design principle. This field is intended for the overarching design principle of the described method. A design principle combines thematically related methods and tools [1]. – Goal/Purpose. Through this module, the goal of the method is described in a free text. This assessment provides for easier, result-oriented access to the methods. – Description. In this section, a brief characterization of the method is presented in the form of a free text. The focus here is summarizing basic information and outlining the approach and step-by-step implementation. – Advantages/chances and disadvantages/risks. Technical, organizational, financial or legal aspects are considered when applying the methods. The potentials and risks that arise during implementation are examined in more detail in the two fields in bullet points. – Supplementary illustrations. In support of the description of the method in the continuous text, this section visualizes the method in summary form in a concept graphic. This is intended to describe the complex interrelationships of the method in graphic form by means of suitable generalization. – Differences to the original method. Compared to classical LPS methods, new challenges, and potentials arise through the use of diverse technologies in LPS 4.0 methods. In this module, these individual changes are mapped to the underlying method of LPS in a bullet point manner. – Maturity level / development status. The maturity level is listed in this field as a result of the evaluated literature source. The classification of the maturity level is mapped by selecting an item of a list field with a concise description of the development status. – References. This field references the literature source used and further reading. Only a single literature source is used for the design of a profile and is listed accordingly. This information should give the user the possibility to deal with the described method in more detail.

4 Systematic Literature Review on LPS 4.0 Methods In the following, the SLR on LPS 4.0 methods carried out is first described with regards to the step-wise operationalization according to the defined research process and second results of the SLR are presented. 4.1 Operationalization of the Systematic Literature Review According to the six-step SLR process mentioned in Sect. 3.1, this section will give details on the operationalization and implementation of the SLR performed. The SLR

Characterization of Digitally-Advanced Methods in Lean Production Systems 4.0

189

was carried out between October and December 2022. English and German search terms were applied to identify the most significant scientific contributions. 1. Definition of the research question. Based on related works presented in Sect. 2, the field of LPS 4.0 was analyzed according to existing approaches for the systematic description of LPS 4.0 methods and gaps in industrial engineering support. This led to the following research question: “How can digitally-advanced methods for Lean Production Systems 4.0 be systematically identified, described and analyzed?” 2. Identification of relevant and suitable databases. Starting with a broad list of scientific databases, criteria for the database selection were defined by the project team: (1) feasibility of individual search terms and criteria; (2) technical engineering focus of the database; (3) amount of databases searched should comply with the study’s pragmatism. As a result, the databases searched are Scopus and Web of Science (WoS). 3. Formulation of appropriate search terms for each database. This SLR had 35 existing LPS methods as its starting point, i.e., each database was to be searched 35 times in the exact same way with different method names. For each of the two databases, the search terms were individually set up of a combination of typical attributes and the according method names. This is an example for a search term (TITLE-ABS-KEY) used in Scopus with method as a placeholder for the method names run through heuristically: (“data based method”) OR (“digital method”) OR (“dynamic method”) OR (“intelligent method”) OR (“smart method”) OR (“virtual method”) OR (“short-term method”) OR (“real-time method”) OR (“method 4.0”). 4. Definition of criteria for the selection process of the searched results. The aim of this SLR is to conduct the evaluation of the databases systematically, completely, and transparently to gain meaningful insights from the results and to filter them by selected criteria. For conducting the literature review, the PRISMA statement method was used [41–43]. This established process was carried out in the stages of identification, screening, assessment, and synthesis of literature according to the defined criteria shown in Fig. 2. 5. Performance and documentation of the literature review. One core reviewer performed the SLR with iterative consultation with the project team. The results for each of the hundreds of searches carried out were documented. 6. Synthesis of the results of the SLR. The synthesis of the SLR results is based on a five-stage maturity model to characterize the applicability of the LPS 4.0 method described. The design criteria for the maturity model itself are in accordance with relevant literature such as [44, 45] and comprises (1) a defined amount of maturity levels, (2) a concise definition/description of each maturity level, (3) a generic description of characteristics per maturity level, (4) problem-oriented fields of action, (5) a generic description for each field of action, (6) a generic description of measures per field of action and each maturity level. Table 1 shows the five maturity levels defined for the assessment of LPS 4.0 methods. In the following chart in Fig. 2, the complete process of the literature review is shown. In the review and evaluation phases, further records were removed based on defined criteria after screening of the title, abstract, and full text, as these are not suitable for the final synthesis of results.

190

S. Schumacher et al. Table 1. Maturity levels for LPS 4.0 method assessment

Maturity level

Description

Level 4 Complete description of the method

In this literature source, a method is described in detail and extensively. Here, formulated features and characteristics enable a holistic view of the method. The description systematics of this method shows similarities to well-known methods of LPS

Level 3 In this literature source, several characteristics More detailed description of several properties of the method are mentioned and described in of the method detail. The source is technically related to the subject area and has several articles with a high information content. The information is of high quality Level 2 Detailed description of a few properties of the method

This literature source describes individual characteristics and features of a specific method in detail. The source technically belongs to the corresponding subject area and provides few contributions with a high information content. Nevertheless, individual descriptions of the method are detailed and offer deep insight

Level 1 Superficial description of a few properties of the method

In this literature source, the method is mentioned, whereby single features and characteristics of the method are mentioned. However, the description of the method is only superficial and not very detailed

Level 0 No description of properties of the method

In this literature source, certain characteristics and features of a method are not explained in detail. Instead, the method is merely mentioned in the body text. The source from which the method originates is technically relevant to the subject area but does not contain any further contributions or information

During the SLR process, the number of all identified records was reduced from the initial total of 706 to 273 papers after filtering of title, keywords and abstract in accordance to the described research focus. Subsequently, another 219 records were removed by the supplementary screening of the full texts, searching for possible descriptions of LPS 4.0 methods. Papers with promising titles, but no further description of a method, were excluded from the analysis. Thus, for the qualitative analysis step, the number of records to be analyzed was 54. This corresponds to the total number of literature sources included in the evaluation procedure according to maturity levels in this study.

Characterization of Digitally-Advanced Methods in Lean Production Systems 4.0

191

Fig. 2. Number of papers analyzed throughout the SLR screening process

4.2 Results of the Systematic Literature Review The SLR results shed light on various new and interesting findings. Even though 54 papers about digitally-advanced LPS methods were identified, the review does not provide any new method for 17 classical LPS methods from [11]. The most qualified literature was found for value stream planning (13) and shopfloor management (11). Table 2 gives an overview of the literature identified per LPS method through the SLR. 4.3 Limitations of the Systematic Literature Review As every literature review, the analysis has limitations. It should be noted that the research design, including the choice of search terms and syntax, have a significant influence on the search results. Since, in this case, the description of the classical LPS methods of [1] was used as the core elements of the search term, the possibilities of finding completely new LPS 4.0 methods were focused but at the same time reduced. No search results were found for multiple methods. On the one hand, this may be due to the relatively young research field, but on the other hand, it may also be due to the chosen research method. Nevertheless, the systematic methodological approach based on the SLR remains valid and sound for the study’s aims. Another important aspect is the quality and origin of the literature sources. The provided information directly influences the content’s scope as well as the quality of the LPS 4.0 methods fact sheets during the final evaluation. In addition, the selection of the methods template’s features and characteristics significantly influences the information content of the document. For new LPS 4.0 methods, additional information such as application examples, use cases, company processes, or the used tools and technologies must be considered to fully describe the methods. The different maturity levels of the methods also have a significant impact on the quality and information content of further processing. Therefore, further research is needed to describe the identified methods in more detail.

192

S. Schumacher et al. Table 2. Quantity of qualified literature identified per LPS design principle and method.

#

Design Principle

Method

Qualified literature

1

Zero defects principle

Autonomation/Jidoka

3

2

Zero defects principle

Poka Yoke

1

3

Zero defects principle

Six Sigma

4

4

Zero defects principle

Statistical process control

3

5

Visual management

Andon

1

6

Visual management

Shop floor management

11

7

Continuous improvement process

Audit

3

8

Continuous improvement process

Cardboard engineering

1

9

Continuous improvement process

Idea management

1

10

Continuous improvement process

PDCA

1

11

Flow principle

Quick changeover (SMED)

1

12

Flow principle

Value stream planning

13

13

Pull principle

Just-in-time/just-in-sequence

3

14

Pull principle

Kanban

4

15

Pull principle

Milk run

1

16

Pull principle

Levelling (Heijunka)

1

17

Avoidance of waste

Waste analysis (Muda)

1

18

Avoidance of waste

Total Productive Maintenance

1

19

Standardization

5S

–

20

Standardization

Process standardization

–

21

Zero defects principle

5 x why

–

22

Zero defects principle

8D report

–

23

Zero defects principle

A3 method

–

24

Zero defects principle

Ishikawa diagram

–

25

Zero defects principle

Short control loops

–

26

Zero defects principle

Worker self-check

–

27

Continuous improvement process

Benchmarking

–

28

Employee orientation & mgmt

Hancho

–

29

Employee orientation & mgmt

Target management

–

30

Flow principle

First in first out (FIFO)

–

31

Flow principle

One piece flow

–

32

Flow principle

U layout

– (continued)

Characterization of Digitally-Advanced Methods in Lean Production Systems 4.0

193

Table 2. (continued) #

Design Principle

Method

Qualified literature

33

Pull principle

Supermarket

–

34

Avoidance of waste

Chaku Chaku

–

35

Avoidance of waste

Low-cost automation

–

5 Analysis Results: Systematic Description of LPS 4.0 Methods As a result of the analysis, a total of 18 LPS 4.0 methods were systematically identified out of 35 LPS methods examined from [1] and transferred into the format of a standardized template. Table 3 consists of a list of all characterized LPS 4.0 methods with their respective maturity level. Figure 3 gives an exemplary overview of a LPS 4.0 method fact sheet for the method Smart Andon. All 18 LPS 4.0 methods are made publicly available in a LPS 4.0 methods catalogue as an attachment to this paper [46]. Table 3. Maturity levels per identified LPS 4.0 method Method according to [1]

LPS 4.0 method

Maturity level

Idea management

Digital idea management [47, 48]

4

Autonomation (Jidoka)

Jidoka 4.0 [3]

4

Kanban

Cyber-physical Kanban 4.0 [48]

4

Andon

(Customized) Smart Andon [49]

4

Value stream planning

Digital value stream twin [50]

3

Audit

Digital Audit [51]

3

Cardboard engineering

Cardboard engineering 4.0 [52]

3

Six Sigma

Six Sigma 4.0 [53]

3

Statistical process control

Dynamic statistical process control [54]

3

Milk run

Dynamic milk run routing [55]

3

Just-in-time/just-in-sequence

Dynamic Just-in-Time [56]

3

Shop floor management

Digital shop floor management [57]

3

PDCA

PDCA 4.0 [58]

2

Poka Yoke

Intelligent Poka-Yoke [59]

2

Quick changeover (SMED)

Intelligent SMED [60]

1

Levelling (Heijunka)

Heijunka 4.0 [61]

1

Waste analysis (Muda)

Digital Muda [62]

1

Total Productive Maintenance

Total Productive Maintenance 4.0 [63]

1

194

S. Schumacher et al.

The resulting 18 fact sheets were assessed based on the above introduction of maturity levels. Four methods only qualified for maturity level 1, two methods reached level 2, eight scored for maturity level 3, and four were classified as complete with maturity level 4. This shows that not all LPS 4.0 methods have yet reached maturity/their full potential, but a majority are already advanced methods ready to be used by industrial engineering. The quality of the fact sheets mainly depends on the available selection of literature sources, and the quality of the information.

Fig. 3. Exemplary LPS 4.0 method fact sheet for the method Smart Andon [49]

Characterization of Digitally-Advanced Methods in Lean Production Systems 4.0

195

6 Evaluation of the First Practical Implementation For initial testing and evaluation, seven out of the 18 LPS 4.0 methods have been implemented in the Toolbox Lean 4.0, a software prototype for an industrial engineering methodical framework containing LPS 4.0 methods [8]. The selection of the first LPS 4.0 methods was made based on the maturity and applicability of the respective methods. Implementing the methods was successful and only required mapping attributes from the developed fact sheets to the toolbox template item fields. The Toolbox Lean 4.0 is being used by applied research to aid industrial engineering in decision-making concerning specific problems, design principles, or solution patterns. It is also part of an industrial working group on LPS 4.0, consisting of 14 industrial firms and over 40 industrial engineering practitioners. The Toolbox Lean 4.0 was presented to users in a very early stage containing the abovementioned seven LPS 4.0 methods from this study. First user feedback was positive with regard to the applicability of the content.

7 Conclusions This paper presents the results of a systematic analysis of the current state of research in digitally-advanced methods of LPS 4.0. The study identified and analyzed 18 new LPS 4.0 methods, whereas the basis for comparison were 35 classical LPS methods from the established Lean methods catalog in [11]. The research methodology follows a sound stepwise process of a systematic literature review. Overall, the elaborated LPS 4.0 methods fact sheets offer a first insight into the developments of LPS 4.0 on the methods level. The applied research process developed in this study is transferable and applicable to other research questions in various fields of engineering and technology. The 18 fact sheets presented in this study can serve as a basis for further research and practical application of LPS 4.0 in industry. As mentioned, this paper only provides a first insight into new LPS 4.0 methods. Nevertheless, the results underline the relevance of the research topic for industrial engineering. One important area to further address is the change at the methods level in LPS. As technologies are now becoming more available for practical use, it is likely that more use cases will emerge in the future, and thus more research contributions will follow. It is, therefore, still necessary to investigate whether and which new methods emerge and how they are shaped and designed. A distinction between modernized and new methods should be analyzed in detail. The introduction of new methods in the field of LPS 4.0 raises questions about how they will establish themselves in production systems practice and what pathways for new methods development exist. In particular, how the human factor will interrelate with technological development. Another important aspect is the relationship between the methods and the adjacent levels of the structure of LPS. In this context, possible changes in existing design principles may occur and an influence on the level of tools is also to be expected. These aspects offer ample potential for further scientific research and will be explored in more detail in future work.

196

S. Schumacher et al.

Acknowledgements. This work was supported by the German Federal Ministry of Education and Research (BMBF) under grant 02L16Y101 (Future Work Lab) and implemented by the Project Management Agency Karlsruhe (PTKA). Credit Author Statement Simon Schumacher – Conceptualization, methodology, writing – original draft. Roland Hall – Conceptualization, methodology, writing – reviewing and editing. Michael Hautzinger – Conceptualization, methodology, writing – original draft. Jan Schöllmann – Visualization, writing – reviewing and editing. Thomas Bauernhansl – Supervision, writing – review and editing.

References 1. VDI : Lean Production Systems: Basic Principles, Introduction, and Review (2870 Part 1) (2012) 2. Dombrowski, U., Richter, T., Krenkel, P.: Interdependencies of industrie 4.0 & lean production systems. Procedia Manufacturing 11, 1061–1068 (2017). https://doi.org/10.1016/j.promfg. 2017.07.217 3. Deuse, J., Dombrowski, U., Nöhring, F., et al.: Systematic combination of Lean Management with digitalization to improve production systems on the example of Jidoka 4.0. Int. J. Eng. Bus. Mgmt. 12, 1–9 (2020). https://doi.org/10.1177/1847979020951351 4. Richter, T.: Vorgehensmodell zur Implementierung von Industrie 4.0 im Kontext Ganzheitlicher Produktionssysteme. Dissertation. Schriftenreihe des IFU, vol 37. Shaker, Düren (2021) 5. Dombrowski, U., Krenkel, P.: Ganzheitliches Produktionsmanagement. Springer, Berlin Heidelberg, Berlin, Heidelberg (2021) 6. Schumacher, S., Hall, R., Bildstein, A., et al.: Lean production systems 4.0: systematic literature review and field study on the digital transformation of lean methods and tools. Int. J. Prod. Res., 1–23 (2022). https://doi.org/10.1080/00207543.2022.2159562 7. Baudin, M., Netland, T.: Introduction to Manufacturing: An Industrial Engineering and Management Perspective. Routledge, New York (2023) 8. Schumacher, S., Hall, R., Bildstein, A., et al.: Toolbox lean 4.0 - development and implementation of a database approach for the management of digital methods and tools. Procedia CIRP 107, 776–781 (2022). https://doi.org/10.1016/j.procir.2022.05.061 9. Dombrowski, U., Richter, T.: The lean production system 4.0 framework. In: Moon, I., Lee, G.M., Park, J., et al. (eds) Advances in Production Management Systems, 536. Springer Int. Publ., Cham, pp. 410–416 (2018) 10. Fink, A.: Conducting Research Literature Reviews: From the Internet to Paper. Sage, Los Angeles (2020) 11. VDI: Lean Production Systems: List of Methods (2870 Part 2) (2013) 12. Nyhuis, P., Heins, M., Pachow-Frauenhofer, J., et al.: Wandlungsfähige Produk-tionssysteme. ZWF 103, 333–337 (2008). https://doi.org/10.3139/104.101285 13. Taylor, F.W.: The Principles of Scientific Management. Harper & Brothers, New York, London (1911) 14. Spath, D.: Ganzheitlich produzieren: Innovative Organisation und Führung. LOG_X Verlag, Wiesbaden (2003) ¯ 15. Ono, T.: Toyota Production System. Productivity Press, New York (1988) 16. Womack, J.P., Jones, D.T.: Lean Thinking. Free Press, New York (2003)

Characterization of Digitally-Advanced Methods in Lean Production Systems 4.0

197

17. Womack, J.P., Jones, D.T., Roos, D.: The Machine that Changed the World: Based on the Massachusetts Institute of Technology 5 Million Dollar 5 Year Study on the Future of the Automobile. Rawson, New York (1990) 18. Liker, J.K.: The Toyota Way: 14 Management Principles from the World’s Greatest Manufacturer, 2nd edn. McGraw-Hill; EBSCO Publ, New York, Ibswich Mass (2021) 19. Shah, R., Ward, P.T.: Defining and developing measures of lean production. J. Oper. Mgmt. 25, 785–805 (2007). https://doi.org/10.1016/j.jom.2007.01.019 20. Kagermann, H., Wahlster, W.: Ten years of industrie 4.0. Sci. 4, 26 (2022). https://doi.org/10. 3390/sci4030026 21. Bauer, D., Schumacher, S., Gust, A., et al.: Characterization of autonomous production by a stage model. Procedia CIRP 81, 192–197 (2019). https://doi.org/10.1016/j.procir.2019.03.034 22. Neugebauer, R.: Digital Transformation. Fraunhofer-Forschungsfokus. Springer Vieweg, Berlin (2019) 23. Schumacher, S., Schmid, F., Bildstein, A., et al.: Lean production systems 4.0: the impact of the digital transformation on production system levels. In: 54th CIRP Conference on Manufacturing Systems. Procedia CIRP 104, pp. 259–264 (2021). https://doi.org/10.1016/j. procir.2021.11.044 24. Meudt, T., Metternich, J., Abele, E.: Value stream mapping 4.0. CIRP Ann. 66, 413–416 (2017). https://doi.org/10.1016/j.cirp.2017.04.005 25. Romero, D., Gaiardelli, P., Powell, D., et al.: Digital lean cyber-physical production systems: the emergence of digital lean manufacturing and the significance of digital waste. In: Moon, I., Lee, G.M., Park, J., et al. (eds) Advances in Production Management Sys-tems. Production Management for Data-Driven, Intelligent, Collaborative, and Sustainable Manufacturing: IFIP WG 5.7 International Conference, APMS 2018, Seoul, Korea, Au-gust 26–30, 2018, Proceedings, Part I. Springer International Publishing, Cham, pp. 11–20 (2018) 26. Rosin, F., Forget, P., Lamouri, S., et al.: Impacts of industry 4.0 technologies on lean principles. Int. J. Prod. Res. 58, 1644–1661 (2020). https://doi.org/10.1080/00207543.2019.1672902 27. Rossini, M., Costa, F., Tortorella, G.L., et al.: Lean Production and industry 4.0 integration: how lean Automation is emerging in manufacturing industry. Int. J. Prod. Res., 1–21 (2021). https://doi.org/10.1080/00207543.2021.1992031 28. Tortorella, G.L., Fettermann, D., Cauchick Miguel, P.A., et al.: Learning organisation and lean production: an empirical research on their relationship. Int. J. Prod. Res. 58, 3650–3666 (2020). https://doi.org/10.1080/00207543.2019.1633028 29. Lorenz, R., Buess, P., Macuvele, J., et al.: Lean and digitalization - contradictions or complements? In: Ameri, F., Stecke, K.E., von Cieminski, G., et al. (eds.) Advances in Produc-tion Management Systems. Springer Int, vol. 566, pp. 77–84. Publ, Cham (2019) 30. Dolgui, A., Bernard, A., Lemoine, D., et al.: Advances in production management systems. Artificial Intelligence for Sustainable and Resilient Production Systems. IFIP Advances in Information and Communication Technology. Springer International Publishing, Cham (2021) 31. Ciano, M.P., Dallasega, P., Orzes, G., et al.: One-to-one relationships between Industry 4.0 technologies and lean production techniques: a multiple case study. Int. J. Prod. Res. 59, 1386–1410 (2021). https://doi.org/10.1080/00207543.2020.1821119 32. Buer, S.-V., Semini, M., Strandhagen, J.O., et al.: The complementary effect of lean manufacturing and digitalisation on operational performance. Int. J. Prod. Res. 59, 1976–1992 (2021). https://doi.org/10.1080/00207543.2020.1790684 33. Langlotz, P., Aurich, J.C.: Causal and temporal relationships within the combination of lean production systems and industry 4.0. Procedia CIRP 96, 236–241 (2021). https://doi.org/10. 1016/j.procir.2021.01.080 34. Hines, P., Tortorella, G.L., Antony, J., et al.: Lean industry 4.0: past, present, and future. Qual. Manag. J. 30, 64–88 (2023). https://doi.org/10.1080/10686967.2022.2144786

198

S. Schumacher et al.

35. Mittal, S., Khan, M.A., Purohit, J.K., et al.: A smart manufacturing adoption framework for SMEs. Int. J. Prod. Res. 58, 1555–1573 (2020). https://doi.org/10.1080/00207543.2019.166 1540 36. Jungkind, W., Könneker, M., Pläster, I., et al.: Handbuch der Prozessoptimierung: Die richtigen Werkzeuge auswählen und zielsicher einsetzen, 1. Auflage. REFA-Kompendium Arbeits organisation, Band 2. Hanser, München (2018) 37. Ullmann, G.: Ganzheitliche Produktionssysteme: IPH-Methodensammlung (2009). https://www.iph-hannover.de/_media/files/downloads/IPH_Lean_Methodensammlung. pdf. Accessed 12 Mar 2023 38. Torraco, R.J.: Writing integrative literature reviews. HR Development Rev. 15, 404–428 (2016). https://doi.org/10.1177/1534484316671606 39. Warschat, J., Schimpf, S., Korell, M. (eds.): Technologien frühzeitig erkennen. FraunhoferVerlag, Stuttgart, Nutzen-potenziale systematisch bewerten (2015) 40. Rummel, S.: Eine bewertungsbasierte Vorgehensweise zur Tauglichkeitsprüfung von Technologiekonzepten in der Technologieentwicklung. Schriftenreihe zu Arbeitswissen-schaft und Technologiemanagement, vol 16. Universitätsbibliothek der Universität Stuttgart, Stuttgart (2015) 41. Antes, G., von Elm, E.: Das PRISMA-Statement - was sollte über systematische Ubersichtsarbeiten berichtet werden? (The PRISMA Statement - what should be reported about systematic reviews?). Dtsch Med Wochenschr 134, 1619 (2009). https://doi.org/10.1055/s0029-1233989 42. Moher, D., Liberati, A., Tetzlaff, J., et al.: Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. PLoS Med 6, e1000097 (2009). https://doi.org/ 10.1371/journal.pmed.1000097 43. Liberati, A., Altman, D.G., Tetzlaff, J., et al.: The PRISMA statement for reporting systematic reviews and meta-analyses of studies that evaluate health care interventions: explanation and elaboration. PLoS Med 6 (2009). https://doi.org/10.1371/journal.pmed.1000100 44. Becker, J., Knackstedt, R., Pöppelbuß, J.: Entwicklung von Reifegradmodellen für das IT-Management. Wirtsch Inform 51, 249–260 (2009). https://doi.org/10.1007/s11576-0090167-9 45. Mettler, T.: Supply-Management im Krankenhaus: Konstruktion und Evaluation eines konfigurierbaren Reifegradmodells zur zielgerichteten Gestaltung. Sankt Gallen, Univ., Diss., 2010, Sierke (2010) 46. Schumacher, S., Hall, R., Hautzinger, M., et al.: LPS 4.0 Methods Catalogue: Attachment to the Paper “Characterization of Digitalized Methods in Lean Production Systems 4.0” (2023). https://owncloud.fraunhofer.de/index.php/s/m7dqQSWvJVihLiS. Accessed 25 Mar 2023 47. Röltgen, A.T., Bernardy, V., Müller, R., et al.: Entwicklung, Einsatz und Evaluation eines Tools für digitales Ideenmanagement. Ein Fallbeispiel. Gr Interakt Org 51, 49–58 (2020). https://doi.org/10.1007/s11612-020-00500-6 48. Stadnicka, D., Bonci, A., Lorenzoni, E., et al.: Symbiotic cyber-physical kanban 4.0: an approach for SMEs. In: 2020 25th IEEE International Conference on Emerging Technologies and Factory Automation (ETFA). IEEE, pp. 140–147 (2020) 49. Lopez-Leyva, J.A., Molina-Inzunza, A., Navarro-Paz, P., et al.: Customized smart andon system to improve the efficiency of industrial departments. JSIR 79, 35–37 (2020). https:// doi.org/10.56042/jsir.v79i1.68111 50. Frick, N., Metternich, J.: The Digital Value Stream Twin. Systems 10, 102 (2022). https://doi. org/10.3390/systems10040102 51. Qiuling, Y.: Innovative practice of digital audit management mode for power grid enterprises. In: 2021 International Conference on Control Science and Electric Power Systems (CSEPS). IEEE, pp. 238–241 (2021)

Characterization of Digitally-Advanced Methods in Lean Production Systems 4.0

199

52. Aufleger, M., Schneider, M.: Cardboard engineering 4.0 – IIoT für jedermann? ZWF 115, 442–445 (2020). https://doi.org/10.3139/104.112361 53. Schäfer, F., Mayr, A., Hess, A., et al.: Six sigma 4.0. ZWF 114, 140–144 (2019). https://doi. org/10.3139/104.112047 54. Cooke, T.A., Howe, W.R.: Dynamic statistical process control limits for power quality trend data. In: 2018 18th International Conference on Harmonics and Quality of Power (ICHQP). IEEE, pp. 1–5 (2018) 55. Adriano, D.D., Montez, C., Novaes, A.G.N., et al.: DMRVR: dynamic milk-run vehicle routing solution using fog-based vehicular ad hoc networks. Electronics 9, 2010 (2020). https:// doi.org/10.3390/electronics9122010 56. Xu, Y., Chen, M.: Improving just-in-time manufacturing operations by using internet of things based solutions. Procedia CIRP 56, 326–331 (2016). https://doi.org/10.1016/j.procir. 2016.10.030 57. Longard, L., Meißner, A., Müller, M., et al.: Digitales shopfloor management – wohin geht die reise? ZWF 115, 645–649 (2020). https://doi.org/10.3139/104.112358 58. Peças, P., Encarnação, J., Gambôa, M., et al.: PDCA 4.0: a new conceptual approach for continuous improvement in the industry 4.0 paradigm. Applied Sci. 11, 7671 (2021). https:// doi.org/10.3390/app11167671 59. Romero, D., Gaiardelli, P., Powell, D.J., et al.: Intelligent poka-yokes: error-proofing and continuous improvement in the digital lean manufacturing world. In: Kim, D.Y., Cieminski, G., Romero, D. (eds) Advances in Production Management Systems. Smart Manufacturing and Logistics Systems: Turning Ideas into Action, 664. Springer Nature Switzerland, Cham, pp. 595–603 (2022) 60. Do˘gan, O., Cebeci, U., Oksuz, M.K.: An intelligent decision support system for SMED and its application in textile industry. In: Proceedings of the International Conference on Industrial Engineering and Operations Management (2018) 61. Kjellsen, H.S., Ramillon, Q.J.L., Dreyer, H.C., et al.: Heijunka 4.0 – key enabling technologies for production levelling in the process industry. In: Dolgui, A., Bernard, A., Lemoine, D., et al. (eds) Advances in Production Management Systems. Artificial Intelligence for Sustainable and Resilient Production Systems, 630. Springer International Publishing, Cham, pp. 704–711 (2021) 62. Alieva, J., Haartman, R.: Digital muda - the new form of waste by industry 4.0. OSCM: An Int J. 13(3), 269–278 (2020). https://doi.org/10.31387/OSCM0420268 63. Valamede, L.S., Akkari, A.C.S.: Lean 4.0: A new holistic approach for the integration of lean manufacturing tools and digital technologies. Int J Math, Eng, Manag Sci. 5, 851–868 (2020). https://doi.org/10.33889/IJMEMS.2020.5.5.066

Synergies Between Industry 4.0 and Lean on Triple Bottom Line Performance Thomas Bortolotti1(B) , Stefania Boscari1 , Willem Grob1 , and Daryl Powell2 1 University of Groningen, PO Box 800, 9700 AV Groningen, The Netherlands

[email protected] 2 SINTEF Manufacturing, Horten, Norway

Abstract. Technological advancements are increasingly integrated into industrial processes. This integration leads to enhanced adaptability and connectivity within the manufacturing sector, commonly referred to as Industry 4.0. As Industry 4.0 remains in its nascent stage, its impact on operational, social, and environmental performance – the Triple Bottom Line dimensions – is relatively understudied. In contrast, Lean has a more established presence in numerous companies. However, potential synergies or conflicts between Lean practices and Industry 4.0 remain largely unexplored. This research aims to examine the effects of Industry 4.0 on the Triple Bottom Line performance dimensions, and the possible synergies when combined with Lean. This study analyzes data from the High-Performance Manufacturing (HPM) database. The findings provide evidence supporting Industry 4.0’s positive impact on the Triple Bottom Line performance dimensions, identifying areas where Lean contributes to synergizing effects. Keywords: Lean · Industry 4.0 · Triple Bottom Line

1 Introduction The business landscape continuously evolves as new technologies are adopted and others discarded. Lean thinking is a crucial strategy for achieving world-class performance [1]. With globalization, increased efficiency and competitiveness have become even more relevant [2]. In recent years, sustainability has gained importance, leading to the Triple Bottom Line (TBL) approach, which considers operational, social, and environmental performance. Studies on Lean and TBL show possible synergies and trade-offs [2]. One significant development in recent years is the emergence of Industry 4.0, powered by advances in Information and Communication Technologies. However, literature on Industry 4.0’s impact on TBL performance is inconclusive [3, 4]. Given the growing importance of environmental and social dimensions alongside operational effectiveness, researchers suggest a broader focus when examining the impacts of Industry 4.0. A key question is how Industry 4.0 fits Lean strategies and how their joint implementation affects performance. Literature lacks empirical research on the combination of Lean and Industry 4.0 [4–7]. Synergies between Lean and Industry 4.0 practices could influence © IFIP International Federation for Information Processing 2023 Published by Springer Nature Switzerland AG 2023 E. Alfnes et al. (Eds.): APMS 2023, IFIP AICT 689, pp. 200–212, 2023. https://doi.org/10.1007/978-3-031-43662-8_15

Synergies Between Industry 4.0 and Lean on Triple Bottom Line Performance

201

TBL performance, and empirical evidence is needed to understand their complementarity [5]. This research aims to investigate the influence of Industry 4.0 on TBL performance and the possible moderating effect Lean has on this relationship. The study will address two research questions: Can Industry 4.0 contribute to TBL performance? And does Lean moderate the effect of Industry 4.0 on TBL performance? Data from the HighPerformance Manufacturing (HPM) database will be used to explore these relationships. The present study contributes to the literature and practice by addressing the scarcity of empirical research on the relationship between Lean and Industry 4.0, providing a more generalizable understanding of the effects of Lean practices on the relationship between Industry 4.0 and TBL performance.

2 Theoretical Background Industry 4.0 is a socio-technical concept encompassing various interacting practices and technologies [3]. Vaidya et al. [8] identified four main technological drivers: Internet of Things (IoT), Industrial Internet of Things (IIoT), Cloud-based manufacturing, and Smart manufacturing. The literature on Industry 4.0 indicates a lack of evaluation and assessment methodologies [9]. Schumacher et al. [10] mention several assessment tools and measurements for Industry 4.0; however, their construction and development details are often lacking. The variety of dimensions used in these maturity models further underscores Industry 4.0’s ambiguity. Lasi et al. [11] provide a starting point for operationalizing Industry 4.0 by dividing it into several pillars (Table 1). Adopting this approach can circumvent issues of reduced research comparability while still maintaining a basis for understanding and analyzing Industry 4.0. Therefore, this research will operationalize Industry 4.0 practices based on the pillars outlined by Lasi et al. [11]. Regarding Industry 4.0 influence on the TBL, Beier et al. [3] assert that while positive effects of Industry 4.0 on sustainability are frequently discussed, empirical evidence is often lacking. Despite the growing interest in corporate sustainability from both shareholders and stakeholders [12], the majority of literature only suggests Industry 4.0’s positive impact on sustainability, but empirical evidence is still scarce [13]. As Industry 4.0 is deemed a socio-technical concept [3, 14], its effects are anticipated to extend beyond operational improvements. However, Margherita and Braccini [4] argue that most literature on Industry 4.0’s potential benefits fails to assess its impact from a socio-technical perspective. To effectively measure the effects of such a concept, a more holistic approach to business performance is needed. Lean, a well-established management approach, has been more widely adopted and has existed for a considerable period compared to Industry 4.0. Shah and Ward [15] propose a model to measure lean, comprising bundles such as JIT, TPM, TQM, and HRM. Bortolotti et al. [16] use six practices for both hard and soft Lean practices. Although numerous studies have reported positive effects of Lean on operational performance [17], research on Lean practices’ impact on environmental and social performance remains scarce [2].

202

T. Bortolotti et al. Table 1. Pillars of Industry 4.0 according to Lasi et al. [11].

Pillars

Explanation:

Smart Factory

The extent to which factories are equipped with sensors, actors, and automated systems

Cyber-physical Systems

Merge physical items such as systems and products with digital representation

New systems in distribution and procurement Connectedness within distribution and procurement New systems in the development of products and services

Due to the increasing individualization of products. Approaches such as open innovation, product- intelligence, and memory

Adaptation to human needs

Focus on human needs in manufacturing systems

Self-organization

Decreased hierarchy within production

Corporate Social Responsibility

Focus on sustainability, resource efficiency when designing industrial manufacturing processes

There is a growing literature that investigates the synergies and conflicts between Lean practices and Industry 4.0. Significant enhancements in organizational competitiveness have been observed as an outcome of Lean implementations, analogous to the potency of IT solutions in driving similar results [18]. The literature, however, oscillates between those proposing a complementary application of these two principles and others suggesting a fundamental incongruity between them. A conflict between Lean and Industry 4.0 is encapsulated in the concept of ‘takt time’ [19], an operational strategy intended to align production rhythm with market demand. In the face of increasingly dynamic demand patterns and high customization, the mechanism of takt time could potentially disrupt the efficacy of Industry 4.0. Furthermore, the study of Tortorella et al. [20] underscores the dual impact on operational performance; process-related technologies were shown to adversely moderate the influence of reduced setup times on operational performance, while product/service-related technologies exerted a positive moderation on the efficacy of flow practices. The evidence derived from these studies suggests that despite potential synergies between Lean and Industry 4.0, there exist tangible tensions as well. Langlotz et al. [6] support the notion of distinct interdependencies between Industry 4.0 and Lean practices. Given the increased propensity for deploying Industry 4.0 technologies within Lean frameworks, the potential impact and interdependencies between these two entities warrant substantial attention. The exploration of synergies between Lean and Industry 4.0 in the context of social and environmental performance remains in its nascent stages. Kamble et al. [21] have proposed that Lean has the capacity to augment the beneficial influence of Industry 4.0 on both social and environmental performance. However, empirical research investigating

Synergies Between Industry 4.0 and Lean on Triple Bottom Line Performance

203

the joint effect of Lean and Industry 4.0 on social performance remains relatively sparse. Prior research, including studies by Henao et al. [2], Silva et al. [22], and Margherita and Braccini [4], have primarily analyzed the impacts of Lean and Industry 4.0 in isolation, eschewing a holistic examination of their synergy. 2.1 Hypotheses A PwC study in 2016 [23] surveying 235 German manufacturing companies revealed optimism about Industry 4.0 technologies’ potential to increase efficiency by 3.3% annually and reduce costs by 2.6%. This optimism spanned multiple industries, and a survey by the American Society for Quality (ASQ) [24] found that 82% of organizations reported increased efficiency after implementing Industry 4.0 technologies. Fettermann et al. [25] highlighted the positive impact of Industry 4.0 technologies on operation performance within companies based on 38 successful case studies. Similarly, Dalenogare et al. [26] conducted a survey-based study, concluding that Industry 4.0 technologies show promise in enhancing industrial performance. In summary, most analyzed literature provides evidence for the positive effects of Industry 4.0 on operational performance. Hypothesis 1a: Industry 4.0 positively affects operational performance. Although scarce, there is empirical evidence supporting a positive effect of Industry 4.0 on social aspects. These include increased morale due to productivity [27], reduction of safety incidents [28], workload reduction and job enrichment [29], and a safer workplace [30]. In contrast, Dalenogare et al. [26] concluded that Industry 4.0 did not offer benefits concerning social performance. They suggest that this may be due to the companies being from relatively underdeveloped countries, where social aspects might be less prioritized than operational performance. The authors propose that benefits could exist in more developed countries. They also note the low big data maturity in Brazilian industry, while Dubey et al. [31] found that predictive big data analytics positively affects social performance. The impact of Industry 4.0 on social performance remains unclear, but it is anticipated that higher levels of Industry 4.0 implementation may yield positive effects on social performance. Hypothesis 1b: Industry 4.0 positively affects social performance. Beier et al. [3] analyzed the environmental benefits of Industry 4.0, reaching similar conclusions as for its social performance benefits: although scarce, there is evidence that Industry 4.0 leads to environmental benefits. Dubey et al. [31] not only provided evidence on the positive influence of predictive big data analytics on social performance but also on environmental performance. Similarly, Bai et al. [32] identified big data analytics as one of the most influential technologies for improving environmental performance. Although some literature highlights the downsides and potential energy or raw material usage increases caused by Industry 4.0 technologies, most studies agree that the overall effect of Industry 4.0 on environmental performance is positive [33]. Hypothesis 1c: Industry 4.0 positively affects environmental performance.

204

T. Bortolotti et al.

Lean aims to minimize waste by reducing process variability, rework, and scrapped materials. It focuses on standardized practices, or hard practices, and continuous improvement, or soft practices [16]. Through streamlining operations, Lean improves efficiency and productivity. Furthermore, lean focus on quality following customer requirements, while enabling flexibility to adapt to changes without cost increase [15]. Academic literature consistently supports the positive relationship between Lean implementation and operational performance enhancement, highlighting the importance of Lean methodologies in improving organizational competitiveness [20]. Hypothesis 2a: Lean positively affects operational performance. Scholarly literature presents evidence in favor of the positive impact of Lean methodologies on social performance [2, 34, 35]. These studies accentuate the extensive benefits Lean delivers across a range of social parameters, bolstering its reputation as a transformative organizational practice. Nevertheless, there is a need to acknowledge the potential challenges it may pose for shop-floor employees, as outlined in Parker [36]. This perspective, however, represents a minority viewpoint within the research on Lean’s influence on social performance. Most studies, including Distelhorst et al. [35] and Silva et al. [22], align in corroborating the positive effects of Lean, thereby reinforcing the argument that Lean practices significantly contribute to enhancing social performance within the organizational context. Hypothesis 2b: Lean positively affects social performance. Although some studies argue that lean and environmental practices differ in terms of objectives, as lean focuses on cost-effective production optimization, while environmental programs prioritize environmental benefits, most of the studies found a positive impact of Lean methodologies on environmental performance [e.g., 2, 34, 37]. A comprehensive literature review conducted by Dieste et al. [38] revealed a consensus within Operations Management (OM) scholarship, attesting to the advantageous implications of Lean in relation to environmental measures and performance. Hypothesis 2c: Lean positively affects environmental performance. Lean manufacturing and Industry 4.0 both aim to improve companies’ competitiveness, but their compatibility remains a topic of debate [7, 18]. Tortorella et al. [20] found both positive and negative effects on operational performance when combining Lean and Industry 4.0 principles. They suggested that managers should strike the right balance, integrating flow-related Lean practices with cloud services, IoT, or big data analytics, for example. Pagliosa et al. [5] explored the relation between Lean and Industry 4.0 practices through a literature review. They identified synergies, such as between Value Stream Mapping (VSM) and Industry 4.0, as VSM reduces waste across the value chain [36] and can be enhanced with real-time data provided by Industry 4.0 technologies [39]. Sanders et al. [40] also viewed Lean as complementary to Industry 4.0, offering solutions for potential implementation issues in Lean manufacturing. Hypothesis 3a: Lean positively moderates the relationship between Industry 4.0 and operational performance. Varela et al. [41] examined the impact of Industry 4.0 and Lean on social performance. Their study concluded that Industry 4.0 had an impact on social performance, whereas Lean did not. Instead, Kamble et al. [21] found that the positive effect of Industry 4.0 on social performance is magnified by the implementation of Lean practices. Thus,

Synergies Between Industry 4.0 and Lean on Triple Bottom Line Performance

205

while Industry 4.0 positively impacts social performance, Lean measures can enhance this effect. Apart from these studies, there are no other empirical studies found that describe the effects of combining Lean and Industry 4.0 on social performance. While literature on the social benefits of Lean and Industry 4.0 separately does exist, the effects of combining the two remain relatively understudied. Based on Kamble et al. [21], we hypothesize that: Hypothesis 3b: Lean positively moderates the relationship between Industry 4.0 and social performance. Kamble et al. [21] found that Industry 4.0 have a positive effect on environmental performance, which was further amplified by Lean practices, similar to the impact on social performance. However, there is a lack of studies investigating the combination of Lean and Industry 4.0 on environmental performance, just as with social performance. Varela et al. [41] found that Industry 4.0 lead to increased sustainable performance, while Lean did not. In contrast, Dalenogare et al. [26] found that Industry 4.0 did not improve sustainable performance. This variation in results highlights the need for more research on the combined effects of Lean and Industry 4.0 on environmental performance. While both Lean practices and Industry 4.0 have been studied extensively for their separate effects on environmental performance, the combination of the two remains understudied. Based on Kamble et al. [21], we hypothesize that: Hypothesis 3c: Lean positively moderates the relationship between Industry 4.0 and environmental performance.

3 Methodology To address the research question, this study utilizes the fourth round of the HighPerformance Manufacturing (HPM) database. The database comprises 330 respondents from Europe, Asia, North America, and South America, representing three distinct industries: automotive, electronics, and machinery. The HPM database is derived from 12 questionnaires administered to participating companies, covering a range of subjects such as human resources, accounting, supply chain management, product development, information systems, quality management, and production planning (due to limited space, we refer to Danese et al. [42] for more information about the fourth round of data collection of the HPM project). Industry 4.0 is a socio-technical concept comprising several practices, as described by Lasi et al. [11]. These practices and their corresponding meanings are summarized in Table 1. To operationalize these practices, specific scales from the High-Performance Manufacturing (HPM) database were selected based on Lasi et al.’s [11] definitions. See Table 2 for the scales related to Industry 4.0 used in this study. To operationalize Lean, twelve practices, each with underlying scales from the HPM database, were derived from Bortolotti et al. [16]. These practices comprise six hard Lean practices and six soft Lean practices (Table 3 reports the list of scales related to Lean used in this study). Performance has been defined using a TBL perspective encompassing operational, social, and environmental performance [43, 44]. In particular, we follow the literature review of Gimenez and Tachizawa [43] and operationalized economic sustainability as operational performance, environmental performance as energy efficiency and waste

206

T. Bortolotti et al.

reduction, and social sustainability as labor conditions and connectedness within and outside the community. Based on these definitions of performance, corresponding scales were chosen from the HPM database to operationalize the three types of performance. Three control variables were selected to account for factors that may influence the study’s outcome. First, Return on Assets (ROA) was chosen, as Margolis et al. [45] highlight the positive relationship between financial and sustainable performance. Second, the location of the headquarters was introduced as a control variable due to the potential influence of geographical location on the relationship between Industry 4.0 and TBL [26]. Lastly, company size was included by taking the natural logarithm of the number of employees, as Gallo and Christensen [46] suggest a positive relationship between a firm’s size and the likelihood of adopting a TBL sustainability policy. The complete list of items is available upon request. Tables 2, 3, and 4 show that the Cronbach’s alphas of the multi-item scale are all above 0.6 (and most above 0.7), indicating internal consistency of the items. The constructs in this study consist of multiple items, and an Exploratory Factor Analysis (EFA) with varimax rotation and the eigenvalue-greater-than-one rule was conducted to assess convergent and discriminant validity for the items measuring each construct. The results show that six factors emerge from the items related to Industry 4.0, 12 from items related to Lean, and four from items related to performance (one factor for operational performance, one for environmental performance, and two factors for social performance: internal and external social performance) All items exhibit high discriminant validity, as they do not load on multiple factors, with cross-factor loadings below 0.4. Additionally, all constructs demonstrate high convergent validity, as items associated with a construct load strongly onto the same underlying factor, with factor loadings exceeding 0.6. The factor loadings are available upon request. Table 2. Industry 4.0 scales and Cronbach’s alphas. Industry 4.0 Pillars

alpha

Smart Factory & Cyber-Physical systems (SFCSP)

0.698

Systems in Distribution and Procurement (SDP)

0.876

Systems in Development of Products/Services (SDPS)

0.955

Adaptation to Human Needs (AHN)

0.810

Self-Organization (SO)

0.818

Corporate Sustainability Responsibility (CSR)

0.653

Synergies Between Industry 4.0 and Lean on Triple Bottom Line Performance

207

Table 3. Lean scales and Cronbach’s alphas. Hard Lean

alpha

Soft Lean

alpha

Process Control

0.761

Top Manag. Leadership

0.888

Just In Time

0.709

Supplier Partnership

0.731

Kanban

0.819

Small-Group Solving

0.648

Layout

0.828

Training Employees

0.772

Autonomous Maintenance

0.607

Customer Involvement

0.642

Setup Time Reduction

0.610

Continuous Improvement

0.659

Table 4. Performance scales and Cronbach’s alphas. Triple Bottom Line (TBL) Performance

alpha

Operational performance

0.836

Environmental performance

0.876

Social performance Social internal performance

0.662

Social external performance

0.634

4 Results In order to test the research hypotheses, regression was performed using SPSS 28. Table 5 shows the results of the regressions. Industry 4.0 is positively related to Social performance (internal) (H1b partially supported) and environmental performance (H1c supported), but not directly related to Operational performance (H1a not supported) and Social performance (external). Interestingly, the direct effect of Lean is positive and significant for operational (H2a supported) and social (external) performance, and not for social (internal) (H2b partially supported) and environmental performance (H2c not supported). When looking at the interaction terms (i.e., the interaction between Industry 4.0 and lean, representing their synergy), synergies between Industry 4.0 and Lean are evident for further improving Operational performance (H2a supported) and Social performance (internal) (H3b partially supported), but not for Social performance (external) and Environmental performance (H3c not supported). Taken together, lean and industry 4.0 have their peculiar effects on specific performance. This shows the importance of leveraging on both Industry 4.0 and Lean to achieve significant improvements in all the TBL dimensions.

208

T. Bortolotti et al. Table 5. Regression results. Operational

Social internal

Social external

Environmental

Controls: ROA

0.030

−0.008

−0.005

Firm size

0.007

0.045

0.072

South America1

0.174

0.132*

−0.217**

North America1

0.006

0.093

0.033

0.042

−0.001

0.096

0.003

0.094

0.302**

0.145

0.253**

Asia1

0.010 0.197** −0.229**

Explanatory: Industry 4.0

0.093

Lean

0.337**

0.062

0.352**

Industry 4.0 x Lean

0.139*

0.133*

0.097

−0.012 0.007

5 Discussion and Conclusions Studies on the effects of Industry 4.0 on the TBL are relatively limited; however, our findings are consistent with the existing literature that predominantly highlights positive outcomes for internal social performance. Notable instances encompass improved morale attributable to heightened productivity [27], a decline in safety incidents [28], reduced workload, job enrichment [29], and a more secure work environment [30]. Conversely, in terms of external practices, Industry 4.0 has not yet demonstrated significant benefits. This observation aligns with Ghobakhloo’s [47] characterization of Industry 4.0 as a maturity system, commencing with human resource development (internal social performance) and subsequently addressing sustainability (environmental performance) and social welfare enhancement (external social performance). As organizations embark on their Industry 4.0 journey, they are likely to prioritize internal social and environmental performance, with external social aspects receiving comparatively less attention. In terms of the positive influence of Industry 4.0 on environmental performance, our study concurs with the findings of Dubey et al. [31]. They contend that big data analytics can also yield social and environmental advantages. Contrarily, our results do not support Ghobakhloo’s [47] concerns regarding potential detrimental environmental effects arising from Industry 4.0 technologies. Consequently, while the literature acknowledges both positive and negative repercussions of Industry 4.0 on TBL performance, our research supports a predominantly positive impact of Industry 4.0 on internal social and environmental performance. As Industry 4.0 continues to evolve, it is crucial for organizations to consider the broader implications of this transformation, particularly in terms of TBL performance. It is also essential to recognize that the journey towards Industry 4.0 is an ongoing process, and as such, the relationship between Industry 4.0 and TBL performance may shift over time. Continued research in this area will be vital for understanding and navigating

Synergies Between Industry 4.0 and Lean on Triple Bottom Line Performance

209

these dynamic interactions, ultimately enabling businesses to capitalize on the potential benefits of Industry 4.0 while mitigating any potential drawbacks. Given the widespread adoption of Lean across various industries, its potential compatibility with Industry 4.0 could have significant practical and managerial implications. From an optimistic standpoint, if Lean exhibits synergistic effects with Industry 4.0, it may render the transition to Industry 4.0 practices more appealing for companies. Concrete empirical evidence could offer valuable corroboration, as the primary barriers to implementing Industry 4.0 are perceived to be high investment costs and uncertain outcomes [23]. To examine the effects of combining Lean and Industry 4.0 practices, a moderation analysis was conducted, which revealed that the impact of Industry 4.0 on social internal and operational performance are significantly enhanced by Lean. Concerning moderation on operational performance, Pagliosa et al. [5] identify possible synergies as well as conflicts. The analysis suggests that potential synergies prevail over conflicts, as the overall effect was determined to be significantly positive. Kamble et al. [21] identified a significant moderating effect of Lean on sustainability. However, their study did not directly measure the effects of Industry 4.0 and Lean moderation on the TBL performance dimensions. This research can partially corroborate Kamble et al.’s [21] findings, as Lean moderation on the impact of Industry 4.0 has the potential to enhance certain performance dimensions, but not others. Moreover, it provides greater insight into which dimensions are significantly influenced. From a managerial perspective, this study substantiates the considerable advantages that Industry 4.0 can offer for TBL performance. This finding helps mitigate the prevailing outcome uncertainty that organizations encounter when considering the implementation of Industry 4.0. Furthermore, the research demonstrates a significant synergistic impact between Industry 4.0 and Lean on social internal and operational performance dimensions, highlighting the augmented positive effects that Industry 4.0 can yield when introduced within a Lean production context. This study revealed that Industry 4.0 positively impacts social internal and environmental performance, while Lean positively moderates the impact of Industry 4.0 on operational and social internal performance. These findings offer evidence for potential synergies between the two at a high level, although specific underlying practices may still present conflicts. This information could help managers focus on specific areas when implementing Industry 4.0 practices within a Lean environment, maximizing potential synergies while avoiding conflicts. Although Industry 4.0 has been criticized for its ambiguity [10, 11], operationalizing it within the constraints of the HPM database introduced certain limitations. While developing a dedicated survey could have allowed for more comprehensive operationalization of Industry 4.0 constructs, it would likely have reduced the number of respondents and geographical diversity. Future iterations of the HPM project could incorporate additional scales reflecting the level of digitalization in operational processes. Future research might also focus on more specific operationalization of Industry 4.0 and a confirmatory approach regarding interactions at the sub-practice level between Industry 4.0 and Lean. Furthermore, this study opted to examine the effects on the TBL performance. Although the TBL concept is widely recognized for offering a comprehensive view of a company’s performance, it lacks concrete methods to operationalize its three dimensions. In this

210

T. Bortolotti et al.

study, the literature was reviewed alongside the potential scales from the HPM database to determine the most relevant operationalization. However, this approach does not yield high levels of research comparability, which Sridhar & Jones [48] identified as one of TBL’s weaknesses.

References 1. Fullerton, R.R., Kennedy, F.A., Widener, S.K.: Lean manufacturing and firm performance: the incremental contribution of lean management accounting practices. J. Oper. Manag. 32(7–8), 414–428 (2014) 2. Henao, R., Sarache, W., Gómez, I.: Lean manufacturing and sustainable performance: trends and future challenges. J. Clean. Prod. 208, 99–116 (2019) 3. Beier, G., Ullrich, A., Niehoff, S., Reißig, M., Habich, M.: Industry 4.0: How it is defined from a sociotechnical perspective and how much sustainability it includes – a literature review. J. Cleaner Prod. 259, 120856 (2020) 4. Margherita, E.G., Braccini, A.M.: Organizational impacts on sustainability of industry 4.0: a systematic literature review from empirical case studies. Digital Business Transformation, pp. 173–186 (2020) 5. Pagliosa, M., Tortorella, G., Ferreira, J.C.E.: Industry 4.0 and lean manufacturing. J. Manufacturing Technology Manage. 32(3), 543–569 (2019) 6. Langlotz, P., Siedler, C., Aurich, J.C.: Unification of lean production and Industry 4.0. Procedia CIRP, 99, 15–20 (2021). https://doi.org/10.1016/j.procir.2021.03.003 7. Bokhorst, J.A., Knol, W., Slomp, J., Bortolotti, T.: Assessing to what extent smart manufacturing builds on lean principles. Int. J. Prod. Econ. 253, 108599 (2022) 8. Vaidya, S., Ambad, P., Bhosle, S.: Industry 4.0 – a glimpse. Procedia Manufacturing 20, 233–238 (2018) 9. Oztemel, E., Gursev, S.: Literature review of Industry 4.0 and related technologies. J. Intelligent Manufacturing, 31, 127–182 (2020) 10. Schumacher, A., Erol, S., Sihn, W.: A maturity model for assessing industry 4.0 readiness and maturity of manufacturing enterprises. Procedia CIRP 52, 161–166 (2016) 11. Lasi, H., Fettke, P., Kemper, H.G., Feld, T., Hoffmann, M.: Industry 4.0. Business & Information Systems Engineering 6(4), 239–242 (2014) 12. Ballou, B., Heitger, D., Landes, C.: The rise of corporate sustainability reporting: a rapidly growing assurance opportunity. J. Account. 202(6), 65–74 (2006) 13. Toth-Peter, A., de Oliveira, R. T., Mathews, S., Barner, L., Figueira, S.: Industry 4.0 as an enabler in transitioning to circular business models: a systematic literature review. J. Cleaner Production, 136284 (2023) 14. Sony, M., Naik, S.: Industry 4.0 integration with socio-technical systems theory: a systematic review and proposed theoretical model. Technology in Society, 61, 101248 (2020) 15. Shah, R., Ward, P.T.: Lean manufacturing: context, practice bundles, and performance. J. Oper. Manag. 21(2), 129–149 (2003) 16. Bortolotti, T., Boscari, S., Danese, P.: Successful lean implementation: organizational culture and soft lean practices. Int. J. Prod. Econ. 160, 182–201 (2015) 17. Danese, P., Manfè, V., Romano, P.: A systematic literature review on recent lean research: state-of-the-art and future directions. Int. J. Manage. Reviews 20(2), 579–605 (2018) 18. Riezebos, J., Klingenberg, W., Hicks, C.: Lean Production and information technology: connection or contradiction? Comput. Ind. 60(4), 237–247 (2009) 19. Sundar, R., Balaji, A.N., Kumar, R.S.: A review on lean manufacturing implementation techniques. Procedia Eng. 97, 1875–1885 (2014)

Synergies Between Industry 4.0 and Lean on Triple Bottom Line Performance

211

20. Tortorella, G.L., Giglio, R., van Dun, D.H.: Industry 4.0 adoption as a moderator of the impact of lean production practices on operational performance improvement. Int. J. Operations & Production Manage. 39(6/7/8), 860–886 (2019) 21. Kamble, S., Gunasekaran, A., Dhone, N.C.: Industry 4.0 and lean manufacturing practices for sustainable organisational performance in Indian manufacturing companies. Int. J. Prod. Res. 58(5), 1319–1337 (2019) 22. Silva, C., Vaz, P., Ferreira, L.M.: The impact of lean manufacturing on environmental and social sustainability: a study using a concept mapping approach. IFAC Proceedings Volumes 46(24), 306–310 (2013) 23. PwC. Global industry 4.0 survey: building the digital enterprise (2016). www.pwc.com/gx/ en/industries/industries-4.0/landing-page/industry-4.0-building-your-digital-enterpriseapril2016.pdf 24. Shrouf, F., Ordieres, J., Miragliotta, G.: Smart factories in industry 4.0: A review of the concept and of energy management approached in production based on the internet of things paradigm. In: 2014 IEEE International Conference on Industrial Engineering and Engineering Management, pp. 697–701. IEEE. (2014) 25. Fettermann, D.C., Cavalcante, C.G.S., Almeida, T.D.D., Tortorella, G.L.: How does industry 4.0 contribute to operations management? J. Industrial and Prod. Eng. 35(4), 255–268 (2018) 26. Dalenogare, L.S., Benitez, G.B., Ayala, N.F., Frank, A.G.: The expected contribution of Industry 4.0 technologies for industrial performance. International J. Prod. Econ. 204, 383– 394 (2018) 27. Lee, C., Lv, Y., Ng, K., Ho, W., Choy, K.: Design and application of Internet of things-based warehouse management system for smart logistics. Int. J. Prod. Res. 56(8), 2753–2768 (2017) 28. Yuan, Z., Qin, W., Zhao, J.: Smart manufacturing for the oil refining and petrochemical industry. Engineering 3(2), 179–182 (2017) 29. Braccini, A., Margherita, E.: Exploring organizational sustainability of industry 4.0 under the triple bottom line: the case of a manufacturing company. Sustainability 11(1), 36 (2018) 30. Arana-Landín, G., Laskurain-Iturbe, I., Iturrate, M., Landeta-Manzano, B.: Assessing the influence of industry 4.0 technologies on occupational health and safety. Heliyon 9(3) (2023) 31. Dubey, R., et al.: Can big data and predictive analytics improve social and environmental sustainability? Technol. Forecast. Soc. Chang. 144, 534–545 (2019) 32. Bai, C., Dallasega, P., Orzes, G., Sarkis, J.: Industry 4.0 technologies assessment: a sustainability perspective. Int. J. Prod. Econ. 229, 107776 (2020) 33. Arana-Landín, G., Uriarte-Gallastegi, N., Landeta-Manzano, B., Laskurain-Iturbe, I.: The contribution of lean management—industry 4.0 technologies to improving energy efficiency. Energies 16(5), 2124 (2023) 34. Wu, L., Subramanian, N., Abdulrahman, M., Liu, C., Lai, K.H., Pawar, K.: The impact of integrated practices of lean, green, and social management systems on firm sustainability performance—evidence from Chinese fashion auto-parts suppliers. Sustainability 7(4), 3838– 3858 (2015) 35. Distelhorst, G., Hainmueller, J., Locke, R.M.: Does lean improve labor standards? management and social performance in the nike supply chain. Manage. Sci. 63(3), 707–728 (2017) 36. Parker, S.K.: Longitudinal effects of lean production on employee outcomes and the mediating role of work characteristics. J. Appl. Psychol. 88(4), 620 (2003) 37. Garza-Reyes, J.A., Kumar, V., Chaikittisilp, S., Tan, K.H.: The effect of lean methods and tools on the environmental performance of manufacturing organisations. Int. J. Prod. Econ. 200, 170–180 (2018) 38. Dieste, M., Panizzolo, R., Garza-Reyes, J.A.: A systematic literature review regarding the influence of lean manufacturing on firms’ financial performance. J. Manuf. Technol. Manag. 32(9), 101–121 (2021)

212

T. Bortolotti et al.

39. Abdulmalek, F.A., Rajgopal, J.: Analyzing the benefits of lean manufacturing and value stream mapping via simulation: a process sector case study. Int. J. Prod. Econ. 107(1), 223–236 (2007) 40. Sanders, A., Elangeswaran, C., Wulfsberg, J.: Industry 4.0 implies lean manufacturing: research activities in industry 4.0 function as enablers for lean manufacturing. J. Ind. Eng. Manage. 9(3), 811 (2016) 41. Varela, L., Araújo, A., Ávila, P., Castro, H., Putnik, G.: Evaluation of the relation between lean manufacturing, industry 4.0, and sustainability. Sustainability 11(5), 1439 (2019) 42. Danese, P., Lion, A., Vinelli, A.: Drivers and enablers of supplier sustainability practices: a survey-based analysis. Int. J. Prod. Res. 57(7), 2034–2056 (2019) 43. Gimenez, C., Tachizawa, E.M.: Extending sustainability to suppliers: a systematic literature review. Supply Chain Management: an International J. 17(5), 531543 (2012) 44. Elkington, J.: Towards the sustainable corporation: win-win-win business strategies for sustainable development. Calif. Manag. Rev. 36, 90e100 (1994) 45. Margolis, J.D., Elfenbein, H.A., Walsh, J.P.: Does it pay to be good? and does it matter? a meta-analysis of the relationship between corporate social and financial performance. SSRN Electronic Journal (2009) 46. Gallo, P.J., Christensen, L.J.: Firm size matters: an empirical investigation of organizational size and ownership on sustainability-related behaviors. Bus. Soc. 50(2), 315–349 (2011) 47. Ghobakhloo, M.: Industry 4.0, digitization, and opportunities for sustainability. J. Cleaner Prod. 252, 119869 (2020) 48. Sridhar, K., Jones, G.: The three fundamental criticisms of the triple bottom line approach: an empirical study to link sustainability reports in companies based in the Asia-Pacific region and TBL shortcomings. Asian J. Bus. Ethics 2, 91–111 (2013)

Driving Sustainability Through a VSM-Indicator-Based Framework: A Case in Pharma SME Zuhara Zemke Chavez1(B)

, Mayari Perez Tay1 , Mohammad Hasibul Islam2 , and Monica Bellgran1

1 KTH Royal Institute of Technology, Kvarnbergagatan 12, 151 36 Södertälje, Sweden

[email protected] 2 American International University - Bangladesh, 1229 Dhaka, Bangladesh

Abstract. Despite the availability of various tools, practitioners often lack a holistic perspective when conducting sustainability-related improvement activities. Existing Value Stream Mapping (VSM) tools that incorporate environmental aspects suffer from a lack of standardization and limited alignment with sustainability reporting standards. This paper introduces a framework for VSM that integrates sustainability indicators to assess and enhance the triple-bottom-line performance of manufacturing processes aligned with the global reporting initiative (GRI) standards. We validate the proposed framework through a case study conducted in a biopharmaceutical production system of a small and mediumsized enterprise (SME) that employs single-use technologies. The study highlights the critical need for collaboration with equipment suppliers to develop production specifications considering operational performance and environmental impact, thereby capturing a comprehensive perspective. By utilizing the proposed framework, practitioners can identify opportunities for improving the design and efficiency of production systems. The case study provides valuable insights into SMEs’ challenges when transitioning to sustainable production, particularly when product-related requirements impose limitations on sustainability improvements. Although the framework’s validation focuses on a biopharmaceutical SME, it can be applied to manufacturing companies across industries to assess and enhance all three aspects of sustainability in their processes. Keywords: E-VSM · circular production system · Lean-green

1 Introduction The increasing energy price, the impacts on climate change, and the strict regulations have resulted in the pharmaceutical industry emphasizing energy usage and emission reduction throughout the entire life-cycle of pharma products [1]–[3]. However, compared to other industries, the heavily regulated nature of the pharmaceutical industry (i.e., regulations related to Food and Drug Acts) has added challenges in managing and implementing Greenhouse gas (GHG) emissions reduction actions [3]. © IFIP International Federation for Information Processing 2023 Published by Springer Nature Switzerland AG 2023 E. Alfnes et al. (Eds.): APMS 2023, IFIP AICT 689, pp. 213–227, 2023. https://doi.org/10.1007/978-3-031-43662-8_16

214

Z. Z. Chavez et al.

Small and medium-sized enterprises (SMEs) are predominant economic actors in economies [4]. As stated by the European Commission, they account for 99% of all businesses in the EU. Within sustainability, their critical role is reinforced, as they account for around 70% of industrial pollution [5]. However, companies require sufficient data and skills to design sustainable and efficient production systems, often lacking by SMEs. For instance, bio-pharmaceutical SMEs often do not have in-use data to make strategic manufacturing decisions [6, 7] regarding whether and when to use single-use disposable technologies (SUTs) or stainless steel (SS) equipment for their production system. Currently, the usage of SUTs is gaining popularity because of their potential advantage in bio-pharmaceutical production systems that include reduction of water usage. However, no arguments regarding safety, cost-effectiveness, or operational efficiency are compelling to choose one technology platform or the other for all applications [6]. Hence, companies working with SUTs require supporting methods and tools to assess and guide the integration of sustainability in their production systems. At present, companies are adopting integrated Lean-Green practices to improve operational performance related to environmental sustainability. Value stream mapping (VSM) is a lean tool widely used to develop manufacturing processes with minimal production flow waste supporting strategic decisions [8]. Researchers (see Table 1) have gradually developed the VSM tool and incorporated other environmental metrics such as energy consumption, CO2 emission, and resource consumption associated with the processes and inventory, commonly referred to as E-VSM. Despite existing literature incorporating environmental aspects within VSM, there is a lack of uniformity in the methods and metrics, making it harder for practitioners to adopt. Additionally, there is often a lack of connection between the VSM-assessed environmental metrics and the companies’ sustainability goals, which can limit the urgency of environmental improvements over efficiency, e.g., managers require suitable indicators to inform their decisionmaking [9]. One of the possible solutions could be incorporating the Global Reporting Initiative (GRI) indicators within VSM. GRI is a global sustainability reporting standard [10], a set of applicable guidelines for reporting on economic, environmental, and social performance for corporations and any business, governmental, or non-governmental organization [11], in alignment with ISO 14001. According to Sulaiman et al. [8], existing literature has only paid attention to calculating efficiency and ignored the possible interaction between desirable, productive, and environmental efficiency. Likewise, an alignment between VSM data and GRI guidelines is encouraged for improvement in the value stream to be guided on measurable facts and reflected on the information disclosed to stakeholders, a clear match between “what is communicated” and “what is done.” To address the identified gap, this study investigates how SMEs can utilize VSM to integrate and assess sustainability aspects for developing their production systems aligned with sustainability reporting standards. We propose an extension of E-VSM that considers sustainability indicators in alignment with global reporting standards. In our framework, KPIs will guide improvements and changes to the desired state in compliance with targets to meet customer requirements while minimizing sustainability impact. A case study was conducted in a bio-pharma SME embedded with SUTs to better understand and validate the proposed framework’s applicability.

Driving Sustainability Through a VSM-Indicator-Based Framework

215

2 Theoretical Background 2.1 Single-Use Technologies Impact SUTs, also known as single-use plastics, started as an innovative alternative to traditional stainless steel equipment and components for the production of biopharmaceutical proteins but have become a strong competitive technology for many parts of the production and processing chain. The SUTs seem to surpass conventional stainless steel in terms of economy, convenience, and quality [12]. However, one issue raised by the throw-away nature of disposable modules is the environmental impact since the plastic modules are discarded after each production run and incinerated [13]. The increasing pressure on capital and operating costs, the risk of product crosscontamination, and the cost of cleaning and validation push manufacturers to embrace the significant advantages of disposable processes for an expanding good manufacturing practice (GMP) in manufacturing facilities. Studies on SUTs [14, 15] show benefits in terms of efficiency, economy, and resource management under various circumstances; however, in practice, they are not the answer for all scales. For instance, the performance of disposable modules sometimes fails to match that of fixed equipment at high volumes [7, 14]. However, there has been an evident trend toward the availability of larger disposable modules, driven by increasing acceptance at more minor scales, such as the case for SMEs. SUTs provide a reduction in economic investments necessary when developing production processes and validation since this approach reduces or eliminates the need for large quantities of steam, process water, and water for injection [14]. SMEs also tend to see economic rewards in the short term, eliminating process bottlenecks imposed by specific assurance processes, for instance, cleaning tasks and inspections in the production processes. However, adopting SUTs introduces new challenges related to the production, distribution, and disposal of single-use components. 2.2 Value Stream Mapping for Sustainable Production The merger of lean and green initiatives helps achieve economic objectives and improve organizational sustainability performance [16, 17]. VSM can be described as a lean tool to identify value-adding and non-value-adding activities needed to take products through all necessary flows, from raw materials to customers [18]. VSM has been intensely applied in manufacturing companies to improve production performance by identifying waste and improvement opportunities. The most recent research efforts are dedicated to merging VSM with green initiatives to deploy operational strategies [19]. Researchers have been proposing diverse versions of the tool (see Table 1), integrating KPIs that claim to allow the measurement and improvement of environmental performance. When utilizing VSM, the first phase is to select a product family and its related quantifiable indicators [20], i.e., the problem under consideration and the aspects that can measure the process performance. A trend observed in the different papers is the selection of KPIs in the three aspects of sustainability, i.e., social, economic, and environmental. Each bottom line is assessed to specific indicators, where social sustainability is the least assessed. There is no standard on the included indicators per bottom line, i.e., social,

216

Z. Z. Chavez et al. Table 1. A summary of identified VSM studies integrating sustainability indicators.

Tool (s), methods integrated

Indicator Social

Application area Design (D), Operations (O), Project (P)

Reference

Economic

Company type (SME, MNC)

Industry/Sector

Environ-mental

VSM and LCA

Y

Y

N

MNC

Construction

O

Salvador R. et al. (2021)

Multi layer VSM

Y

Y

N

MNC

Construction

O

Rabnawaz & Zhang (2021)

MED VSM, Sustainable Setup Stream Mapping (3SM)

Y

Y

Y

Not specfied

Home appliance

O

Ebrahimi, Khakpour, & Saghiri (2021)

DMAIC-based approach with VSM

Y

N

N

Not specfied

Metal processing

O

Gholami H. et al. (2021)

Sust. Indicators for VSM

Y

Y

Y

MNC

Gears

O

Swarnakar V (2021)

VSM and root cause analysis

N

Y

N

SME

Food

O

Ahmad, R. et. Al. (2022)

Green Lean Six Sigma (GLSS) framework

Y

Y

Y

SME

No case study validation

P

Yadav V., Gahlot P. (2022)

Sustainable VSM (SVSM) & Arena simulation

Y

Y

N

MNC

Clothing

O

Mishra A.K. et al. (2020)

Green performance map, E-VSM, Waste flow mapping, LCA

Y

Y

Y

MNC

Metal processing

O

Shahbazi S. et al. (2019)

Sus-VSM and MTM (Methods-Time Measurement)

Y

Y

N

Not specfied

Automotive

O

Sunk, et al. (2017)

Energy-focused VSM

Y

Y

N

MNC

Steel

O

Svensson A.& Paramonova S. (2017)

Sus-VSM

Y

Y

Y

Not specfied

Satellite television dishes

O

Faulkner W.& Badurdeen F. (2014)

Eco-Function matrix-integrated VSM

Y

Y

Y

None

No case study validation

O, P

Vinodh S. et al. (2011)

Extended VSM to support simulations

Y

Y

Y

MNC

Consumer goods

O

Bait, S. et al. (2020)

environmental, and economical, or an agreement on the selection and KPI calculation methods. However, there is a trend to link the original VSM lean KPIs to the economic bottom line, including productivity, overall equipment efficiency (OEE), and labor cost [17, 21]. Guidelines to integrate sustainability improvements through VSM seem to be missing. A study [17] contributed to minimizing this gap by providing a detailed solution approach to sustainability improvements, identifying an empirically tested list of sustainability indicators, and developing a conceptual method for sustainability assessment

Driving Sustainability Through a VSM-Indicator-Based Framework

217

where some listed indicators are linked to the 300 and 400 GRI standards. Though, the conceptual method was primarily applied in existing production lines, i.e., the operations phase. 2.3 Sustainability Reporting Impact Sustainability reporting is a communication tool, and voluntary practice companies utilize to state their commitment to sustainable development [22]–[24]. It includes nonfinancial information to stakeholders, e.g., sustainable practices and impact [25]. The GRI, introduced in the late 1990s, is the most widely used worldwide standard for sustainability reporting, according to several researchers [23, 25]. GRI reports contain information about a company’s economic, environmental, and social aspects. Under GRI, a company’s overall performance can be measured regarding its contribution to economic prosperity, environmental quality, and capital [23]. A report under GRI contains three types of information; i. plan and company profile, ii. Management focus, and iii. Performance indicators. The performance indicators enable comparing information regarding an organization’s economic, environmental, and social aspects. The GRI also provides guidelines for data collection regarding performance indicators and designing indicators to measure efficient performance in different periods; in this way, companies can compare their performance with other firms’ performances in a specific sector [23]. According to GRI guidelines, companies should report environmental-related information, such as the total amount of waste generated, the percentage of waste recovered by reusing, recycling, or other means, and the percentage of energy saving, among others [26]. A common practice observed in companies utilizing sustainability reporting is setting targets for improvement regarding their GHG emissions in compliance with the Sciencebased targets initiative (SBTi) [22]. Science-based targets are greenhouse gas emissions reduction targets that align with the level of decarbonization required to meet the goals of the Paris Agreement. They suggest paths towards more long-term and corporate ‘netzero’ targets, which we see most commonly adopted by large companies. In recent years attention has been given to targets that include scope 3, where multinational companies’ (MNCs) role is critical to driving emissions measurement and management throughout the entire value chain. SMEs often act as suppliers of MNCs and, in this context, are often requested to provide information regarding sustainability indicators and GHG emissions from their products. A lack of standardization of calculation methods and standards makes it hard for SMEs to understand where to start and how to account for sustainability performance and GHG emissions specifically.

3 Method The present study employs a qualitative approach executed by a single-case study [27]. Our methodology consists of three key steps 1. Conceptual framework coming from a systematic literature review, 2. Framework validation in a single case selected by purposeful sampling [28] and 3. Framework analysis for improvement opportunities. The

218

Z. Z. Chavez et al.

systematic literature review was conducted according to the method by Smart, Tranfield, & Denyer [29] to analyze the advance of VSM for sustainable development in the manufacturing industry. We utilized the Scopus database to perform the search from 2009 until 2022 employing the initial search query: Include from title, abstract, and keywords (value AND stream AND map*, OR mapping, OR vsm, OR e-vsm) AND (pharma OR pharmaceutical OR industry OR production) AND (sustainable AND development, OR sustainab*). The search returned a total of 109 papers. After applying selection and exclusion criteria to the title, abstracts, and keywords, we obtained a core list of 55 selected papers. The selection and exclusion criteria considered journal papers, books, book chapters, and papers written in English. We read the 55 papers in full and subsequently applied snowballing, which resulted in the inclusion of additional papers in the analysis. We reviewed a total of 73 papers to extract studies that proposed extensions, variations, or applications of the VSM tool for improving sustainability. In Table 1, we summarize the different studies that describe the development and application of VSM as a tool for enhancing sustainability. We developed the triple bottom line value stream mapping (TBL-VSM) framework based on the literature review. To test the framework’s applicability, we conducted a case study in a bio-pharmaceutical SME, focusing on assessing the impact of using SUTs. Thoroughness in case selection involves explicitly picking cases that are congruent with the study purpose and that will yield data on major study questions [28]. Therefore, the case selection was based on location criteria, access to data, and management interest and support. 3.1 Data Collection and Analysis Data was collected through interviews, presentations, two onsite visits, and discussions with the case company; the details of the different data collection forms are described in Table 2. All events were recorded and transcribed. Triangulation improves the case study’s accuracy and credibility of the research findings [30, 31]. Therefore, at least two researchers were part of all semi-structured interviews. During the VSM mapping sessions, three researchers took part. Two of the researchers analyzed interview transcripts and notes from the meetings separately. The documentation of on-site observation was used to complement the data already captured through interviews and VSM mapping meetings and served as a guide to the discussion of improvements and opportunities session with the production scientist and the rest of the team. The TBL-VSM framework was developed and validated by the three researchers.

4 A Conceptual Framework for Sustainability Assessment and Reporting Using Value Stream Mapping Our conceptual framework using VSM integrates indicators from three core streams i. Lean, ii. GRI reporting initiative standards, and iii. SBTi and GHG protocol, each element with a specific progressive purpose. First, the classical VSM methodology, the lean indicators such as processing times, changeover, lead times, etc., enable the understanding of the context, essentially how things work, and help spot initial improvements in the value stream driven by waste elimination. Next, GRI indicators can be a guide to

Driving Sustainability Through a VSM-Indicator-Based Framework

219

Table 2. Data collection at LS case Company role(s)

Data type

Scientific officer

2 x 60 min interview

Production scientist 1

1 x 60 min interview, 3 x 45 min working sessions for VSM mapping, 1x 60 discussion session about improvement opportunities including GRI reporting standards

Production scientist 2

1 x 60 min interview, includes a short presentation from the interviewee

Company CEO, SHE manager, Plant manager, Communication manager

1 x 120 min, 1 x 60 min presentations, discussions about VSM and improvement opportunities, including GRI reporting standards (one of them including the communication manager)

2 x 240 min Factory visits Bi-weekly project meetings over one year with multiple employees of the company

assess and measure improvements in the three aspects of sustainability. Once there is an initial understanding of the production processes, and the organization starts detecting areas of improvement, science-based targets can guide the definition of relevant targets that contribute to a zero-carbon transformation. The framework is linked to GRI indicators because of the relevance of already accepted indicators complying with global reporting standards and regulations to facilitate the work and understanding of SMEs. Figure 1 presents the TBL-VSM conceptual framework steps. The left side of the figure includes the six steps for creating a VSM map; the right part focuses on defining the TBL indicators assessment and improvement. The framework can support SMEs to integrate sustainability concerns earlier in the design of the production system and help them set relevant targets that contribute to minimizing GHG emissions. 4.1 Practical Application in the Development of Sustainable SMEs Case Background. The case company is a life science SME located in northern Sweden, in the following called LS case. The company specializes in developing diseasemodifying drugs and is currently developing two main products and conducting clinical trials. The company used to outsource production to a different company overseas. However, during the covid-19 pandemic, the company decided to own its production. Therefore, they decided to purchase the production processes and started a tech transfer to a new facility in Sweden. The tech transfer means that standards operating procedures are set according to the acquired processes and must follow GMP standards. Though, manufacturing equipment was sourced from new suppliers. The LS case acquired equipment that uses single-use devices referred to as SUTs, i.e., bags, filters, and tubing made of plastic, all of them used a single time and must be disposed of afterward. At the

220

Z. Z. Chavez et al.

Fig. 1. TBL-VSM conceptual framework.

moment of the data collection, the LS case was installing the production equipment, and testing was ongoing to ensure scaling up production was possible. For the LS case, scaling the process meant going from an R&D lab volume production to about 200 l without compromising the product quality. LS case is a company that has sustainability ambitions. They are part of an ongoing research project focused on applying AI already during design and construction to minimize GHG emissions of the plant when in the operations phase. However, at the start of the project and during this study, they had not formulated specific sustainability goals, visions, or targets to work with. Still, they would like to know the environmental impact of future production operations, such as carbon footprint and waste amount. LS

Driving Sustainability Through a VSM-Indicator-Based Framework

221

management stated that they see a benefit in starting with their sustainability reporting (aligning with GRI standards) in the near future, as most of their suppliers and competitors already do it. The company does not have a sustainability department at the moment, and it is under the plant manager’s responsibility to oversee all relevant matters related to health and safety. TBL-VSM Framework Application. Figure 2 includes a VSM map created in collaboration with LS personnel following our TBL-VSM framework. All sensitive data has been excluded or altered to ensure secrecy and comply with company policies. Information about production equipment and SUTs was extracted from the specification sheets publicly available on suppliers’ websites. In the TBL-VSM map, some inventory days between processes are marked as N/A, given that there are considerations to make regarding the testing particular to the case. In the future, capturing the actual waiting times and setup times between processes will be relevant as these have implications for energy use and efficiency. The TBL-VSM lean data provided an understanding of the processes and material and information flows that allowed building the GRI data mapping, i.e., waste of SUTs per process. Obtaining process-level data is a critical step that can be difficult for practitioners to succeed without a “go and see” approach, which in our case, capturing was only possible due to support from the production scientist during the mapping. Table 3 includes some examples of the improvement opportunities identified in the case on the current state and the sustainability dimension they will impact. These improvements were discussed together with the team and limited to indicators 306 and 403 as a starting point for the LS case. The improvements were determined from data that can be easily collected without disturbing installation and development activities at the facility. Simultaneously, they provide an example of how improvements for minimizing LS company’s carbon footprint in the operations phase can already be set at these early development phases despite regulations and process restrictions. For instance, in improvement number 2, the SME has already started conversations with their leading SUTs supplier to perform a more detailed study on optimizing the SUTs distribution and exploring alternatives for more sustainable packaging material. At the bottom of the TBL-VSM map, we indicate GRI KPIs connected to mostly 306 standards. In this case, no amount of SUTs waste was recycled or recovered. However, these indicators were used to provide feedback to the company on development (step 6 of mapping TBL-VSM, Fig. 1). Currently, the LS case has set a provisional area to separate the packaging of SUTs (mainly as a reactive action). The carton and plastic are being sent to a local waste management company for recycling and recovery. In the future, the LS case may need to designate an area for SUTs disassembly as some items include different pieces that need to be separated before sending to a waste management company. For instance, an identified case is a specific type of bag that includes a metal component, for which the supplier has provided a prototype tool to separate the component from the bag. Neither the LS case nor the SUT supplier has identified a waste management company that can recycle these components, although they have started conversations looking for alternatives. The previous has also implications for the internal flow of material. LS case needs to decide the most optimal arrangement to collect and segregate SUTs

Fig. 2. TBL-VSM current state LS case.

222 Z. Z. Chavez et al.

Driving Sustainability Through a VSM-Indicator-Based Framework

223

without disturbing GMP processes quality while minimizing the non-value adding time that waste handling implies.

5 Discussion Integrating sustainability indicators into VSM without consensus or a standard method leads to a lack of harmonization [32] and makes it harder for companies to implement. On the other hand, for companies to adopt a particular tool, it must help reduce environmental impacts and show the connection to improving business and economic performance. Our proposed TBL-VSM framework considers the scenario of processes not fully developed. It provides SMEs with a simple way to assess sustainability impact early in the design of the production system. It can help SMEs assess ongoing development activities and aid in the decision-making towards actions that can positively impact operations in alignment with the sustainable development goals, such as the example in our case study. An essential task in step 3.2 of the TBL-VSM framework is for the team to identify suitable indicators. We recommend starting with standards 301 to 306 and 403, which involve collecting less complex data that may be readily available at the companies for other purposes, e.g., health and safety, quality, and assurance metrics. Our analysis from the case study revealed that, as stated by previous researchers, manufacturing processes are often identified as sustainable from the customers’ perspective based on lean metrics, i.e., cycle time, takt time, and lead times, but not sustainable from the TBL metrics perspective [17]. SUTs were perceived as advantageous to the case study company for reducing cleaning times and risks for assurance; however, through the TBL-VSM, the LS case became aware of the environmental constraints and additional tasks for waste handling imposed by the nature of these technologies. Such constraints were not considered when deciding on this type of technology against stainless steel. The latest literature (see Table 1) shows that the application of VSM for sustainability is still tightly connected to the operations phase. TBL-VSM framework provides a step-based approach for companies to start the journey and shows business value by connecting with GRI. Given the limitations of GMP and development activities, the improvement suggestions for the LS case are on high-return, low-risk areas. Still, the VSM shows that the workload and processing times are not balanced by design and may significantly impact operational performance and other environmental metrics, e.g., OEE and energy consumption. The proposed framework could aid in identifying current and future states based on customer requirements and global sustainability standards to design production equipment accordingly to suppliers’ available budgets. Our framework is suggested as a starting point for sustainability assessment at SMEs. Due to its lean core, it can also guide improvements that benefit productivity performance; the natural and well-tested progression within sustainability is applying life cycle assessment to quantify CO2 emissions. The type of questioning a TBL-VSM mapping triggered in the LS case is relevant for any company undergoing design and installation activities for assessing the sustainability performance of the production system and taking proactive actions towards embedding sustainability and circularity in the production system.

224

Z. Z. Chavez et al. Table 3. Improvement opportunities identified in TBL-VSM in LS case study. Dimension

Indicator

Implications

1

Economic, Environment

306–3 kg waste, 306–4 total amount diverted from disposal, and 306–5 amount directed to disposal

Reduce additional costs for handling extra packaging material, including personal space, and outsource waste handling services for repacking, sorting, and disposing of SUTs Standard procedures for how to separate SUTs for disposal. No standards are available at the moment

2

Environment

306–3 kg waste, 306–4 total amount diverted from disposal, and 306–5 amount directed to disposal

Reduce packaging material and void in SUTs incoming boxes; a suggestion to implement returning containers involves negotiations with suppliers for long-term impact

3

Social

403–9 the number and rate of recordable work-related injuries and the number of hours worked

Reduce work stress from irregular work shifts during testing phases, reduce long shifts and excessive overtime during testing phases, and prevent work-related injuries

6 Conclusion This research contributes to the growing Lean-Green literature by developing a framework for E-VSM that could measure and improve a wide range of performance indicators related to sustainability under global reporting standards. Furthermore, this paper provides a case study on the applicability of the proposed TBL-VSM framework for assessing the design of a biopharmaceutical production system working with single-use plastics. SMEs often rely on the equipment suppliers’ expertise to design their production system. The empirical findings emphasized that equipment suppliers must work on developing the specification of single-use materials considering the overall operational performance, including material flow and environmental aspects. Thereby, practitioners can use the proposed TBL-VSM tool to identify improvement opportunities in the design of production systems, which will, in the long term, minimize operations impact such as the ones identified in the LS case. While our framework has been validated only at one pharmaceutical SME, we firmly believe that any manufacturing company can benefit from its utilization regardless of its size (MNC or SME). The framework is designed to assist decision-makers, practitioners, and managers in assessing sustainability within their manufacturing processes and analyzing improvement opportunities that align with global targets. The case study provides a unique example of the challenges SMEs may

Driving Sustainability Through a VSM-Indicator-Based Framework

225

face in transitioning to sustainable production, where product-related requirements and decisions often are stated to limit sustainability process improvements. Future research will further improve and validate the proposed framework in different scenarios, aiming to provide comprehensive guidelines for effectively implementing the framework in existing and new processes. Acknowledgments. The authors gratefully acknowledge the support of the case company; without their support, this research would not have been possible. This work was supported by Sweden’s Government Agency for Innovation VINNOVA Programme, ALISTAIR project [grant number 2020–03404].

References 1. Schneider, J.L., Wilson, A., Rosenbeck, J.M.: Pharmaceutical companies and sustainability: an analysis of corporate reporting. Benchmarking 17(3), 421–434 (2010). https://doi.org/10. 1108/14635771011049371 2. Ding, B.: Pharma Industry 4.0: literature review and research opportunities in sustainable pharmaceutical supply chains. Process Saf. Environ. Prot. 119, 115–130 (2018). https://doi. org/10.1016/j.psep.2018.06.031 3. Low, Y.S., Halim, I., Adhitya, A., Chew, W., Sharratt, P.: Systematic framework for design of environmentally sustainable pharmaceutical supply chain network. J. Pharm. Innov. 11(3), 250–263 (2016). https://doi.org/10.1007/s12247-016-9255-8 4. Adams, D., Donovan, J., Topple, C.: Achieving sustainability in food manufacturing operations and their supply chains: key insights from a systematic literature review. Sustain. Prod. Consum. 28, 1491–1499 (2021). https://doi.org/10.1016/j.spc.2021.08.019 5. Acar, M.F., Aktas, E., Agan, Y., Bourlakis, M., Fatih, M., Aktas, E.: Does sustainability pay ? evidence from the food sector. J. Food Serv. Bus. Res. 22(3), 239–260 (2019). https://doi. org/10.1080/15378020.2019.1597672 6. Rogge, P., Müller, D., Sch, S.R.: The single-use or stainless steel decision process: a CDMO perspective. Bioprocess Int. 13, January 2015 7. Shirahata, H., Hirao, M., Sugiyama, H.: Decision-support method for the choice between single-use and multi-use technologies in sterile drug product manufacturing. J. Pharm. Innov. 12(1), 1–13 (2017). https://doi.org/10.1007/s12247-016-9264-7 8. Sulaiman, M.A., et al.: Cleaner production value stream mapping at a chromium plating plant: a case study. Int. J. Agil. Syst. Manag. 12(3), 245 (2019). https://doi.org/10.1504/ijasm.2019. 10022800 9. Domingo, R., Aguado, S.: Overall environmental equipment effectiveness as a metric of a lean and green manufacturing system. Sustainability 9031–9047 (2015). https://doi.org/10. 3390/su7079031 10. Narula, S., et al.: Applicability of industry 4.0 technologies in the adoption of global reporting initiative standards for achieving sustainability. J. Clean. Prod. 305 (2021). https://doi.org/10. 1016/j.jclepro.2021.127141 11. Hedberg, C.-J., von Malmborg, F.: The global reporting initiative and corporate sustainability reporting in Swedish companies. Corp. Soc. Responsib. Environ. Manag. 10, 153–164 (2003) 12. Shukla, A.A., Gottschalk, U.: Single-use disposable technologies for biopharmaceutical manufacturing. Trends Biotechnol. 31(3), 147–154 (2013). https://doi.org/10.1016/j.tibtech.2012. 10.004 13. Scheper, T., Eibl, R., Eibl, D.: Disposable Bioreactors (2009)

226

Z. Z. Chavez et al.

14. Pietrzykowski, M., Flanagan, W., Pizzi, V., Brown, A., Sinclair, A., Monge, M.: An environmental life cycle assessment comparison of single-use and conventional process technology for the production of monoclonal antibodies. J. Clean. Prod. 41, 150–162 (2013). https://doi. org/10.1016/j.jclepro.2012.09.048 15. Budzinski, K., et al.: Streamlined life cycle assessment of single use technologies in biopharmaceutical manufacture. N. Biotechnol. 68, 28–36 (2022). https://doi.org/10.1016/j.nbt. 2022.01.002 16. Chiarini, A.: Sustainable manufacturing-greening processes using specific Lean Production tools: an empirical observation from European motorcycle component manufacturers. J. Clean. Prod. 85, 226–233 (2014). https://doi.org/10.1016/j.jclepro.2014.07.080 17. Swarnakar, V., Singh, A.R., Antony, J., Kr, A., Cudney, B.: Development of a conceptual method for sustainability assessment in manufacturing. Comput. Ind. Eng., August 2021. https://doi.org/10.1016/j.cie.2021.107403 18. Rother, M., Shook, J.: Learning to see: value stream mapping to create value and eliminate MUDA. Lean Enterprise Institute, Cambridge, MA, June 1998 19. Garza-Reyes, J.A.: Green lean and the need for Six Sigma. Int. J. Lean Six Sigma 6(3), 226–248 (2015). https://doi.org/10.1108/IJLSS-04-2014-0010 20. Rabnawaz Ahmed, R., Zhang, X.: Multi-layer value stream assessment of the reverse logistics network for inert construction waste management. Resour. Conserv. Recycl. 170, 105574, September 2020, 2021. https://doi.org/10.1016/j.resconrec.2021.105574 21. Salvador, R., Barros, M.V., dos Santos, G.E.T., van Mierlo, K.G., Piekarski, C.M., de Francisco, A.C.: Towards a green and fast production system: Integrating life cycle assessment and value stream mapping for decision making. Environ. Impac (2021). https://www.scopus. com/inward/record.uri?eid=2-s2.0-85097739938&doi=10.1016%2Fj.eiar.2020.106519& partnerID=40&md5=afac3fc326c478cb444c942089405632 22. Bjørn, A., Lloyd, S., Matthews, D.: From the Paris agreement to corporate climate commitments: evaluation of seven methods for setting ‘science-based’ emission targets. Environ. Res. Lett. 17(3) (2022). https://doi.org/10.1088/1748-9326/ac548e 23. Marimon, F., Alonso-Almeida, M.D.M., Rodríguez, M.D.P., Cortez Alejandro, K.A.: The worldwide diffusion of the global reporting initiative: what is the point? J. Clean. Prod. 33, 132–144 (2012). https://doi.org/10.1016/j.jclepro.2012.04.017 24. Liesen, A., Hoepner, A.G., Patten, D.M., Figge, F.: Does stakeholder pressure influence corporate GHG emissions reporting? empirical evidence from Europe. Account. Audit. Account. J. 28(7), 1047–1074 (2015). https://doi.org/10.1108/AAAJ-12-2013-1547 25. Benameur, K.B., Mostafa, M.M., Hassanein, A., Shariff, M.Z., Al-Shattarat, W.: Sustainability reporting scholarly research: a bibliometric review and a future research agenda (2023) 26. Global Sustainability Standards Board: Global Reporting Initiative Standards (2020). www. globalreporting.org/standards 27. Yin, R.: Case Study Research: Design and Methods, 5th ed. Sage publications, London, United Kingdom (2013) 28. Patton, M.Q.: Enhancing the quality and credibility of qualitative analysis. Health Serv. Res. 34(5), 1189–1208 (1999). https://doi.org/10.4135/9781412985727 29. Smart, P., Tranfield, D., Denyer, D.: Towards a methodology for developing evidenceinformed management knowledge by means of systematic review. Br. J. Manag. 14(3), 207–222 (2003). https://doi.org/10.1111/1467-8551.00375 30. Morgan, S.J., Pullon, S.R.H., MacDonald, L.M., McKinlay, E.M., Gray, B.V.: Case study observational research: a framework for conducting case study research where observation data are the focus. Qual. Health Res. 27(7), 1060–1068 (2017). https://doi.org/10.1177/104 9732316649160

Driving Sustainability Through a VSM-Indicator-Based Framework

227

31. Yin, R.K.: Case Study Research: Design and Methods, 4th edn. SAGE, London (2009) 32. Swarnakar, V., Singh, A.R., Antony, J., Tiwari, A.K., Cudney, E.: Development of a conceptual method for sustainability assessment in manufacturing. Comput. Ind. Eng. (2021). https://www.scopus.com/inward/record.uri?eid=2-s2.0-85107037576&doi=10. 1016%2Fj.cie.2021.107403&partnerID=40&md5=0a1a32060a99959e28250afbe3757caf

Design and Application of a Development Map for Aligning Strategy and Automation Decisions in Manufacturing SMEs Malin Löfving1(B) , Peter Almström2 , Caroline Jarebrant3 , and Magnus Widfeldt3 1 School of Engineering, Jönköping University, Jönköping, Sweden

[email protected]

2 Chalmers University of Technology, Göteborg, Sweden 3 RISE, Mölndal, Sweden

Abstract. The research in this paper focused on (i) to increase the knowledge about the strategic role of automation technology and (ii) how automation technology investment decisions can be aligned with the strategy in manufacturing SMEs. This paper describes the design, application, and evaluation of the Strategic Development Map (SDM) in three settings, whereof two manufacturing SMEs. The SDM was also applied in more than 500 SMEs in a Swedish initiative called the Robot Lift. The SDM was derived from existing literature about Toyota management principles and thereafter adapted to the topic. The process of designing, testing, and evaluating the tool was inspired by design science research. It also contains explanations of the outcomes and mechanisms that lead to the successful adaptation and use of the SDM. The SDM provides a structured guidance process to understand the strategic role of automation technology investments in SMEs. This process can support companies towards a deeper understanding of the current state, target state and obstacles and challenges to reach the target with a focus on automation technology investments. The SDM was emphasized by companies and external coaches as very simple to understand and use. Using the SDM increased involvement and knowledge about strategic role of automation technology. Keywords: automation investment · Toyota management principles · small and medium-sized enterprises

1 Introduction Industrial automation technologies form an integral part of today’s manufacturing companies [1]. Still, many small and medium enterprises (SMEs) rely on manual labour [1]. The machines themselves in SMEs have been automated for decades, but the tasks of feeding machines are often manual if there is a diversity of products with high variation in quantities. Barriers that prevented SMEs from investing in automation technologies were large investment cost, uncertainty about investment and associated cost, lack of expertise as well as lack of suitable automation for the manufacturing operation [1]. © IFIP International Federation for Information Processing 2023 Published by Springer Nature Switzerland AG 2023 E. Alfnes et al. (Eds.): APMS 2023, IFIP AICT 689, pp. 228–241, 2023. https://doi.org/10.1007/978-3-031-43662-8_17

Design and Application of a Development Map

229

However, this is changing as new flexible automation technologies brings new opportunities even for SMEs with high variation production [1–3]. Today there exist new types of more flexible automation. These are less expensive to purchase and have new smart safety equipment as well as simple programming [2, 3]. This study focuses on SMEs, a group of companies that are often sub-suppliers to larger companies, consist of between 10–249 employees [4]. SMEs tend to be flexible and adapt to new circumstances, even though they are limited in terms of financial and human capital [5]. Even though the threshold is getting lower for SMEs to invest in automation technology, this decision still has a degree of uncertainty. The automation investment also leaves the company dependent of the new automation technology for long time, and this can affect both stakeholders as internal operations. Previous studies reported that many of the problems raised in the investment of advanced manufacturing technologies, such as automation, are related to an unawareness of the strategic role of the technology and no alignment between strategy and automation decision [6–9]. For less uncertain automation decisions, the automation decision should be aligned with strategy. The decision-making process in SMEs is often a non-structured, non-formal and centered to one person [10]. According to [2], decision makers in SMEs lack expertise in automation and do always properly assess the capabilities of automation technology or predict associated costs. Decision makers in SMEs need guidance in making these important decisions, but the guidance need to be in a format and with a scope that the decision makers can understand and embrace. In literature, there exists a vast array of literature on Industry 4.0 and advanced manufacturing technology decision tools including fixed industrial robots and automation [for example 11, 12, 13]. Only a few of them were developed and applied at manufacturing SMEs or include SMEs in the study [14, 15]. The description of the tools mainly focuses on the content, and less effort has been put on the process, i.e., how to use the tools. Few of them also emphasize the strategic role of automation technology or include how automation decision should be aligned with strategy. Moreover, most of these existing tools are conceptual and the application and practical relevance of these existing tools is questioned [11]. Based on this, the goals of this research were (i) to increase the knowledge about the strategic role of automation technology and (ii) how automation technology investment decisions can be aligned with the strategy in manufacturing SMEs. One important step to fulfil these goals has been the development of a practical tool for structuring and focusing the strategic discussion in SMEs. The tool is called “Strategic Development Map” (SDM) and it was developed in two research projects between 2017 and 2019 (see Sect. 2.1). Later, from 2019 to 2021, it was used in a national development program called the Robot Lift (Robotlyftet) [16]. The Robot Lift systematically evaluated the automation potential in over 500 SMEs in Sweden. The later success of the application of the SDM in Robotlyftet motivated this paper. The purposes of this article are to: 1. Describe the research process that led to the “Strategic Development Map”. 2. Describe the “Strategic Development Map” and how it is used. 3. Explain the mechanisms that lead to the successful adaptation and use of the “Strategic Development Map”.

230

M. Löfving et al.

The paper is outlined based on the Context–Intervention–Mechanism–Outcome (CIMO)- logic [17]. The context of the study is described in the introduction chapter. Section 2.1 present the background of the study and the research process. The data collected derives from both theoretical and empirical data, where the first phase consists of a literature review and the design process of the tool, presented in Sect. 2.2. The data collection and analysis for the applications of the tool are then presented in Sect. 2.3. Section 3 describes the SDM. Section 4 contains descriptions of three applications of the tool. The applications consist of descriptions of the interventions and outcomes of each application. In the presentation of each application there is also description of the application object, data collection and analysis. Section 5 presents an evaluation of the tool that were derived from the people participating the in applications. Section 6 contain another application that was conducted after initial applications were conducted. Thereafter the mechanisms were identified and discussed. Finally, next step is presented in the conclusion Sect. 8.

2 Research Design 2.1 Research Process The study of this paper constitutes parts of two collaborative research projects, namely Swedprod (2016–2018) and LoHiSwedprod (2018–2020). The two projects were carried out in collaboration between one research institute, one university, one union, one automation technology supplier, one industrial development center and five SMEs. The research projects aimed at increase knowledge of flexible and moveable automation technology solution for low volume production. One of the objectives of the project was to develop guidelines for automation technology decision including independent tools. A result from these projects was an automation guide directed to SMEs [18, 19]. One part of the guide included the independent tool called Strategic Development Map (SDM), that is the focus of this paper. The study in this paper focuses on solving a practical problem by describing how a tool was designed and applied for automation decision in an iterative process. The study is framed as an application of Design Science Research [20]. According to [21, p. 67] DSR seeks to “(1) create new innovative and generic artifacts to solve problems (2) to explain this explorative process, and (3) to improve the problem-solving process”. In this research strategy designed artifacts, such as tools, are expected to contribute to new knowledge and to contribute to academia. The data collected in this study derives from both theoretical and empirical data. In this study, the data collection and analysis are divided into two phases. The first phase consists of the design of the tool and is presented in Sect. 2.2. The second phase consists of data collection and analysis of the applications, and these are presented in Sect. 2.3. 2.2 The Design Process of the Tool This part describes how the tool was developed and designed. This part also includes a description of how data was collected and analyzed to be able to design the tool.

Design and Application of a Development Map

231

An interview study was conducted with two aims. First to identify the requirements of flexible automation suitable for low volume and high mix production, and the second to understand how automation decisions were made in the companies. The first step in the interview study was to develop a structured interview guide. Subsequently, one to three representatives of each company (five in total) participated in an interview. The respondents were decision makers and had roles such as owner, CEO, production manager, sales manager, and production technician. At least two of the authors of this paper participated at each interview and the interviews lasted between 60 to 120 min. The interviews were recorded and one of the authors took notes. After each interview, the data was analyzed by all in the research team, the authors of this paper, together. After all interviews were conducted, the data from each company was compared and analyzed further to identify themes and requirements. The requirements were identified of flexible automation suitable for low volume and high mix production. One of the identified requirements was to include both strategy and automation decision. At the same time a literature review was also conducted to understand what tools existed that included both strategy and automation. Keywords for the literature review were “strategy”, “advanced manufacturing technology”, “automation investment”, “tool” and synonyms to them. Databases used were Scopus and Web of Science. Data was also searched for in Google Scholar, and popular science literature, such as books, reports, practical courses, and lecture notes. In total 30 unique tools were identified. Evaluation criteria was derived from previous literature [for example 2] the result from the interview study, and experience in the research team. The evaluation criteria for the tools studied were the following: – Be easy to understand and use for production managers or CEOs in SMEs with low volume and high mix production. – Be based on and stimulate participation. – Be transparent and visual. – Aim for long term gain and sustainability. – Should be built on existing knowledge that can be adapted and adjusted to the topic. – Include both strategy and automation decision. – Be developed for or tested at manufacturing SMEs. The 30 identified tools were evaluated against criteria. Most of the identified tools did not fulfill more than 50% of the criteria. However, one tool was identified as promising as it could potentially fulfill all these criteria if adapted further to suit the topic of this study. This tool had previously been used in a course given for industry practitioners by the Research Institutes of Sweden (RISE). This tool had already been adapted once to suit the course and the first version of the tool was used in a national production development program called Produktionslyftet (Production Lift) [22]. In the practical course it was used to increase the understanding about how to improve the material flow using automation in manufacturing SMEs. The overall course plan was a 2x2 matrix, based on Toyota Management Principles. Each course occasion focused on one square in the matrix. The course and 2x2 matrix resulted in a strategic support for how to solve each companies’ problem.

232

M. Löfving et al.

Instead of developing a new tool, it was decided to use the tool, hereon called Strategic Development Map (SDM), and adapt it to the context of this study. The tool was further adapted in this study in close collaboration with the participating companies in the research projects. As the automation investment needs to be properly planned and analyzed before it is implemented, the SDM focus on Planning phase. Due to this, the fifth kata [23] question that aims at the Do phase, i.e., execution of plans is the investment of automation technology, was not included in the SDM. After the first tests and evaluations of the SDM, conceptual modifications of the process of the map were done inspired by policy deployment. Policy deployment included for example cascading of objectives into the whole organization [24]. Moreover, both policy deployment and kata aim at iterating the same pattern in the organization to ease communication with co-workers. This were added to the SDM process and presented for and evaluated by the participating companies in the research projects. The tool is presented in Sect. 3. The applications and evaluation of the tool are described in occurring order in Sects. 4, 5 and 6. 2.3 Data Collection and Analysis for the Applications of the Tool After the tool had been identified, adaptations were made. Adaptations were made both for the process and content of the 2 x 2 matrix. The content that included questions and descriptions were adapted to the context of this study, i.e., automation technology investment decisions in SMEs. For example, questions regarding current knowledge and use of automation were included. The process was described further and visualized to suit production managers in SMEs. The 2 x 2 matrix was also adapted to be more transparent and easier to understand. Thereafter it was applied in different settings to evaluate the applicability and usefulness. In the different applications data was collected with different techniques due to setting of each application. However, the unit of analysis was the SDM process. In the first application (see Sect. 4.1), the authors of this paper acted as participant observers [25] at project meetings. The project meetings lasted for 4 h and were conducted twice a year between 2017 and 2020. The companies that participated in the research projects were involved at these meetings. The data collected during these project meetings were documents in terms of presentation notes and meeting notes. In the application at Company A (see Sect. 4.2), two of the authors of this paper acted as participant observers [20] and notes were taken. The notes were rewritten after each meeting. In the application at Company B (see Sect. 4.3), one of the authors of this paper acted as participant observer [25] and took notes during the meeting. Photos were also taken of whiteboard notes. Notes were transcribed into a written document that also included the photos. After each application of the tool, semi-structured interviews with the respondents were conducted, see Table 1. The interviews focused on evaluating the process and content of the SDM. In the application Robot Lift (see Sect. 6), one of the authors participated in the development of the Robot Lift process and report template. This author also participated in the execution of the Robot Lift as an expert, evaluated data collected by the coaches in companies and recommended the companies’ next step. The results from Robot Lift

Design and Application of a Development Map

233

Table 1. Application and respondents of the evaluation of the SDM. Application

Respondents

4.1 SDM used as facilitator at project meetings

Five representatives from five companies

4.2 SDM used at Company A

Sales manager and part-owner of the company

4.3 SDM used at Company B

External coach, production manager, production technician

were presented in an independent evaluation of a consulting firm [16]. This report was analyzed by the authors of this paper. The collected empirical data was analyzed in an iterative process by all the authors of this paper. Data from each application (company) was analyzed separately using content analysis and resulted in the outcomes for each application. The data that included evaluation for each application was also analyzed and summarized in Sect. 5. The data from Robot Lift was analyzed separately and presented in Sect. 6. Finally, a thematic analysis was conducted for data from all applications to identify mechanisms.

3 Description of the Strategic Development Map The tool is called “Strategic Development Map” (SDM) as it provides a structured guidance process to understand the strategic role of automation technology investments. This process can support companies towards a deeper understanding of the current state, target state and obstacles and challenges to reach the target with a focus on automation technology investments. The SDM consists of a 2 x 2 matrix that represents four steps, see Fig. 1. Each step requires information and includes questions that needs to be answered to understand the next step. The four steps derive from the main ideas of improvement kata and kata questions [23, p. 155]. Each step is presented below and questions for each step is presented in Fig. 1. (1) Describe the company´s current state. Make an analysis of the current state based on customers, products, manufacturing resources and current competitive priorities. (2) Identify a target state for the whole company, with a focus on customer need in the future, competitive priorities for the company in the future and what the core competences in manufacturing are in a target state. (3) Identify and describe obstacles and challenges preventing the company to reach the target state with a focus on manufacturing and automation strategy. Here different automation solutions are identified. They are also evaluated based on objective facts and calculations to understand effects and consequences of investing in each of them. (4) Next step is to prioritize and decide next step. Deadlines are also decided for the parts in the next step, for example what is next step for the coming 6 months. The order of number 1 and 2 in the list above is often reversed in improvement kata. However, [23] states that number 1 and 2 can be reversed when the target state has not yet been established.

234

M. Löfving et al.

Fig. 1. The Strategic Development Map.

Conceptually, the SDM process presented in Fig. 1, could be iterated down and up in the organization using same pattern. The dotted line in Fig. 1 is representing the iteration. Number of iterations depend on organizational size, but three iterations are recommended: 1. First use the SDM on a strategic level for the company. 2. The information in the first step is deployed down one level to manufacturing strategy and automation strategy. Next step in terms of interesting operations to automate are identified. 3. Third step focus on one operation in manufacturing and if and which automation technologies are suitable for this specific operation.

4 Applications of the Strategic Development Map 4.1 Strategic Development Map Used as Facilitator at Project Meetings The SDM was first applied as a facilitator at one research project meeting in the research project Swedprod. The research team used the tool to present the result from five case studies in the participating companies. The result from each case study was presented in the same way. This enabled discussions among the companies regarding current states, target states and how to make successful automation decisions. After the presentation, the participants evaluated the SDM, and in dialogue, they decided to use the SDM as facilitator at all project meetings during the project Swedprod. Outcomes of this application were: – A shared view of the progress of the project and progress for automation decisions in the participating companies with low volume and high mix production. – Verification of collected empirical data.

Design and Application of a Development Map

235

– Joint analysis with the research team and participating companies – Dialogue among the participating companies about strategies and automation decisions. 4.2 Strategic Development Map Used at Company A Company A had low volume and high mix production with small orders that recurred on an irregular basis. It had under 50 employees and was one of two companies in a small private company group. Investment decisions of manufacturing technology including automation technology were made by the owners, based on customers’ needs and requirements. Reasons for investing in automation technology solutions were to increase the capacity in their manufacturing and to minimize monotonous working tasks. The participant from the company was the sales manager as well as owner. He/she was introduced to the SDM in paper format. Thereafter, he/she worked with the SDM him/herself and two of the authors that supported this application acted as participant observers [20]. First, the sales manager/owner answered the questions in Fig. 1. However, the result focused on overall business level and business strategy and automation on an overall level was emphasized. Due to this, another iteration of the SDM process was conducted, here with focus on manufacturing system that included automation in whole manufacturing system. To understand which operation should be suitable for automation, a third iteration focused on automation specific operations. However, this led to more questions regarding suitability of automation in specific operations and it was also decided that the company should work with the SDM themselves. Here, the sales manager/owner involved both production managers and operators in the SDM process to be able to identify suitable automation investments in relation to business strategy. Vice versa, the sales manager/owner also communicated the target state to all employees so they could get an understanding of the company´s target state. The company´s work with the SDM was followed up on a regular basis for 2 years. The main outcome of the use of the SDM in Company A, was not new investments in new automation technology, but the knowledge about strategic roles of manufacturing technologies and how to connect long term plans and manufacturing technology investments. Moreover, another outcome was the increased involvement of employees in automation technology decisions and investments. 4.3 Strategic Development Map Used at Company B Company B was both a subcontractor of wooden components as well as manufacturer of its own wooden products with low volume and high mix production. It had under 50 employees, was part of a larger company group, but managed as a small manufacturing company. The company had a written strategy including manufacturing that emphasized automation investments. The application of SDM in Company B was conducted as the final test of the SDM. An external coach was invited to test and verify the process of SDM. The coach had long experience working with lean principles and had knowledge of the wood industry. The coach was responsible for the process of how to use the SDM.

236

M. Löfving et al.

Participants from the company were the production manager and production technician. When needed, operators were also involved and answered questions (production technician went to manufacturing and asked the operators for information in certain areas). The aim with the application was to identify the company’s next step towards automation. The SDM was visualized on a whiteboard and both the external coach and company participants wrote on the whiteboard. To understand and to be able to describe the current state using objective facts, the external coach recommended the company participants to conduct a value stream mapping in the manufacturing plant. In this case the value stream mapping aimed at increasing knowledge about the processes and flows in the manufacturing system. The production technician was responsible for the value stream mapping and the result were written down on the whiteboard. Thereafter the manufacturing objectives were identified, and obstacles were discussed. In the last step, the participants identified next step and made a prioritization of tasks in the next step. Here, automation was not identified as a priority, as it became evident for the participants that Company B needed to work with increasing the value-added time and remove waste in the system before investing in automation. Photos were taken of the whiteboard and transcribed into a written document. After the meeting, the transcribed map was used by the representatives to illustrate and communicate the outcomes of the workshop to the CEO. The company was followed up during the final project meeting. The authors of this paper analyzed the iterative data collected during the project. Outcomes from this application were: – No automation technology investments were done for 6 months as the company focused on improvements in the manufacturing system identified in the application. – Coaches with lean principles experience seem to easily understand and use the map.

5 Evaluation of the Strategic Development Map Table 2 summarizes the evaluation of the SDM according to the evaluation criteria described in Sect. 2.2. In general, the feedback about both process and content were very positive. The respondents emphasized that the map was easy to understand and use in different settings. SDM also was found to be easy to explain to coworkers not involved in the first introduction of the SDM process. Moreover, respondents were positive about how fast they could begin to use the map. In Company A for example, the introduction of the map took around 10 min, and the first results from using the map took around 30 min. Time for introduction of tools is essential for the usability in SMEs, as SMEs have limited time and human resources [5]. Here the SDM seems suitable for SMEs. The use of the SDM stimulated participation and involvement of other co-workers than the decision makers. For example, in Company B, operators were involved as they had the knowledge about manufacturing processes and in Company A, the decision maker communicated the results in the map to all employees.

Design and Application of a Development Map

237

Table 2. Summary of the evaluation of the SDM. Criteria

Main results

Process: Easy to use and understand for production managers or CEOs in SMEs

Respondents highlighted that the SDM was easy to understand and use for all employees in a company

Process: Stimulate participation

The process of using the SDM emphasized involvement from other co-worker than the decision maker

Process: Transparent and visual

The SDM was transparent and visual as it only contained 4 squares

Content: Aim for long term gain and sustainable decisions

The respondents identified a target state and next step to reach the target state. In one case, the decision was not to automate, but focus on improvements that lead to sustainable decisions

Content: Align strategy and automation decision

The respondents were positive to get knowledge about how to align automation decisions with strategy and not to make automation decisions in isolation from the strategy or long-term plans When asking the questions in the SDM, it became clearer for the company what to prioritize for successful automation investments

6 Application in the Robot Lift The application of SDM in the Robot Lift is in this study regarded as an extension of the study and show the practical use of the SDM. This application was conducted after the initial research study. One of the co-authors of this paper was involved in the design, development, and execution of the Robot Lift, Robotlyftet in Swedish [16, 26]. The Robot Lift was a Swedish national initiative aiming at increasing manufacturing SMEs’ knowledge of automation technologies and increasing investments in automation technologies in these companies. The Robot Lift was ongoing between the years 2018–2021 and funded by “The Swedish Agency for Economic and Regional Growth”. The Robot Lift was managed and executed by IUC Sweden AB, Automation Region, Robotdalen, RISE IVF, Swedish Industrial Robot Association. Experienced practical coaches from a national network of industrial development centers were responsible for visiting and supporting SMEs. The fundamental part of Robotlyftet and for the coaches, was to analyze SMEs conditions to automate. The coaching process was created for the Robot Lift in collaboration with the partners mentioned above and was based on the SDM process. The information collected by the coaches were documented and summarized in a report template. The template was based on the SDM process. The SDM was also enclosed in the template. Thereafter the process of SDM was extended to also contain an evaluation meeting. The evaluation meeting aimed at recommending a next step regarding automation solution investment at the SME. The evaluation was made by preferably three persons (from

238

M. Löfving et al.

coaches, external experts). One of the co-authors of this paper acted as external expert at these evaluation meetings. These meetings resulted in recommendations for a next step. The recommendations were divided into three decisions and one of the alternatives were chosen for each company: – Proceed with an in-depth study and apply for an Automation check (financial support for continuing with a deeper automation pre-study prior to an investment) – Do not proceed with an in-depth study but ensure unclear questions and/or strategy. – Do not proceed with automation solution if other recommendation was given. The decision for each company was followed by written recommendations of the next step, e g the most promising automation technology to proceed with for the company. This procedure, including the SDM, was used by the external coaches at over 500 manufacturing SMEs in Sweden. The outcome of this application was that over 170 of these companies had already or would invest in automation within a year. Several of them invested in new types of flexible automation technologies. One example is from a small manufacturing company. This company had been thinking about automating operations for a while but had not identified a suitable automation technology for their needs regarding high variations in manufacturing and budget. An external coach from the Robot Lift got in touch with the company and together they analyzed the current state, target state and the need for automation. The recommendation for the company was to invest in a new type of flexible and movable automation that suited the company´s needs. The outcome of this was the first automation technology investment for this company. According to [16], 97% of the 500 SMEs involved in the Robot Lift stated that their knowledge increased about automation.

7 Discussion An outcome for the companies when using the SDM was a deeper understanding of the strategic role of automation technology investments. For sustainable automation technology investments, the decision should not be taken in silo without alignment to the long-term plans [6, 7]. One outcome of the application in Company B also showed that they prioritized improvements of existing manufacturing before investing in automation. This was also shown in the Robot Lift where the evaluation led to a “go or no go” decision regarding automation technology investments. When the decision was “not to go”, other priorities, more beneficial for the company, were recommend [16]. For automation technology investments in SMEs, it is essential to question how the investment supports the strategic objectives in the company. Here, the SDM may be a powerful tool to enable discussions about future directions, challenges and how to get there, not only focusing on the automation technology investment. The use of the SDM led to an awareness of current state in manufacturing as well as target state and the gaps between them. As the SDM process led to increased involvement and communication in Company A, the use of SDM seems to result in an increased strategic awareness. One conclusion for the applications of using the SDM was, that it seems to be a successful and suitable tool for manufacturing SMEs as a support for automation technology investments. Below is an explanation of the mechanisms that lead to the successful adaptation and use of the SDM in the SME context:

Design and Application of a Development Map

239

First, the topic, automation, is an urgent and important issue for SMEs [16]. The degree of automation in SMEs with low volume and high mix has traditionally been low, for many years under 1% of the manufacturing operations. During the last years, new types of automation technologies have been developed that are more suitable and less expensive for SMEs. Due to this, the interest of getting more knowledge and investing in automation is increasing in SMEs. Second, in all the applications, independent external coaches or researchers introduced the SDM to the companies and had a dialogue with the company regarding the SDM as well as automation technologies. The dialogue with this independent external part led to new knowledge about the topic and added value for the progress of automation investments. This mechanism was also mentioned as a success factor by the project manager of Robotlyftet: “Common to the participating companies is that they highlight the value of a complementary discussion with an independent part” [26]. Moreover, in the evaluation report of Robot Lift, it was also emphasized that coaches were not salespersons from automation integrators or automation manufacturers, but merely independent coaches supporting the companies’ progress in making the right automation investment. This was similar in the case of Company B, where an independent external coach supported the process. These two first mechanisms lead to a willingness to invest more time in evaluating and solving the issue, in this case automation investments. The SDM was also a facilitator for further understanding of the strategic role of automation. Third, it is likely that automation investments have been processed and discussed in the companies prior to the interventions. The SDM process accumulated quantitative facts that clarified previous assumptions and focused on objectivity. Objectivity means that the information to answer the questions needs to be based on quantitative facts and details, not from assumptions [23, 24]. Automation in SMEs tends to be large investments that can affect the whole manufacturing system and competitiveness, and it is beneficial if automation decisions are not made based on assumptions, but on quantitative facts. The SDM and underlying process were good guides of the necessity of being objective and base decisions on facts rather than on assumptions when making automation decisions. Fourth, the SDM was transparent, easy to understand and use. All steps are shown in one page, all text is written on this page, no hidden ideas or hidden text. As the SDM was iterated in a repetitive process, the decision maker recognized the steps. This led to an ease using the SDM when involving co-workers and communicating automation investments and long-term plans.

8 Conclusion This paper describes the design and application of the “Strategic Development Map” and explains the outcomes and mechanisms that lead to the successful adaptation and use of the SDM. The tool provides a structured guidance process to understand the strategic role of automation technology investments. This has been proven by the extensive application in Robot Lift, where external coaches supported users in a process to make a suitable automation investment decision. Four mechanisms were identified and explained that can be summarized as: Automation investment is an urgent and important issue for SMEs

240

M. Löfving et al.

and with support from independent external coaches they learnt how to base automation investment decisions on fact instead of assumptions in an easy and transparent process. The decision could also be to not invest in automation in its current state. This study contributes to improved practices of how manufacturing SMEs can align automation technology investment with strategy by using a structured and systematic tool. The theoretical contribution in this study is twofold. The study contributes to the automation field as this study is a step towards deeper understanding about automation technology needs and investment in manufacturing SMEs. The study also contributes to Toyota Management principles, and it exemplifies how the principles can be adapted and applied in practice in SMEs. This study was conducted in the context of automation decisions in manufacturing SMEs. The SDM has a generic nature, that was not tested and evaluated in this study. Next step of the research is to apply the SDM in another context to verify the generalizability of the tool. This will be done in a new research project starting September 2023. Acknowledgement. The authors wish to thank VINNOVA, the Swedish Governmental Agency for Innovation Systems for their financial support that enabled the research studies. The first author also wishes to thank KK-foundation for their financial support in the call “Recruitments”. The authors owe deep gratitude to the companies and external coach that participated in these research projects as well to Robotlyftet.

References 1. Buerkle, A., et al.: Towards industrial robots as a service (IRaaS): flexibility, usability, safety, and business models. Robot. Comput.-Integr. Manufact. 81, 102484 (2023) 2. Löfving, M., Almström, P., Jarebrant, C., Wadman, B., Widfeldt, M.: Evaluation of flexible automation for small batch production. Procedia Manufacturing 25, 177–184 (2018) 3. Perzylo, A., et al.: SMErobotics: smart robots for flexible manufacturing. IEEE Robot. Autom. Mag. 26(1), 78–90 (2019) 4. OECD: SMEs, Entrepreneurship and Innovation, Organisation for Economic Co-operation and Development, Paris (2010) 5. Ates, A., Garengo, P., Cocca, P., Bititci, U.: The development of SME managerial practice for effective performance management. J. Small Bus. Enterp. Dev. 20(1), 28–54 (2013) 6. Säfsten, K., Winroth, M., Stahre, J.: The content and process of automation strategies. Int. J. Prod. Econ. 110(1–2), 25–38 (2007) 7. Ortega, C.H., Garrido-Vega, P., Dominguez Machuca, J.A.: Analysis of interaction fit between manufacturing strategy and technology management and its impact on performance. Int. J. Oper. Prod. Manag. 32(8), 958–981 (2012) 8. Thomassen, M.K., Sjøbakk, B., Alfnes, E.: A strategic approach for automation technology initiatives selection. In: Grabot, B., Vallespir, B., Gomes, S., Bouras, A., Kiritsis, D. (eds.) Advances in Production Management Systems. Innovative and Knowledge-Based Production Management in a Global-Local World. APMS 2014. IFIP Advances in Information and Communication Technology, vol. 440. Springer, Heidelberg (2014). https://doi.org/10.1007/ 978-3-662-44733-8_36 9. Afolayan, A.O., De la Harpe, A.C.: The role of evaluation in SMMEs’ strategic decisionmaking on new technology adoption. Technol. Anal. Strategic Manag. 32(6), 697–710 (2020)

Design and Application of a Development Map

241

10. Hudson Smith, M., Smith, D.: Implementing strategically aligned performance measurement in small firms. Int. J. Prod. Econ. 106(2), 393–408 (2007) 11. Iakymenko, N., Alfnes, E., Thomassen, M.K.: A differentiated approach for justification of advanced manufacturing technologies. Adv. Manufact. 4, 257–267 (2016) 12. Hamzeh, R., Xun, X.: Technology selection methods and applications in manufacturing: a review from 1990 to 2017. Comput. Ind. Eng. 138, 106123 (2019) 13. Borges, L.A., Tan, K.H.: Incorporating human factors into the AAMT selection: a framework and process. Int. J. Prod. Res. 55(5), 1459–1470 (2017) 14. Granlund, A., Jackson, M.: Managing automation development projects: a comparison of industrial needs and existing theoretical support. In: Azevedo, A. (eds.) Advances in Sustainable and Competitive Manufacturing Systems. Lecture Notes in Mechanical Engineering. Springer, Heidelberg (2013). https://doi.org/10.1007/978-3-319-00557-7_63 15. Epping, K., Zhang, H.: A sustainable decision-making framework for transitioning to robotic welding for small and medium manufacturers. Sustainability 10(10), 3651 (2018) 16. Leandersson, C.: Slututvärdering Robotlyftet, (in Swedish) (2022). https://www.automatio nregion.com/positiva-resultat-fran-robotlyftet/. Accessed 18 Apr 2023 17. Denyer, D., Tranfield, D., Van Aken, J.E.: Developing design propositions through research synthesis. Organ. Stud. 29, 393–413 (2008) 18. Swedprod guide. www.edig.nu/sweprod (in Swedish). Accessed 20 Apr 2023 19. Löfving, M., Almström, P., Jarebrant, C., Wadman, B., Widfeldt, M.: Guide for automation of low volume production. In: Proceedings of the Swedish Production Symposium, vol. 13, pp. 13–23 (2020) 20. Simon, H.A.: The Science of the Artificial. MIT, Cambridge (1969) 21. Holmström… 22. Produktionslyftet. www.produktionslyftet.se. Accessed 19 Apr 2023 23. Rother, M.: Toyota Kata: Managing People for Improvement, Adaptiveness, and Superior Results. MGH, New York (2010) 24. Liker, J.K.: Toyota Way: 14 Management Principles from the World’s Greatest Manufacturer. McGraw-Hill, NY (2021) 25. Vinten, G.: Participant observation: a model for organizational investigation? J. Manag. Psychol. 9(2), 30–38 (1994) 26. Robotlyftet. https://swira.se/robotlyftet-en-framgangsrik-modell-for-utveckling. Accessed 14 June 2023

Using the Lean Approach for Improving Eco-Efficiency Performance: A Case Study for Plastic Reduction Matteo Ferrazzi(B)

and Alberto Portioli-Staudacher

Department of Management, Economics and Industrial Engineering, Politecnico di Milano, Via Lambruschini 4, b, Milano, Italy [email protected]

Abstract. In recent decades, there has been a growing awareness of reducing plastic consumption. Manufacturing companies are a major contributor to plastic consumption, during production. In the coming years, the transition to production models with less impact on the environment will be central for manufacturing companies. One possible solution, to address environmental sustainability needs, is through the development of a Lean approach. The Lean approach for improving eco-efficiency performance has been gaining increasing attention, from academia and companies. This article, through the analysis of a case study of an Italian company in the food sector, shows how the use of the Lean approach specifically the A3 tool, enabled the company to improve its eco-efficiency performance. In particular, the company was able to reduce plastic consumption in its production process by half by identifying and eliminating actions with no added value for the customer but responsible for plastic waste. At the final point of the study, a conclusion was done stating limitations and future research opportunities to expand the field of the work. Keywords: Lean manufacturing · Environmental sustainability · Eco-efficiency · Sustainability

1 Introduction Environmental sustainability is increasingly the focus of interest both in academia and for professionals. Some effects, such as rising energy prices, increased environmental pollution, and market demand for green practices, have prompted companies to rethink their strategies from a green perspective [1]. Especially in manufacturing companies, environmental sustainability initiatives aim to transform an input into an output using as few resources as possible [2]. The most important outcome for companies implementing sustainability strategies is the reduction of their environmental impact. Particularly manufacturing companies in recent years have become increasingly aware of their environmental impact. For this reason, especially in manufacturing, it is increasingly crucial to analyze and monitor eco-efficiency performance. Eco-efficiency indicators aim to © IFIP International Federation for Information Processing 2023 Published by Springer Nature Switzerland AG 2023 E. Alfnes et al. (Eds.): APMS 2023, IFIP AICT 689, pp. 242–252, 2023. https://doi.org/10.1007/978-3-031-43662-8_18

Using the Lean Approach for Improving eco-efficiency Performance

243

monitor the resources used during a production cycle. They are calculated by relating the value of production and the quantification of the effect on the environment (in terms of resources used or pollution generated) [3]. In fact, the most common performances monitored by companies using these indicators are natural resource use (energy, water, diesel fuel), pollution emissions (carbon footprint, CO2 emissions), material use/consumption (raw materials, auxiliary materials, packaging), and non-productive outputs (production waste, defects, production auxiliary materials). For several years, the need for production systems with increasingly better environmental performance has prompted academics to investigate how the lean approach is used for problems related to environmental sustainability. Nowadays, the lean manufacturing approach is considered the most influential paradigm in manufacturing. Empirical evidence shows that companies applying this approach benefit in terms of cost reduction, process improvements, and waste elimination while also increasing customer satisfaction [4]. 1.1 The Impact of the Lean Approach on Environmental Sustainability Performance In today’s literature, there is an increasing number of publications highlighting how the lean manufacturing approach in companies can have an impact on sustainable environmental performance. A strong relationship between lean manufacturing and environmental sustainability clearly emerges in the literature [5]. Many articles discuss how sustainable practices in manufacturing settings are strongly linked to lean manufacturing principles [6]. In particular, many articles highlight the positive effect of lean manufacturing practices by analyzing their impact on eco-efficiency performance. In fact, there are several studies in the literature that highlight how different lean practices have been successfully used to have better eco-efficiency performance. The lean practices that are most present to address issues related to environmental sustainability are Value Stream Map, Total productive maintenance, Kaizen, Just-in-time, and 5s. These lean practices have been used in manufacturing settings with the goal of improving various eco-efficiency performances. Some representative examples are VSM for energy consumption [7, 8] for CO2 decrease [9] and SMED for carbon footprint reduction [10]. At the same time, there are studies that highlight how the use of lean practices for improving eco-efficiency performance has not led to the desired results. In fact, articles such as [11] explain how using continuous improvement programs for water and energy consumption instead led to worse performance. While studies such as [12] point out how after the application of Human Resource Management, eco-efficiency performance to energy consumption remained unchanged, highlighting an implementation effort without succeeding in achieving the desired results. The interaction between of lean manufacturing and environmental sustainability share common goals, so integration between these paradigms is natural. Nevertheless, some scholars point out that manufacturing companies often face obstacles in implementing lean production processes for better environmental performance. This implementation involves several challenges and requires appropriate actions to overcome internal and external barriers. Some articles propose lean manufacturing as a suitable approach to support the transformation to production systems that also put attention on the environmental aspect. In particular, the article by Colombo et al. 2020 [13] presents an inductive analysis of a textile company

244

M. Ferrazzi and A. Portioli-Staudacher

that, while operating under lean manufacturing principles, has recently initiated a green transformation project. The results confirm that green transformation is a continuous process. In addition, the study suggests that the distinctive features of lean manufacturing can help organizations overcome specific barriers that arise at different stages of the green transformation process. Another example of a study is reported by Kaswan et al. 2021 [14]. This article proposes a decision-making model to help companies prioritize barriers and manage the important and causal relationships among them. The results reveal that management-related barriers rank first, followed by environmental and organizational barriers. Some interesting thoughts may emerge from this brief review of the literature. As presented earlier, the lean manufacturing approach, and in particular its practices [15], certainly has a positive influence on eco-efficiency performance. In fact, the lean manufacturing approach originated in manufacturing contexts to try to eliminate waste during production. In the literature review study of [16], which considered a review of papers from 1993 till 2010, the result of the investigation led to the conclusion that the causal relationship between lean and eco-efficiency is highly positive and few have some drawbacks. Consequently, the connection between the lean approach and environmental sustainability seems natural. It remains unclear, how in some contexts, the application of lean manufacturing does not lead to improvements in environmental performance. As reported earlier, some studies show how the application of lean practices for improving eco-efficiency performance can lead to no benefit or even a deterioration in performance. Despite the growing interest in environmental sustainability issues related to the lean approach, there is still a lack of understanding of what features of lean manufacturing are capable of positively influencing the sustainable performance of companies. In fact, in today’s literature, it is still unclear what success factors enable companies to achieve better eco-efficiency performance using the lean approach. This paper, through the analysis of a single case study, seeks to understand what are the success factors of the lean approach to achieve better eco-efficiency performance (RQ). In the following chapters, the case study of an Italian company in the food sector will be exposed. Then it will be shown how through the application of the lean approach, using the A3 problem-solving tool, the company was able to reduce plastic consumption for its packaging. Finally, it will be shown what were the success factors of the lean approach that enabled it to achieve better eco-efficiency performance.

2 Methodology The following chapter will describe the case study of an Italian company in the food sector. The main objective of this case study analysis is to understand how the use of a lean approach can contribute to increasing the company’s sustainable environmental performance. Then, how the company used the A3 problem-solving tool to address the problem of plastic consumption during the packaging of finished products will be presented. The problem analyzed in the case study was addressed using the lean approach of the A3 problem-solving framework [17]. The A3 methodology is a structured approach to problem-solving aimed at continuous improvement, first used by Toyota [18] and typically leveraged by lean manufacturing practitioners [19]. However, another

Using the Lean Approach for Improving eco-efficiency Performance

245

scholar Dave LaHote in his presentation [20], and John Shook [21] demonstrated A3 as a storyboard of steps of problem-solving and how A3 works in real life. The following will explore how the company addressed environmental performance through the PDCA (plan, do, check, and act) steps of the A3 model (see Fig. 1).

Fig. 1. A3 Problem Solving: PDCA steps

2.1 Presentation of the Case Study The case study analyzed is that of an Italian multinational company that operates in the processing of food into finished packaged products. One of the company’s strategic goals is to succeed in decreasing its impact on the environment by improving its ecoefficiency performance. Specifically in the production plant taken under analysis, the company aims to reduce the consumption of plastic used for packaging finished products. The plant taken into analysis consists of three packaging lines. During the past year, the company has been investing in continuous improvement programs through teams composed of managers, with high knowledge of Lean management. Accordingly, to address the problem of plastic consumption during production, the company decided to create a two-person Lean team. The Lean team through the use of A3 problem-solving tools will go on to develop a continuous improvement project with the goal of reducing plastic usage.

246

M. Ferrazzi and A. Portioli-Staudacher

2.2 Problem Statement and Problem Background The first step in the continuous improvement project was to identify the problem to be addressed. During processing, there is a large consumption of plastic. This is evidenced by the high number of kilograms of plastic that are thrown away during production. The wasted plastic is not used for the final packaging of the finished product and consequently must be disposed of by the company. This results in high operating costs, increased congestion in the plant’s internal logistics, and higher procurement costs. During the problem definition phase, the Lean team was able to ascertain that most of the wasted plastic came from finished product packages considered unfit for sale. As a result, packages not suitable for sale are considered plastic waste. Packages not adequate for sale were categorized into two macro-categories: finished product packages with label printing errors and empty finished product packages (during the packaging process they suffered from process errors, resulting in wasted plastic for packaging without product content). The result of this preliminary analysis was to focus more precisely on the problem. In fact, the Lean team decided to address the problem of reducing plastic consumption by focusing on the material waste generated by unfit-forsale packaging. Once the problem was clarified, the company moved on to the problem decomposition phase, analyzing the AS-IS situation of the three packaging lines that make up the department. During this phase of the project through data extraction from the packaging line monitoring system, the Lean team was able to analyze the current situation regarding plastic waste. The monitoring focused on understanding how much unfit-for-sale packaging affects plastic waste in the department. The monitoring phase lasted four weeks, so as to have a total rotation of the production mix at least two times. The first out of this phase is shown in Fig. 2. Figure 2 highlights how unfit-for-sale packages impact the department’s plastic waste. Considering the sum of the types of packages unfit for sale, it can be seen that 10% on Line 1, 7% on Line 3, and 8.6% on Line 4 of product packages cannot be sold, and consequently become a waste of plastic material. In particular, it can be seen from the graph that the percentage of empty packages is the effect with the highest magnitude on the department’s plastic consumption (7.9% on Line 1; 5.1% on Line 3; 6.8% on Line 4). At this point in the project, the company has set targets to guide the improvement process. Analyzing the data from the AS_IS situation, the Lean team decided to focus on reducing empty packages unfit for sale. The packaging lines produce an average of 1,500 pieces per day. The Lean team aims to reduce the number of empty packages on all lines by at least 2%. This means a reduction in plastic waste of about 10.5 tons per year. 2.3 Root Causes Identification, Devolving Countermeasures, Implementation, and Monitoring The last step before developing countermeasures to reduce the number of empty packages produced daily is to define the root causes that influence this phenomenon. Through the use of tools such as the Ishikawa diagram and especially through the involvement of departmental operators, the Lean team has identified a number of root causes. Root causes related to the high number of empty packages produced were classified into three

Using the Lean Approach for Improving eco-efficiency Performance

247

Fig. 2. AS-IS situation of plastic waste

distinct moments during production. The moments in production responsible for the highest number of empty package creations are line set up at the beginning of a shift, change of product to be packaged during production, and running out of plastic film reels for packaging. In addition, another root cause identified concerns the difficulty in communication between the two department managers. This is due to the fact that the production lines are separated and are developed in two different environments (sterilized aseptic chamber and end of line) separated by glass containment. Once the root causes responsible for the high percentage of empty packages produced daily were identified, the Lean team developed and implemented tailored countermeasures for each root cause. Again, the involvement of line operators was crucial at this stage. In fact, the mixed top-down and bottom-up approach to countermeasure development was one of the keys to the project’s success. In addition, for each countermeasure developed, the impact of reducing the percentage of empty packages was quantitatively estimated. The countermeasures developed and implemented after the monitoring phase is: – – – –

formalization of standard procedures for setting up the start of shift formalization of standard procedures for finished product changeover formalization of standard procedures for running out of plastic film reels Introduction of a walkie-talkie for communication between department managers.

248

M. Ferrazzi and A. Portioli-Staudacher

The indicated countermeasures were implemented gradually over two weeks after the four-week monitoring phase. The proposed graph below shows the trend in the percentages of empty packages produced daily over three months. It can be seen that after the implementation phase, the percentage of empty packages produced daily decreased on all 3 packaging lines. Through the standardization of performance monitoring processes by having a greater awareness of the data collected, the Lean team is still able to act in case the production process returns to unsuitable sustainable performance.

3 Results The outcomes proposed by the analysis of this case study are multiple. Through the use of the A3 problem-solving tool, the company was able to decrease the number of empty trays produced daily in production. The company started from an initial situation of 7.9% empty trays produced daily on Line 1, 5.1% for Line 3, and 6.8% for Line 4. Annual production in 2022 was 20.144 million packs of pieces produced. Considering the current percentage of the number of empty trays on the three lines (Line 1: 3.9%, Line 3: 3.2%, Line 4: 3.2%) (see Fig. 3), the projection for production in 2023 sees a decrease of 690,000 packs, corresponding to a reduction of 11.5 tons of plastic per year. The presence of the Lean team within the plant has also resulted in the identification of quality benefits for the plant. First of all, the decrease in the environmental impact of the plant must be highlighted as an effect that is not at all marginal in the application of the project. This first benefit allowed for other benefits at the production level, such as the reduction of 11.5 tons of plastic per year, the decrease in plastic disposal costs, and the simplification of internal and external logistics related to plastic handling, thus being understood as a reduction in the complexity of processes. In addition, it is important to emphasize how operators have become more aware of the issue of plastic consumption and the causes that generate waste of this material. Lastly, the presence of the Lean team and the development of the project has enabled the plant and the resources working in it to gain greater awareness of the data processed. To be more definite, the Lean team within the A3 pathway, identified what actually the problem is within the domain of environmental performance and clarified the areas of challenges of plastic disposal through problem breakdown and data shown. Consequently, it supported a clearer definition of the target to reduce the percentage of empty packages in 3 lines.

Using the Lean Approach for Improving eco-efficiency Performance

Fig. 3. Decrease % of empty packages during time: line 1, line 3 and line 4

249

250

M. Ferrazzi and A. Portioli-Staudacher

4 Discussion In this session it will be discussed what factors of the Lean approach enabled the achievement of the performance shown above. This fact was not only reported by the actors who were involved in the project, but also by the fact that plastic consumption in the production process was decreased by half. From this assumption, the authors looked to identify what were the success factors of the Lean methodology that enabled the achievement of the performance shown above. Analysis of the case study revealed what the success factors of the lean approach were that enabled the company to achieve the reduction of plastic consumption. 1. Defining and breaking down a complex problem in a continuous and granular way. Understand and discover the problem in a philosophy or clearer approach and refine it through demonstration of data. 2. Create teams dedicated to continuous performance improvement. The company found it beneficial to include a dedicated team with an understanding of the lean approach to focus on the project as a third party. 3. Develop quantitative metrics to guide the development of the solution. Through continuous performance monitoring, the team was able to have objective feedback on the interventions made. 4. Implement tailored countermeasures to the identified problems. The identification of root causes enabled the company to build measures perfectly related to the issues that emerged. As a result, avoiding wasting time and resources by implementing generic countermeasures that would have brought minimal benefits. 5. Involvement of stakeholders at different hierarchical levels. Interestingly, throughout the project, different stakeholders, regardless of their higher authority and hierarchical level in the company, had to be involved in the planning and action phase of countermeasures. In fact, many of the issues that emerged were identified by line operators, who experience the problems of the production department on a daily basis. 6. Monitoring the process to ensure that the results obtained are not lost. Recursive monitoring of plastic consumption has enabled the company not to nullify its achievements but, on the contrary, to foster a culture of continuous improvement. The factors outlined above were able to make the company achieve lower plastic use during the production process without worsening the operational performance of the plant.

5 Conclusion In conclusion, the case study showed how the lean approach proved essential in addressing the company’s eco-efficient performance problems. By applying the lean approach, particularly by using the A3 problem-solving tool, the company was able to decrease its environmental impact by reducing the consumption of plastic used during production. This was achieved without affecting the plant’s production levels. The quantitative results shown in Sect. 3 were achieved through some of the success factors of the lean approach. Then in Sect. 4, the authors brought out what the most significant factors of the

Using the Lean Approach for Improving eco-efficiency Performance

251

lean approach were. During the improvement project, the factors exposed in Sect. 4 led the company to improve its eco-efficiency performance. Specifically, the success factors that emerged from the case study are: defining and breaking down a complex problem, creating dedicated Lean teams, developing quantitative metrics, implementing tailored countermeasures, involving different stakeholders, and monitoring the process. Through the analysis of this case study we have further evidence of how the lean approach can positively influence environmental performance. We also have a greater awareness of what lean success factors manufacturing companies can leverage to improve their ecoefficiency performance. There is evidence in the literature of how the lean approach, particularly with the use of A3, can help companies in continuous improvement processes by impacting operational performance. While it is still little explored how the same approach can improve environmental performance. From an academic perspective, this case study was useful for a preliminary analysis of what the success factors of the lean approach are for improving sustainable performance. The major limitations of this article’s research are related to the use of a single case study. In fact, further empirical evidence studied in different contexts is needed to fully understand what the success factors of the lean approach are for achieving better eco-efficiency performance. Future research will seek to explore this issue through the analysis of multiple case studies. In particular, it is interesting to go out and conduct interviews with actors with different hierarchical levels. Through coding these interviews identify how the success factors of the lean approach can influence sustainable performance improvement, and how these factors are perceived by different hierarchical levels. To have a better impact of the results, multiple case studies will be selected: in different industries, to investigate different eco-efficiency performance, and in companies with different level of lean implementation. Through the collection of heterogeneous case studies, one can extrapolate which lean success factors are common and which ones differ.

References 1. Gaikwad, L., Sunnapwar, V.: Integrated lean-green-six sigma practices to improve the performance of the manufacturing industry. In: Concepts, Applications and Emerging Opportunities in Industrial Engineering (2021) 2. Benabdellah, A.C., Zekhnini, K., Cherrafi, A., Garza-Reyes, J.A., Kumar, A.: Design for the environment: an ontology-based knowledge management model for green product development. Bus. Strateg. Environ. 30(8), 4037–4053 (2021) 3. Maxime, D., Marcotte, M., Arcand, Y.: Development of eco-efficiency indicators for the Canadian food and beverages. J. Clean. Prod. 14, 636–648 (2006) 4. Womack, J.P., Jones, D.T. Roos, D.: Machine that changed the world. Simon and Schuster (2007) 5. Garza-Reyes, J.A., Kumar, V., Chaikittisilp, S., Tan, K.H.: The effect of lean methods and tools on the environmental performance of manufacturing organizations. Int. J. Prod. Econ. 200, 170–180 (2018) 6. Garza-Reyes, J.A.: Lean and green-a systematic review of the state of the art literature. J. Clean. Prod. 102, 18–29 (2015) 7. Estrada González, I.E., Adolfo, P., Guerrero García Rojas, H.: Decreasing the environmental impact in an egg-producing farm through the application of lca and lean tools. Appl. Sci. 10(4), 1352 (2020)

252

M. Ferrazzi and A. Portioli-Staudacher

8. Antomarioni, S., Bevilacqua, M., Ciarapica, F.E.: More sustainable performances through lean practices: a case study. In: 2018 IEEE IEEE International Conference on Engineering, Technology and Innovation ICE/ITMC 2018 - Proceedings, pp. 1–8 (2018) 9. Wu, P.: Monitoring carbon emissions in precast concrete installation through lean production – a case study in Singapore. J. Green Build. 9(4), 191–212 (2015) 10. Leme, R.D., Nunes, A.O., Message Costa, L.B., Silva, D.A.L.: Creating value with less impact: Lean, green and eco-efficiency in a metalworking industry towards cleaner production. J. Clean. Prod. 196, 517–534 (2018) 11. Chen, P.K., Fortuny-Santos, J., Lujan, I., Ruiz-de-Arbulo-López, P.: Sustainable manufacturing: exploring antecedents and influence of total productive maintenance and lean manufacturing. Adv. Mech. Eng. 11(11), 1–16 (2019) 12. Resta, B., Dotti, S., Gaiardelli, P., Boffelli, A.: How lean manufacturing affects the creation of sustainable value: an integrated model. Int. J. Autom. Technol. 11(4), 542–551 (2017) 13. Colombo B., Gaiardelli P., Dotti S.: Overcoming barriers of green transformation through the adoption of lean manufacturing: a case study (2020) 14. Kaswan M.S., Rathi R., Reyes J.A.G., Antony J.: Exploration and investigation of green lean six sigma adoption barriers for manufacturing sustainability (2021) 15. Shah, R., Ward, P.T.: Lean manufacturing: context, practice bundles, and performance. J. Oper. Manag. 21(1), 129–149 (2003) 16. Moreira, F., Alves, A.C., Sousa, R.M.: Towards eco-efficient lean production systems. In: Balanced Automation Systems for Future Manufacturing Networks: 9th IFIP WG 5.5 International Conference, BASYS 2010, Valencia, Spain (2010) 17. Torri, M., Kundu, K., Frecassetti, S., Rossini, M.: Implementation of lean in IT SME company: an Italian case. Int. J. Lean Six Sigma 12(5), 944–972 (2021) 18. Shook, J.: Toyota’s secret. MIT Sloan Manag. Rev. 50(4), 30 (2009) 19. Rossini, M., Audino, F., Costa, F., Cifone, F.D., Kundu, K., Portioli-Staudacher, A.: Extending lean frontiers: a kaizen case study in an Italian MTO manufacturing company. Int. J. Adv. Manuf. Technol. 104(5–8), 1869–1888 (2019) 20. LaHote, D.: A3 Story Boards. https://www.lean.org/wpcontent/uploads/2021/08/277.pdf. Accessed 19 Apr 2023 21. Shook, J.: Managing to learn: using the A3 management process to solve problems, gain agreement, mentor, and lead. Lean Enterprise Institute (2008)

Work Pattern Analysis with and without Site-Specific Information in a Manufacturing Line Takeshi Kurata1(B) , Rei Watanabe2 , Satoki Ogiso1 , Ikue Mori1 , Takahiro Miura1 , Karimu Kato1 , Yasunori Haga3 , Shintaro Hatakeyama3 , Atsushi Kimura3 , and Katsuko Nakahira2 1 Human Augmentation Research Center, AIST, Kashiwa, Japan

[email protected]

2 Faculty of Engineering, Nagaoka University of Technology, Nagaoka, Japan 3 DENSO Corporation, Nishio, Japan

Abstract. In recent years, there has been an increasing number of technological developments and practical applications for efficient and quantitative work analysis by measuring workers’ flow lines. However, there may be a delay in starting work analysis if it is started after preparing site-specific information, such as each worker’s role and typical work areas. This paper reports on a case study of work analysis using geospatial intelligence techniques with and without such sitespecific information. First, this paper introduces the work site targeted in this case study, the purpose of the analysis, and the data measured and collected at the work site. Next, it describes a series of methods such as indoor positioning, generation of work area transition model, clustering of work area transition instances, and exception extraction of the instances for the purpose of analyzing work patterns. A comparison of the clustering results with each worker’s role and an analysis of non-routine work patterns are also reported. Keywords: Geospatial intelligence · Indoor positioning · Work area transition model · Cluster analysis · Non-routine work pattern

1 Introduction To realize health and productivity management (HPM) [1–3], it is necessary to simultaneously improve productivity and quality of work (QoW) [4, 5]. QoW consists of health, workability, and meaningfulness of work, and was proposed by the Council on Industrial Competitiveness (COCN) in Japan. To improve both labor productivity and QoW, which is closely related to occupational health, quality of working life (QWL), decent work, and social capital, in a balanced manner, it is necessary to incorporate an engineering approach and a concept such as Digital Transformation (DX). On the other hand, it has been reported that approximately 60 to 80% of information is related to location information [6] and that humans spend approximately 90% of their time indoors [7]. Given © IFIP International Federation for Information Processing 2023 Published by Springer Nature Switzerland AG 2023 E. Alfnes et al. (Eds.): APMS 2023, IFIP AICT 689, pp. 253–266, 2023. https://doi.org/10.1007/978-3-031-43662-8_19

254

T. Kurata et al.

these facts, Geospatial Intelligence (GSI), especially indoor GSI [8–10], which supports problem solving by linking geospatial information including location information with other information, is considered to be an effective means to promote DX and HPM. Work analysis using GSI has been conducted at work sites in various industries and business categories to support HPM to date. In each of those cases, the analysis was initiated after prior investigation of site-specific information, such as each worker’s role, shift, typical flow lines and typical work areas. In addition, the site floor map had to be divided into multiple work areas for analyzing work area transition [11–14]. However, such a way of proceeding would have delayed in starting the analysis and obtaining preliminary and final results of the analysis due to the personnel and time costs involved in on-site coordination and preparation as shown in Fig. 1 (A). This paper reports on a case study of work analysis using GSI with and without such site-specific information as shown in Fig. 1 (B). First, Sect. 2 outlines the purpose of the analysis in this case study, the targeted work sites, and the types of data used in the analysis. Next, Sect. 3 introduces the indoor positioning system and method for measuring flow lines of workers. Then, Sects. 4 and 5 describe the work area transition model generation and non-routine work pattern extraction, respectively. Finally, Sect. 6 discusses the results of the work pattern analysis without site-specific information and the elaborated analysis with site-specific information.

Fig. 1. Analysis started with/without site-specific information (ground truth).

Work Pattern Analysis with and without Site-Specific Information

255

2 Basics: Purpose, Work Site and Data The case study reported in this paper is about on-site measurements performed on a manufacturing line at the DENSO CORPORATION factory in September 2021. DENSO CORPORATION is promoting activities aimed at improving both productivity and QoW, especially the working environment, such as the thermal environment. Part of these activities is being undertaken in a joint research project with AIST. In this joint research, a multifaceted analysis of productivity and QoW is being conducted using various acquired data mentioned below, while utilizing indoor GSI technology (Fig. 2) [15].

Fig. 2. Example of visualization of room temperature heat map and flow line (exposed temperature and activity of each worker).

The site targeted in this analysis is responsible for the manufacturing process of automotive work-in-process. The main area for analysis was a manufacturing line that fits within a rectangle of approximately 1,400 m2 , but the surrounding area and a break room were also included for measurement. An indoor positioning system (outlined in the next section) and more than 30 temperature and humidity sensors were installed to continuously collect data to understand each worker’s work behavior and environmental conditions, and to monitor their changes over time. Also, noise measurements were taken once under steady-state operating conditions, although no time variation was obtained. In addition, each worker completed questionnaires on temperature, fatigue, and noise twice, once during meal breaks and once at the end of each shift. The measurement period was five days, and the number of workers to be measured was 10, five each for the day shift (first shift) and night shift (second shift). The roles of the day shift workers were leader, deputy leader, receiving, visual inspection, and internal inspection, while the night shift workers were leader, deputy leader, receiving, visual inspection, and assembly. From the multifaceted analysis in this joint study, this paper focuses on a work analysis using flow line data for capturing a comprehensive view of the actual work behavior of each worker on the manufacturing line. In particular, the research question

256

T. Kurata et al.

here is to examine whether it is possible to start the analysis even before site-specific information is available.

3 Indoor Positioning The following is an overview of the indoor positioning system installed at the site of this analysis and the applied positioning method [16]. The hardware of the system consists of a mobile device (Android-smartphone-like device for business use) and a BLE (Bluetooth Low Energy) beacon (Fig. 3). When the measurement range is wide and the number of moving objects (measurement targets) is small, it is more cost-effective to place low-cost transmitters such as BLE beacons in the field and relatively expensive receivers such as smartphones on the moving object side. This type of system is called a “SB (Stationary Beacon) type system” in [9]. In this case study, the SB-type system configuration was adopted. BLE beacons were installed at 48 locations along and near the manufacturing line, and at 8 locations to monitor the flow line to the break room and other locations.

Fig. 3. Indoor positioning system.

A mobile device placed in a pocket of a work uniform measures acceleration, angular velocity, magnetic field, and RSSI (Received Signal Strength Indicator) of BLE beacons placed in the environment. First, the relative speed and moving direction are estimated from the acceleration, angular velocity, and magnetic field using PDR (Pedestrian Dead Reckoning) [17–19]. Next, these are used to calculate the predicted position of each particle in the particle filter. By comparing the RSSI expected at these predicted positions with the actually measured RSSI, the likelihood is calculated, and the positions are updated. However, if the particle filter is used for a long time, particles may disappear around the actual location. Therefore, if the marginalized likelihood of a particle is below a certain value and the RSSI of the strongest BLE beacon is above a certain value, the particle is placed around the BLE beacon to correct the approximation of the posterior distribution of its position.

Work Pattern Analysis with and without Site-Specific Information

257

4 Work Area Transition Model and Instances The indoor positioning system described above provides detailed flow line data. Although the descriptive granularity of such flow line is effective for microscopic analysis, in this analysis, a state transition model is automatically generated as in [20] to understand the overarching trend of on-site work. In this paper, we refer to this model as “work area transition model.” The procedure for generating the work area transition model used in this analysis (Fig. 4) is outlined as follows. (1) Divide flow lines into moving segments and staying segments. (2) Obtain the “staying plot” which is the representative position of the spatial distribution in each staying segment. (3) Obtain the “work area” which is the cluster of all staying plots based on non-hierarchical clustering. (4) Obtain the “staying node” which is the representative position of each working area and connect between each of the staying nodes with the directed “moving edge.” (5) Determine “staying length levels” by clustering the lengths of all staying segments. In (1), tentative moving segments are first obtained by simple filtering, and if the tentative moving segment contains moving velocity above a threshold, it is selected as a moving segment; otherwise, it is selected as a staying segment. In this analysis, the length of the tentative moving segment was set to at least 5 s, and the threshold for the velocity was set to 0.7 m/sec. For the staying plot in (2), the position of the center of mass of the spatial distribution was used as the representative position. In (3) and (4), the k-means++ method [20] was applied for the non-hierarchical clustering, and the number of clusters (staying nodes) was determined based on the combined criteria of the elbow and silhouette methods. Clustering in (5) also employed the k-means++ method, and the number of clusters (level of staying time length) was determined based on the elbow method. Thus, the work area transition model is generated. By registering each feature in this work area transition model as shown in Fig. 5 and as described in (6) and (7) below, it is able to obtain a “work area transition instance”. (6) Register the rate of the number of staying segments corresponding to each stay length level for each staying node relative to the number of all staying segments as the “staying rate feature.” (7) Register the rate of the number of transitions (number of moving segments) corresponding to each moving edge relative to the number of all transitions as the “moving rate feature” for each edge. In such a way, work area transition instances can be obtained from the flow line data of each worker. Furthermore, using a non-hierarchical clustering method (k-means ++ method), the instances are portioned into clusters. Those clusters are regarded as “work patterns.”

258

T. Kurata et al.

Fig. 4. How to generate the work area transition model used in this analysis.

Fig. 5. How to obtain a work area transition instance from the work area transition model.

5 Exception Extraction of Work Patterns The quantification of work patterns is foundational to work analysis and is expected to have a variety of applications. As part of its application, this study addressed the identification of non-routine work (exception) and their analysis. For the purpose, an extraction

Work Pattern Analysis with and without Site-Specific Information

259

method of work pattern exception was developed based on typical work patterns as shown in Fig. 6.

Fig. 6. How to extract exceptions in this analysis.

The following is a summary of the method for exception extraction. (1) Compute representative values of each feature in the work area transition instances included in the same work pattern, and generate the representative instance for the work pattern, whose elements are those representative values.

260

T. Kurata et al.

(2) Compare a work area transition instance with a representative instance of the work pattern to which the instance belongs and calculate the exception indicator (the residual sum of squares (RSS) of the features in each dimension). (3) Determine a unified threshold of exception indicators for all work patterns and extract exceptions by thresholding exception indicators. (4) Extract work patterns with only one work area transition instance as exceptions. Here, the representative value in (1) was set as the median. In (2), each RSS is calculated within each work pattern as an exception indicator, while in (3), thresholding is performed with a threshold obtained using exception indicators of all work patterns. Specifically, the exception indicators of all work patterns are sorted in ascending order, and Q3 + 1.5IQR (Q3 : third quartile, IQR: interquartile range) is used as the threshold as in box-and-whisker diagram drawing. In this way, it is expected that it is more likely to achieve both individuality by calculating each exception indicator with the representative instance for each work pattern and stability by determining one unified threshold from all exception indicators of all work patterns. 6 Result and Discussion. 5.1 Analysis Without Site-Specific Information A series of analysis methods described in Sects. 4 and 5 is designed and developed so that the analysis can be started only with measurement data just as in unsupervised learning. To verify the feasibility in the actual site, the analysis methods were applied according to the analytical process as shown in Fig. 1 (B). Figure 7 shows the work area transition model generated in the target site. The number of staying nodes, which are the representative positions of each working area, was 21 in total: 13 on the manufacturing line, 7 in the surrounding area, and 1 outside the targeted site. The degree of length of stay was graded into the four levels as in Fig. 4. As a result, the model consisted of 84-dimensional staying rate features, which are combinations of staying nodes (21 nodes) and the degree of staying time length (4 levels), and 441-dimensional moving rate features (441 is the number of moving edges between each of the 21 staying nodes) for a total of 525 dimensions. As for working patterns, 46 work area transition instances were portioned into 12 clusters with the k-means++ method according to the criteria for determining the number of clusters used in (3) in Sect. 4. It means that 12 work patterns were found in this site. Furthermore, 7 exceptions were found as non-routine works out of the 46 work area transition instances. Consequently, it was confirmed that the work patterns can be extracted without using site-specific information, such as each worker’s role, shift and typical work areas. In practice, it was necessary to finish the following preparations prior to flow line measurement: Obtaining the number of people to be measured and the floor map, determining the location of BLE beacons, and installing the beacons into the site. However, it is not necessary to conduct detailed interviews with site managers to make those preparatory works, and previous cases have confirmed that it is relatively easy to proceed with them [10–14].

Work Pattern Analysis with and without Site-Specific Information

261

Fig. 7. Work area transition model generated in this case study.

Table 1. Clustering results of work area transition instances and their relationship to roles. The numbers in this table are the work pattern IDs (cluster ID of the work area transition instances). For the 4 work patterns with one instance, the work pattern IDs are indicated as 0.

5.2 Elaborated Analysis with Site-Specific Information In addition, the above analysis was elaborated as site-specific information became available. Table 1 shows the clustering results, organized by the role of the worker on each shift. From these results, the following tendencies were observed. • The work patterns of the leader and deputy leader are generally the same for both the day shift and night shift. • Receiving and visual inspection have different work patterns between the day and night shifts even in the same role, but the work patterns are the same within each of the day and night shifts.

262

T. Kurata et al.

Figure 8 shows that example state transition diagrams of work area transition instances included in the typical work patterns for each role. On the other hand, Fig. 9 shows that the assumed deployment information for each role provided by the site manager. These comparisons indicated that, overall, the work patterns were quantitatively as expected by the site manager. In addition, it was found that the leader and the deputy leader were mainly responsible for the right and left sides of the manufacturing line, respectively. Subsequently, it was confirmed by additional interview with the site manager that there was such an assignment of duties between the leader and the deputy leader. In this way, these were findings that could not have been obtained simply by interviewing site managers without any hypothesis. As mentioned above, seven exceptions were found as non-routine works out of 46 work area transition instances. Figure 10 shows the extracted exceptions with the unified threshold for all work patterns. Out of seven exceptions, three were extracted by thresholding. The remaining four exceptions were extracted because they were the only work area transition instance in the cluster to which each belonged. Table 2 shows where these exceptions are in Table 1.

Fig. 8. Example state transition diagrams of work area transition instances in the typical work patterns for each role.

Work Pattern Analysis with and without Site-Specific Information

263

Examples of the non-routine works extracted are shown in Fig. 11. These relate to visual inspection workers. In the typical work pattern, the worker stays most at the right-most staying node, and stays and moves mostly between the rightmost to the center of the manufacturing line. In the case of Exception A, as in the typical pattern, the worker stayed most at the right-most staying node, but the difference was that the worker stayed throughout the entire manufacturing line. In Exception B, on the other hand, the difference was that the worker stayed most at the second staying node from the right.

Fig. 9. Assumed deployment information for each role provided by the site manager.

Table 2. Exceptions of work area transition instances in Table 1

264

T. Kurata et al.

Fig. 10. Extracted exceptions of work area transition instances along with their exception indicators and the threshold.

Fig. 11. State transition diagrams of exceptions A and B for visual inspection.

6 Conclusions This paper reported on a case study of work analysis using GSI technologies conducted on a manufacturing line. In the case study, 12 work patterns were found without collecting site-specific information containing each worker’s role, shift and typical work areas of each worker. After such site-specific information was collected, it was confirmed that

Work Pattern Analysis with and without Site-Specific Information

265

the work patterns were described appropriately and quantitatively. It was for the first time in this site. Those work patterns were, in general, as expected by the site managers. On the other hand, their quantification made the site virtually more tangible. This led to findings that could not have been obtained simply by interviewing on-site managers without any hypothesis. In addition, we were able to extract when and by which worker the non-routine works were performed. These are only possible through analysis based on on-site measurements and modeling, as in this study. This type of analysis would also serve as the basis for implementing a relatively low-cost methodology for understanding the site without sacrificing breadth or depth. For instance, a detailed survey including interviews with workers based on automatic screening from big data and is expected to be widely used in the future [9, 10]. Our research goal is to establish a methodology to support health and productivity management by balancing productivity and QoW. To this end, the following future developments are planned based on the findings of this case study. • Integrated analysis for comprehensive understanding of the site with operational data (takt/lead time, OEE: Overall Equipment Effectiveness), environmental exposure data (room temperature, humidity, noise), physical work data (flow line, behavior), vital sign data (heart rate, body temperature), and subjective data [15, 21]. • Pre-evaluation of site improvement/design ideas by simulation based on digital twinning of the actual site utilizing methods such as work area transition model generation [14, 22]. • Support for designing factory environments [23] that can achieve both productivity improvement and QoW improvement, especially in the working environment, such as the thermal environment. Acknowledgment. We would like to express our sincere appreciation to all factory workers who participated in this study.

References 1. Goetzel, R.Z., Guindon, A.M., Turshen, I.J., Ozminkowski, R.J.: Health and productivity management: establishing key. J. Occup. Environ. Med. 43(1), 10–17 (2001) 2. Takahashi, H., Nagata, M., Nagata, T., Mori, K.: Association of organizational factors with knowledge of effectiveness indicators and participation in corporate health and productivity management programs. J. Occup Health 63(1),(2021). https://doi.org/10.1002/1348-9585. 12205 3. Mori, K., et al.: Development success factors, and challenges of government-led health and productivity management initiatives in Japan. J. Occup. Environ. Med. 63(1), 18–26 (2021) 4. QoW, http://www.cocn.jp/report/thema91-L.pdf. Accessed 7 Apr 2023 5. Taylor, M. P., Boxall, P., Chen, J. J.J., Xu, X., Liew, A., Adeniji, A.: Operator 4.0 or Maker 1.0? Exploring the implications of Industrie 4.0 for innovation, safety and quality of work in small economies and enterprises, Comput. Ind. Eng., 139, 105486 (2020) 6. Hahmann, S., Burghardt, D.: How much information is geospatially referenced? Networks and cognition. Int. J. Geogr. Inf. Sci. 27(6), 1171–1189 (2013)

266

T. Kurata et al.

7. Klepeis, N. E., et al.: The national human activity pattern survey (NHAPS): a resource for assessing exposure to environmental pollutants, Lawrence Berkeley National Laboratory (2001) 8. Openshaw, S., Openshaw, C.: Artificial Intelligence in Geography, Chichester. John Wiley & Sons, UK (1997) 9. Kurata, T., et al.: IoH technologies into indoor manufacturing sites. APMS, IFIP AICT 567, 372–380 (2019) 10. Kurata, T.: Geospatial intelligence for health and productivity management in Japanese restaurants and other industries. APMS, IFIP AICT 632, 206–214 (2021) 11. Kurata, T., Harada, M., Nakahira, K., Maehata, T., Ito, Y., Aso, H.: Analyzing operations on a manufacturing line using geospatial intelligence technologies. APMS, IFIP AICT 663, 69–76 (2022) 12. Shinmura, T., Ichikari, R., Okuma, T., Ito, H., Okada, K., Nonaka, T.: Service robot introduction to a restaurant enhances both labor productivity and service quality, CIRP ICME ‘19 (2019) 13. Ichikari, R., et al.: A case study of building maintenance service based on stake-holders’ perspectives in the service triangle, Proc. ICServ 2018, 87–94 (2018) 14. Myokan, T., et al.: Pre-evaluation of Kaizen plan considering efficiency and employee satisfaction by simulation using data assimilation - Toward constructing Kaizen support framework. Proc. ICServ 2016, 7 (2016) 15. Nakae, S., et al.: Geospatial intelligence system for evaluating the work environment and physical load of factory workers. In: 45th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC), (2023) (to appear) 16. Ogiso, S., et al.: Integration of BLE-based proximity detection with particle filter for daylong stability in workplaces. In: IEEE/ION Position Location and Navigation Symposium (PLANS), pp.1060–1065 (2023) 17. Kourogi, M., Kurata, T.: Personal positioning based on walking locomotion analysis with self-contained sensors and a wearable camera. Proc. ISMAR, 103–112 (2003) 18. Kourogi, M., Ishikawa, T., Kurata, T.: A method of pedestrian dead reckoning using action recognition. In: IEEE/ION Position, Location and Navigation Symposium, pp. 85–89 (2010) 19. Kourogi, M., Kurata, T.: A method of pedestrian dead reckoning for smartphones using frequency domain analysis on patterns of acceleration and angular velocity. In: IEEE/ION Position, Location and Navigation Symposium-PLANS, pp. 164–168 (2014) 20. Arthur, D., Vassilvitskii, S.: K-means++: the advantages of careful seeding. In: Proceedings of the Eighteenth Annual ACM-SIAM Symposium on Discrete Algorithms (SODA ‘07). Society for Industrial and Applied Mathematics, pp. 1027–1035 (2007) 21. de la Riva, J., Garcia, A.I., Reyes, R.M., Woocay, A.: Methodology to determine time allowance by work sampling using heart rate. Procedia Manuf. 3, 6490–6497 (2015) 22. Mouayni, I.E., Etienne, A., Lux, A., Siadat, A., Dantan, J.-Y.: A simulation-based approach for time allowances assessment during production system design with consideration of worker’s fatigue, learning and reliability. Comput. Ind. Eng. 139, 105650 (2020) 23. Romero, D., Stahre, J., Taisch, M.: The Operator 4.0: towards socially sustainable factories of the future. Comput. Ind. Eng. 139, 106128 (2020)

Digital Transformation Approaches in Production Management

Digital Transformation Towards Industry 5.0: A Systematic Literature Review Jelena Crnobrnja1

, Darko Stefanovic1 , David Romero2 and Ugljesa Marjanovic1(B)

, Selver Softic3

,

1 Faculty of Technical Sciences, University of Novi Sad, Trg Dositeja Obradovica 6,

21000 Novi Sad, Serbia {darko.stefanovic,umarjano}@uns.ac.rs 2 Tecnológico de Monterrey, Mexico City, Mexico 3 IT and Business Informatics, CAMPUS 02 University of Applied Sciences, 8010 Graz, Austria [email protected]

Abstract. A Fifth Industrial Revolution has emerged, referred to as Industry 5.0, which represents a paradigm shift towards the integration of advanced technologies and the latest innovations in manufacturing processes and systems with a new hallmark in mind characterized by three main features: human-centricity, sustainability, and resilience. In the current scientific literature, there is still a lack of efforts to systematically review the state-of-the-art of Digital Transformation towards Industry 5.0. This systematic literature review aims to address this research gap by investigating the academic progress at the crossroads of Digital Transformation and Industry 5.0. The systematic review covered the analysis of journal articles and conference papers within the Industry 5.0 domain that were published during the last five years until March 2023. An overview of Digital Transformation and Industry 5.0 research fields is firstly captured by enumerating a list of prominent journals and international conferences for publishing on Industry 5.0-related content, and secondly by illustrating the significance of Digital Transformation and Industry 5.0 based on keywords classification. Keywords: Digital Transformation · Industry 5.0 · Human-Centric Systems · Cyber-Physical Systems · Sustainability · Resilience

1 Introduction A Fifth Industrial Revolution has emerged, referred to as “Industry 5.0”, which represents a paradigm shift towards the integration of advanced information and operational technologies and the latest innovations in manufacturing processes and systems with a new hallmark in mind characterized by three main features: human-centricity, sustainability, and resilience [1, 2]. Industry 5.0 is a term that refers to the next phase of industrial transformation, following Industry 4.0, which had limited consideration of the triple-bottom-line in economic and technological development activities [3, 4]. It envisions a human-centered approach to manufacturing systems that combines the benefits © IFIP International Federation for Information Processing 2023 Published by Springer Nature Switzerland AG 2023 E. Alfnes et al. (Eds.): APMS 2023, IFIP AICT 689, pp. 269–281, 2023. https://doi.org/10.1007/978-3-031-43662-8_20

270

J. Crnobrnja et al.

of advanced automation solutions and new digital and smart technologies with human expertise and creativity to enhance productivity, innovation, and resilience in manufacturing processes while caring about the workforce and the environment [5]. Thus, Industry 5.0 aims to integrate the latest advancements in Autonomous and Collaborative Robotics (AMRs/Cobots), Artificial Intelligence (AI), Augmented and Virtual Realities (AR/VR), and the Internet of Things (IoT), among other modern technologies with the knowledge and ingenuity of human workers to create more “sustainable” manufacturing ecosystems in compliance with the triple-bottom-line [6]. The emphasis is on creating a harmonious balance between economic and technological developments, social prosperity, and the environmental limits of nature to achieve optimal results in terms of productivity, quality, innovation, and sustainability in all industrial sectors [6, 7]. Despite the recent emergence of Industry 5.0 as a novel research paradigm, its current level of exploration remains at its “nascent” stage, with limited systematic inquiry and a scarcity of empirical evidence [2]. Therefore, this paper aims to analyze the research topics at the crossroads of “Digital Transformation” and “Industry 5.0” covered by academic journals and conference proceedings in the last five years until March 2023. This systematic literature review will also provide information about future research directions and opportunities in the Industry 5.0 domain. More specifically, three research questions will be addressed by this systematic review: • RQ1: Who is working on Industry 5.0, where, and when? • RQ2: What approaches and methods are used in Industry 5.0 research? • RQ3: What are the main research directions for Industry 5.0? The rest of this work is outlined as follows: Sect. 2 presents the systematic literature review method; Sect. 3 shows and discusses the obtained results from the basic data analysis of included journal articles and conference papers and the specific data analysis corresponding to each research question, and Sect. 4, concludes the paper with a synthesis of all the above discussions.

2 Research Methodology One of the defining features of academic research is the systematic review of existing literature, which serves as the foundation for advancing knowledge in a given field [8]. In a narrower sense, a systematic literature review is seen as a form of research that deals with existing publications and applies a “systematic” methodology to synthesize data that has already been published [9]. Through a comprehensive analysis of previously published works, researchers can identify research gaps in the body of knowledge and gain valuable insights that inform their investigations. By building upon existing knowledge, academic researchers can expand the frontiers of their respective fields, pushing the boundaries of what is known and shedding new light on previously unexplored areas of inquiry. In essence, the review of literature is a critical component of the academic research process, as it enables researchers to locate the border of knowledge and identify the most promising avenues for future investigation. The proliferation of knowledge in diverse management and engineering sciences is accelerating rapidly. The condition of reliability and originality of research underlies all investigation activities. Therefore, literature review as a research method is more important than ever [10].

Digital Transformation Towards Industry 5.0

271

For this study, the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) method was used [11]. This method aims to help authors improve the construction of systematic literature reviews, which was adopted. This method addresses questions that individual studies have not answered, identifying research questions and theories [11]. This method is illustrated by applying three phases: (i) identification, (ii) screening, and (iii) inclusion. These steps are necessary to bring the article portfolio closer to the objective of this study and are widely used in different research fields [2]. The PRISMA flow chart that reports the different phases of this systematic literature review is shown in Fig. 1.

Fig. 1. Systematic Literature Review PRISMA Flow Chart

272

J. Crnobrnja et al.

2.1 Publications Collection In this systematic review, the literature was sought and analyzed using two online metadatabases: Scopus and Web of Science (WoS). These two meta-databases were selected due to their comprehensive coverage of scientific disciplines, high-quality indexing of journals and conference proceedings, citation tracking features, advanced search options, and reputation and credibility within the scientific community. Also, both have advanced search options and offer search options that allow the refinement of search criteria and tailor search to specific needs. To identify a comprehensive set of journal articles and conference papers, the search string was constructed through the combination of the operator ‘OR’ in between each two of the following four terms: ‘Digital Transformation’, ‘Fifth Industrial Revolution’, ‘5th Industrial Revolution’, and ‘Industry 5.0’. This inclusive search yielded 42 Scopus publications and 71 WoS publications (up to March 2023). Furthermore, narrowing the search field will bring the results closer to the research context, and therefore an extra term was added to the search string: ‘Manufacturing’, which now yields 33 Scopus publications and 24 WoS publications. Moreover, by changing the ‘Manufacturing’ additional (contextual) term for the ‘Production’ term (a potential synonym) in the search string, the following results were obtained 32 Scopus publications and 25 WoS publications. Both research strings were preserved and then analyzed because ‘Manufacturing’ and ‘Production’ are two related terms that are often used interchangeably but they have distinct meanings and implications, which were confirmed and duplicated publications (47.3% duplicates) were eliminated: (Digital Transformation + Industry 5.0 + Manufacturing) and (Digital Transformation + Industry 5.0 + Production). In summary, manufacturing refers to the entire process of creating a product, from design to distribution, while production specifically refers to the physical act of making or creating something. Manufacturing is a broader term that encompasses production as one of its components. After the screening phase, 31 publications were selected for further analysis (i.e., 20 journal articles, five conference papers, and one book chapter). After reading each publication, three conference papers were excluded from the final results since these did not address “Industry 5.0” as their main topic, and the book chapter was excluded too due to its focus solely on the “Economics of Industry 5.0”. Finally, 26 eligible publications were studied in detail and classified into diverse sub-categories of the inclusion criteria. 2.2 Publications Data Collection For each included publication, two types of information were collected. The first type is basic data about the publications, which include: (i) titles, (ii) keywords, (iii) online database, (iv) document type (e.g., journal article or conference paper) and related information. For journal articles: titles, publisher names, and years of publications were collected. Meanwhile, for conference papers: conference names and conference years were gathered. Moreover, each publication was also classified according to the following four content-based categories: (i) Review – this type of publication presented a review related to Industry 5.0 as its main content; (ii) Theoretical Solution – this type of publication

Digital Transformation Towards Industry 5.0

273

identified Industry 5.0 research problems and proposed some conceptual or theoretical solutions but without any laboratory experiments or industrial applications, and (iii) Practical Solution – this type of publication provided conceptual or theoretical solutions as well as a laboratory experiment or an industrial application as part of its contribution. The second type is specific data related to the three declared Research Questions (RQs), which include: • For RQ1 – ‘Who is working on Industry 5.0, where, and when?’ – the data collected from included publications are: (i) authors and (ii) publication years. • For RQ2 – ‘What approaches and methods are used in Industry 5.0 research?’ – the data collected from included publications are: (i) methods and (ii) scope of research. • For RQ3 – ‘What are the main research directions for Industry 5.0?’ – the data collected from included publications are: (i) research purpose(s) (e.g. digital transformation models, human centrality, and maturity models), and (ii) domain focused by each included publication. 2.3 Publications Data Analysis The collected data were analyzed by applying both “qualitative” and “quantitative” methods. Microsoft Office Excel and FreeWordCloudGenerator.com tools were used to process the data in further research.

3 Industry 5.0 Characterization in Digital Transformation This section first presents the basic data analysis to provide an overall picture of included publications and then gives corresponding answers to each research question. 3.1 Basic Data Analysis 26 publications were finally included in this systematic literature review. 65% (17 publications) were found in the Scopus meta-database and 35% (9 publications) in the WoS meta-database. According to the content-based categories, 50% (13 publications) are classified as “literature review works”, 27% (7 publications) as “theoretical research works”, and 23% (6 publications) as “practical research works”. 3.2 Analysis of Resulting Publications Further investigation was carried out to analyze the included journal articles and conference papers. Of the 26 publications included in this systematic review, 22 (84%) were published in journals and only four in conference proceedings. According to the publication years, as can be seen in Fig. 2(a), there is a peak in the number of publications related to “Industry 5.0” in 2022 (13 publications). As for the name of the publisher, it was found that, among the 26 included publications, the highest number was published by MDPI (7 publications), closely followed by Elsevier (6 publications) and Springer (5 publications). Figure 2(b) depicts the number of publications by publishers.

274

J. Crnobrnja et al.

Fig. 2. Basic Data Analysis including Journal and Conference Publications

According to the journal titles, it was found that, among the 22 included journal articles, three were published in the Sustainability journal, two in the Journal of Manufacturing Systems, and all others in different 17 journals (see Fig. 2c).

Digital Transformation Towards Industry 5.0

275

The four conference papers were published in the International Conference on Optimization, Learning Algorithms and Applications (October-2022-Portugal), the International Conference on Quality Engineering and Management (July-2022-Portugal), the International Conference on Computer Communication and Informatics (January2021-India), the International Scientific Conference on Digital Transformation on Manufacturing, Infrastructure and Service (October-2018-Russia). 3.3 The Analysis of Publications Keywords The identification of relevant keywords serves as a crucial indicator of the underlying concepts and themes addressed within a given publication. To gain insight into the research focuses, methods, domains, and Industry 5.0-related features, a comprehensive analysis of explicitly mentioned keywords within the included publications was conducted. This preliminary examination provides a foundational understanding of the research landscape and helps to contextualize the subsequent analyses. The 154 collected keywords were first examined and grouped into corresponding clusters. For example, the keywords ‘Human-Centricity’, ‘Human-Cyber-Physical Systems’, and ‘Human-Machine Coexistence’ were grouped into the “Human-Cyber-Physical Systems” cluster. Clusters were identified by grouping keywords with closely related meanings. The Top 5 clusters are “Industry 5.0” (19 keywords), “Digital Transformation” (13 keywords), “Human-CyberPhysical Systems” (10 keywords), “Industry 4.0” (8 keywords), and “Sustainability” (8 keywords), which cover 37.6% of the total keywords. Figure 3 portrays these keywords in a word cloud that was generated using FreeWordCloudGenerator.com to present these and review the most frequently used ones.

Fig. 3. World Cloud of Keywords used in Industry 5.0 Publications

276

J. Crnobrnja et al.

3.4 Data Analysis with Specific Purposes: Answering Research Questions To find the answer to RQ1: ‘Who is working on Industry 5.0, where, and when?’ – a summary of the authors involved in the “Industry 5.0” research field in recent years is presented. This investigation discovered that Russia, Brazil, and the USA are the top promoters of the Industry 5.0 paradigm. More specifically, the answer to this question is divided into the following three perspectives. From a contributor’s perspective, from the 26 publications included in this systematic review, there are 83 different authors (all of them authored at least one paper). From a geographical location perspective, 53% of the 26 publications compromising this systematic review involved institutions from the Europe region, followed by Asian institutions with 26.5% of the total, and with some lag the regions of Americas with 19.3% and Oceania with 1.2%, respectively. In particular, as can be seen in Fig. 4, the top country in the Europe region is Russia with 13 authors, followed by Austria (6) and Italy (6). Despite this, it was discovered that, as the world manufacturing centers, China and India have also shown a great interest in Industry 5.0 domain. By contrast, the countries from the Americas and Oceania regions, with exceptions made for the United States, Brazil, and Australia, are just currently starting to draw attention to the Industry 5.0 paradigm.

Fig. 4. Geographical Distribution of Publications

From the year of publication perspective, even though the announcement of the “Industry 5.0” concept is recent, it began to attract attention two years ago. As can be seen in Fig. 2(a), there is a dramatic increase in the number of Industry 5.0 publications in 2022 (13 papers). It is expected that the number of publications will increase in the next five years.

Digital Transformation Towards Industry 5.0

277

For answering RQ2: ‘What approaches and methods are used in Industry 5.0 research?’ – a qualitative analysis was performed. Among the 26 publications included in this investigation, 13 publications were considered “literature reviews”, and from these 12 used a “Systematic Literature Review” (SLR) method. The dominant method for SLR was “PRISMA”. Four publications used the PRISMA methodology. For the “theoretical research works”, qualitative descriptive research was used (7 publications). Only six publications conducted empirical research. The dominant method was a “case study” (3 publications), and a “questionnaire” combined with the survey as a data collection method (3 publications). The data analysis for answering RQ3: ‘What are the main research directions for Industry 5.0?’ – is mostly based on results from the SLR and practical publications. Based on the scope of the publication as well as the main findings following research directions were proposed from the literature review papers: (i) Human-centricity, Sustainability, and Resilience of Industry 5.0 (12 publications), (ii) Maturity Models for Industry 5.0 (3 publications), (iii) Human Skills for Industry 5.0 (3 publications), (iv) Models for Digital Transformation Technology Adoption (2 publications), and (v) Quadruple and Quintuple Helix Innovation Systems in Industry 5.0 (2 publications).

4 Implications – Towards a Research Agenda for Industry 5.0 Industry 5.0 represents a new paradigm in manufacturing that is emerging as the latest trend in industrial transformation. This paradigm shift presents numerous opportunities and challenges for researchers, who must explore the implications of Industry 5.0 to develop new technologies and techniques that enable the integration of advanced technologies with human skills and abilities. By doing so, researchers can help create a new era of manufacturing featuring human-centricity, resilience, and sustainability characteristics. To identify the knowledge gaps and propose future research from a context and research effort perspectives in the crossroad of “Digital Transformation” and “Industry 5.0”, a qualitative and quantitative analysis was conducted and presented in Sect. 3. The results of these analyses suggest that five key areas warrant further research: (i) Human-centricity, Sustainability, and Resilience of Industry 5.0, (ii) Maturity Models for Industry 5.0, (iii) Human Skills for Industry 5.0, (iv) Models for Digital Transformation Technology Adoption, and (v) Quadruple and Quintuple Helix Innovation Systems in Industry 5.0, which will be discussed separately in the continuation of this paper. 4.1 Human-Centricity, Sustainability, and Resilience of Industry 5.0 Out of all the publications examined in this systematic literature review, the authors seem to agree that Industry 5.0 represents a paradigm shift from a technology-centred approach to a human-centred approach in the design of manufacturing systems. Rather than solely relying on advanced automation technologies to drive manufacturing processes, Industry 5.0 recognizes the importance of integrating human skills and experience with modern technologies to create smarter and more resilient manufacturing systems. As a result, new research efforts will be focused on studies that promote workforce well-being and

278

J. Crnobrnja et al.

engaging and rewarding work and work environments that prioritize workers’ physical and mental health, autonomy, privacy, and dignity [3, 12]. On the other hand, achieving sustainability in manufacturing requires the active involvement and co-responsibility of multiple stakeholders, including manufacturers, suppliers, customers, and regulatory bodies [1]. In addition, the current global landscape is characterized by frequent disruptive events that can significantly impact the operations of manufacturing systems, highlighting the importance of building resilience capabilities as a key facet of sustainability. To achieve this, it is essential to further explore the synergies between human-centricity, sustainability, and resilience in the manufacturing arena by leveraging the power of multi-stakeholders collaboration working together to develop sustainable solutions that are resilient to disruptive conditions [1, 13]. To do this, research scientists need to develop a common Industry 5.0 framework that combines all these three factors. 4.2 Maturity Models for Industry 5.0 To answer the RQ3 about the main research directions in Industry 5.0, it was identified the importance of the readiness of organizations to adopt this new industrial transformation paradigm through specialized Maturity Models (MMs) for this Fifth Industrial Revolution. Developing and refining these models is crucial to enabling a smooth transition to Industry 5.0. MMs can be a valuable tool for shaping this transition, as they provide a framework for assessing an organization’s current capabilities and identifying areas for improvement. However, existing MMs for Industry 4.0 may not fully address the specific requirements of Industry 5.0, which places a greater emphasis on “humancentricity” and the integration of disruptive technologies supporting “sustainability” and “resilience” [14]. Therefore, further research is needed to develop MMs that are specifically tailored to the demands of Industry 5.0. 4.3 Human Skills for Industry 5.0 Based on the scope of the reviewed publications, Industry 5.0 will require closer cooperation between industry and academia. Hence, further research must identify the key skills that are necessary for current and future workers in Industry 5.0 and develop the proper training and educational programs to ensure that the workforce possesses these in the present and the future [9]. As Industry 5.0 continues to evolve, education systems must transform to meet their demands, giving rise to “Education 5.0” and therefore a novel era of education (and training) [14]. According to the SLR results, researchers have already introduced the concept of “Education 5.0” as a response to the emergence of Industry 5.0, creating a new avenue for exploration. To prepare workers for the demands of Industry 5.0, Education 5.0 must be purposefully designed to equip them with the required competencies (i.e. knowledge, skills, and values) [15]. This requires a departure from conventional lecture-based instruction in favor of experiential and collaborative learning approaches. Additionally, education systems must be agile and adaptable to accommodate the evolving needs of industry, facilitating lifelong learning and ongoing (digital) skills development for workers [16, 17].

Digital Transformation Towards Industry 5.0

279

4.4 Models for Digital Transformation Technology Adoption Models for Digital Transformation Technology Adoption are crucial for enabling the successful integration of advanced technologies in Industry 5.0 [18, 19]. Through this SLR, “sustainable business models” and “agility innovation models” were identified as relevant models used in theory and practice. Overall, both types of models are becoming increasingly important in modern manufacturing systems. By adopting these types of models, companies can reduce their environmental impact while also improving efficiency and meeting their customers’ demands for more sustainable and personalized products [20, 21]. In future research, it must be considered the factors that influence the adoption of these technologies and develop new models to guide their implementation with the triple bottom line in mind. 4.5 Quadruple and Quintuple Helix Innovation Systems in Industry 5.0 Lastly, the Quadruple and Quintuple Helix Innovation Systems in Industry 5.0 present an area of opportunity that requires further research and practical application since the research works identified in this SLR have a “theoretical” nature. The Quadruple Helix model emphasizes the involvement of four key stakeholders: (i) academia, (ii) industry, (iii) government, and (iv) civil society [22]. In this multistakeholder model, each stakeholder contributes with its unique perspective, expertise, and resources to a collaborative innovation process. On the other hand, the Quintuple Helix model expands on the Quadruple Helix model by adding a fifth stakeholder: (v) the public. This fifth stakeholder is seen as fundamental in driving innovation through “participatory engagement” and “co-creation of new knowledge” as well as through the promotion of “social innovation” that addresses societal challenges [22, 23]. Overall, the Quadruple and Quintuple Helix Innovation Systems represent powerful frameworks for driving innovation and progress in the manufacturing industry, particularly in the context of Industry 5.0. By fostering co-innovation and value co-creation among a diverse group of stakeholders, these models can help unlock the full potential of emerging technologies and drive sustainable growth and development.

5 Conclusions As the “Industry 4.0” and “Society 5.0” paradigms continue to gain momentum, Industry 5.0 emerges as a new industrial transformation paradigm. This paper provides a comprehensive review of the evolution of “Digital Transformation towards Industry 5.0” by defining its characteristics and its potential impact on the design, engineering, and management of manufacturing systems and ecosystems. An overview of Digital Transformation and Industry 5.0 paradigms is firstly captured in this systematic literature review by enumerating the list of prominent journals and international conferences for publishing Industry 5.0-related content, and secondly by illustrating “Digital Transformation” and “Industry 5.0” terms significance based on keywords classification. Then, special attention was given to the following elements of analysis: (a) publishers and publications in the last five years, (b) the geographical

280

J. Crnobrnja et al.

locations of authors that are involved in Industry 5.0 research in the last five years, and (c) the identification of existing research efforts and also the research areas that are being neglected recently in the current literature. After conducting a systematic review of the literature, a research agenda was developed to summarize the implications of this review for future research related to Industry 5.0 initiatives. The results of this analysis suggest that five key areas warrant further research: (i) Human-Centricity, Sustainability, and Resilience of Industry 5.0, (ii) Maturity Models for Industry 5.0, (iii) Human Skills for Industry 5.0, (iv) Models for Digital Transformation Technology Adoption, and (v) Quadruple and Quintuple Helix Innovation Systems in Industry 5.0. In light of the findings presented in this study, it is important to acknowledge several limitations. Firstly, the scope of the review was confined to a selection of publications obtained from prominent abstract and citation databases, namely Scopus and Web of Science. Although efforts were made to include a diverse range of publications, the analysis focused on a relatively small sample size. To enhance future investigations, it is recommended that research endeavors shift towards a broader examination of aggregated dimensions rather than exclusively scrutinizing individual or interconnected dimensions. Secondly, it is noteworthy that the search criteria employed in this review encompassed keywords specifically related to “Digital Transformation and “Industry 5.0” within the past five years. Consequently, research about this domain utilizing distinct keywords such as human-centricity, sustainability, and resilience, as separate entities, was inadvertently excluded. To ensure a more comprehensive understanding, future research should not be limited to a specific five-year timeframe and should encompass all relevant keywords associated with the Industry 5.0 domain. By addressing these limitations and broadening the scope of inquiry, researchers can advance our insights in the field of Digital Transformation and Industry 5.0, thus facilitating the development of more robust and holistic frameworks for analysis and application.

References 1. Camarinha-Matos L.M., Rocha, A.D., Graça P.: Collaborative approaches in sustainable and resilient manufacturing. J. Intell. Manuf. (2022). https://doi.org/10.1007/s10845-02202060-6 2. Leng, J., Sha, W., Wang, B., et al.: Industry 5.0: prospect and retrospect. J. Manuf. Syst, 65, 279–295 (2022) 3. Xu, X., et al.: Industry 4.0 and Industry 5.0: inception, conception and reception. J. Intell. Manuf. 61, 530–535 (2021) 4. Salunkhe, O., Fast-Berglund, Å.: Industry 4.0 enabling technologies for increasing operational flexibility in final assembly. Int. J. Ind. Eng. Manag. 13(1), 38–48 (2022) 5. Nahavandi, S.: Industry 5.0 a human-centric solution. Sustainability 11(16), 4371 (2019) 6. Maddikunta, P.K.R., et al.: Industry 5.0: a survey on enabling technologies and potential applications. J. Ind. Inf. Integrat. 26, 100257 (2022) 7. Lalic, D.C., et al.: Digital transformation in the engineering research area: scientific performance and strategic themes. IFIP, AICT 664, 196–204 (2022) 8. Spasojevi´c, I., et al.: Research trends and topics in IJIEM from 2010 to 2020: a statistical history. Int. J. Ind. Eng. Manag. 12(4), 228–242 (2021) 9. Kolade, O., Owoseni, A.: Employment 5.0: the work of the future and the future of work. Technol. Soc. 71, 102086 (2022)

Digital Transformation Towards Industry 5.0

281

10. Lenart-Gansiniec, R.: The dilemmas of systematic literature review: the context of crowdsourcing in science. Int. J. Contemp. Manag. 58(1), 11–21 (2022) 11. Page, M.J., et al.: The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. Int. J. Surg. 88, 105906 (2021) 12. Alves, J., Lima, T.M., Gaspar, P.D.: Is industry 5.0 a human-centred approach? a systematic review. Processes 11(1), 193 (2023) 13. Raja Santhi, A., Muthuswamy, P.: Industry 5.0 or industry 4.0? introduction to industry 4.0 and a peek into the prospective industry 5.0 technologies. Int. J. Interact. Des. Manuf. 17, 947–979 (2023) 14. Hein-Pensel, F., et al.: Maturity assessment for industry 5.0: a review of existing maturity models. J. Manuf. Syst. 66, 200–210 (2023) 15. Gürdür Broo, D., Kaynak, O., Sait, S.M.: Rethinking engineering education in the age of industry 5.0. J. Ind. Inf. Integrat. 25, 10031 (2022) 16. da Silva, L.B.P., et al.: Evolution of soft skills for engineering education in the digital era. Commun. Comput. Inf. Sci. 1754, 640–653 (2022) 17. Lalic, D.C., et al.: How project management approach impact project success? from traditional to agile. Int. J. Managing Proj. Bus. 15(3), 494–521 (2022) 18. Rakic, S., Pero, M., Sianesi, A., Marjanovic, U.: Digital servitization and firm performance: technology intensity approach. Eng. Econ. 33(4), 398–413 (2022) 19. Rakic, S., et al.: Transformation of manufacturing firms: towards digital servitization. IFIP, AICT 631, 153–161 (2021) 20. Gracanin, D., et al.: The Impact of Lean Improvements on Cost-time Profile. Procedia Manuf. 38, 316–323 (2019) 21. Sofic, A., et al.: Smart and resilient transformation of manufacturing firms. Processes 10, 1–13 (2022) 22. Carayannis, E.G., Campbell, D.F.J.: Towards an emerging unified theory of helix architectures (EUTOHA): focus on the quintuple innovation helix framework as the integrative device. Triple Helix 6, 1–11 (2022) 23. Carayannis, E.G., Morawska-Jancelewicz, J.: The futures of Europe: society 5.0 and industry 5.0 as driving forces of future universities. J. Knowl. Econ. 13, 3445–3471 (2022)

Industry 5.0 and Manufacturing Paradigms: Craft Manufacturing - A Case from Boat Manufacturing Bjørnar Henriksen1(B) and Maria Kollberg Thomassen2 1 SINTEF Digital, Trondheim, Norway

[email protected] 2 SINTEF Digital, Oslo, Norway [email protected]

Abstract. Industry 5.0 is emerging as a novel approach in manufacturing, with focus on sustainability, resilience, and human centricity. This study investigates the complex and comprehensive Industry 5.0 concept, its relevance for the craft manufacturing paradigm and the role of technology in the industry transition towards Industry 5.0. Findings are based on empirical data from a research and development project with Norwegian leisure boat manufacturers, and results of previous research projects within the boat building industry. The boat building industry is a traditional hand-craft industry with high degree of manual labour. Findings show that this industry have several suitable conditions for developing strong capabilities related to sustainability, resilience, and human centricity. The investigated case companies, characterized as craft manufacturers, seem to have limited knowledge of both Industry 4.0 and Industry 5.0. Despite that previous research projects have included technological elements, advanced technologies are still scarce in production processes. Keywords: Industry 5.0 · Technology Maturity · Craft Manufacturing

1 Introduction Manufacturing companies must deal with increasingly complex business environments and challenges that require novel approaches. European policymakers and organizations have therefore introduced Industry 5.0 on the agenda. This is driven by severe implications of unforeseen events and crises such as the pandemic, war in Europe, disruptions in supply chains, the climate and environmental crisis, and the urge to put people at the centre. Industry 5.0, which provides a future vision for the European industry, is gaining increased attention. Still, its impacts and practical implications on industries, companies, people and society are to be defined and exploited [1]. Since companies are still engaged in Industry 4.0, or in earlier industrial revolutions, it can be questioned whether Industry 5.0 should be positioned on a timeline of “Industrial Revolutions” based on innovations from technology, Industry 1.0 (mechanization from © IFIP International Federation for Information Processing 2023 Published by Springer Nature Switzerland AG 2023 E. Alfnes et al. (Eds.): APMS 2023, IFIP AICT 689, pp. 282–296, 2023. https://doi.org/10.1007/978-3-031-43662-8_21

Industry 5.0 and Manufacturing Paradigms

283

water- and steam power) and to Industry 4.0 (Cyber physical systems from internet of things, artificial intelligence, etc.) [2]. Instead, innovations that push manufacturing towards Industry 5.0 are not solely limited to technological ones, but rather within socialor human–machine system-sphere [3]. Furthermore, according to Kraaijenbrink [1], the idea of Industry 5.0 is not limited to industry only, but applies to a wide range of sectors and organizations. This means that its applicability is significantly wider than Industry 4.0. Manufacturing paradigms may be related to and even overlap with industrial revolutions. The development of paradigms is associated with changing customer needs [4], but also changing capabilities such as enabling technology. However, manufacturing paradigms are not necessarily a function of time or technology. Thus, “low tech” manufacturing could still meet market requirements and imply relevant approaches for companies. With a starting point in the view that Industry 5.0 is a complement to the Industry 4.0 technological paradigm, this transition is technology dependent and emphasizing the importance of technology for the Industry 5.0 transition. It can thus be assumed that significant technological investments are required if companies are to make the transition towards Industry 5.0. However, the Industry 4.0 transition has been mainly aligned with the optimisation of business models and economic thinking, supporting the current digital economy, based on a winner-takes-all model that creates technological monopoly and giant wealth inequality [5]. This may question to what extent existing technologies that were once implemented with purposes of increasing productivity and economic benefits only, may be utilized to support the new transition. The Industry 5.0 vision suggests that the traditional financial objectives of technology investments, such as productivity and profitability, should be replaced with goals related to resilience, sustainability, and human-centricity. Even though Industry 4.0 has been well known in many industries for several years, its adoption rate still is slow. Moreover, manufacturing companies may be close to the Industry 5.0 vision, without major investments in new digital technologies. This means that companies may have developed strong sustainable, resilient, and human-centric abilities independent of heavy technological investments. Thus, the starting point for this research is that the Industry 5.0 vision does not necessarily have to rely on enabling technologies and that technology advancement should not be a goal. There is a need for a more nuanced picture of the role of technology for the Industry 5.0 transition. This study investigates sustainability, resilience, and human centricity to bring further understanding the relevance of the Industry 5.0 vision for manufacturing paradigms with a technology maturity level different from we find in Industry 4.0. It seeks to bring further understanding of the complex and comprehensive Industry 5.0 concept, and to the specific role of technology in the industry transition towards Industry 5.0. Findings are based on empirical data from case companies in the boatbuilding industry, which is a traditional hand-craft industry with high degree of manual labour.

284

B. Henriksen and M. K. Thomassen

2 Theoretical Perspectives 2.1 Industry 5.0 – A Vision for the European Industry According to the European Union (EU) [5], the Industry 5.0 concept provides a vison of industry that look beyond efficiency and productivity as the sole goals, and reinforces the role and the contribution of industry to society and a particular focus on the wellbeing of the workers. The use of modern technologies should provide prosperity beyond jobs and growth while respecting the production limits of the planet. The concept further promotes the transition towards a sustainable, human centric and resilient European industry. The Industry 4.0 technological paradigm is mainly centred around cyber-physical objects, promising enhanced efficiency through digital connectivity and artificial intelligence [5]. In order to Industry 4.0 often refers to the fourth industrial revolution, with focus on digital transformation in the manufacturing industry for faster delivery times, more efficient and automated processes, higher quality and customised products [6]. However, this paradigm does not fit in the context of climate crisis, planetary emergency, and deep social tensions. Industry 5.0 complements the existing Industry 4.0 paradigm by highlighting research and innovation as drivers for a transition to a sustainable, human-centric and resilient European industry [7]. It moves focus from shareholder to stakeholder value, with benefits for all concerned, and attempts to capture the value of new technologies, providing prosperity beyond jobs and growth, while respecting planetary boundaries, and placing the wellbeing of the industry worker at the centre of the production process [7]. Industry 5.0 moves the focus from shareholder to stakeholder value, with benefits for all concerned. Industry 5.0 is about interaction between people and technology, with the employee as a positive driver for development of company, but where the objectives go beyond company profit, heading for sustainability and positive contribution to society. According to the European Union [7], Industry 5.0 is defined as follows; “provides a vision of industry that aims beyond efficiency and productivity as the sole goals, and reinforces the role and the contribution of industry to society.” and “It places the wellbeing of the worker at the centre of the production process and uses new technologies to provide prosperity beyond jobs and growth while respecting the production limits of the planet.” It complements the Industry 4.0 approach by “specifically putting research and innovation at the service of the transition to a sustainable, human-centric and resilient European industry.” Industry 5.0 has three key pillars: human-centric, resilient, and sustainable, but with a comprehensive approach these pillars are seen in context and building culture, principles, technology, and solutions, so that these elements are intertwined and create synergies. • Human centricity: Places essential human needs and interests at the centre of the manufacturing and is a shift from seeing people as means (e.g., as in human resources) to seeing people as ends. Or, in other words, a shift in perspective from people serving organizations, to organizations serving people.

Industry 5.0 and Manufacturing Paradigms

285

• Sustainability: The company must face the challenges (and opportunities) outside the factory. Aiming for positive societal development beyond purely (short-term) business targets (profit). Rather than only reducing a company’s negative impact, sustainable companies focus on increasing their positive impact. • Resilience: Manufacturing systems should have the ability to withstand demanding situations and to accommodate disruptions without significant additional costs. 2.2 Enabling Technologies There are a wide range of Industry 4.0 technologies enabling this digital transformation, including for instance additive manufacturing, augmented and virtual reality, automation and industrial collaborative robotics, big data analytics, blockchain, cloud data and computing, cybersecurity, cyber-physical systems, internet of things, artificial intelligence, and simulation and modelling (see for instance literature reviews by [6] and [8]). These technologies can also have a wide range of applications in various processes in manufacturing companies, including product life cycle management, supply chain management and production and operations management [6]. Also, similar enabling technologies are recognized within the Industry 5.0 concept [9]. However the abilities of technologies as co-drivers in producing resilience and sustainability on a wider scale are often uncertain and context-dependent [10]. Today’s technologies are primarily designed and used with a focus on maximizing organizational/operational performance instead of human well-being and societal growth and prosperity, prioritizing full automation of tasks per se, that sometimes creates trust issues rather than empowering the human in the task [11]. However, researchers have started to investigate how technologies such as for example digital twins and artificial intelligence may help to improve worker situations and enhance human centricity [12, 13]. 2.3 Manufacturing Paradigms As the focus and perspectives in manufacturing have developed over time, the changes have often been described as shifts in paradigms, “a set of beliefs that guides action” [14]. These paradigms are not absolute in terms of complete frameworks for how to conduct business or organize manufacturing. What they represent are more coherent sets of principles and methods that inspire companies. Strategies will often have elements from different paradigms even if it is claimed they have adopted only one. Defining paradigms is not an exact science; rather, to a considerable extent it is a question of choosing a set of criteria supporting the purpose of categorization. Paci et al. [15] have presented criteria in what they call a “Production Paradigms Ontology” (PPO) which focuses on knowledge- and innovation. PPO is based on the NEST (Nature, Economy, Society, Technology) context defined by Jovane et al. [16]. Jovane et al. use these criteria to identify and describe five manufacturing paradigms. While industrial revolutions as described above mainly focus on innovations, technology, and other enablers for radical shift in manufacturing, manufacturing paradigms are more about what you want to obtain, purpose and focus areas in manufacturing, described as “Society needs.” Although it is tempting to challenge this list of paradigms from 2003, referring to many changes and innovations, we consider that these paradigms are basically those we can identify in the broad context [14]:

286

B. Henriksen and M. K. Thomassen

• Craft Manufacturing—which means to make exactly the product that the customer asks for, usually one product at a time in a “pull-type business model”: sell (get paid)—design-make-assemble. The processes have a low level of automation, but use skilled and flexible workers • Mass Manufacturing—means to manufacture high quantities of identical products, selling them to customers and markets that will absorb what is manufactured. High volume means low costs and cheaper products, thereby increasing the market, which implies a “push-type business model”. The moving assembly line is an enabler for this paradigm, requiring standardized processes and specialized workforces • Flexible manufacturing—has been the answer to the increased complexity and uncertainty in the business environment. In the 1970s overproduction and the demand for more diversified products resulted in decreased lot size and the requirement for shorter time-to-market. The manufacturing still followed principles from mass manufacturing but more module-based to meet the demand for variation. • Mass customization and personalization—is a society driven paradigm due to customers asking for greater variety in products, and because of globalization creating a huge excess of production capacity. This situation has put customers in power and the manufacturer aims to manufacture a variety of almost customized products at mass production prices. • Sustainable manufacturing—is based on societies’ needs for ‘clean’ products and product-life-cycle management related to clean products. Nano-, bio- and material technology are regarded as enablers for sustainable manufacturing. All the above listed manufacturing paradigms are more-or-less relevant today. It all depends on products- and market requirements and the context within the companies operates. And there is not a clear cut between them so we will see a myriad of combinations for example related to “sustainable manufacturing.” 2.4 Industry 5.0 Maturity A company seeking to realize the Industry 5.0 vision, may consider its capabilities regarding human-centric, resiliency, and sustainability. The development of Industry 5.0 capabilities in industry requires a thorough understanding of the concept as well as relevant guidelines and showcases for practitioners [17]. However, since Industry 5.0 is a fairly new concept, models for measuring and assessing Industry 5.0 maturity or advancement are still lacking [18]. In contrast, a wide range of maturity assessments models and tools have been developed recently for measuring advancements related to Industry 4.0 (see literature reviews by [19–21] and [22]). Even though industry 4.0 maturity assessment models may serve as an important starting point for assessing Industry 5.0 maturity, these models rely heavily on technology and are to a considerable extent focused on measuring the digital readiness in companies. Yet, digital readiness and Industry 4.0 maturity models may be relevant for companies to support digital transformation and the transformation process towards Industry 5.0 [18]. Moreover, several digital maturity models include Industry 5.0 aspects related to human-centricity, sustainability and resilience [18]. This means that the Industry 5.0 maturity level of a company can be assessed based on measures related to the three

Industry 5.0 and Manufacturing Paradigms

287

basic elements of Industry 5.0. However, if considering the interdependencies between ongoing shifts, of both Industry 5.0 transformation and digital transformation, it may be relevant to investigate opportunities for combining human-centric, resiliency and sustainability maturity measures with digital measures that specifically reflect the digital readiness in accordance with these three elements. An assumption for this research is that despite that Industry 4.0 is considered technology-driven and digital enablers are recognized within Industry 5.0, the technology-dependency of the Industry 5.0 transformation can be questioned.

3 Methodological Considerations This research is applied research, conducted in collaborative manner with three case companies. It is based on a mixed methods approach, combining empirical data collected by a questionnaire and literature studies, as well as information from previous research projects involving companies in the Norwegian boatbuilding industry. Most research conducted in various previous projects in this industry has been based on a combined approach involving action research and case research strategies. This has implied an extensive understanding of this industry and in-depth insights to several of the companies, among involved researchers. Empirical data on enabling technologies are collected through a questionnaire filled out by representatives of three companies. The companies are selected because they participate in a research project, the TEL project. The questionnaire is developed by the authors based on a brief review of relevant literature on digital enabling technologies and Industry 5.0. It includes thirteen questions. For twelve of the questions, ratings are conducted by a 5-point Likert scale ranging from “Strongly agree” to “Strongly disagree”, in addition to the alternative “Don’t know”. One of the questions, related to prioritization of the key qualities and expertise, involves six suggested response alternatives to choose from. Respondents could also add open comments in a dedicated text box after each question. SINTEF researchers evaluated the questionnaire before it was distributed to the companies. It was sent as an attached word document by e-mail to the main contact person of each case company. Answers were returned from all the companies. To further investigate the relevance of enabling technologies and Industry 5.0 in the Norwegian boat building industry, a mapping of previous relevant R&D projects is conducted. The mapping is based on information collected from the final reports of the research projects. Project research results and innovations are mapped in relation to digitalization and Industry 5.0 key dimensions. The projects were selected since researchers and case companies were previously involved in these projects. 3.1 The TEL Research Project and Case Companies This research is conducted as part of a four-year research project, Transformation to electric propulsion (TEL), which is partly funded by the Norwegian Research council. The project focus on electrification of boats, but also has on production efficiency. Environmental and sustainability aspects for leisure and professional boats are recognized to have major potential for improvements. The main objective of the TEL-project is

288

B. Henriksen and M. K. Thomassen

to realize this potential. By evaluating and testing user demands, industrial fabrication standards, available technology and several other aspects from a commercial standpoint, development of fabrication, new flexible hulls, propulsion, and electrification/hybrid technologies will be conducted. The use of electrification is presently the most promising technology to achieve environmental and sustainability potentials. The automotive industry has taken immense electrical innovation leaps over a brief period, and now the marine industry is expected to achieve the same. There is however a major difference in requirements, demands and solutions. For example, the laws and regulations for the implementation of electric propulsion in the maritime industry are not available and clear. Therefore, the project also will collaborate closely with the authorities on this topic. To provide the knowledge needed for a new industry standard, there is a push from the project to present products that may be the start of a new generation of boats. New drivelines, a lower speed range and technology advancement will provide new hull designs. The project also addresses the relation between range, speed and use-profiles will affect the user demands and how this influences commercial alternatives. The TEL project includes three boatbuilders Ibiza Boats AS, Viknes Båt og Service AS and Nor-Dan Composites AS. In addition, the project includes the technology providers Farco AS and SeaDrive AS, and the R&D-partners SINTEF Digital and Inventas Kristiansand AS. The boat companies produce boats between 20 to 37 feet to customers in different market segments. One of them focuses on professional boats, primarily coastal fishing, while the others focus on the private leisure boat segments. The companies are SMEs and are all companies with a history that goes back several decades. Several TEL-project partners have participated in larger R&D projects, both national with co-funding from the Research Council of Norway and the European Union (EU) research funding programs. The most relevant projects are shown in Table 1. 3.2 Craft Manufacturing The leisure boat manufacturers and production of small workboats are competing on global markets that are very fluctuating, vulnerable for the financial situation and priorities among households. Traditionally, these companies have been small or medium sized (SME) but the last two decades there has been a tendency towards larger more industrialized operations with larger production volumes and reduced costs. This has to some extent rugged the industrial structure, increased competition and challenged traditional craft manufacturing principles [23]. The Norwegian leisure boat industry has long traditions and a broad culture for quality products. The design, production and after sales activities have been characterized by craft manufacturing and to a large extent based on tacit knowledge in development, improvement and design [24]. However, several research and development projects have to some extent changed that bringing in elements from new technologies, digitalization and modularization [25]. The craft manufacturing of leisure boats is characterized by engineer-to-order manufacturing and projects. Even though the design and production of boats is characterized by wide number of quite demanding requirements, processes are manual. However, a wide variety of digital tools and equipment are used in boats and in manufacturing processes in this industry.

Industry 5.0 and Manufacturing Paradigms

289

One of the main advantages in craft manufacturing is that the employees have a strong ownership to their products and put a lot of effort in finding good ad-hoc solutions to problems. For example, when a supplier has delivered a part with minor quality problems, the craftsmen can often fix the problems themselves instead of initiating administrative processes. The craftsman represents a capability for adjusting and practical solutions in line with customer expectations. This also tends to create a customer relationship that are appreciated and could lead to “loyal” customers. A general characteristic of quality and quality improvement in leisure boat companies is that their employees constantly work to optimize quality.

4 Empirical Findings 4.1 Industry 5.0 and Enabling Technologies in the Case Companies The TEL-project aims at developing transfer leisure boats (and small craft vessels) to more environmentally friendly propulsion solutions, which also includes hydro dynamics and material technology. Putting all these elements together has required application of new simulation methods and digital tools. As these electric and hybrid solutions are more expensive, an important part of the project has been focusing on cost reduction and novel approaches to production. The concepts Industry 4.0 and 5.0 have not been explicitly introduced to the project and its partners. So, when the companies were asked to answer a simple questionnaire, it was to get picture of their knowledge of the concepts. The aim was also to get an indication of to which extent they comply with industry 5.0 and the underlying elements and especially the technology. The knowledge of Industry 4.0 and Industry 5.0 is low when it comes to the concrete content, and what it means for a company at strategic as well as operational levels. However, when they were asked more specific on the basic elements of it the picture change: • Sustainability is a central element in two of the company’s overall strategies, plans and goals, where they also had a focus on enabling digital technology. The third company responded neutral to these questions • Resilience is an essential element in the three companies’ strategies and plans. Digital technology is an important part of it. • Human centricity in terms of focus on the workers, involvement in improvement and development are important in all three companies. The answers vary when it comes to which skills are considered most important, but “solution-oriented” gets the highest priority followed by “structured” and “independent” • The companies lack a clear strategy on digitalization, and they have different answers on the degree they use digital tools in production improvement and product development. The results of the questionnaire show that there are major variations between the answers from the companies. Even though the knowledge of Industry 4.0 and 5.0 is limited, and the “digital maturity” is low, it looks different when the questions are on the underlying elements of Industry 5.0. Sustainability, resilience and in particular human centricity are by no means neglected, where enabling technologies are used by the companies.

290

B. Henriksen and M. K. Thomassen

4.2 Industry 5.0 and Enabling Technologies in Previous R&D Projects Table 1. Overview of results of previous R&D projects Project title and period

Description of project results

Industrialised Small Scale Boat Production, ISB (2008–2012)

Modularization, manual adjustments are minimized; Production-, product- and market-knowledge for design and production strategies; Supply chain integration for customized production; Production that addresses environmental issues and making the leisure boat industry an attractive workplace

Boat Management, BOMA (2011–2014)

Software platform for data handling, analysis, etc. tied to production / documentation /DSS mm. Intelligent Universal Marine Gateway installed that collects, analyses, transmits operating data from boat to manufacturer, service partners, and the boat owner

Planning and Management of Small Series, PLAN (2011–2014)

Better utilization of resources of partners from coordinated planning and management; Differential control based on partners’ mass custom profiles; Automating logistics and technical information in the supply chain through better application of existing technology

EcoBoat MOL, EBM (2011–2015)

Technical solutions in prototypes and test solutions for data acquisition, analyses, and transmission of data/results from real-life running of individual boats; Sensors adapted to the need of the boat producer; Improvement related to actual systems data entry and use of data for product development

Marine Platform, MAP (2012–2016)

Development of flexible and standardized structural platforms as the basis for well-defined interfaces for a boat’s various modules; Environmentally friendly material solutions regarding weight, reuse, and recycling

Innovative Connected Vessels, LINCOLN (2013–2019)

Methodologies and tools for the development of new vessels concepts through dynamic simulation model testing; IT customized tools to enable the acquisition and usage of field data, from an IoT platform; Introduced High Performance Computing Simulation to verify the resulting concept design of boats (continued)

Industry 5.0 and Manufacturing Paradigms

291

Table 1. (continued) Project title and period

Description of project results

Robust Industrial Transformation, RIT (2018–2021)

Design dashboard that can handle the display and analysis of real test data, and critical information in product development based on new hardware for data capture from real operation of the products, and innovative algorithms that allow you to, for example, transform a measured force impact; Modularization

Radically Improve Costly Development, RADDIS (2018–2021)

Visualization technologies, i.e., virtual product and factory twins; Regulations opening new opportunities for quality development and documentation; Avatar with checklists and module based BOM solutions; Updated product and process information throughout the boat’s life cycle

Table 2. Overview of linkages between results of R&D projects and Industry 5.0 elements Project name

Industry 5.0 dimensions and elements Sustainability STY

PCI

ISB

x

x

BOMA

x

x

PLAN

Resilicene PDI x

DIG

STY

PCI

x

x

Human centricity PDI

DIG

x

x

x

x

x

x

x

x

x

x

x

MAP

x

x

x

x

x

LINCOLN

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

RADDIS

x

PCI

x

x

x

EBM

RIT

STY

x

DIG x

x x x x x x

x

STY = Strategy, PCI = Process improvements, PDI = Product improvements, DIG = Digitalization.

Several R&D projects have been conducted in the Norwegian boat building industry since 2008 (see Table 1). These have aimed at increasing efficiency in production with linkages to product development, seeking to increase performance while keeping the advantages of the craft manufacturing tradition. The table below (Table 2) shows an overview of the results of the previous projects in view of Industry 5.0 dimensions and elements including digitalization.

292

B. Henriksen and M. K. Thomassen

5 Is There Industry 5.0 Without the 4.0 (Digital) Maturity? 5.1 Overall Perspective and Transition The transition of a business is the process of bringing a company into its next chapter [26]. Transition is about strategies to create new or better capabilities for major changes and improvements. A business or industry may progress in the right direction towards a major shift. However, such movements do not necessarily imply a transition. In general, leisure boat manufacturers seem to have moved forward in line with new demands and requirements related to societal changes. Our findings show that they have also conducted a number of R&D projects that have included elements from both Industry 4.0 (technology) and Industry 5.0. (i.e., sustainability). However, it seems like Industry 4.0 and Industry 5.0 are still unfamiliar concepts in this industry. This may indicate that leisure boat manufacturers lack explicit objectives and strategies for green and digital transition. Instead, changes have implied more of inherent moves aligned with general industry expectations rather than well defined actions that are part of a long-term strategic transition plan. Still, our findings suggest that several elements of resilience, sustainability and human centricity are highly relevant for this industry, although these have limited foundation in the Industry 5.0 vision. Regarding technological progress, findings indicate that the digital technology maturity level in the industry is low, despite several previous R&D projects on advanced technologies. 5.2 Industry 5.0 Dimensions Sustainability. Findings show that sustainability is critical for this industry. Sustainability has been linked to digitalisation. For example, data captured from boat operations may be used for analysis and simulation. Such solutions can provide more data and knowledge about a boat’s life cycle and areas of use, use conditions, enabling optimization of design solutions to reduce material thickness, weight, and hydrodynamics, which affect resource requirements also in relation to propulsion. New production methods and technology, for example the use of vacuum injection in casting, also contribute to less environmentally harmful raw materials. Another example from one of the case companies is a new digital milling machine, which is expected to have positive effects in terms of reduced waste, increased quality, and precision. In addition, leisure boat manufacturers work purposefully for recycling of plastic boats and explore possibilities for using recycled materials, for instance from plastic bottles. Resilience. The leisure boat market face major fluctuations and is exposed to changes in the global economy that affect demand, such as economic crisis, energy crisis, as well as events like geoeconomic confrontations or pandemics. These conditions increase the need for strategies and solutions for enhanced resilience. There has been a major restructuring of the leisure boat manufacturing industry in Norway and internationally during the last decades, which has resulted in fewer companies and more challenging market competition. Despite this, the companies have had an ability to survive and adapt to these major changes in competition and business conditions. Our findings also confirm that resilience has high priority in this industry.

Industry 5.0 and Manufacturing Paradigms

293

A specific feature of the leisure boat industry is that products and production processes are often modularized. This implies flexibility and adaptability to changes in market conditions and user requirements. Also, it enables scalability as modules can be outsourced to networks in periods of high demand or lack of own capacity, for instance. Moreover, computerized simulation has contributed to reduce costs and risks of product development. Digitalization of bills of materials, visual assistant systems, and digital twins are examples of technologies that have been explored in the industry with a positive impact on resilience. However, to enhance resilience in the industry, companies also need employees have competence capabilities that enable them to change, improvise and navigate within complex and turbulent business environments. Human Centricity. The craft-based production of leisure boats has a strong focus on humans, especially production workers and customers (end users). This means that there is often close dialogue and collaboration between production workers, other employees, and customers. This constitutes a solid foundation for developing a strong commitment to products and the entire company among workers. Workers have high responsibility and are expected to find solutions to problems occurring in production and assembly and to product related challenges. The companies also focus on sharing tacit knowledge and involving employees in improvement and development work. There has been an increased focus on improved working conditions the last decades with better health, safety, and environment (HSE) procedures and equipment. Also, manual casting processes of hand-laid fiberglass mats have been substituted in boatbuilding by cleaner injection moulding and vacuum processes. Investments in a new milling machine may further improve the dusty and uncomfortable cutting and grinding process at the shop floor. Even though boatbuilding companies have appreciated practical skills, there has been an increased focus on support systems, such as digitized bills of materials and visualised descriptions and procedures. Human centricity is thus considered a major element for competitiveness. However, findings indicate that human centric aspects only have had little attention in previous R&D projects.

5.3 Industry 4.0 Technology Gap There is a large amount of literature on Industry 4.0 technologies. It typically includes descriptions of advanced cyber physical systems such as smart factory, smart machines, digital equipment, which often requires resource demanding digitization processes and/or a specific manufacturing context. Industry 4.0 technologies are more suitable for mass production and mass customization contexts, compared to craft manufacturing. This may imply that some sectors of the manufacturing industries with products, processes or scales that may be unsuitable for Industry 4.0 technologies, risk to be left behind in the digital transition. However, cognitive technologies are one exception of a type of technology that may be exploited in this industry. Findings show a low level of utilization of Industry 4.0 technologies in boat production. This may be explained by contextual factors. Companies are typically small and medium sized enterprises (SMEs) with limited ability to make large investments in digitization. Also, small production volumes with high degree of customization and manual processes imply limited advantages of automation and digitalization. At the same time,

294

B. Henriksen and M. K. Thomassen

findings show that companies have a higher digital technology level on the product side. The boats constitute advanced products with digital equipment and modern state of the art technologies, enabling data capture and analysis of condition and use. These technologies have the potential to improve processes and products, and boat builders also seem to focus more on digital support systems in product development, production and after sales. Industry 5.0 implies a different technology perspective, where technology could enable sustainability, resilience, and human centricity. In craft manufacturing, it seems like the Industry 5.0 dimensions may be enabled by other means. Even though the investigated companies do invest in modern technology, building capabilities that enable employees to be a driving force towards Industry 5.0 may be even more important. Requirements regarding working conditions, qualification, responsibilities, organisation and involving management, but also cognitive support systems and other digital tools, therefore need further attention.

6 Conclusion Craft manufacturing still represents an important paradigm in the manufacturing industry. The ability to adapt to customers and the flexibility through well-educated and motivated employees can provide decisive capabilities for competitiveness and survival also in high-cost countries such as Norway. The boat building companies involved in the TEL project demonstrates this ability. The low technological maturity in relation to Industry 4.0 in the investigated companies is to some extent adapted to the context of the companies. Even though the case companies lack profound knowledge of Industry 5.0 as a concept or vision, they still have strong focus on sustainability and resilience. The companies further show high focus on human centricity, which is a particularly important dimension in Industry 5.0. Hence, the companies have prerequisites for setting a course towards and working in line with the Industry 5.0 vision, especially when adopting relevant technology. Previous revolutions in manufacturing have been based on innovative technology: Industry 1.0, mechanization from steam power and weaving loom; Industry 2.0, mass production and assembly line from electrical energy; Industry 3.0, automation from computers and electronics; Industry 4.0, cyber physical systems from internet of things (IoT) and networks. With Industry 5.0, this technological focus of revolutions may be shifting, representing a uniquely different revolution based on other perspectives than technological that drive major changes in manufacturing. This research shows how companies in the boat building sector, representing craft manufacturing, have developed strong Industry 5.0 capabilities without adopting advanced digital technologies. These findings from the boat building sector add to the limited body of research on the Industry 5.0 concept. It contributes with further empirical insight into the role of technology and the relation between enabling technologies and the dimensions of Industry 5.0 in craft manufacturing. Furthermore, it contributes to research dealing with craft manufacturing, by a better understanding of the relation between digital technology and Industry 5.0 dimensions in this sector. The results of this research may be relevant for practitioners, especially managers in boatbuilding companies and other companies within craft manufacturing. The study

Industry 5.0 and Manufacturing Paradigms

295

may help them to better understand the concept of Industry 5.0 and its dimensions, and the role of technological advancement in craft manufacturing settings. Even though these results are based on limited empirical evidence from a specific industry, they may be relevant for other craft industries. Further research that involves a wider range of craft industries and contexts is however needed. Also, the questionnaire used in this research may be further developed to collect data from a larger sample of companies. The results of this study may constitute a starting point for the further development of a general framework for Industry 5.0 capabilities and digital technologies in craft manufacturing. Also, findings indicate that specific contextual aspects influence the role of digital technologies for building Industry 5.0 capabilities. Further research is therefore suggested that develop further understanding of how digital technology can help to enhance Industry 5.0 capabilities in different empirical settings of craft manufacturing. Acknowledgements. The Research Council of Norway partly fund this research. We are grateful for the valuable contributions from the reviewers that have helped to improve the paper.

References 1. Kraaijenbrink, J.: What is industry 5.0 and how it will radically change your business strategy (2022). https://www.forbes.com/sites/jeroenkraaijenbrink/2022/05/24/what-is-industry50-and-how-it-will-radically-change-your-business-strategy/?sh=7402650320bd 2. Lee, J., Kao, H.-A., Yang, S.: Service innovation and smart analytics for industry 4.0 and big data environment. Procedia CIRP 16, 3–8 (2014). https://doi.org/10.1016/j.procir.2014. 02.001 3. n.d.: Human-machine systems. https://encyclopedia2.thefreedictionary.com/Human-mac hine+systems (2002) 4. Thareja, P.: Manufacturing paradigms in 2010. In: Proceedings of National Conference on Emerging trends in Manufacturing Systems, JMIT, Radaur (2005) 5. European Commission, Directorate-General for Research and Innovation, Renda, A., et al.: Industry 5.0, a transformative vision for Europe : governing systemic transformations towards a sustainable industry. Publications Office of the European Union (2022). https://doi.org/10. 2777/17322 6. Zheng, T., Ardolino, M., Bacchetti, A., Perona, M.: The applications of Industry 4.0 technologies in manufacturing context: a systematic literature review. Int. J. Prod. Res. 59, 1922–1954 (2021). https://doi.org/10.1080/00207543.2020.1824085 7. European Commission, Directorate-General for Research and Innovation: Industry 5.0. https://research-and-innovation.ec.europa.eu/research-area/industrial-research-and-inn ovation/industry-50_en. Accessed 21 Mar 2023 8. Pacchini, A.P.T., Lucato, W.C., Facchini, F., Mummolo, G.: The degree of readiness for the implementation of Industry 4.0. Comput. Ind. 113, 103125 (2019). https://doi.org/10.1016/j. compind.2019.103125 9. Müller, J.: Enabling technologies for Industry 5.0 (2020) 10. Markard, J., Raven, R., Truffer, B.: Sustainability transitions: an emerging field of research and its prospects. Res. Policy 41, 955–967 (2012) 11. Van Kuiken, S.: Tech at the edge: trends reshaping the future of IT and business. https://www.mckinsey.com/capabilities/mckinsey-digital/our-insights/tech-at-theedge-trends-reshaping-the-future-of-it-and-business#/. Accessed 21 Mar 2023

296

B. Henriksen and M. K. Thomassen

12. Kim, G.-Y., et al.: Human digital twin system for operator safety and work management. In: Kim, D.Y., von Cieminski, G., Romero, D. (eds.) Advances in Production Management Systems. Smart Manufacturing and Logistics Systems: Turning Ideas into Action: IFIP WG 5.7 International Conference, APMS 2022, Gyeongju, South Korea, September 25–29, 2022, Proceedings, Part II, pp. 529–536. Springer Nature Switzerland, Cham (2022). https://doi. org/10.1007/978-3-031-16411-8_61 13. Bousdekis, A., Mentzas, G., Apostolou, D., Wellsandt, S.: Evaluation of AI-based digital assistants in smart manufacturing. In: Kim, D.Y., von Cieminski, G., Romero, D. (eds.) Advances in Production Management Systems. Smart Manufacturing and Logistics Systems: Turning Ideas into Action: IFIP WG 5.7 International Conference, APMS 2022, Gyeongju, South Korea, September 25–29, 2022, Proceedings, Part II, pp. 503–510. Springer Nature Switzerland, Cham (2022). https://doi.org/10.1007/978-3-031-16411-8_58 14. Rolstadås, A., Henriksen, B., O’Sullivan, D.: Manufacturing outsourcing: a knowledge perspective. Springer (2012). https://doi.org/10.1007/978-1-4471-2954-7 15. Paci, A.M., Chiacchio, M.S., Lalle, C.: Production Paradigms Ontology (PPO): a response to the need of managing knowledge in high-tech manufacturing. In: Bernard, A., Tichkiewitch, S. (eds.) Methods and tools for effective knowledge life-cycle-management, pp. 227–240. Springer Berlin Heidelberg, Berlin, Heidelberg (2008). https://doi.org/10.1007/978-3-54078431-9_13 16. Jovane, F., Koren, Y., Boer, C.R.: Present and future of flexible automation: towards new paradigms. CIRP Ann. 52, 543–560 (2003) 17. Henriksen, B., Røstad, C.C., Thomassen, M.K.: Industry 5.0 – making it happen in the Agri industry. the core product service platform. In: Kim, D.Y., von Cieminski, G., Romero, D. (eds.) Advances in Production Management Systems. Smart Manufacturing and Logistics Systems: Turning Ideas into Action: IFIP WG 5.7 International Conference, APMS 2022, Gyeongju, South Korea, September 25–29, 2022, Proceedings, Part I, pp. 424–431. Springer Nature Switzerland, Cham (2022). https://doi.org/10.1007/978-3-031-16407-1_50 18. Hein-Pensel, F., et al.: Maturity assessment for Industry 5.0: a review of existing maturity models. J. Manuf. Syst. 66, 200–210 (2023). https://doi.org/10.1016/j.jmsy.2022.12.009 19. Felch, V., Asdecker, B., Sucky, E.: Maturity models in the age of industry 4.0–do the available models correspond to the needs of business practice? (2019) 20. Mittal, S., Khan, M.A., Romero, D., Wuest, T.: A critical review of smart manufacturing & Industry 4.0 maturity models: implications for small and medium-sized enterprises (SMEs). J. Manuf. Syst. 49, 194–214 (2018) 21. Schumacher, A., Erol, S., Sihn, W.: A maturity model for assessing industry 4.0 readiness and maturity of manufacturing enterprises. Procedia CIRP 52, 161–166 (2016). https://doi. org/10.1016/j.procir.2016.07.040 22. Sony, M., Naik, S.: Key ingredients for evaluating Industry 4.0 readiness for organizations: a literature review. Benchmarking: Int. J. 27(7), 2213–2232 (2019). https://doi.org/10.1108/ BIJ-09-2018-0284 23. Henriksen, B., Røstad, C.: Norwegian leisure boats in front (2014) 24. Henriksen, B., Landmark, A., Røstad, C.C.: What product developers really need to knowcapturing the major design elements. In: Advances in Production Management Systems. Towards Smart Production Management Systems: IFIP WG 5.7 International Conference, APMS 2019, Austin, TX, USA, September 1–5, 2019, Proceedings, Part II. pp. 447–454. Springer, Cham (2019). https://doi.org/10.1007/978-3-030-29996-5_52 25. Henriksen, B., Røstad, C.C., von Stietencron, M.: Development projects in SMEs. In: Lödding, H., Riedel, R., Thoben, K.-D., von Cieminski, G., Kiritsis, D. (eds.) APMS 2017. IAICT, vol. 514, pp. 193–201. Springer, Cham (2017). https://doi.org/10.1007/978-3-319-66926-7_23 26. The Bay Area Urban Manufacturing Initiative: Transition strategies for your manufacturing business (2021)

Industry 4.0 Readiness Assessment of Enterprises in Kazakhstan Dinara Dikhanbayeva1 , Malika Aitzhanova1 , Yevgeniy Lukhmanov1 , Ali Turkyilmaz2(B) , Essam Shehab1 , and Idriss El-Thalji3 1 School of Engineering and Digital Sciences, Nazarbayev University, Astana, Kazakhstan 2 University of Stavanger, UIS Business School, Stavanger, Norway

[email protected] 3 Faculty of Science and Technology, University of Stavanger, Stavanger, Norway

Abstract. Industry 4.0 is a complex and broad concept that affects many parts of the business, such as processes, customers, and business models, through various enablers, advanced technologies, and practices. Despite the high investment cost of Industry 4.0, many developed countries and large enterprises are already taking advantage of it. In contrast, developing countries are only exploring this concept and making the first transformation steps towards it. This study aims to identify the state of Industry 4.0 readiness of enterprises in Kazakhstan, a developing country in Central Asia, based on the information collected from 205 enterprises across the country. The study also investigates common barriers and weaknesses faced by these organizations. Based on the analysis, the study proposes practical recommendations for companies to support their transformation toward Industry 4.0. Keywords: Industry 4.0 · Digital Transformation · Readiness Assessment · Kazakhstan

1 Introduction Industry 4.0 (I4.0) is an ongoing global tendency of enterprises to implement advanced information communication technologies (ICT) to enhance their business performance. I4.0 organizations seek to implement interconnectivity, data transparency, decentralization, and technical assistance in their factories and processes [1]. These design principles benefit I4.0 stakeholders with increased productivity, flexibility, sustainability, improved customer experience, and new services. Achieving fully automated yet flexible production necessitates technological solutions like big data analytics, IoT, cloud manufacturing, etc., as well as changes in corporate strategy, management, and organization (e.g., supply chain, innovation, lean management) [2]. Companies worldwide face challenges in transitioning to I4.0. Its broad scope, complex technologies, and costs pose uncertainties for managers and employees [3]. Due to these reasons, companies, especially in developed countries, often rely on outdated business strategies. Assessing their current state and identifying strengths, © IFIP International Federation for Information Processing 2023 Published by Springer Nature Switzerland AG 2023 E. Alfnes et al. (Eds.): APMS 2023, IFIP AICT 689, pp. 297–310, 2023. https://doi.org/10.1007/978-3-031-43662-8_22

298

D. Dikhanbayeva et al.

weaknesses, and bottlenecks is crucial. However, not all countries provide sufficient infrastructure, tools, and guidance for a competent transition to I4.0 [2]. Governments and non-governmental organizations (NGOs) around the globe are launching various programs to support the digital transformation of their economies. The “Digital Kazakhstan” program was launched in 2017 to improve the digitization level of the Kazakhstan industry [4]. Kazakhstan has a developing economy with a GDP per capita of $8,500 and is one of the largest raw materials producers with its rich natural resources. Despite possessing valuable natural resources, the country’s industry is underdeveloped [5]. Kazakhstan lacks a comprehensive quantitative analysis of Industry 4.0 readiness, which creates a literature gap in this area. Therefore, this study focuses on the following research questions: 1. What are the common barriers and challenges faced by Kazakhstani enterprises in adopting I4.0 practices? 2. What are the key findings and implications of the I4.0 readiness assessment conducted in Kazakhstan? A newly developed comprehensive maturity assessment model (COMMA 4.0) was utilized to address these questions with data collected from 205 enterprises.

2 Theoretical Background The term “Industry 4.0” refers to the digital transformation of production and service systems, synonymous with the Fourth Industrial Revolution, Smart Manufacturing, or Digital Manufacturing [3]. This era of digitization and automation is marked by the integration of digital technologies into various industries and service systems, which allows for critical features such as automation and self-optimization of the entire supply chain, from product development to manufacturing, logistics, and post-production processes [2]. I4.0 combines physical and virtual worlds using technologies like IoT, cyber-physical systems, cloud computing, big data analytics, digital twins, machine learning, autonomous robots, additive manufacturing, and AI. These technologies enable real-time data management in interconnected systems [6]. I4.0 brings industry benefits, transforming organizations and introducing innovative business models throughout the supply chain. Integration of physical and virtual systems enables automation, reliable data, transparency, and autonomous decision-making [6]. Further, strategic combinations of technologies like IoT and robots enhance operational flexibility, transforming traditional assembly systems into adaptive, human-centric systems that quickly respond to market demands [7]. Adopting I4.0 is challenging for organizations due to its complexity and lack of unified implementation methodologies. The main barriers identified in the literature include a low level of awareness about I4.0 and its strategic importance, lack of technical standards, economic conditions, and lack of digital culture, qualified personnel, information security infrastructure, management support, knowledge of specific technologies, and proper ICT architecture [6]. To achieve successful digital transformation, a comprehensive management approach is required that focuses on business strategies, processes, operations, projects, and

Industry 4.0 Readiness Assessment of Enterprises in Kazakhstan

299

resources, in addition to technology adaptation [8]. The transformation towards Industry 4.0 also entails a shift in the human resource management paradigm, emphasizing the need for continued learning and competency development in response to increasingly complex production processes [9]. Maturity assessment models are widely used to assist companies in their digital transformation. These models evaluate a company’s business processes from various dimensions and with a set of critical business measures, highlighting the company’s strengths and weaknesses and providing improvement plans to reach the desired maturity stage [10]. They can also be used as a benchmarking tool to compare a company’s performance with other companies in the same industry or country. According to [11], who reviewed 30 maturity models, the most common dimensions identified were technology, people, strategy, leadership, process, and innovation, and these dimensions are the crucial indicators for enhancing the digital transformation performance in organizations, regardless of their size or industry. Each dimension is measured using multiple indicators, or sub-dimensions, and the maturity level is calculated based on these indicators [12]. Maturity models may also include a recommendation mechanism or roadmap that suggests relevant actions for the bottleneck indicators/dimensions, assisting the organization in moving from the current stage to a higher maturity level [12]. Table 1 presents a list of the most recognized Industry 4.0 maturity models along with their corresponding dimensions. Table 1. Existing maturity models and their dimensions. Maturity model

Main dimensions

IMPULS - Industrie 4.0 Readiness [13]

Strategy & Organization; Smart factory; Smart operations; Smart products; Data-driven services; Employees

Reference Architecture Model for the Industry 4.0 [14]

Product life-cycle and Value stream; Location/functional hierarchy; 6 Functional layers: Business, Functional, Information, Communication, Integration, Asset

Industry 4.0 Maturity Model [15]

Strategy; Leadership; Customers; Products; Operations; Culture; People; Governance; Technology

Digital Operations Self-assessment (PWC) [16]

Business Models, Products & Service Portfolio; Market & Customer Access; Value Chain & Processes; IT architecture; Compliance, security, legal & tax; Organization & Culture

SIMMI 4.0 (System Integration Maturity Model Industry 4.0) [17]

Vertical integration; Horizontal integration; Digital product development; Cross-sectional technology criteria

Singapore Smart Industry Readiness Index (SIRI) [18]

Process, Technology, and Organization

ACATECH Industrie 4.0 Maturity Index [19]

Resources; Information Systems; Organizational Structure; Culture

300

D. Dikhanbayeva et al.

3 Methodology The maturity model, COMMA 4.0, used in this research was developed using the design science research (DSR) approach, a well-established methodology for developing frameworks, including I4.0 maturity models [20]. A systematic literature review was conducted to extensively analyze the I4.0 concepts and examine the existing I4.0 maturity models (MMs) in terms of their objectives, specialization, maturity stages, dimensions, items, corresponding design principles, assessment tools, recommendation system availability, and comprehensiveness. The findings were then evaluated thoroughly by the selected experts in academia and industry for their relevance and applicability in the context of emerging economies. The developed model consists of five main I4.0 dimensions and 32 sub-dimensions, as given in Table 2. It was pilot tested with 25 manufacturing firms for its correctness, wording, and design and authorized by the Institutional Research Ethics Committee for implementation. Table 2. Research model structure Main dimensions

Sub-dimensions

Strategy and Organization

Strategy implementation; Business performance management; State of ICT function; ICT systems budget management; Leadership support for I4.0; Innovation management system; Collaboration with stakeholders

Workforce Development

Digital competency; Support for employees’ development; Change acceptance level

Smart Factory

Automation of production system; equipment upgradability; M2M communication (data exchange) level; Supply chain communication/integration; Digitization of enterprise data; Observability of production; ICT Architecture; Cloud services utilization

Smart Processes

Standardization and improvement of BPs; Use of automation and business information systems (MIS/ERP); Data-driven decision-making; Maintenance approach; Quality management systems

Smart Products and Services

Intellectual property development; Digital products; Digital services; Product customization degree; Frequency of product/service upgrades; Customer data utilization; sales channels used; Revenue from data-driven services; Plans for digital features

COMMA 4.0 includes five maturity levels ranging from Level 1 “Entrant” to Level 5 “Expert”. Additionally, it includes an Expert System, which represents an automatic mechanism that 1) calculates the maturity level of the company, 2) generates a report

Industry 4.0 Readiness Assessment of Enterprises in Kazakhstan

301

with score-sensitive recommendations, and 3) identifies the main bottlenecks and drivers of the company and highlights them in the report [21]. Data Collection In total, 430 companies, selected from the public databases, were invited to the assessment, of which 205 have completed the survey. Computer-aided telephone interview (CATI) was the primary method for collecting the data coupled with the survey material provided to firms during the interview. Additionally, the research team accompanied respondents throughout the assessment and clarified I4.0 terms used in the questionnaire. The diverse geography, sizes, and industry types of the assessed companies allowed the authors to perform a comprehensive analysis. 100% of companies were operating in Kazakhstan, and, more precisely, located in cities such as Kyzylorda (9%) and Aktobe (8%), followed by Almaty (8%) and Nur-Sultan (5%). All survey respondents had the necessary qualifications, which ensured the reliability of the collected data. First, respondents’ awareness regarding the company’s status was sufficient. 51% of respondents were senior managers, and 31% were middle-level managers. Secondly, the median education level of respondents was high. 80% of respondents held a bachelor’s degree, while 58% had a master’s degree. Furthermore, more than 75% of respondent companies are considered as small and medium enterprises with a number of employees between 10– 200, and almost 25% of companies are large companies with more than 200 employees (Table 2). 76% of respondents are represented by product companies, while the rest 24% are service companies. The majority of product companies operate in the manufacturing and construction industry (141, 69%), and the majority of service companies are from the financial and real estate activities industry (45, 22%).

4 Results and Discussion The average I4.0 readiness level is 1.86 out of a 0 to 5 scale, which suggests that I4.0 is already present in Kazakhstani companies at some point. The overall results of the whole data set classified by the maturity levels and structured by product and service orientation, separately highlighting manufacturing companies as the biggest sample within product orientation, are presented in Table 3. Overall, the results dictate that there are more companies with readiness factors at the entrant level than at any other level (31%), which assumes the lowest level of I4.0 activities so far. Further, in an almost similar share, companies are spread over beginner and learner levels (25% and 24%). Beginner companies tend to be aware of the I4.0 concept and are at the beginning of their way to implementing I4.0 elements. Learner-level companies tend to have the I4.0 strategy formulated and already have digitized some processes. 13% of companies report that their I4.0 readiness is at the integrator level, meaning that I4.0 is partially in implementation mode and investments into I4.0 are made. Lastly, 7% of companies are at an expert level and have made significant investments into their I4.0 transformation and implemented the I4.0 strategy across the company. Within the overall tendency, the results of service companies are slightly higher. It is clear that most of the companies are concentrated among levels 1, 2, and 3. However, taking into account that 20% of companies are spread over Integrator (level 4) and Expert (level 5) levels reflects that some companies

302

D. Dikhanbayeva et al.

are ahead of others in digitization and I4.0 transformation. In fact, service companies among all are scoring better results in levels 4 (integrator) and 5 (expert). Table 3. Industry 4.0 Readiness of Kazakhstani Companies All (205)

Manufacturing (156)

Service (49)

Level 1: Entrant

31%

30%

33%

Level 2: Beginner

25%

26%

22%

Level 3: Learner

24%

25%

22%

Level 4: Integrator

13%

12%

15%

Level 5: Expert

7%

7%

8%

Average readiness (0 to 5 scale)

1,86

1,83

1,94

According to the analysis of maturity readiness by company size, large companies score higher readiness results in comparison to medium or small enterprises across all dimensions except for smart products and services (Fig. 1). This can be explained by the fact that initial investments are required to implement I4.0 elements, and large companies have a stronger incentive to implement I4.0 practices, more work subject to automation, and more funds for it [22]. Figure 1 shows no statistically significant difference between medium and large companies. However, significant differences were found among subsets of small and medium companies, as well as small and large companies (at a 0.05 level), when compared by companies and main dimensions. A similar analysis of 84 Turkish enterprises indicated little to no correlation between company size and their I4.0 strategies and budgets [23]. The low score of smart products and services can be explained by several facts, such as the heavy focus of the national economy on raw materials extraction and the low level of development of industries such as manufacturing or processing. According to the results given in Fig. 1, the highest readiness score is obtained by workforce development dimension (2.46), followed by smart factory (1.91), smart processes (1.88), smart product and services (1.55), and finally, by strategy and organization (1.47). In particular, the strategy and organization and smart product and services dimensions are identified to be the most common weaknesses for Kazakhstani companies. To confirm a statistical difference for the sample companies grouped by the average readiness scores for all five dimensions, an analysis of variance (ANOVA) has been performed. The results confirm a statistical difference between the average readiness scores among all five dimensions (p-value < 0.005).

Industry 4.0 Readiness Assessment of Enterprises in Kazakhstan

303

Fig. 1. Industry 4.0 readiness by company size.

4.1 Strategy and Organization The average readiness score for the strategy and organization dimension in Kazakhstan received the lowest result among all dimensions and is equal to 1,47. Figure 2 provides a detailed analysis of the strategy and organization-related criteria, which may explain the main figures behind this low score. The highest share of companies placed at the entrant level, which assumes that 37% of companies do not have any I4.0 strategies supported with business performance management, strong leadership, ICT system, and dedicated budget, as well as that no innovation system is present. Another critical weakness of the assessed enterprises is about their collaboration with their stakeholders regarding I4.0 development. Despite the importance of technologies in the era of I4.0, strategic vision and correct organizational architecture are often missed and underrated. Moreover, a long-term strategic vision and agile organizational structure are required to minimize the risks of failures and increase the benefits of I4.0 [8]. The leadership support in such cases and their change management skills play an essential role [13], as the decision-making in developing countries is mostly centralized [9]. 4.2 Workforce Development The overall success of digitization projects depends on workers’ ability to accept and cope with change. According to the results, the I4.0 readiness score for the workforce development dimension is the highest among all dimensions and is equal to 2.46. Within that, 7% of companies identified themselves as experts (Fig. 3), where employees have advanced levels of digital competency and are able to use advanced packages. In addition, employees take proactive actions if changes are implemented and even demand involvement. Despite that, 8% of the firms are at an entrant level, and nearly one-fifth of the companies are at a beginner (19%) level. Companies at the entrant level have a low digital competency level of employees, i.e., no digital skills, experience, or not required during the work process. On top of that, the company does not provide any special budget for training for qualification enhancement. A more detailed analysis shows that 63% of

304

D. Dikhanbayeva et al.

Fig. 2. Industry 4.0 readiness of strategy and organization dimension.

the companies indicated that the acceptance level of change by employees is moderate in this level, which means that if benefits and the need for change is clarified and communicated correctly, the workforce is eager to accept them. 54% of the companies stated to have a moderate level of digital competency, assuming that employees are freely able to use most of the common digital devices: active user of the computer, work-related programs such as 1C ERP, etc. One of the weaknesses identified is the low level of support from employers for the training of employees. While, the workforce needs to be competent enough, continuously learning and developing, possess the knowledge, and improve qualifications in order to contribute to the success of the enterprise [15].

Fig. 3. Industry 4.0 readiness of workforce development dimension.

4.3 Smart Factory Smart Factory provides an environment where human involvement in operations is minimal, as production processes can be carried out by automated machines and equipment [15]. However, in the Kazakhstan case, results reveal a low level of data exchange level between equipment. The general level of equipment modernization is low on a national level, except for very few cases of multinational enterprises (MNEs) [24].

Industry 4.0 Readiness Assessment of Enterprises in Kazakhstan

305

The average readiness score for the smart factory dimension is 1.91. The detailed results in Fig. 4 show that 27% of companies are at entrant and beginner levels in this dimension, which assumes no or low level of automation of production and systems processes, the same as for machine-to-machine (M2M) communication. Companies use only basic communication with suppliers and customers, while data collection is still manual. Conversely, 9% of companies are classified as level expert companies, which is the highest result among all dimensions for expert level. Companies at this level are assumed to have full automation, a high level of M2M communication, global supply chain integration, real-time production monitoring, and a common ICT architecture with stakeholders. Automation of production systems, supply chain communication, and digitization of enterprise data received comparatively better results in this dimension, one of the reasons for which can be the legislative obligation for companies to increase digitalization, as well as follow special forms of financial reporting, which become easier if ERP systems are in place for example.

Fig. 4. Industry 4.0 readiness of Smart Factory dimension.

4.4 Smart Processes Smart Processes are critical for actualizing and maximizing the value of the smart factory concept. Technology or automation is not enough for the correct and proper I4.0 implementation if there is no structure and system to properly integrate them with the business processes [22]. Additionally, a comprehensive smart process approach allows, by changing the focus from individual processes to the entire value chain, to improve the operations, product lifecycle, and supply chain [25]. The average I4.0 readiness score of the smart processes dimension is 1.88. Quality management systems, maintenance approach, use of automation and business information systems, and standardization of business processes are identified as issues in Kazakhstani enterprises (Fig. 5). 26% of companies are placed at the entrant level, which assumes the absence of standardization of business processes, manual jobs prevail, and data is not utilized for decision-making. Quality management level is presented fully in the vast majority of the companies; however, only 6% of the firms implement it within the entire organization and manage

306

D. Dikhanbayeva et al.

Fig. 5. I4.0 readiness of companies for smart processes.

digitally. The maintenance approach has the highest percentage of companies located at the entrant and expert levels (35% and 15%, respectively); at the entrant level, corrective maintenance is applied (maintenance is performed at the time of the occurrence), while at an expert level is prescriptive when maintenance is implemented using a real-time data, which enables the self-diagnosis and scheduled maintenance of the production equipment. 4.5 Smart Product and Services Smart products and services are essential components of I4.0 systems to collect, exchange and manage data through equipped tools such as RFID, sensors, and other technologies [13, 17]. The average I4.0 readiness score for this dimension for Kazakhstani companies is 1.55. Only one-fifth of the companies indicated themselves as integrators and experts (Fig. 6). At these levels, customization possibilities of the products are high, as well as the level of upgrade of offered products and services in order to meet customer demands. However, the results in Fig. 6 show a small level of availability of enterprises to support product customization. The low level of digital products and services along with intellectual property development depicts the national strategy oriented to heavy extraction of raw materials. The traditional way of sales prevails in the country, along with physical stores, without the development of online services. This fact may help to explain the low level of revenue from data-driven services. At the same time, customers’ needs for smart products, in case of a designed process of communication of manufacturers with customers, allow the creation of new services, which gives new possibilities for revenue, which is already introduced in many countries [26]. Thus, growing demand from customers for more individualized products represents the need to establish a high level of digitalization and integration in all stages of the product cycle.

Industry 4.0 Readiness Assessment of Enterprises in Kazakhstan

307

Fig. 6. I4.0 readiness of companies for smart products and services.

5 Conclusion and Implications Overall, the study successfully addresses the previously mentioned research questions. By the criteria of overall I4.0 maturity, Kazakhstan companies demonstrated a “beginner” readiness level achieving 1.86 on the 0–5 scale. Similar results have been observed in several other countrywide studies of developing economies, such as South Africa, Turkey, and Brazil. For example, 45% of South African enterprises are on the emerging level, where I4.0 is either completely absent or only pilot initiatives are launched [27]. The key reason for the low I4.0 maturity is numerous barriers, including legal, cultural, technical, and financial [22, 23]. These findings illustrate the presence of strong barriers to I4.0 and the need for governmental support. The analysis of I4.0 readiness by dimensions revealed that on average, the most developed aspect is workforce development. Possibly, the long-lasted culture of employee training has achieved overwhelming support and was adopted by many companies even before the I4.0 revolution began. On the other hand, commonly, the least developed dimensions are strategy and organization, and smart products and services. This trend indicates that companies in Kazakhstan have struggled to align their strategic vision with the demands and opportunities presented by I4.0. Similarly, Yüksel [23] found out that 69% of Turkish companies do not have an I4.0 strategy and plan. The major reason for the absence of strategy could be the lack of awareness regarding the I4.0 concept, tools, and benefits. This is further complicated by the often centralized decision-making process in developing countries, where a disconnect between top management and operational staff can hinder the effective formulation and implementation of I4.0 strategies. Furthermore, a number of studies also revealed that smart products and data-driven services have worse performance than other aspects, such as smart factories and processes. The low readiness score in smart products and services dimension can be attributed to the national economic structure, which is heavily focused on raw materials extraction. As a result, industries such as manufacturing or processing that typically develop and offer smart products and services have a relatively low level of development. Another possible explanation for this is the hierarchical relationship between the dimensions. Smart products and services require the infrastructure provided by smart

308

D. Dikhanbayeva et al.

processes, such as IoT and big data analytics [28]. Given the current low levels of automation and standardization of business processes, Kazakhstani firms may be encountering difficulties in developing and offering smart products and services. Hence, products and services are expected to inherently lag behind factory and process dimensions. Going forward, it would be crucial for companies in Kazakhstan to focus on improving their readiness in these lagging dimensions to ensure a more balanced and comprehensive approach to their I4.0 transformation. As a result, it is obvious that depending on the goal, different scenarios of the I4.0 introduction in companies are possible. Based on the literature review and the findings of this study, the following suggestions for improvements are offered separately for businesses and governmental entities: For companies: • To develop a deep understanding of the I4.0 concept, components, benefits, and challenges and formulate an I4.0 strategy that covers step-by-step adaptation of ICT technologies and practices. • To establish a dedicated unit with an allocated budget for digital transformation and implementing the developed strategy. • To enhance infrastructure and resources for automation of production and systems and collaborate with I4.0 providers and users to facilitate the transition. • To improve and standardize business processes to leverage captured data. • To invest in digital skills development through training and workshops and foster leadership that supports employee resilience and embraces change. For governmental entities: • To promote awareness and understanding of the I4.0 concept among companies and provide support and guidance in formulating I4.0 strategies. • To facilitate collaboration between companies and I4.0 providers/users. • To implement policies and regulations that encourage I4.0 adoption and standardization and foster an innovation-friendly environment through these supportive policies and initiatives. • To establish digital skill development programs and provide financial incentives to support companies in their I4.0 transformation. In conclusion, increasing Industry 4.0 readiness presents a multitude of benefits for Kazakhstani companies. Improved readiness would lead to more efficient operations, increased productivity, enhanced product and service quality, and heightened customer satisfaction. However, the transition will require concerted efforts in strategy formulation, technological upgrading, process standardization, and customer-centric innovation. The study’s limitations include sample size and reliance on self-reported data, potentially limiting generalization and introducing bias. Future research could address these limitations by expanding the sample size across industries and using a combination of research methods for more reliable and valid findings. This could involve integrating surveys, interviews, and other data collection methods to enhance understanding of I4.0 readiness.

Industry 4.0 Readiness Assessment of Enterprises in Kazakhstan

309

References 1. Hermann, M., Pentek, T., Otto, B.: Design principles for industrie 4.0 scenarios. In: 2016 49th Hawaii International Conference on System Sciences (HICSS), pp. 3928–3937 (2016). https://doi.org/10.1109/HICSS.2016.488 2. Klingenberg, C.O., Borges, M.A.V., Antônio, J.A.V.: Industry 4.0: what makes it a revolution? A historical framework to understand the phenomenon. Technol. Soc. 70, 102009 (2022). https://doi.org/10.1016/j.techsoc.2022.102009 3. Wagire, A.A., Joshi, R., Rathore, A.P.S., Jain, R.: Development of maturity model for assessing the implementation of Industry 4.0: learning from theory and practice. Production Planning & Control. 32, 603–622 (2021) 4. Ministry of Digital Development of the Republic of Kazakhstan: Digital Kazakhstan. https:// egov.kz/cms/en/digital-kazakhstan 5. Daribay, A., Serikova, A., Ukaegbu, I.A.: Industry 4.0: Kazakhstani industrialization needs a global perspective. Procedia Comput. Sci. 151, 903–908 (2019). https://doi.org/10.1016/j. procs.2019.04.125 6. Vaidya, S., Ambad, P., Bhosle, S.: Industry 4.0 – a glimpse. Procedia Manuf. 20, 233–238 (2018). https://doi.org/10.1016/j.promfg.2018.02.034 7. Moeuf, A., Lamouri, S., Pellerin, R., Tamayo-Giraldo, S., Tobon-Valencia, E., Eburdy, R.: Identification of critical success factors, risks and opportunities of Industry 4.0 in SMEs. Int. J. Prod. Res. 58(5), 1384–1400 (2020). https://doi.org/10.1080/00207543.2019.1636323 8. Salunkhea, O., Berglunda, Å.: Industry 4.0 enabling technologies for increasing operational flexibility in final assembly. Int. J. Ind. Eng. Manag. 13, 38–48 (2022). https://doi.org/10. 24867/IJIEM-2022-1-299 9. Dikhanbayeva, D., Tokbergenova, A., Lukhmanov, Y., Shehab, E., Pastuszak, Z., Turkyilmaz, A.: Critical factors of industry 4.0 implementation in an emerging country: empirical study. Future Internet. 13, 137 (2021). https://doi.org/10.3390/fi13060137 10. Bastos, T., Salvadorinho, J., Teixeira, L.: UpSkill@Mgmt 4.0 – a digital tool for competence management: conceptual model and a prototype. Int. J. Ind. Eng. Manage. 13(4), 225–238 (2022). https://doi.org/10.24867/IJIEM-2022-4-315 11. Fettermann, D.C., Cavalcante, C.G.S., Almeida, T.D.de, Tortorella, G.L.: How does Industry 4.0 contribute to operations management? J. Ind. Prod. Eng. 35, 255–268 (2018) 12. Hizam-Hanafiah, M., Soomro, M.A., Abdullah, N.L.: Industry 4.0 readiness models: a systematic literature review of model dimensions. Information 11, 364 (2020). https://doi.org/ 10.3390/info11070364 13. Schumacher, A., Nemeth, T., Sihn, W.: Roadmapping towards industrial digitalization based on an industry 4.0 maturity model for manufacturing enterprises. Procedia CIRP 79, 409–414 (2019). https://doi.org/10.1016/j.procir.2019.02.110 14. Lichtblau, K., et al.: Industrie 4.0 readiness. IMPULS-Stiftung for Mechanical Engineering, Plant Engineering, and Information Technology (2015) 15. Adolphs, P., et al.: Reference architecture model industrie 4.0 (rami4. 0). ZVEI and VDI, Status report (2015) 16. Schumacher, A., Erol, S., Sihn, W.: A maturity model for assessing industry 4.0 readiness and maturity of manufacturing enterprises. Procedia Cirp. 52, 161–166 (2016). https://doi. org/10.1016/j.procir.2016.07.040 17. Geissbauer, R., Vedso, J., Schrauf, S.: Industry 4.0: building the digital enterprise. London: PwC (2016) 18. Leyh, C., Bley, K., Schäffer, T., Forstenhäusler, S.: SIMMI 4.0 – A maturity model for classifying the enterprise-wide IT and software landscape focusing on Industry 4.0. In: 2016 Federated Conference on Computer Science and Information Systems (FEDSIS), pp. 1297– 1302. IEEE (2016)

310

D. Dikhanbayeva et al.

19. Singapore Economic Development Board: The Singapore Smart Industry Readiness Index: Catalysing the Transformation of Manufacturing (2020) 20. Schuh, G., Anderl, R., Gausemeier, J., ten Hompel, M., Wahlster, W.: Industrie 4.0 maturity index. Managing the Digital Transformation of Companies. Munich: Herbert Utz (2017) 21. Hevner, A., Chatterjee, S., Hevner, A., Chatterjee, S.: Design science research in information systems. Design Res. Inf. Syst.: Theory Pract., 9–22 (2010). https://doi.org/10.1007/978-14419-5653-8_2 22. Lukhmanov, Y., Dikhanbayeva, D., Yertayev, B., Shehab, E., Turkyilmaz, A.: An advisory system to support industry 4.0 readiness improvement. Procedia CIRP 107, 1361–1366 (2022). https://doi.org/10.1016/j.procir.2022.05.158 23. Dikhanbayeva, D., Aitzhanova, M., Shehab, E., Turkyilmaz, A.: Analysis of textile manufacturing SMEs in Kazakhstan for industry 4.0. Procedia CIRP. 107, 888–893 (2022). https:// doi.org/10.1016/j.procir.2022.05.080 24. Yüksel, H.: An empirical evaluation of industry 4.0 applications of companies in Turkey: the case of a developing country. Technol. Soc. 63, 101364 (2020). https://doi.org/10.1016/j.tec hsoc.2020.101364 25. Abdikarimova, K.: About Digitalization of the Industry of Kazakhtan. Sci. Bull.: Finance, Banks, Investments 1, 167–174 (2018) 26. Akdil, K.Y., Ustundag, A., Cevikcan, E.: Maturity and readiness model for industry 4.0 strategy. In: Industry 4.0: Managing the digital transformation. SSAM, pp. 61–94. Springer, Cham (2018). https://doi.org/10.1007/978-3-319-57870-5_4 27. Agca, O., Gibson, J., Godsell, J., Ignatius, J., Davies, W., Xu, O.: An Industry 4 Readiness Assessment Tool; University of Warwick: Coventry. University ofWarwick, Coventry, UK (2017) 28. Maisiri, W., van Dyk, L.: Industry 4.0 readiness assessment for South African industries. S. Afr. J. Ind. Eng. 30(3),(2019). https://doi.org/10.7166/30-3-2231 29. Meindl, B., Ayala, N.F., Mendonça, J., Frank, A.G.: The four smarts of Industry 4.0: evolution of ten years of research and future perspectives. Technol Forecasting Soc. Change 168, 120784 (2021)

Critical Factors for Selecting and Integrating Digital Technologies to Enable Smart Production: A Data Value Chain Perspective Natalie Agerskans(B) , Mohammad Ashjaei , Jessica Bruch , and Koteshwar Chirumalla Mälardalen University, Hamngatan 15, SE 631 05 Eskilstuna, Sweden [email protected]

Abstract. With the development towards Industry 5.0, manufacturing companies are developing towards Smart Production, i.e., using data as a resource to interconnect the elements in the production system to learn and adapt accordingly for a more resource-efficient and sustainable production. This requires selecting and integrating digital technologies for the entire data lifecycle, also referred to as the data value chain. However, manufacturing companies are facing many challenges related to building data value chains to achieve the desired benefits of Smart Production. Therefore, the purpose of this paper is to identify and analyze the critical factors of selecting and integrating digital technologies for efficiently benefiting data value chains for Smart Production. This paper employed a qualitative-based multiple case study design involving manufacturing companies within different industries and of different sizes. The paper also analyses two Smart Production cases in detail by mapping the data flow using a technology selection and integration framework to propose solutions to the existing challenges. By analyzing the two in-depth studies and additionally two reference cases, 13 themes of critical factors for selecting and integrating digital technologies were identified. Keywords: Industry 5.0 · Digital Transformation · Smart Manufacturing · Technology Integration · Technology Selection · Production Development

1 Introduction To stay competitive in Industry 5.0, manufacturing companies are developing towards Smart Production. In Smart Production, all available resources and devices within the production system are interconnected to generate valuable insights about production processes in order to support the human workers in their work tasks and enable optimized and more sustainable production [1, 2]. For instance, production equipment will communicate with each other, automated processes, humans, and other resources to allow for proactive and timely actions in production [3, 4]. To utilize the potential offered through Smart Production, manufacturing companies need to use digital technologies to collect and link data from multiple sources and convert this data into valuable knowledge [1, © IFIP International Federation for Information Processing 2023 Published by Springer Nature Switzerland AG 2023 E. Alfnes et al. (Eds.): APMS 2023, IFIP AICT 689, pp. 311–325, 2023. https://doi.org/10.1007/978-3-031-43662-8_23

312

N. Agerskans et al.

5]. In previous research, this can be referred to as the data value chain [6–8]. The data value chain specifies the complete data lifecycle of a system from a technical point of view, which includes data sources and collection, data communication, data processing, data storage, data visualization, and data usage [5, 6, 8]. Thus, there is a need to combine several synergizing digital technologies to generate value. However, manufacturing companies are facing many challenges when selecting and integrating a combination of digital technologies [9–11], and enabling data value chains for Smart Production is particularly problematic [12]. Rather than having a holistic perspective on the entire data value chain, a common approach is to select and integrate one digital technology at a time [13]. Selecting the technologies in isolation may result in issues such as mismatches among the technologies, interoperability issues, and integration issues. A holistic selection approach can potentially reduce the mentioned issues and challenges [14]. However, a holistic digital technology selection approach requires a systematic framework to help both researchers and practitioners in selecting suitable technologies considering the requirements of companies [9]. Therefore, the purpose of this paper is to identify and analyze the critical factors of selecting and integrating digital technologies for efficiently benefiting data value chains for Smart Production. The critical factors are based on a technology selection and integration framework which was identified in our previous work [33]. Moreover, the critical factors are identified using a case study design of four independent manufacturing companies in different industries and of different sizes. The outline of this paper is as follows: Sect. 2 presents a literature review, while Sect. 3 presents the research methodology applied in this paper. The case study findings and discussions are described in Sects. 4 and 5, respectively. Finally, Sect. 6 concludes the paper by indicating some future work.

2 Literature Review 2.1 Smart Production Smart Production considers a desired state that manufacturing companies aim to achieve in terms of sustainability, value-driven production, and efficiency by integrating digital technologies into the production system [1, 4]. Accordingly, Smart Production requires a process innovation, i.e., “…the implementation of a new or significantly improved production or delivery method. This includes significant changes in techniques, equipment and/or software.” [14, p. 49]. Second, changes in production require a holistic approach. Therefore, production can be seen from a system perspective involving a combination of different elements such as people, production equipment, processes, information systems, and networks. In order to produce products, all these mentioned elements are interrelated in an organized but complex way [16]. For this reason, when making changes in one part of the system, it affects other parts of the system as well as sub-processes [17]. This highlights the complex nature of production and shows that the development towards Smart Production requires consideration of many parameters and their relationship to each other. Third, Smart Production is related to Industry 5.0, which focuses on human-centric and value-driven production [2]. Therefore, manufacturing

Critical Factors for Selecting and Integrating Digital Technologies

313

companies integrating digital technologies without addressing human aspects risk failing to achieve full integration of the technologies [18]. Examples of important human aspects include active participation of key stakeholders in problem-solving activities and developing capabilities to explore opportunities of using new digital technologies [19]. Fourth, Smart Production involves digitalization and connectivity to achieve a high level of smartness [5]. This requires using technologies with capabilities of managing digitalized and real-time data, i.e., digital technologies must be used [20–22]. Moreover, digital technologies must be integrated with existing elements in the production system to collect relevant data and transform this data into valuable insights to support actions and decisions in production [23, 24]. However, it is not enough to only have digital technologies installed in the production; it is how they are integrated within the production system to manage the key resource, which is the data, that will determine what benefits will be achieved [1, 12, 25]. Therefore, Smart Production requires having a holistic perspective on the entire data value chain, which includes the following phases [5]: – Data sources and collection. Refers to the points of origin where data is being generated and the methods used for collecting the data. – Data communication. Refers to the transportation of data from their sources to the places where data will be stored, processed, visualized, and used. – Data processing and storage. Data processing refers to the operations conducted to extract information from data. – Data visualization and usage. Refers to the techniques used to present data to endusers and the context in which data is applied to fulfill a purpose. 2.2 Digital Technology Selection and Integration In this paper, we use the term digital technology to cover the wide range of software and hardware technologies related to Industry 5.0 that enable digitalized and automated handling of data to optimize processes and increase value-creating activities [21]. The term covers both simple and extremely complex digital technologies where each digital technology possesses different capabilities [12] and requires consideration of different integration factors [21]. Examples of digital technologies include Cloud Computing, Artificial Intelligence, Digital Twin, Augmented Reality, and Simulation [12, 21, 26, 27]. However, manufacturing companies need more internal knowledge about what technologies to select and how they should be integrated to achieve desired benefits in production [28, 29]. Furthermore, several technologies must be combined with existing production equipment and systems to constitute the different functions in the data value chain – from data generation to data usage – in order to generate value [5, 12]. Previous research studies have analyzed the use of digital technologies for Smart Production from different perspectives. However, these studies present different ways of clustering and defining the same digital technologies [12]. The lack of consistency on what is meant by different terms, the capabilities of different technologies, and application areas, may lead to confusion. Moreover, the integration of digital technologies in production systems when building data value chains requires a high investment regarding people, processes, and technology both at a corporate level [29] and a supply chain level [30] in order to generate

314

N. Agerskans et al.

the desired deep and positive impact on business processes [31]. For this reason, decisions related to investments in digital technologies are strategic decisions that must be aligned with the business strategy [32]. However, the combination of the variety and complexity of evolving digital technologies, lacking internal understanding of digital technologies, and high financial investments, may lead to unprofitable investments if inappropriate digital technologies are selected [12]. A key challenge for companies is, therefore, to define a certain business case [21, 29, 30]. Consequently, it is challenging to make strategic decisions regarding the selection of digital technologies to ensure a desired positive impact on business processes [33].

3 Research Methodology This study follows a multiple case study design [34] based on two reasons. First, case studies are suitable when exploring novel subjects where increased theoretical insights are needed. Second, in case studies, empirically rich data are used, including the context and experience of practitioners. This was considered important to increase the practical relevance of the findings in this paper. A multiple case study involving companies from different industries and of different sizes was chosen to cover different contextual aspects that may contribute to the results (see Table 1 for more information about the companies and data collected). This was considered important to ensure results applicable to different types of companies in the manufacturing industry. Two of the cases have been followed over time, while two cases have been defined as reference cases. Combining the longitudinal cases with reference cases provided the opportunity to compare and contrast the findings with those from similar cases, allowing a more comprehensive understanding of the phenomenon studied. The data was collected using semi-structured interviews and workshops. To supplement the interviews and workshops, secondary data were collected from internal Table 1. Overview of the studied companies and data collected Company A

Company B

Company C

Company D

Railway

Casted components

Electronics

Company size Large

Large

SME

SME

Number of employees

348

58

45

Reference case

Reference case

Longitudinal case

Company overview Type of products

Heavy vehicles

4475

Longitudinal Longitudinal case case or reference case Data collection Interviews

11

3

3

23

Workshops

2

-

-

8

Critical Factors for Selecting and Integrating Digital Technologies

315

documents and the companies’ web pages. This helped expand the knowledge of the companies’ practices and validate the findings. The empirical data was analyzed using thematic analysis, which is a systematic approach for capturing patterns in qualitative datasets. The data analysis included an iterative comparison of the empirical data and literature by following six concurrent activities: (i) data familiarization, (ii) generation of codes, (iii) searching for themes, (iv) reviewing themes, (v) defining and naming themes, and (vi) writing the paper [35]. In the analysis, empirical data was analyzed and sorted using the four categories identified in the literature: data sources and collection, data communication, data processing and storage, and data visualization and usage.

4 Empirical Findings The empirical analysis presented in [14] showed that manufacturing companies are currently facing several challenges related to the selection and integration of digital technologies for Smart Production. To help manufacturing companies overcome these challenges, the concept of data value chain was proposed as an approach to map the current data flows in the production system and identify opportunities or improvements. The empirical analysis for this paper showed that the data value chain concept can be extended to a more hands-on framework to support manufacturing companies in selecting and integrating digital technologies. In this section, we first present the technology selection and integration framework, which is the outcome of the study in [14] but has been further developed in this study. Then we map two cases using the framework to illustrate some practical aspects of using the framework. Finally, we present identified critical factors for selecting and integrating digital technologies from the two cases studied in detail, including two reference cases. 4.1 Selection and Integration Framework The framework is composed of four technology clusters (TC): Data application, data communication, data storage, and Data processing. Each cluster represents a set of technologies with dedicated functions associated with the cluster definition. On the basis of our findings, the clusters need to be as abstract as possible and technology-agnostic to ensure that they encompass existing, developing, and potential future technologies. Our analysis shows that technologies for all four clusters must be carefully selected. These four clusters are the results of identified challenges within these four categories. TC 1 - Data application. This cluster represents a set of digital technologies that assist in not only generating and collecting data from the shop floor but also applying data as input to the machines, equipment, and operators on the shop floor. Some examples of digital technologies include sensors, valves, relays, and smartphones. TC 2 - Data communication. This cluster represents a set of digital technologies that transfer the data among units, e.g., from devices to other devices or systems. Depending on the environment and requirements imposed by the systems, different communication technologies can be used for this purpose. Examples of communication technologies in industry include WiFi, 5G, Ethernet, Industrial Ethernet, and short distance communication systems, e.g., USB and Bluetooth.

316

N. Agerskans et al.

TC 3 - Data storage. This cluster represents a set of technologies that store the data. Any data that is collected and communicated should be stored in a place before it can be used. In addition, any data that will be applied on machines, i.e., will be transferred to the machines, should also be stored in a place first. Therefore, having the technology to assist in the storage of data in a data value chain is essential. Examples of these technologies include local servers, edge servers, and cloud computing servers. TC 4 - Data processing. This cluster represents a set of technologies that perform an action on the data to generate valuable insight to support actions. Moreover, the technologies in this cluster consider aspects of the presentation and visualization of the data. Examples of these technologies include Artificial Intelligence (AI) and Machine Learning (ML) techniques for smart decision-making, traditional control functions on the control systems, data visualization, statistical analysis, and big data analytics. The presented clusters are contributing to implementing a full-fledged data value chain in smart production by handling generated data from the source until a place to use the data for visualization and decision-making. Therefore, to have a complete data value chain, at least one technology from each cluster should be selected.

4.2 Using the Framework In this section, two cases are presented to illustrate some practical aspects of using the digital technology selection and integration framework. Case company A describes an assembly line for circuit boards as an SME, while Case company D presents a production cell for components for heavy vehicles. Case Company A – Overall Equipment Effectiveness (OEE) Company A was a large company in the heavy vehicle industry with a focus on the design and manufacturing of construction equipment. The studied case considered the measurement of Overall Equipment Effectiveness (OEE), i.e., the measurement of production productivity, in order to support decisions regarding machine utilization and faults. Figure 1 shows the mapping of technologies and methods used in the current situation, current challenges, and the future state of this case. In the current situation, data was logged manually by the operators after each shift which increased the workload for the operators as well as the risk of human errors. Further, the logging was often incomplete and not always performed by each shift. Therefore, the quality of the collected data could be misleading, resulting in inaccurate calculations of OEE calculations. A contributing factor to incomplete and inaccurate data was the insufficient work culture on the shop floor. Since there had not been enough dialogue with operators on the importance of collecting data, other work tasks were often prioritized, which resulted in lacking data logging. A standardization of how to log data such as faults was also missing. Some operators first documented data using pen and paper before filling in the OEE software Shift log, while others documented directly in Shift log when the fault occurred. When documenting OEE data in Shift log, there was a free text box where the operators could write freely. However, the perception of machine fault was subjective and, therefore, hard to analyze. Further, the Shift log software was considered unreliable since it often crashed or did not respond properly. Moreover, it happened that operators accidentally

Critical Factors for Selecting and Integrating Digital Technologies

317

created a work log for the wrong machine or wrong shift. As a consequence, enormous amounts of data containing blank work logs were generated and stored, which could not be used. Due to this, there was a need to manually clean the stored data from these logs, which was time-consuming. Further, the visualization of OEE data, such as downtime, machine performance, machine utilization, and availability, was not clearly presented.

Fig. 1. Mapping of technology clusters for Case Company A - Overall Equipment Effectiveness (OEE).

To address the challenges identified in the current state, the IIoT platform PTC Thingworx1 was implemented to communicate, store, and process data. New machines with built-in compatible sensors and actuators enabled automated data collection. Further, some legacy machines remain in operation, and for these machines, additional Kepware2 and brownfield technology were integrated as middleware to enable interoperability with the IIoT platform. The data collected was pre-processed and filtered in order not to store data that cannot be used. To solve this, data lakes and Microsoft can be used to achieve a successful pre-processing of data. Even if OEE data was the focus of this study, the data lake could also be used for other purposes within the organization. The IIoT-platform was expected to generate a large amount of data. However, not all data needed to be stored, meaning that a lot of data would just be viewed for analytical purposes and not be stored over time. Case company A has a private cloud on-premises and with their servers on-site, meaning that the data generated was stored on the company’s own servers and, therefore, also protected from cyber-attacks from outside the company network. Further, real-time updates on relevant production data were available, as well as specified historical data from e.g., the last shift. Data processing was enabled by a Big Data Analytics tool provided by the IIoT platform. Moreover, the processed data generated by the IIoT platform was presented in a visualization application in the IIoT platform. Real-time updates were provided regarding the total number of produced 1 https://ttpsc.com/en/solutions/thingworx-iiot-platform/. 2 https://www.ptc.com/en/products/kepware.

318

N. Agerskans et al.

articles, total number of scraps, obtained quality, total running time of machines as well as total downtime. This data was structurally displayed by color coding the different statuses of a specific machine and the KPI performance. For example, an OEE percentage within the red spectrum was considered a bad OEE, yellow was acceptable, and green was desirable. Moreover, processed data was also visualized by various charts and graphs. Here, the user could choose what type of equipment or production area that should be displayed. Case Company D – Product Quality Case company D was an SME in the electronics industry, which specialized in both design and production of circuit boards. Production involved both surface mount and through-hole mount. This case study focused on decision-making related to product quality in the assembly process, which was the assembly process for surface-mounted products. More specifically, decisions related to product errors were studied. Product errors considered components that had not been assembled correctly. Figure 2 shows the mapping of technologies and methods used in the current situation, current challenges, and the future state of this case. In the current state, data was collected using an Automated Optical Inspection (AOI) machine and manually by operators. Problems related to the automated data collection by the AOI machine were data quality problems in terms of accuracy, timelines, and consistency. Accuracy problems were related to many false product errors generated by the AOI machine. This problem was caused by non-standardized data, the use of nonstandardized components in the assembly process, and the development process of AOI programs. Data quality problems related to timeliness were also identified. If there was a systematic error in the assembly process, it was considered optimal to stop the process immediately to avoid the re-work of products remaining in the work order. However, due to accuracy problems, there was a need to manually process and correct the data before it could be used to support decisions. Since this correction was performed after a production order was finished, this data was not delivered in time to support decisions in real time. Consistency problems were caused by the use of different terminologies and error codes used by the data generated from the AOI machine compared to the operators when manually logging data. The use of different terminologies could make it difficult for the decision-makers to interpret the data, which could lead to an increased time necessary for understanding the data or, in the worst case, misinterpretation of the data. Further, the AOI data was communicated using ethernet and displayed in real-time on a TV screen in the production area and was automatically stored in a SQL database. The AOI data was displayed in real-time, considered one product at a time, and, therefore, the operators had to monitor and keep track of the AOI data while performing other work tasks. However, since the data from the AOI involved many so-called false calls, meaning that the data provided by the AOI was not accurate, no analysis was done on the data. Manual data collection was performed by operators while doing visual inspections. This data was documented using pen and paper and, at a later stage, logged into an Excel sheet. However, manual data collection was time-consuming and could generate inaccurate data due to the human factor. A data quality problem related to manual data collected was incompleteness since the decision makers did not consider the data

Critical Factors for Selecting and Integrating Digital Technologies

319

complete for making decisions related to the root cause of a problem, i.e., important data was missing. In the case of a specific problem, data was studied by the operators. However, since data was presented in a non-user-friendly format in Excel, the operators found it challenging to understand the meaning of the data.

Fig. 2. Mapping of technology clusters for Case Company D – Product Quality

To address the challenges related to data application in the current situation, the implementation of new equipment would provide opportunities for collecting data at different stages of production. If data were to be collected after each process step, errors could be detected earlier in the assembly process, which could reduce re-work costs. This could be achieved by the implementation of a solder paste inspection (SPI) machine and a second AOI machine to replace the manually performed data generation and collection. To collect data during the production processes, sensors and actuators could be installed in existing production equipment. This would also provide quality control after each process step which would help discover quality problems as soon as they occur and facilitate faster identification of the cause of the problem. Further, to address the problems related to incorrect data generated from the AOI machine, the process of developing the AOI programs should be improved, and the possibility of using standardized data and components should be investigated. This way, all necessary data can be collected with the right quality. Regarding data communication, both WiFi and Ethernet will be used. To store the data, the existing SQL database and local server will continue to be used as there are no current problems related to these storage technologies. However, Cloud will be used for some of the data to enable easier data sharing. In order to convert the data collected into valuable insights that can be used for decision making, there is a need to integrate other technologies for data processing. For this, data analytics techniques can be used to extract useful information from collected data. To better visualize the data, statistic software can be used instead of Excel, which is reported to be not user-friendly. Suitable charts and graphs should be evaluated to ensure usability before implementation.

320

N. Agerskans et al.

4.3 Critical Factors of Selecting and Integrating Digital Technologies The empirical analysis showed that several critical factors should be considered related to the selection and integration of digital technologies for Smart Production. The identified critical factors were classified into different themes, which in turn were sorted into the following categories: (i) data sources and collection, (ii) data communication, (iii) data processing and storage, (iv) data visualization and usage, and (v) general critical factors. Data sources and collection includes critical factors related to generating and collecting data from its source of origin, where three main themes of them include the following. The first theme is defining data collection requirements. As presented in Sect. 4.2, many challenges related to the quality of the data were identified, and consequently, the data could not be used. Before selecting technologies and methods for data generation and collection, it is therefore important that manufacturing companies define the requirements of the data that should be collected. Aspects related to accuracy, completeness, and timeline should be considered. The second theme is to consider the people. An identified challenge was that operators were not comfortable with all methods for data collection, for instance, the use of cameras. Therefore, a critical factor is to involve the people who will be affected by new technologies in the investigation of suitable technologies. Another challenge was that the human factor could affect the quality of the data. For this reason, the human factor should be eliminated by having automated processes or mistake-proof systems for manual processes. The third theme was to evaluate limitations and constraints. Depending on what data is collected, there might be regulations that must be considered. For instance, if data is collected from operators as an object, regulations such as GDPR must be considered and evaluated to ensure the privacy of operators is not compromised. Further, standards related to data formats should be considered when exchanging data among equipment and systems. Standardization of the data format helps in the integration of technologies. The second category is data communication which covers themes related to transmitting the data from their sources of origin to the places where they will be stored, processed, and visualized. The first theme considers defining data communication requirements, which is essential to ensure that the company meets its needs in terms of data communication requirements. Four critical factors should be considered. The first factor is the required speed of the network. Second, interoperability requirements, i.e., what is needed to enable data exchange between systems and equipment. Third, accessibility should be considered, meaning when and from where data communication is needed. The fourth factor is structured data communication, which refers to attaching information such as time labels on the data. Another theme of data communication is related to how to enable interoperability. The empirical findings showed that it was important to involve external partners such as machine suppliers and system integrators. For aged production equipment, also called legacy systems, middleware interfaces, and brownfield technology could be used to enable communication among legacy systems. It can also be helpful to map current communication protocols used. Further, defining routines in case of disturbances was another theme. As it might happen that the communication network is down, hindering access to data, a critical factor is to define routines for how to proceed if the communication network is not working. For instance, how to restart the

Critical Factors for Selecting and Integrating Digital Technologies

321

system or who to contact internally and externally for support. Further, if the network cannot be used, there is a need to define routines for how to operate without the network. The third category is data processing and storage, which includes themes related to the operations conducted to extract information from data and the place where data is stored. The first theme is defining data storage requirements, where three critical aspects should be analyzed. Accessibility refers to those people who should be able to access the data. This may include people from different parts of the organization as well as external partners. This is an important factor, especially for identifying the authorization levels for various people involved. Another aspect is the duration of storing the data, which can vary depending on the use of the stored data. For instance, some historical data should be preserved for longer than a year, while some operation data can be stored for a shorter time. Further, security aspects of how data is stored is critical to ensure only authorized people have access to the data. This is also crucial to prevent any adverse attacks from collecting sensitive data. The second theme is defining data processing requirements, where the operations conducted to prepare data for analysis, also called pre-processing, are critical factors. Other critical factors include quality aspects of the analysis: accuracy, completeness, timeline, usefulness, and consistency of results. Visualization and usage of data is the fourth category which refers to themes related to how data is presented to end-users and the production context in which data is used. The first theme is defining visualization requirements. Fundamental for this theme is to define what data should be presented, i.e., the data content. This will depend on what decisions or actions should be taken based on the data. Our analysis also showed that all necessary data could be available. However, it was not presented in a format that was understandable for the end user. For this reason, a critical factor to consider is the techniques used to visualize the data. This can include graphs and charts, but digital technologies such as Virtual Reality and Augmented Reality can also be used. However, it is important to avoid “overloading” with too much data since this can obstruct the understanding of what the data means. The data, techniques, and technologies used should therefore be user-friendly. Furthermore, when considering the visualization of the data, it is important to consider if this data should be presented in real-time or not since this affects the selection of techniques and technologies. The second theme is to define data usage requirements. In order to make sure that data is applied to add value to production, three critical factors should be considered. First, a process or routine should be defined for how to use the data. Second, a timeline should be developed for when the data should be used. Third, it should be defined who is responsible for using the data in production. The last category is general critical factors which covers themes that could be related to all four previously mentioned categories yet not directly connected to only one category. The first theme is related to developing a data-driven organization. The empirical findings revealed that even if the technology is mature and ready to be integrated in the factory, a challenge is that people do not have the mindset and knowledge to work with the technology and data generated. Critical factors are, therefore, to have a clear strategy, the right competence, and leadership to develop a data-driven culture where people are curious to explore what can be achieved with digital technology and data. Furthermore, the way of working in Smart Production projects is another theme.

322

N. Agerskans et al.

Our findings showed that it is not enough to involve competence in technology. Rather, it is important to combine domain experts with technology experts with the support from good management, i.e., having cross-functional teams. If needed, the cross-functional teams can include external partners such as technology providers. Another important factor is to accept and learn from mistakes. When exploring new technologies or new areas to apply technologies, the integration will many times not go according to plan. Instead, many times unexpected challenges will be faced. Making mistakes and learning from them should therefore be seen as a part of the journey, both from management and those working in the Smart Production teams. This also requires agility in the way of working, which is another critical factor. The last theme is identifying areas for Smart Production, which refers to the factors to consider when deciding on what parts of the production should be digitalized. The first critical factor is to create an overview of current systems and technologies from a data value chain perspective. This will help get a holistic perspective of the current situation. The second factor is to have a clear problem definition describing what challenges should be addressed. Third, it should be clearly defined what benefits should be achieved in the desired future state. Fourth, before integrating digital technologies, Lean methods should be used to eliminate waste. Fifth, financial aspects such as the budget for the Smart Production project and estimated costs should be considered.

5 Discussion In this paper, we extend the work presented in [14] by using the proposed framework to map two cases in order to identify current problems and define a desired future state with a full-fledged data value chain. The framework was used in three steps. First, existing technologies and methods were mapped to get an overview of the current data value chain. Second, challenges related to existing technologies and methods were listed. The third step was to analyze identified challenges and propose solutions. In this step, it was important to have a holistic perspective of all technology clusters to ensure a synergizing effect of the different technologies in order to generate value. By having both a technical, process, and people perspective, we gained a deeper understanding of current challenges within this topic. Based on this, we provide a structure for managers, shop floor workers, and competence from IT to work together in cross-functional teams when selecting and integrating digital technologies to enable data value chains for Smart Production. As highlighted in previous studies on Smart Production, it is not enough for people within the IT department to select and integrate the technology, as domain expertise is also important. Our empirical analysis, therefore, showed that the framework can be used to help these individuals in creating a common and holistic overview of current digital technologies and systems used to manage data. This can help the practitioners to identify problems related to a possible mismatch between current technology and desired benefits or the way technologies are integrated within an existing production system. Moreover, the framework can be used for both a top-down and bottom-up approach for new digital technologies. As an outcome of using the framework to map two cases and analyzing two reference cases, 13 themes of critical factors for selecting and integrating digital technologies

Critical Factors for Selecting and Integrating Digital Technologies

323

were identified and analyzed in detail. The critical factors were related to technology, people, and processes, which highlights the importance of having a holistic perspective of the entire production system and its elements [16]. Since Smart Production is about digitalization and connectivity to achieve a high level of smartness [5], many technologyrelated aspects were included. Two of the more important technology-related themes of critical factors are enabling interoperability and defining data processing requirements. Furthermore, as Smart Production requires a human-centric perspective, the full potential of digital technologies cannot be achieved without considering aspects related to people [18]. To address this problem, several themes of critical factors are related to people aspects. Some examples of themes include the involvement of people affected by data collection and defining requirements for data usage. If these critical factors are considered, it can help ensure that data will be a useful resource to support actions and decisions [23, 24]. Process-related critical factors consider how end-users and people involved in Smart Production should work to enable a process innovation [15], i.e., Smart Production. Some examples of process-related themes include ways of working in Smart Production projects and identifying areas for Smart Production.

6 Concluding Remarks The research presented in this paper contributes to the ongoing discussion of selecting and integrating digital technologies to enable Smart Production [9, 13] by identifying current critical factors and presenting two Smart Production cases. More specifically, 13 themes of critical factors were identified related to the selection and integration of digital technologies, which can provide managers with more hands-on guidelines on what to consider when selecting and integrating digital technologies. These themes also showed that it is important to have a holistic perspective on all four technology clusters in the entire data value chain and have clearly defined goals and requirements for each technology cluster. In this study, both large and SME manufacturing companies from different industries were selected to show that the findings are applicable for different manufacturing companies. Planned future research involves a more detailed analysis of how different digital technologies should be selected and combined to achieve Smart Production.

References 1. Schuh, G., et al.: Industrie 4.0 maturity index: managing the digital transformation of companies. Munich (2020) 2. Zizic, M.C., Mladineo, M., Gjeldum, N., Celent, L.: From industry 4.0 towards industry 5.0: a review and analysis of paradigm shift for the people, organization and technology. Energies 15(14), 5221 (2022). https://doi.org/10.3390/en15145221 3. Khan, W., et al.: Industrial internet of things: recent advances, enabling technologies, and open challenges. Comput. Electr. Eng. 81 (2019) 4. Maddikunta, P.K.R., et al.: Industry 5.0: a survey on enabling technologies and potential applications. J. Ind. Inf. Integr. 26, 100257 (2022) 5. Tao, F., et al.: Data-driven smart manufacturing. J. Manuf. Syst. 48, 157–169 (2018)

324

N. Agerskans et al.

6. Curry, E.: The big data value chain: definitions, concepts, and theoretical approaches. In: Cavanillas, J.M., Curry, E., Wahlster, W. (eds.) New Horizons for a Data-Driven Economy, pp. 29–37. Springer International Publishing, Cham (2016). https://doi.org/10.1007/978-3319-21569-3_3 7. Miller, H.G., Mork, P.: From data to decisions: a value chain for big data. IT Professional 15(1), 57–59 (2013) 8. Faroukhi, A.Z., et al.: Big data monetization throughout big data value chain: a comprehensive review. J. Big Data 7(1), 3 (2020) 9. Silva, J., et al.: Decision making in the process of choosing and deploying industry 4.0 technologies. Gestão & Produção, 29 (2022) 10. Hamzeh, R., Xu, X.: Technology selection methods and applications in manufacturing: a review from 1990 to 2017. Comput. Ind. Eng. 138, 106123 (2019) 11. Santos, R.C., et al.: Selection of manufacturing technologies in the context of digital transformation: a systematic review. IEEE Eng. Manage. Rev., 1–15 (2023) 12. Klingenberg, C.O., M.A.V. Borges, Antunes, J.A.V.: Industry 4.0 as a data-driven paradigm: a systematic literature review on technologies. J. Manuf. Technol. Manage. (2019) 13. Choi, T.-M., et al.: Disruptive technologies and operations management in the industry 4.0 era and beyond. Prod. Oper. Manage. (2022) 14. Agerskans, N., Bruch, J., Chirumalla, K., Ashjaei, M.: Enabling smart production: the role of data value chain. In: Kim, D.Y., von Cieminski, G., Romero, D. (eds.) Advances in Production Management Systems. Smart Manufacturing and Logistics Systems: Turning Ideas into Action: IFIP WG 5.7 International Conference, APMS 2022, Gyeongju, South Korea, September 25–29, 2022, Proceedings, Part II, pp. 477–485. Springer Nature Switzerland, Cham (2022). https://doi.org/10.1007/978-3-031-16411-8_55 15. OECD and S.O.O.T.E. Communities, Oslo Manual (2005) 16. Bellgran, M., Säfsten, K.: Production development: Design and operation of production systems. Production Development: Design and Operation of Production Systems. 2010: Springer London. 1–340 https://doi.org/10.1007/978-1-84882-495-9 17. Gopalakrishnan, S., Bierly, P., Kessler, E.H.: A re-examination of product and process innovations using a knowledge-based view. J. High Technol. Manage. Res. 10(1), 147–166 (1999) 18. Lassen, A.H., Waehrens, B.V.: Labour 4.0: developing competences for smart production. J. Global Oper. Strat. Sour. 14(4), 659–679 (2021). https://doi.org/10.1108/JGOSS-11-20190064 19. Tortorella, G.L., Vergara, A.M.C., Garza-Reyes, J.A., Sawhney, R.: Organizational learning paths based upon industry 4.0 adoption: an empirical study with Brazilian manufacturers. Int. J. Prod. Econ. 219, 284–294 (2020). https://doi.org/10.1016/j.ijpe.2019.06.023 20. Berger, S., Denner, M.S., Röglinger, M.: The nature of digital technologies - Development of a multi-layer taxonomy. In: 26th European Conference on Information Systems: Beyond Digitization - Facets of Socio-Technical Change, ECIS 2018. (2018) 21. Ghobakhloo, M.: Determinants of information and digital technology implementation for smart manufacturing. Int. J. Prod. Res. 58(8), 2384–2405 (2020) 22. Parhi, S., et al.: Factors affecting Industry 4.0 adoption – a hybrid SEM-ANN approach. Comput. Ind. Eng. 168, 108062 (2022) 23. Manimuthu, A., et al.: Design and development of automobile assembly model using federated artificial intelligence with smart contract. Int. J. Prod. Res. 60(1), 111–135 (2022) 24. Tripathi, V., et al.: Development of a data-driven decision-making system using lean and smart manufacturing concept in industry 4.0: a case study. Math. Probl. Eng. 2022, 3012215 (2022) 25. Cioffi, R., et al.: Artificial intelligence and machine learning applications in smart production: progress, trends, and directions. Sustainability 12(2), 492 (2020)

Critical Factors for Selecting and Integrating Digital Technologies

325

26. Zheng, T., et al.: The applications of industry 4.0 technologies in manufacturing context: a systematic literature review. Int. J. Prod. Res. (2020) 27. Xu, L.D., Xu, E.L., Li, L.: Industry 4.0: state of the art and future trends. Int. J. Prod. Res. 56(8), 2941–2962 (2018) 28. Machado, C.G., et al.: Industry 4.0 readiness in manufacturing companies: challenges and enablers towards increased digitalization. Procedia CIRP (2019) 29. Sjödin, D.R., et al.: Smart factory implementation and process innovation. Res. Technol. Manag. 61(5), 22–30 (2018) 30. Kache, F., Seuring, S.: Challenges and opportunities of digital information at the intersection of big data analytics and supply chain management. Int. J. Oper. Prod. Manag. 37(1), 10–36 (2017) 31. Burmeister, Lüttgens, Piller: Business model innovation for industrie 4.0: why the ‘industrial internet’ mandates a new perspective on innovation (2015) 32. Bharadwaj, A., et al.: Digital business strategy: toward a next generation of insights. MIS Q.: Manage. Inf. Syst. 37(2), 471–482 (2013) 33. Kazemargi, N., Spagnoletti, P.: IT investment decisions in industry 4.0: evidences from SMEs. In: Lecture Notes in Information Systems and Organisation (2020) 34. Yin, R.K.: Case study research and applications: design and methods. Los Angeles: SAGE (2018) 35. Braun, V., Clarke, V.: Using thematic analysis in psychology. Qual. Res. Psychol. 3, 77–101 (2006)

Business Process Reengineering in Agile Manufacturing – A Mixed Method Research Khadija Lahlou1 , Khaled Medini2(B) , Thorsten Wuest3 , and Qussay Jarrar2 1 Mines Saint-Etienne, Saint-Etienne, France 2 Mines Saint-Etienne, Univ Clermont, INOP Clermont Auvergne, CNRS, UMR 6158 LIMOS,

42023 Saint-Etienne, France [email protected] 3 Industrial and Management Systems Engineering, West Virginia University, Morgantown, WV 26505, USA

Abstract. Business process reengineering has been successfully deployed to improve efficiency, reduce cost, and improve customer satisfaction and organizational agility. The aim of this paper is to provide an overview of the applications of BPR in the context of agile manufacturing and to identify the barriers and drivers of BPR adoption in practice. To this end, a mixed method is used combining literature review and interviews. Drivers, and technical and human barriers identified from the literature are assessed and extended with insights from practitioners. The results indicate also a general agreement on the relevance of BPR to agile manufacturing. The findings contribute to a better understanding of the opportunities and challenges associated with the implementation of BPR in agile manufacturing, and inform practitioners and decision-makers in their efforts to improve organizational agility and performance. Keywords: Business process reengineering · agile manufacturing · industry 4.0 · smart manufacturing · lean manufacturing

1 Introduction Business Process Reengineering (BPR) has been widely recognized as a process management approach for improving organizational performance by redesigning business processes [1–3]. With an increasing emphasis on agility across businesses [4], interest grows to integrate BPR with agile manufacturing. BPR has shown success in improving efficiency, reducing cost, improving customer satisfaction and organizational agility. However, the implementation of BPR initiatives is heavily influenced by a variety of organizational factors, such as culture [1], flexibility, and agility [5]. The aim of this paper is to provide an overview of the applications of BPR in the context of agile manufacturing as well as to identify the barriers and drivers of BPR adoption in practice. The results of this study will contribute to a better understanding of the potential benefits and limitations of BPR in the context of agile manufacturing and © IFIP International Federation for Information Processing 2023 Published by Springer Nature Switzerland AG 2023 E. Alfnes et al. (Eds.): APMS 2023, IFIP AICT 689, pp. 326–338, 2023. https://doi.org/10.1007/978-3-031-43662-8_24

Business Process Reengineering in Agile Manufacturing

327

will inform practitioners and decision-makers in their efforts to improve organizational performance. The remainder of the paper is organized as follows: Sect. 2 provides an overview of background literature. Section 3 presents the research methodology and material. Sections 4 and 5 report respectively on literature and practitioners’ insights into drivers and barriers. Results are discussed in Sect. 6. The paper ends with concluding remarks in Sect. 7.

2 Background Literature 2.1 Concepts Overview BPR is defined as “the fundamental rethinking and radical redesign of business processes to achieve dramatic improvements in critical, contemporary measures of performance, such as cost, quality, service and speed” [6]. BPR as a discipline has been around since the early 90s, promoted notably by Hammer (1990) and reinforced by the book by Hammer and Champy [7]. From a practical point of view, BPR is a structured approach to examine current processes, identify areas for improvement, and implement changes to create a more efficient and effective business operation [8]. This approach enables analyzing and redesigning business processes to improve organizational performance [2]. The ultimate goal of BPR is to enhance organizational agility by creating a flexible and adaptable business process architecture [5, 9]. Agility is defined as the ability of an organization to respond quickly and effectively to changing circumstances [5]. Agility as a production philosophy emphasizes flexibility, responsiveness, and adaptability to change [10]. This is in line with agile manufacturing’s early definition by Yusuf et al. [4], as a concept that involves quickly adapting to changing market demands and customer needs by improving the flexibility and responsiveness of manufacturing processes. Agile manufacturing promotes a focus on customer needs and a willingness to change business processes and organizational structures [5]. It involves the use of advanced technologies, such as real-time monitoring and control systems, to enable rapid and efficient production of customized products, with a focus on reducing lead times and improving customer satisfaction [10]. 2.2 Applications of BPR in Agile Manufacturing BPR has been used in agile manufacturing context with a focus on improving efficiency, reducing wastes, as well as enhancing customer satisfaction through improved product quality and faster delivery times. Some of the specific applications of BPR in agile manufacturing include enterprise information systems improvement, operations management, and smart manufacturing [2, 11, 12]. BPR has been widely applied in the enterprise information systems field [8]. Enterprise information systems are an essential component in the implementation of advanced manufacturing systems, however, they have some limitations such as lack of continuous integration, business intelligence, value based processes, and dynamic optimization. Qu

328

K. Lahlou et al.

et al. [8] promote the use of BPR to ensure the alignment of enterprise information systems with the requirements of intelligent manufacturing systems, such as autonomous operations, and sustainable values. BPR has several other application areas in operations management. In fact, it supports streamlining procurement and logistics processes, resulting in faster delivery times and reduced inventory costs within supply chains. BPR also improves scheduling and reduces lead times, resulting in faster product delivery to customers, as well as quality management (QM). Within QM, it supports the reduction of defects and improves product quality through the reengineering of quality control processes. In the area of lean-agile manufacturing, BPR helps identify and eliminate unnecessary activities, reduce cycle time, improve flexibility, and increase efficiency in manufacturing processes [2]. In the context of smart manufacturing, BPR was used to integrate RFID technology for warehouse management [12]. The authors put forth the benefits of BPR in terms of efficiency improvement, cost reduction, information system integration, communication and collaboration enhancement, automation of tasks and processes to increase productivity and reduce errors. Additionally, the authors note that BPR can be used to redesign organizational structures and workflows, which can help to optimize resource allocation and improve overall business performance. In a recent study, Al-Anqoudi et al. [11] underlined the role of machine learning as a novel solution approach to automate BPR.

3 Research Methodology and Materials The paper’s research methodology relies on a combined approach of literature review providing theoretical insights and interviews to confirm and expand on theoretical results and collect practitioners’ feedback. The review of the current literature aims to explore published BPR applications, drivers and barriers in agile manufacturing. To ensure rigor and transparency, this work is following Systematic Literature Review (SLR) methodology and utilizes PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines [13]. SLR method was chosen due to its ability to provide a systematic and unbiased overview of the available evidence on a given topic. According to Kitchenham and Charters [14], SLR is a research method that seeks to systematically identify, evaluate, and synthesize all available evidence on a specific research question or topic of interest. SLR involves a comprehensive search of relevant literature, critical appraisal of the quality of the studies, and a synthesis of the findings to answer the research question. The application of SLR involved several steps, including defining the research question, developing search queries, selecting studies based on inclusion and exclusion criteria, and critically assessing the quality of the selected studies [15]. This work aims to address the following question: “What are the barriers and drivers of BPR adoption in agile manufacturing?”. Search queries were build based on a combination from each of the following groups of keywords {“business process reengineering”, “BPR”, “agile manufacturing”} and {“implementation”, “barriers”, “success factors”}. The papers were retrieved from Scopus. Additionally, a search technique was employed that combines closely related terms such as “business process management” or “business process modeling” and “agility” or “flexibility”.

Business Process Reengineering in Agile Manufacturing

329

The resulting articles were screened and selected based on the defined inclusion criteria, i.e., published studies between 2011 and February 2023, studies focusing on BPR applications, barriers or drivers in agile manufacturing, studies using primary data collection methods such as surveys, case studies, experiments or interviews, and exclusion criteria, i.e., non-English articles, studies with no empirical investigation, studies not available in full-text format, duplicates or articles with significant overlap with other included studies. The initial search resulted in 36 articles. However, when considering the inclusion criterion of agile manufacturing context, the results yielded eight relevant articles, which is a relatively low number of relevant studies. In fact, the initial search returned studies that focused on agile BPR or how to use BPR with an agile method, rather than the application of BPR in agile manufacturing. To address this issue, we decided to use snowballing technique to expand our search beyond the initial set of studies. This technique involves examining the reference lists of relevant studies and using them to identify additional studies that meet our inclusion criteria [16]. This ultimately resulted in a total of 14 relevant articles. Figure 1 shows the location distribution of the identified papers. More than half (54%) of the identified relevant papers report on research conducted in Europe (e.g., France, UK). Asia represents around one third (31%) of the identified papers. The search resulted only in few papers stemming from the Americas and Africa, which each represent less than 10%. Publicaons distribuon 8% 7%

31%

54%

Europe Asia America Africa

Fig. 1. Publications distribution among continents

Complementarily, semi-structured interviews were used to confront the results of the literature review and to obtain additional information about the barriers and drivers of BPR adoption from practitioners and other experts. To this end, a questionnaire was built and used during the interviews. It is structured in three main sections, i) respondent profile (i.e., job category, activity sector), ii) drivers’ assessment, and iii) barriers’ assessment. The questionnaire was spread using social media networks (LinkedIn) as well as using emails among practitioners in manufacturing (e.g., production, design, supply chain, information system, quality). Respondents were asked to select the drivers and barriers they believe to have a significant impact on BPR adoption in agile manufacturing. In order to gain further insights into BPR relevance to agile manufacturing from a practitioners’

330

K. Lahlou et al.

points of view, a general question was added about the role of a process approach in improving agility within manufacturing organizations. To address this question, the respondents rate their level of agreement on a Likert scale ranging from 1 (“strongly disagree”) to 5 (“strongly agree”). The questionnaire results were analyzed using descriptive statistics, while the interview results were analyzed using content analysis. By triangulating the results from the literature review and interviews, the resulting understanding of the drivers and barriers to BPR adoption can be considered more complete. In total 14 responses were collected with respondents’ job titles include the following: consultant, project manager, functional team manager, and equivalent jobs. All respondents are involved directly or indirectly in the manufacturing sector. This sample size is assumed to be sufficient for the current study providing preliminary insights into the topic.

4 Literature Insights into Drivers and Barriers 4.1 Drivers to BPR Adoption Incorporating BPR into smart manufacturing shows promising results with regard to cost savings, increased efficiency, and reduced cycle time [6, 8, 9, 11, 12, 17]. BPR provides managers with an easy way to manage, implement, and accommodate internal and external changes, financial environment, requirements, and technological changes. BPR addresses changes in systems and workflows, with ontologies representing organizational knowledge. It develops a process-based cost/value and proposes strategic information in order to plan the strategic goals for the long term setting [3, 5, 8, 11, 12]. The application of BPR could eliminate wastes, unwanted activities, unwanted transportation, according to Six Sigma approach it reduces variance. BPR supports also Kaizen methodology in particular small and incremental changes continuously [11]. Among the drivers identified by Hussein et al. [3] is also increased competition within the marketplace To remain competitive, organizations must continuously improve their operations and processes. This, in turn, leads to pressure to improve efficiency and reduce costs. The desire to improve customer satisfaction is another significant driver for BPR. Organizations recognize that meeting and exceeding customer expectations is critical to their success. Furthermore, the need to comply with regulations and standards is driving factor for organizations to reengineer their processes. Additionally, the recognition of the potential benefits of technology and process improvements can be considered a driver for BPR. By leveraging technology and process improvements, organizations can enhance their operations, reduce costs, and improve customer satisfaction [3]. The drivers of BPR in the food and beverage industry (from a case study in Nigeria) include the need to improve organizational performance, enhance customer satisfaction, increase profitability, improve quality, and increase efficiency. Other drivers of BPR identified in the study include the need to respond to changing market demands, increase competitiveness, and address organizational problems such as duplication of efforts, poor communication, and lack of coordination among departments [18] (Table 1).

Business Process Reengineering in Agile Manufacturing

331

Table 1. Drivers of BPR adoption Drivers

References (examples)

Cost reduction

[2, 6, 8, 9, 11, 12]

Efficiency

[2, 6, 8, 9, 11, 12]

Time saving

[2, 6, 8, 9, 11, 12]

Change and adaptation support

[3, 5, 8, 11, 12]

Systems and workflow change support

[3, 5, 8, 11, 12]

Strategic planning support

[3, 5, 8, 11, 12]

Eliminate non-added value activities

[11]

Variance reduction

[11]

Support small and incremental changes

[11]

Organizational performance improvement

[3, 18]

Customer satisfaction improvement

[3, 18]

Profitability

[3, 18]

Quality

[3, 18]

Responsiveness to market evolution

[3, 18]

Competitiveness

[3, 18]

Problem solving at organizational level

[3, 18]

4.2 Barriers to BPR Adoption BPR is a complex and challenging endeavor that can be hindered by a variety of barriers. These barriers can be classified into two main categories based on results identified in the literature: i) technical barriers and ii) human barriers [1]. BPR initiatives face several technical barriers including lack of reliable advanced information technology (IT), unclear strategy and inadequate business case, unrealistic/unreasonable/unjustifiable expectations from the BPR project, incomplete BPR implementation, and insufficient authority given to the BPR team [17]. Furthermore, measuring the results of BPR initiatives can be challenging. The lack of appropriate tools and technologies can hinder the effectiveness of BPR initiatives [1]. Process complexity and regulations may also hinder BPR adoption [5, 11]. Other technical barriers identified by Qu et al. [8], include issues related to information systems and technology infrastructure, such as incompatible systems, lack of integration, and data quality problems (Table 2). Human barriers may include a lack of readiness, ineffective communication, poor leadership style, lack of top management commitment, poor collaboration in work, and resistance to change [2, 3]. Employee resistance to change, lack of understanding of BPR concepts and methodology, lack of employee involvement, and inadequate communication and training impede the success of BPR initiatives. In addition, there may be opposition from internal staff and union members, as well as a lack of interest from

332

K. Lahlou et al. Table 2. Technical barriers to BPR adoption Technical barriers

References (examples)

Lack of advanced IT

[1, 3]

Unclear strategy and inadequate business case

[1, 3]

Unrealistic/unreasonable expectations

[1, 3]

Incomplete BPR implementation

[1, 3]

Insufficient authority

[1, 3]

Difficult to measure results

[1, 3]

Complexity

[5, 11]

Regulations

[5, 11]

Lack of integration and data quality

[8]

many users, which can make it difficult to achieve buy-in and engagement with the initiative [1, 3]. Internal resistance hinders the work making it difficult to analyze and model several processes, which adds uncertainty to subsequent redesign proposals and related structural changes. CEOs may also level political interests with objective goals, leading to a misalignment of priorities. Moreover, the gradual shift of operational teams focus from core business process redesign to problem-solving and automation of existing processes may obstruct the implementation of BPR. Finally, there may be a tendency to hide failure signs, consciously or unconsciously, which can undermine the effectiveness of BPR initiatives [19] (Table 3). Table 3. Human barriers to BPR adoption Human barriers

References (examples)

Lack of preparation

[2, 3]

Ineffective communication

[2, 3]

Lack of management commitment

[2, 3]

Poor collaboration

[2, 3]

Resistance to change

[2, 3]

Lack of understanding

[1, 3]

Insufficient training

[1, 3]

Internal opposition

[1, 3]

Lack of users interests

[1, 3]

Change of priorities

[1, 3]

Business Process Reengineering in Agile Manufacturing

333

5 Practitioners Feedback on Drivers and Barriers 5.1 Practitioners’ Feedback on the Drivers to BPR Adoption Practitioners’ feedback on the drivers is summarized in Fig. 2. There is an agreement on quality improvement and time savings as driving factors of BPR implementation in agile manufacturing (86%). Efficiency (79%) and cost reduction (71%) are also seen as important drivers of BPR adoption. This is consistent with the literature indicating the relevance of BPR to process improvement [3]. The practitioners however less emphasize customer satisfaction than the literature suggests (64%). Similarly, only 57% of the respondents believe that change and adaptation are among BPR adoption drivers, while only 36% assume systems and workflow change support is among the drivers. Unexpectedly, only 21% see BPR as means for improving responsiveness to market evolution. These insights and discrepancies might be rooted in contextual information with the companies which may limit the use of BPR. Quality Time saving Efficiency Cost reducon Customer sasfacon improvement Profitability Organizaonal performance improvement Change and adaptaon support Eliminate non-added value acvies Variance reducon Strategic planning support Compeveness Systems and workflow change support Responsiveness to market evoluon Problem solving at organizaonal level Support small and incremental changes

14% 14% 0%

21%

20%

36% 36%

43% 43%

50%

40%

57% 57% 57%

64%

71%

60%

79%

86% 86%

80%

100%

Fig. 2. Perceived drivers to the Adoption of BPR in agile manufacturing

Overall, a majority of 84% of the respondents agree on the relevance of BPR in agile manufacturing which can be seen in Fig. 3. This validates the initial assumption of the author team.

79% 7% 5

4

0%

0%

7%

3

2

1

Fig. 3. Perceived impact of process approach on organizational Agility

334

K. Lahlou et al.

5.2 Practitioners Feedback on the Barriers of BPR Adoption Practitioners’ feedback on technical barriers is summarized in Fig. 4. The results highlight some of the key challenges that organizations face when implementing BPR initiatives in agile manufacturing. The extreme complexity of processes (36%), with unnecessary activities, underscores the need for a streamlined and efficient approach to process design and execution. The lack of a clear strategy (36%) and incomplete implementation (29%) indicate a lack of planning and preparation before engaging BPR projects. The lack of suitable advanced IT was cited as one important barrier which suggests that many companies are struggling to keep up with the rapid pace of technological advances and may be falling behind their competitors. Other barriers include not suited authority to conduct BPR projects and unrealistic expectations (14%) which is in line with the inadequate business case (36%) as barrier for BPR. 36%

Complexity

36%

Unclear strategy and inadequate business case 29%

Incomplete BPR implementaon

29%

Lack of advanced IT 14%

Insufficient authority

14%

Unrealisc/unreasonable expectaons Regulaons

7%

Difficult to measure results

7% 7%

Lack of integraon and data quality 0%

10%

20%

30%

40%

Fig. 4. Perceived technical barriers to the adoption of BPR in agile manufacturing

Practitioners’ feedback on human barriers is summarized in Fig. 5. The survey results highlighted the significant human barriers that organizations face when implementing process redesign initiatives in agile production. The lack of user understanding and interest (57%) underscores the need to better communicate the benefits of process redesign and how it can improve processes and drive business success. It is essential to involve employees affected by the project in the majority of the stages, including the design of the solution, in order to take their views into account and integrate them into the project, rather than limiting their involvement to the implementation phase. This would help tackle the problems of low users’ interest and poor collaboration highlighted respectively by 50% and 36% of the respondents. Unsurprisingly, ineffective communication (50%) and insufficient management involvement (50%) are among the drivers in this category. This underlines the importance of strong leadership and effective communication strategies to ensure that all stakeholders are on board with BPR initiatives and working toward common goals. Lack of training (21%) and therefore not prepared staff (36%) barriers should be also addressed in order to overcome the problems of resistance to change (36%) and internal opposition (14%) and ensure successful BPR implementation projects.

Business Process Reengineering in Agile Manufacturing

335

57%

Lack of understanding Ineffecve communicaon

50%

Lack of management commitment

50% 50%

Lack of users interests Poor collaboraon

36%

Lack of preparaon

36% 36%

Resistance to change 21%

Insufficient training 14%

Internal opposion 0%

Change of priories 0%

10%

20%

30%

40%

50%

60%

Fig. 5. Perceived human barriers to the adoption of BPR in agile manufacturing

During the interviews, experts and consultants highlighted the presence of internal and external environment barriers as well as economic factors in addition to the barriers identified in the literature. Internal environment barriers relate to ergonomic issues in the workplace, such as poorly designed workspaces and inadequate equipment. External environment barriers include regulatory restrictions, such as the General Data Protection Regulation (GDPR), environmental regulations, and laws that restrict business practices. Addressing these environmental barriers during the planning phase of BPR initiatives is critical to ensure that the implementation process is not hindered by unexpected barriers. The cost associated with implementing BPR initiatives can be a major barrier for many organizations, particularly in small and medium-sized enterprises (SMEs). In some cases, the cost of implementing BPR can outweigh the benefits, which can make it difficult to convince decision makers to support the initiative. BPR requires a significant investment of resources, including time, money, and personnel, which can be difficult for companies with limited resources. In addition, the benefits of BPR may not be immediately apparent, which can make it difficult to justify the cost of the initiative to stakeholders. All these factors are critical and interrelated. Addressing them requires a multifaceted approach that involves the participation and engagement of various stakeholders. By understanding these barriers and their underlying causes, organizations can develop effective strategies to overcome them and ensure the successful implementation of BPR initiatives. Organizations need to focus on careful planning, strategic decisionmaking, and the adoption of the appropriate technologies to overcome these barriers and successfully implement BPR initiatives.

6 Discussion This paper contributes towards the capitalization on the drivers and barriers of BPR implementation in agile manufacturing. This theoretical contribution is complementary with the literature, generally focusing on service industry, by specifically examining the context of agile manufacturing. Understanding the drivers and barriers will pave

336

K. Lahlou et al.

the ground for developing well-suited approaches, methods and tools in order to aid BPR implementation towards agility and performance improvement in manufacturing. Example of relevant works in this area are the process model fragments used by Erasmus et al. [20] to model manufacturing processes. Although research involving business process management in manufacturing is relatively scare, an emerging research stream can be identified in the literature [20, 21]. Industry 4.0 technologies and the growing body of literature addressing smart manufacturing are among the drivers of this orientation. Given the alignment of objectives of BPR and Smart Manufacturing, there are varieties of avenues for future research that are possible. To-date, there is a lack of tangible data-backed metrics that allow for a quantification of BPR drivers and barriers. Given the data-driven nature of smart manufacturing technologies and processes, this presents an opportunity to align the ability for data collection, data availability, and process execution with the existing results presented here to develop such metrics. Providing quantifiable and measurable metrics are expected to support the adoption as companies prefer data-backed insights when making decisions. Complementarily, another perspective involves the exploration of the synergies between knowledge management and BPR. For instance knowledge based systems could be used to improve efficiency of BPR projects. These later could lead in turn to enrich the knowledge base with insights regarding business processes, best practices and know-how. From a practical point of view, the results provide insights to decision makers in manufacturing organizations to better prepare for and execute BPR projects. For instance, BPR requires a comprehensive approach and collaboration between different departments and stakeholders as well as integration and orchestration efforts [21]. Understanding and prioritizing the drivers and barriers is expected to help optimize material and human resources use by targeting most critical improvement areas allowing for significant savings for manufacturing companies. The results provide also valuable input to risk identification during BPR projects planning by underlying pain points related to technical and human aspects. Looking at the results of the interviews of practitioners, some observations are worthy of discussion. Overall, it is interesting to note that the overall agreement of practitioners with the literature identified human barriers seems to be higher than it was for technical barriers. At the same time, the agreement with the identified drivers exceeds both. There are several possible reasons for these results. One possible explanation might be rooted in the implementation stage of the practitioners BPR journey. Assuming several are in the early planning stages, they will be more familiar with the drivers as they are relevant for decision making whether or not to engage in BPR in the first place. The barriers emerge during the later stage of the implementation and also depend on the context and role of the interviewee. We are not able to conclusively answer the question of why these discrepancies exist at this point and suggest further research is needed to better understand these discrepancies and the root causes. This is relevant also for developing adequate support mechanisms moving forward and establish a solid foundation for future research in the field.

Business Process Reengineering in Agile Manufacturing

337

Future research could help address the research perspectives identified in this study as well as overcome its limitations. For instance, the sample size of the survey respondents does currently not provide statistical validity of the derived insights.

7 Conclusion The paper combines a literature review and interviews with practitioners to derive insights into drivers and barriers for BPR adoption in agile manufacturing. Generally, the results highlight the importance of BPR in achieving agility in manufacturing. Implementing BPR is however hindered by several barriers which need to be addressed. The interviews revealed additional barriers which may indicate that further research is still needed to provide a more comprehensive overview of challenges, drivers and barriers ahead of BPR adoption towards more agile manufacturing. Acknowledgments. The authors would like to thank the interviewees for their time and valuable input. This work is partly supported by the German-French Academy for the Industry of the Future through RAMP-UP II project.

References 1. Omidi, A., Khoshtinat, B.: Factors affecting the implementation of business process reengineering: taking into account the moderating role of organizational culture (Case Study: Iran Air). Procedia Econ. Finance 36, 425–432 (2016) 2. Bhaskar, H.L.: A study of business process reengineering (BPR) in manufacturing industry perspective. Int. J. Manage. Behav. Sci. 1, 239–247 (2012) 3. Hussein, B., Bazzi, H., Dayekh, A., Hassan, W.: Critical analysis of existing business process reengineering models: towards the development of a comprehensive integrated model. J. Project Program Portfolio Manage. 4(1), 30 (2013) 4. Yusuf, Y.Y., Sarhadi, M., Gunasekaran, A.: Agile manufacturing: the drivers, concepts and attributes. Int. J. Prod. Econ. 62(1–2), 33–43 (2000) 5. Gong, Y., Janssen, M.: From policy implementation to business process management: principles for creating flexibility and agility. Govern. Inf. Q. 29(1), 561–571 (2012) 6. Siregar, I.W.: A literature review on business process reengineering. Am. Int. J. Bus. Manage. 7(9), 1–11 (2021) 7. Hammer, M., Champy, J.: Reengineering the Corporation: A Manifesto for Business Revolution. Harper Collins, New York (1993) 8. Qu, Y., et al.: An integrated framework of enterprise information systems in smart manufacturing system via business process reengineering. Proc. Inst. Mech. Eng. Part B: J. Eng. Manuf. 233(11), 095440541881684 (2018) 9. Meymand, M., Mollaei, E.: The effect of business process re-engineering factors on organizational agility using path analysis: case study of ports & maritime organization in Iran. Asian Econ. Fin. Rev. 4, 1849–1864 (2014) 10. Llamas, V., Coudert, T., Geneste, L., Romero-Bejarano, J.C., de Valroger, A.: Proposition of an agile knowledge-based process model. IFAC-PapersOnLine 49(12), 1092–1097 (2016) 11. Al-Anqoudi, A., Al-Hamdani, A., Al-Badawi, M., Hedjam, R.: Using machine learning in business process reengineering. Big Data Cogn. Comput. 5, 61 (2021)

338

K. Lahlou et al.

12. Bevilacqua, M., Ciarapica, F.E., Marchetti, B.: Integration of BPR and RFID technology in a process industry: spare parts warehouse management analysis. Int. J. RF Technol. 2, 205–223 (2011) 13. Wilson, P., et al.: How evidence-based is an ‘evidence-based parenting program’? A PRISMA systematic review and meta-analysis of Triple P. BMC Med. 10, 130 (2012) 14. Kitchenham, B., Charters, S.: Guidelines for performing systematic literature reviews in software engineering, Technical report EBSE 2007-001, Keele University and Durham University Joint Report (2007) 15. Fernández-Sáez, A.M., Genero Bocco, M., Romero, F.P.: SLR-Tool - a tool for performing systematic literature reviews. In: Proceedings of the 5th International Conference on Software and Data Technologies, vol. 2 (2010) 16. Wohlin, C.: Guidelines for snowballing in systematic literature studies and a replication in software engineering. In: Proceedings of the 18th International Conference on Evaluation and Assessment in Software Engineering, pp. 1–10 (2014) 17. Bhaskar, H.L.: Business process reengineering framework and methodology: a critical study. Int. J. Serv. Oper. Manage. 29(4), 527–556 (2018) 18. Olajide, O.A., Idowu, O.O.: Effects of business process reengineering on organizational performance in the food and beverage industry in Nigeria. J. Bus. Manage. Res. 3(1), 57–74 (2020) 19. Kock, N.F., McQueen, R.J., Baker, M.: BPR in the public sector: a case of successful failure. In: Gray, V., Llanes, V. (eds.) Proceedings of AIBSEAR Conference, pp. 485–490 (1996) 20. Erasmus, J., Vanderfeesten, I., Traganos, K., Grefen, P.: Using business process models for the specification of manufacturing operations. Comput. Ind. 123, 103297 (2020) 21. Reijers, H.A.: Business process management: the evolution of a discipline. Comput. Ind. 126, 103404 (2021)

Service-Oriented Architecture for Driving Digital Transformation: Insights from a Case Study Omid Maghazei1(B) , Marco Messerli2 , Thomas Gittler3 and Torbjørn Netland2

,

1 University of Bath, Bath, UK

[email protected]

2 ETH Zurich, Zürich, Switzerland 3 Maxon Group, Sachseln, Switzerland

Abstract. Organizations use digital technologies to stay competitive but it poses a range of challenges regarding the selection and integration of technologies. Having a clear enterprise architecture supports organizations choose the right technology, integrate new technologies, and capture latent values. This study applies the enterprise architecture framework TOGAF when designing and implementing two baseline service-oriented architectures for Maxon, a Swiss manufacturer of precision drives. The first use case aims to improve quality management. It describes an architecture for data analytics using Microsoft Azure cloud services, which exports the working dataset from a local SQL database to a cloud-based SQL server to further process the data. The second use case aims to improve warehouse logistics processes. It proposes the introduction of citizen development where non-IT employees developed a warehouse logistics application for smartphones. This study demonstrates how companies can leverage enterprise architecture to improve business operations. More specifically, this paper shows how serviceoriented architecture in a well-designed enterprise architecture enables use case development and scaling. Considering the trend towards (cloud) service adoption, this paper provides useful guidelines for the transformation. It presents a structured approach to design an enterprise architecture that aligns transformation efforts within the organization. Keywords: Enterprise Architecture · Service-oriented Architecture · Cloud Technology · Use Case Approach

1 Introduction Digital technologies, some of which are labeled as Industry 4.0 technologies, are increasingly offering companies new solutions to improve business performance. These technologies are driving companies to continuously innovate and transform their operations to remain competitive in an increasingly dynamic market [1, 2]. For instance, digital technologies could enhance manufacturing operations through increased efficiency, © IFIP International Federation for Information Processing 2023 Published by Springer Nature Switzerland AG 2023 E. Alfnes et al. (Eds.): APMS 2023, IFIP AICT 689, pp. 339–356, 2023. https://doi.org/10.1007/978-3-031-43662-8_25

340

O. Maghazei et al.

cost-effectiveness, flexibility, and reduced waste [3]. Leveraging digital technologies can also enable companies to respond more quickly to changing customer demands by reducing the time it takes to bring products to market [4]. Additionally, as markets become more dynamic and product lifecycles become shorter, increased agility is required of enterprises in order to stay competitive [5]. Firms face the challenge of how to effectively align the use of digital technology with their business strategy [6]. The market of available software is diverse, and each supplier offers unique features with different characteristics, which poses distinct challenges to select appropriate systems with different impacts on investment decisions. With increasing system complexity, it is possible for certain redundancies to arise within an organization, where two systems that achieve the same objective are in use. “Enterprise architecture” can help to align information technology (IT) with business strategy and unite efforts towards digital transformation by addressing these challenges and ensuring that the use of digital technology aligns with the overall objectives of the organization [7]. Enterprise architecture has a long pedigree in information systems (IS) [8]. This architecture defines and interrelates data, hardware, software, and communications resources, as well as the supporting organization required to maintain the overall physical structure required by the architecture” [9, p. 386]. Enterprise architecture is “an instrument to articulate an enterprise’s future direction, while also serving as a coordination and steering mechanism toward the actual transformation of the enterprise” [10, p. 7; emphasis in original]. Enterprise architecture takes a holistic view of an enterprise, including both IT and non-IT components. The general principles of enterprise architecture have been demonstrated to be effective in improving business performance, as the success of an enterprise is increasingly dependent on the design and integration of its various components [10]. Notwithstanding the potential of enterprise architecture, its application in operations management has not received sufficient research. The extant literature rather investigates enterprise architecture for IT adoption, integration, and standardization [11, 12]. More critically, with the proliferation of cloud technologies, the role of enterprise architecture is even more central to navigate the adoption and implementation process of new IT systems as well as leveraging the existing IT systems to extract new insights, which could aid companies achieve their business objectives. To benefit from the enterprise architecture to its full extent, it should be fully integrated in business operations and fully aligned with business objectives [13]. This study investigates how companies could combine and make a better alignment between enterprise architecture and service-oriented architecture (SOA). In particular, this work aims to address the following research questions: “How can the concepts of enterprise architecture and SOA be leveraged to achieve business goals?” And, “What are the opportunities and challenges involved in exploiting the SOA?”

2 Theoretical Background The beginnings of enterprise architecture can be traced back to the 1960s. John Zachman is considered the first to introduce enterprise architecture with a structured framework approach when he published his framework in 1987. The framework was a response to the

Service-Oriented Architecture for Driving Digital Transformation

341

emergence of new technology and was designed to support enterprises to best leverage them [14]. An enterprise architecture “is a conceptual framework that describes how an enterprise is constructed by defining its primary components and the relationships among these components” [15, p. 106]. More specifically, an enterprise architecture describes the relationships and dependencies between business processes and IT systems [16]. An enterprise architecture can be split into three major sub-architectures: (1) the business architecture responsible for defining strategy, processes, and functional requirements, (2) the IS architecture providing a foundation of applications to fulfill the business requirements, and (3) the technology architecture supporting the IS with the necessary infrastructure including hardware and software platforms, networks, and development platforms [13]. The process of creating value within an enterprise follows an Input-Process-Output model. An enterprise takes a given input from the external environment (e.g., customer demand), uses its architecture and processes to transform it, and produces an output usually in the form of products and/or services [15]. The goals associated with enterprise architecture could be classified in four categories: (1) creating a baseline to gain transparency and remove redundancies, (2) managing complexity by defining a new target architecture while future complexities are kept to a minimum, (3) driving transformation by providing a roadmap for the future, and (4) supporting innovations by providing a better alignment between the business and IT [17]. After all, the architecture forms an enterprise-wide standard, applicable to all relevant areas of business, ultimately improving efficiency and effectiveness of an organization. The enterprise architecture could provide direct benefits such as increased standardization and reduced costs that are quantifiable as well as improved decision making that is non-quantifiable [18]. Many of the enterprise architectures’ benefits are indirectly related to enterprise architecture such as improved alignment with partners and many benefits could also be achieved through the advisory services (and governance) enabled by the enterprise architecture [18, 19]. The organizational benefits of enterprise architecture could also be categorized into technology-related benefits such as reduced IT costs and increased IT responsiveness, and business-related benefits such as managerial satisfaction and long-term strategic gains like product leadership [20]. Despite this arguably convincing list of benefits from enterprise architecture, the technology management literature does not sufficiently discuss the use cases and the mechanisms through which it could enhance business operations, particularly when combined with cloud technologies. Enterprise architecture only conceptually defines how to align business and IT. SOA is an emerging methodology to facilitate the application architecture in enterprise architecture [21]. While enterprise architecture provides a strategic alignment between business and IT, SOA provides technological alignment [21]. SOA emerged as a solution to cope with the increasing complexity of diverse software systems [22]. Integrating a new application into the IT system landscape of an enterprise can be a substantial challenge, as traditionally, there is a need for interfacing the new system with all other existing ones [22]. The basic idea of SOA is that, to any user, all functions are handled as a service, and do not differ from application to application. A service is the basic building block

342

O. Maghazei et al.

of SOA, and has three features: an implementation (the code that defined its function), a contract (the terms and conditions under which it operates), and a service interface (the interface through which it is invoked) [23]. The role of SOA benefits the leveraging of data in two ways. First, it provides the means for simple data exchange and availability, thereby abolishing the need for extensive interfacing of applications and databases. Furthermore, it increases the accessibility of data for people without extensive IT knowledge, which is also known as democratization of data. Second, the broader accessibility of data together with the emergence of low- or no-code applications fosters greater collaboration between employees, such that everyone gets to contribute to improving productivity. The flexibility of services can also have drawbacks. Due to being web-based, all disadvantages are mainly related to being network-dependent. For instance, having millions of services exchanging data over the network can lead to certain performance and latency issues [24]. Other disadvantages may include the necessity of a large initial monetary or labor investment and complicated service management [24]. Services are standalone functions that operate as black boxes to consumers. They take in an input and return an output without the consumer being aware of the internal processing of the input. Services are commonly implemented using web services and communication using protocols such as Simple Object Access Protocol or Representational State Transfer, although these are not always required. While the individual objectives of each enterprise implementing SOA may be different, SOAs are characterized by interrelated properties such as discoverable and dynamic, loosely coupled, locationally transparent, diversely owned, and self-healing [25]. In essence, implementing a SOA simplifies the interconnection of applications within an enterprise. To ensure an architecture is used optimally to reach the business goals, a governance model should be established. Governing over an SOA can be seen similarly to IT governance, but in a tighter scope. The SOA governance mainly focuses on the problem areas including the compliance to legislation and audit requirements, budgeting, consequences of changing services, ensuring service quality, and change in attitude [26]. Creating an appropriate governance model is important to best control the usage and deployment of existing and new services. Despite these benefits, using SOA to drive digital transformation has not been adequately discussed. While organizations are striving for gaining competitive advantage through digital technologies, the diversity of solutions is posing challenges for companies to decide what technologies to adopt. There is also a lack of guidance on how the technological solutions could be linked with businesses’ strategies and enhance business operations. This study unpacks the role of SOA by building new use cases and discusses opportunities and challenges involved.

3 Research Method This study focuses on combining the tools of enterprise architecture and SOA for operations at Maxon Group (hereafter named Maxon for brevity). Maxon is a worldwide manufacturer and supplier of high-precision drive systems based in Switzerland. Maxon motors are used in various areas including industrial automation, mobility solutions, medical technology, and aerospace.

Service-Oriented Architecture for Driving Digital Transformation

343

3.1 About Maxon and Its Business Objectives Maxon is currently seeking to increase speed to market. Due to the tight governance of IT processes, certain processes such as approval processes and development cycles in the company can be unnecessarily lengthy. Maxon aims to innovate and develop new products or software quickly and efficiently to achieve a competitive advantage. Maxon’s second business objective is to improve operational productivity using digital technologies. Many manufacturing companies including Maxon already produce or have the potential to produce large amounts of data, which could be used to increase productivity. However, many are missing the base IT architecture to accommodate the analyses. 3.2 Research Approach To provide specific recommendations for Maxon and produce actionable plans for the proposed architectures, this study follows a use case-oriented approach to develop small examples of what a larger architecture should look like. Use cases are practical examples that illustrate how a concept can be applied to understand its results, which could also provide a reliable basis for business case analysis of emerging technologies [6]. To build the use cases, this study follows three steps [6]. First, the uses are identified though brainstorming with relevant actors. Second, the use cases are prioritized based on their ease of implementation and operational fit as well as their potential economic value and financial impact. Third, for the prioritized use cases, it produces a blueprint such that the use case could be replicated and scaled at other sites. Following this procedure, in collaboration with Maxon, two use cases were selected to help attain the set business objectives. 3.3 Selected Enterprise Architecture Framework: TOGAF There are various frameworks to design the enterprise architecture. Ideally, a framework should provide a method for describing both a baseline and target state [27]. The choice of framework depends on the requirements of an architect and the project’s characteristics. In this study, the TOGAF framework is selected, which is introduced and maintained by The Open Group and has been widely used by various companies. The Open Group is a global consortium of over 890 members including IBM, Intel, Huawei, and Fujitsu, as well as government agencies such as the U.S. Department of Defense, and the U.S. National Aeronautics and Space Administration [27]. The TOGAF Standard rests on three pillars: The Enterprise Continuum, the Architecture Repository, and the Architecture Development Method (ADM). The Enterprise Continuum is a categorization of relevant source material [27] and is used as a reference for standards in the industry. The Architecture Repository is the enterprise’s own documentation of existing architecture elements, models, and patterns; also including work already done by the enterprise [27]. The ADM follows an iterative process that can be repeated for each new element in the enterprise architecture and includes eight phases [27]. Table 1 provides the details and objectives of the ADM’s each phase. These phases can be customized to fit the needs of an enterprise, so they are not necessarily a universal approach. The enterprise architecture is developed by repeatedly executing the ADM, adding more elements to the architecture [27].

344

O. Maghazei et al. Table 1. ADM phases’ details

Main phases

Objectives

Phase A: Architecture Vision

Develop a vision of capabilities and business value to be delivered by the architecture, and to obtain approval for that vision

Phase B: Business Architecture

Define the target business architecture, begin creating architecture roadmap based on current and target state

Phase C: Information Systems Architectures Develop the target information systems architecture enabling the business architecture from phase B, which involves defining the target data architecture and application architecture Phase D: Technology Architecture

Create a technology architecture that supports the information system, data, and application architectures

Phase E: Opportunities and Solutions

Create initial draft of completed architecture roadmap. Conduct initial implementation planning and identify deliverables for the architecture defined in the previous phases

Phase F: Migration Planning

Determine how to move from the current state to the target state, which involves finalizing architecture roadmap and creating implementation/migration plan

Phase G: Implementation Governance

Ensure conformance with target architecture. Identify necessary skills and resources to facilitate and guide the development and trace implementation quality

Phase H: Architecture Change Management Establish procedures for managing change to the new architecture and ensure enterprise architecture capabilities meet requirements

4 Use Case Development This section draws on the two selected use cases’ development process following the TOGAF framework. The first use case was data analytics using Azure cloud services. The second use case was citizen development using Microsoft Power Apps. 4.1 Use Case 1: Data Analytics using Azure Cloud Services The first use-case described an architecture for data analytics using Microsoft Azure cloud services, more precisely the implementation of a report containing a statistical analysis of motor speed measurements. The architecture exports the working dataset from a local Structured Query Language (SQL) database to a cloud-based SQL server, where Azure services such as Databricks and Data Factory further process the data.

Service-Oriented Architecture for Driving Digital Transformation

345

The first use case was intended to support the business goals of improved productivity and technology leadership. Currently, employees at Maxon conduct various data analyses using scripting languages such as Visual Basic (Excel). These solutions are unstandardized and are usually locally stored on the author’s computer. The drawback is that the collaboration between people is cumbersome and in the case of the departure of an employee, there is a great risk that all the knowledge contained in their reports is lost, and knowledge must be reacquired or written off. The goal of this use case was to define a standardized way of analyzing data that is contained in the organization’s various databases. Having an enterprise-wide standard architecture allows Maxon to come closer to becoming a technology leader and leverage the architecture for machine learning and artificial intelligence (AI) applications in the future. See Table A-1 in the appendix for the details of the first use case development process. Example Analysis: Motor Speed Measurement Report This section showcases the example report created to demonstrate the capabilities of the proposed architecture. In a Python notebook in Azure Databricks, motor speed measurement data from the last seven days is used to calculate statistical measures (standard deviation, mean, median, skewness, kurtosis) grouped by article and production order. The resulting figures are written into an Azure SQL table and visualized in Power BI. Only the measurements which lie in the specified tolerance to pass the test are visualized, as those that are outside the tolerance were discarded regardless. The goal is to find anomalies in the distribution of measurement values, which in the past has led to large quantities of motors being recalled due to an issue that could not be identified by looking at pass/fail rates alone. The statistical measures were calculated to aid in the selection of production order or article. An example usage of the report would be to sort the results by ascending kurtosis, and analyzing those groups with the lowest values, as they represent the flattest distributions. To further analyze a chosen group of values, the user executes a drillthrough in the report, opening the page showing the details of the selected group, shown in Fig. 1.

Fig. 1. Scatterplot of all measurements over time, outlier measurements are marked red. (Color figure online)

The purpose of this report was to address a previous issue involving product recalls. The need to address these recalls necessitated the operationalization of this report, which

346

O. Maghazei et al.

justifies its business value. This example demonstrates the capabilities of Azure services in manufacturing operations. The report’s clear and easily understandable content has contributed to its approval within the organization. 4.2 Use Case 2: Citizen Development using Microsoft Power Apps The second use-case proposed the introduction of citizen development to Maxon, where non-IT employees are given control over small-scale application development. To give an example of this, a warehouse logistics application for smartphones was developed. This app contained four functions to improve warehouse logistics processes for the Parvalux factory in the UK. The second use case was intended to address the business goal of increasing speed to market, together with technology leadership. There is a growing interest in creating small-scale applications. Traditionally, users that want such an application developed must file their requests with the IT department, which then commissions a software developer to implement the solution. This middle-man situation combined with the limited resources of software development and high volume of requests can cause a severe slowing of solution deployment and slowing speed to market. Proposing an architecture that enables non-IT specialists to develop their own apps (i.e., citizen development) could remove these limitations. Citizen development focuses on small to medium size applications. The chosen platform for citizen development was the Microsoft Power Platform, more specifically Power Apps. It was determined that implementing a Power App for warehouse logistics at Parvalux (a Maxon subsidiary located in the UK) would serve as a valuable and impactful case for the use of Power Apps. See Table A-2 in the appendix for the details of the second use case development process. Example Power App: Parvalux Warehouse Logistics This section presents the warehouse logistics application utilized by Parvalux to streamline their factory migration process. The app does not introduce any new functionality that is not already present in the ERP system. However, it significantly streamlines processes by eliminating unnecessary information, reducing the time required to use the ERP system, and making the application accessible on smartphones, eliminating the need to physically walk back and forth between a computer and the warehouse. As defined in the requirements, the app has four functions accessible from the “Home” screen: inventory count, inventory disposal, warehouse part transfer, and work order picking.

5 Discussion This study shows that there is potential to drive digitalization without a large investment of time and money, just by properly leveraging the existing systems for previously unused features. Even though the shown examples were very specific, the underlying architectures are customizable to fit into diverse situations, due to standardizations in the way data sources are connected.

Service-Oriented Architecture for Driving Digital Transformation

347

5.1 Opportunities Many organizations like Maxon are aiming at utilizing the available data, as well as generating new data to make informed decisions. However, these efforts are fragmented and inconsistent, which results in progress being limited to the individuals who are developing their own solutions rather than benefiting the organization as a whole. SOAs could coordinate such efforts. For instance, with the implementation of the Azure architecture, a scalable, flexible, and universally accessible platform to unite Maxon’s analytics activities was developed, which created a foundation that can accommodate future additions of cutting-edge techniques such as machine learning and artificial intelligence. The implementation of further analytics technologies will enable Maxon to make more data driven decisions, which can result in fewer quality issues, better utilization of machines and staff, or other productivity-related improvements. Utilizing cloud resources is in line with industry trends, which point towards increased adoption of cloud services. In particular, for small to medium-sized organizations like Maxon, cloud services relieve the already busy IT department of many responsibilities, and the costs of renting resources in the cloud are expected to be lower than installing local hardware and hiring local staff to maintain it. The risks of storing company data in cloud services can be mitigated by carefully controlling what data is stored and governing over the access rights. Low-code development technologies are also gaining traction across the worldwide market. These technologies include both workflow automation services like Power Automate, and low-code application platforms like Power Apps, enabling faster delivery of customizable business applications. Such technologies will lead to a higher digital literacy among non-IT professionals, empowering them to automate their own processes and deliver value using digital technologies. The implementation of Power Apps provides potential for streamlining many processes involving ERP handling similar to the Parvalux logistics app. Currently, the ERP system is most often used with a desktop client, which contains many functionalities. These clients are not meant to be replaced, as certain situations still require full control and information. Regular tasks involving data handling do not require the complex interface a full client contains. By packaging the necessary information into lightweight (mobile) apps, employees could carry out their tasks much more quickly and efficiently, as has been observed at Parvalux. A side benefit is that using the application programming interface (API) to access the ERP as opposed to individual licenses brings potential cost savings, as it removes the need for each employee to have a license for potentially only a handful of transactions they use. This could enable setups, where there is shared hardware running a Power App. 5.2 Challenges The implementation of such use cases also poses certain security and compliance/audit concerns. For instance, during testing of the API connectivity to Power Apps, the IT department was hesitant to allow non-personalized access to the data due to security issues. Allowing non-personalized access to the data poses certain threats, such as the possibility of unauthorized users hiding behind anonymity or the inability to track individual department activity. To address these security issues, it is necessary to implement

348

O. Maghazei et al.

a well-designed access control concept, and to reimagine auditing through the use of tools like the center of excellence, where apps and connectors can be organized into groups and monitored collectively. Another potential challenge of implementing new technologies is the increased complexity of available tools to the employees. In the context of the cloud analytics use case, it can be argued that moving the analytics efforts to the cloud is only a minimal increase in complexity that brings more benefits than drawbacks. In the case of citizen development, there might be a risk that the decentralization of development can lead to a confusing environment for employees, where there are many scattered tools all over the place. To this end, the counterargument would be that the existence of localized tools creates a situation that already has a similar nature to the stated problem. Aggregating every tool on a universal platform, allows for easier governance, and is generally beneficial for overseeing all available solutions. 5.3 Practical Implications Conducting data analysis requires access to both data and skilled employees [28]. In seeking a suitable use case for analytics, several challenges were encountered related to data quality and accessibility. These challenges highlight the importance of ensuring that the necessary data is available and in a usable state and format before any useful analyses can be conducted. In addition, the use of advanced techniques such as artificial intelligence is dependent on the quality of the available data. Poor quality data can lead to incorrect or misleading insights, which can have unintended consequences. Moreover, if an organization lacks employees who are educated in advanced technology, access to advanced analytics will be limited. To address these challenges, organizations may need to consider options such as training existing staff or recruiting new employees with the necessary skills. The feasibility of these options will need to be carefully evaluated. An advantage of the proposed architectures is that they do not require an entire reimagination of day-to-day processes. Analytics allows shopfloor supervisors to make changes to their planning with the help of data driven decisions, but it does not change the manufacturing processes as such. Power Apps can help improve or replace the way employees currently interact with systems like ERP, but the underlying data is the same. This means that the organization is not entirely dependent on these tools, which in case of downtime or failure of the cloud services does halt all operations. On the other hand, the supplementary nature of these systems could be seen as an extra burden for the organization to maintain. New employees would need to be introduced to another platform, adding on to the complexity of their jobs. Because the IT management stays in charge of the administration, security, and audit of IT resource usage, it is imperative to have their support when introducing new projects. Without IT buy-in, implementing digital solutions is an up-hill battle at best, if not impossible. In the context of this study, the lack of IT buy-in led to the failure of accessing the productive ERP system, which forced a rethink in security aspects of using APIs. To achieve the highest level of adoption within the organization, it is recommended to prioritize implementing use cases that will cause the highest possible improvement to the business. First, focusing on maximum impact directs resources towards projects

Service-Oriented Architecture for Driving Digital Transformation

349

with the highest potential return on investment, which also helps to stay within resource constraints such as time and budget. Second, the most impactful use cases tend to involve a larger number of individuals in the organization, which promotes the propagation of knowledge and adoption of the platform. This also fosters collaboration and innovation, as more people contribute to a larger pool of ideas and knowledge. The siloed nature of the organization causes the IT department to lack knowledge of processes, and operations to lack knowledge of digital technologies. To enable a transition towards more collaboration between the two departments requires staff that is both knowledgeable enough in digital technology and is also able to understand the requirements of operations. Deploying transformation talent could include creating a team that solely focuses on aiding employees in the organization to improve their processes by leveraging Power Apps. This team should be familiar with the technology behind Power Apps, but also gather requirements from the business to develop solutions in a highly agile fashion. Being able to both communicate with IT management and shopfloor employees at Parvalux contributed to the successful process improvement in a short time. The task of these “digitalization leaders” is to bridge the gap between the current silos by fostering capability building within the organization, as well as being able to communicate with different levels in the organizational hierarchy. 5.4 Limitations and Outlook This study only implemented rather simple use cases, which may not fully capture the complexity of larger projects employing systems that do not fulfill all dependencies. This work was conducted from the perspective of a medium-sized organization, which means that the findings may not be fully representative of the experiences of other organizations that differ in size or sector. Lastly, there is no quantification of the benefits for implementing the architectures. Questions relating to the monetary value of the technologies were only qualitatively described. There are also several aspects that were not discussed in-depth, including considerations relating to cyber security, the development of a concrete business case, the importance of service lifecycle management, first/second level support of applications, and the potential impact of external factors (e.g., the rate of technological change and the availability of skilled employees) on the success of such digital transformation initiatives. It is suggested that future research continue to explore these aspects in greater depth.

6 Conclusion This study has resulted in the implementation of two baseline architectures that have the potential to improve existing processes in various settings. In addition to enhancing processes, new process opportunities were identified that had previously been unexplored. This creates the opportunity to not only make existing processes more efficient, but also find new process opportunities. By using an enterprise architecture framework such as TOGAF, a robust architectural foundation was created that forced the consideration of all necessary aspects for the implementation of the chosen technologies. This is important, as it forces the organization to not focus solely on the individual technologies,

350

O. Maghazei et al.

but the architecture as a whole and how it is integrated into the existing organization. Documenting the exact architectures allows them to be reproduced and expanded in the future.

Appendix: Use Case Development Phases See Tables A-1 and A-2. Table A-1. Data analytics use case process development. Main phases

Description

Architecture Vision

In collaboration with the management at Maxon it was decided to focus on a singular analytics application. After brainstorming for possible applications, one promising application was chosen concerning the analysis of end-of-line quality data, more specifically motor speed measurements. The architecture vision was to implement an analytics architecture leveraging Microsoft Azure cloud services

Business Architecture

Because analytics at Maxon is currently an unstandardized activity, there is no exact business architecture per se. However, analyzing data generally follows a similar pattern: data acquisition, data analysis, and applying certain models or calculations to the dataset. The implementation of the use case mainly affects the processes of analysis and consumption of the data

Information Systems Architectures In the current state, the generation phase of the process yields a set of measurements taken by employees using specialized Q-Tools on industrial PCs. These measurements are subsequently stored in SQL tables in an on-site SQL server. To facilitate an analysis, working datasets are extracted from the SQL tables using SQL queries and exported as CSV or Excel files. Data analysts then utilize various scripting languages, such as Python or VBA, to analyze the data. The results of the analyses are often shared with others using diagrams in Excel or PowerPoint, and occasionally also paper based (continued)

Service-Oriented Architecture for Driving Digital Transformation

351

Table A-1. (continued) Main phases

Description In the target state, for greater standardization, the “analyze” and “consume” processes will be changed to leverage Microsoft Azure cloud services. The on-site master database of measurement data will remain established as the master data storage for security purposes and master data governance. The new element in the “store” process is that some of the working data is now connected to Azure SQL databases. In this situation, migration of existing data is not necessary, because the master data stays in place, while the cloud only handles working data, which can be restored using the connection to the master database. Storage in the Azure cloud seamlessly connects to the other used services, namely Azure Data Factory for ingestion (replacing manual exports), and Azure Databricks for Analysis (instead of local Python scripts)

Technology Architecture

The technology used in this architecture falls into three categories. First, On-Site Technology, which represents the currently available technologies used. Second, Cloud Services, which perform operations on data using Microsoft Azure resources. Third, Connecting Services, which form a link between two application services

Opportunities and Solutions

The most important mindset that must change within the organization is to move away from localized solutions toward an enterprise-wide, uniform solution. Companies should use the cloud services provisioned for the organization. The main advantages of this approach include increased collaboration opportunities by removing the need for installation and maintenance of software, scalability when demand for resources increases, and flexibility from taking an incremental approach to grow the architecture

Migration Planning

The implementation of the new analytics architecture will be a greenfield project, as there are no pre-existing systems or software to modify or migrate. Migration to the Azure architecture should be executed incrementally. The main task is setting up an appropriate Azure environment to facilitate the incoming analytics solutions (continued)

352

O. Maghazei et al. Table A-1. (continued)

Main phases

Description

Implementation Governance

Because cloud resources do not have to be maintained by in-house staff, there is no direct cost associated with salaries or workforce. However, the usage of cloud resources is billed based on the type and duration of resources used. Besides the programming skills for authors of analyses, a key skill that may not yet be present in the organization is the know-how on provisioning, configuring, and using cloud resources

Architecture Change Management The value behind using this cloud architecture is the standardization and centralization of all analytics activities in the organization. With one platform, people in the organization know where to conduct their projects. These solutions can be efficiently scaled up and are more performant than localized solutions. Azure services meet industry cyber security standards, which simplifies compliance concerns. The first risk concerning the usage of Azure is the cost. Although it is difficult to judge whether using cloud resources provides more benefits than costs increased use should result in lower average costs per solution. Secondly, because using cloud resources that are managed and maintained off-site, there is the risk of operational downtime. Thirdly, although Azure implements various industry-standard security features, data privacy and cyber security must be ensured, particularly concerning sensitive data storage in the cloud. Finally, the risk within the organization is failing to properly adopt the new architecture and the associated technology, which can be mitigated by establishing a comprehensive communication plan and driving the adoption with dedicated staff and culture

Service-Oriented Architecture for Driving Digital Transformation

353

Table A-2. Citizen development use case process development. Main phases

Description

Architecture Vision

The vision for the introduction of Power Apps into Maxon includes the integration of business systems and customized app development. The stakeholder groups for this use case were the IT department, operational management, and all employees

Business Architecture

At Maxon, the process for obtaining custom tools in operations involves submitting a request to the IT department. Then the IT department commissions a software developer to implement the requirements. However, due to the limited capacity and the high volume of tasks in the IT department, this process can have lengthy lead times. Introducing Microsoft Power Apps could enhance small-scale application development. Operations and IT will no longer be siloed, when the development cycle is firmly in the hands of the IT department

Information Systems Architectures Power Apps can connect to various data sources using custom connectors, and do not alter the underlying data architectures of these systems. Custom connectors use services, which are programmed by the supplier of the software. As a result, data management, migration, and governance practices for the connected systems remain unchanged. It is then not necessary to revise the data architectures of enterprise systems, such as ERP or CRM for this use case. However, when connecting to services such as APIs, it is important to ensure that their design preserves and uphold the business logic of the systems. Regarding the application architecture, the Microsoft Power Platform provides an integrated ecosystem of cloud applications and tools for developing, testing, and deploying apps Technology Architecture

The technology taxonomy for this use case consists of three categories: data sources (e.g., ERP), data connections, and application services. The Microsoft Power Platform includes most of the technology architecture, except for APIs of data sources. It is important to ensure the API availability and quality of the system to be connected. This means that the API must contain the necessary functions and uphold the business logic as if the transactions were done in the system itself. Modern ERP systems might come with a standard API that simply needs to be enabled by an administrator

(continued)

354

O. Maghazei et al. Table A-2. (continued)

Main phases

Description

Opportunities and Solutions

The major corporate change attributes relate to the willingness of the organization to adopt the Microsoft Power Platform. First, it is important to secure buy-in from top management. Second, effective communication through engagement and collaboration as well as training are crucial for the successful adoption of the platform. Third, flexibility is important for successful adoption, which includes being open to an initial investment of time and resources, as well as the willingness of employees to acquire new capabilities and responsibilities

Migration Planning

This migration plan was created for the implementation of the Parvalux logistics app. This plan includes: 1) defining process and assessing business needs and objectives, 2) planning Power App solution, 3) setting up dependencies, 4) building app, 5) deploying, and 6) monitoring and governance

Implementation Governance

To use Power Apps, the most important technical resource is APIs. The desired system should have an active API. From a financial standpoint, deploying certain Power Apps may involve additional licensing costs. Despite Power Apps’ low-code development environment, certain technical skills for those developing the apps are needed, which can be developed quicker than true programming skills

Architecture Change Management Power Apps can provide value to an organization in several ways. First, Power Apps can help achieve the business goal of increasing speed to market for software solutions. Second, Power Apps can improve user satisfaction by offering tailored solutions that streamline processes and simplify workflows, potentially setting the organization apart from its competitors as a technology leader. Third, there may be opportunities for cost savings using APIs, which can eliminate the need for individual licenses for systems such as ERP. A central risk in using Power Apps and APIs is cyber security. When dealing with sensitive data, there is a risk of unauthorized access. Another risk is the increasing complexity of having many business functions contained in Power Apps. It can occur that certain apps get forgotten, or the development of a similar or identical app happens within the organization. This risk should be mitigated with effective governance. By having the right documentation within the organization and a proper governance setup, unused and duplicate apps can be identified and removed

Service-Oriented Architecture for Driving Digital Transformation

355

References 1. Fitzgerald, M., Kruschwitz, N., Bonnet, D., Welch, M.: Embracing digital technology: a new strategic imperative. MIT Sloan Manag. Rev. 55(2), 1 (2014) 2. Romero, D., et al.: Advances in production management systems: issues, trends, and vision towards 2030. In: Goedicke, M., Neuhold, E., Rannenberg, K. (eds.) Advancing Research in Information and Communication Technology. IAICT, vol. 600, pp. 194–221. Springer, Cham (2021). https://doi.org/10.1007/978-3-030-81701-5_8 3. Maghazei, O., Netland, T.: Implementation of Industry 4.0 technologies: what can we learn from the past? In: Lödding, H., Riedel, R., Thoben, K.D., von Cieminski, G., Kiritsis, D. (eds.) Advances in Production Management Systems: The Path to Intelligent, Collaborative and Sustainable Manufacturing, vol. 513, pp. 135–142. Springer, Cham (2017). https://doi. org/10.1007/978-3-319-66923-6_16 4. Feng, T., Sun, L., Zhu, C., Sohal, A.S.: Customer orientation for decreasing time-to-market of new products: it implementation as a complementary asset. Ind. Mark. Manage. 41(6), 929–939 (2012) 5. Sambamurthy, V., Bharadwaj, A., Grover, V.: Shaping agility through digital options: reconceptualizing the role of information technology in contemporary firms. MIS Q. 27, 237–263 (2003) 6. Maghazei, O., Lewis, M.A., Netland, T.H.: Emerging technologies and the use case: a multiyear study of drone adoption. J. Oper. Manag. 68(6–7), 560–591 (2022) 7. Goerzig, D., Bauernhansl, T.: Enterprise architectures for the digital transformation in small and medium-sized enterprises. Procedia CIRP 67, 540–545 (2018) 8. Kotusev, S.: The history of enterprise architecture: an evidence-based review. J. Enterp. Archit. 12(1), 29 (2016) 9. Richardson, G.L., Jackson, B.M., Dickson, G.W.: A principles-based enterprise architecture: lessons from Texaco and Star Enterprise. MIS Q. 14, 385–403 (1990) 10. Greefhorst, D., Proper, E., Greefhorst, D., Proper, E.: The Role of Enterprise Architecture. Springer, Cham (2011) 11. Hazen, B.T., Bradley, R.V., Bell, J.E., In, J., Byrd, T.A.: Enterprise architecture: a competencebased approach to achieving agility and firm performance. Int. J. Prod. Econ. 193, 566–577 (2017) 12. Kreuter, T., Kalla, C., Scavarda, L.F., Thomé, A.M.T., Hellingrath, B.: Developing and implementing contextualised S&Op designs–an enterprise architecture management approach. Int. J. Phys. Distrib. Logist. Manag. 51(6), 634–655 (2021) 13. Pereira, C.M., Sousa, P.: A method to define an enterprise architecture using the Zachman framework. In: Proceedings of the 2004 ACM Symposium on Applied Computing, pp. 1366– 1371 (2004) 14. White, S.K.: What Is Enterprise Architecture? A Framework for Transformation. https://www.cio.com/article/222421/what-is-enterprise-architecture-a-framework-fortransformation.html. Accessed 6 Jan 2023 15. Rood, M.A.: Enterprise architecture: definition, content, and utility. In: Proceedings of 3rd IEEE Workshop on Enabling Technologies: Infrastructure for Collaborative Enterprises, pp. 106–111 (1994) 16. Caetano, A., Silva, A.R., Tribolet, J.: A role-based enterprise architecture framework. In: Proceedings of the 2009 ACM symposium on Applied Computing, pp. 253–258 (2009) 17. Lange, M., Mendling, J.: An experts’ perspective on enterprise architecture goals, framework adoption and benefit assessment. In: IEEE 15th International Enterprise Distributed Object Computing Conference Workshops, pp. 304–313 (2011)

356

O. Maghazei et al.

18. Niemi, E.: Enterprise architecture benefits: perceptions from literature and practice. In: The Proceedings of the 7th IBIMA Conference Internet & Information Systems in the Digital Age, pp. 1236–1615; 18 (2008) 19. Shanks, G., Gloet, M., Someh, I.A., Frampton, K., Tamm, T.: Achieving benefits with enterprise architecture. J. Strateg. Inf. Syst. 27(2), 139–156 (2018) 20. Ross, J.W.: Enterprise Architecture: Driving Business Benefits from IT (2006) 21. Kistasamy, C., Van Der Merwe, A., De La Harpe, A.: The relationship between service oriented architecture and enterprise architecture. In: The 14th IEEE International Enterprise Distributed Object Computing Conference Workshops, pp. 129–137 (2010) 22. Channabasavaiah, K., Holley, K., Edward, M., Tuggle, J.: Migrating to a ServiceOriented Architecture. https://public.dhe.ibm.com/software/info/openenvironment/G2247298-00_Final.pdf. Accessed 9 Nov 2022 23. Amazon Web Sevices, What Is SOA (Service-Oriented Architecture)? https://aws.amazon. com/what-is/service-oriented-architecture/. Accessed 3 Feb 2023 24. Nolle, T.: Service-Oriented Architecture (SOA). https://www.techtarget.com/searchapparchit ecture/definition/service-oriented-architecture-SOA. Accessed 19 Feb 2023 25. Windley, P.J.: Service Oriented Architectures. https://www.windley.com/docs/2004/SOA.pdf. Accessed 8 Jan 2023 26. Schepers, T.G., Iacob, M.-E., Van Eck, P.A.: A lifecycle approach to SOA governance. In: The Proceedings of the ACM Symposium on Applied Computing, pp. 1055–1061 (2008) 27. The Open Group, The Togaf® Standard. https://www.opengroup.org/. Accessed 16 Feb 2023 28. Acerbi, F., Assiani, S., Taisch, M.: A methodology to assess the skills for an Industry 4.0 factory. In: Ameri, F., Stecke, K., von Cieminski, G., Kiritsis, D. (eds.) Advances in Production Management Systems. Towards Smart Production Management Systems: IFIP WG 5.7 International Conference, APMS 2019, Austin, TX, USA, Proceedings, Part II, vol. 567, pp. 520–527. Springer, Cham (2019). https://doi.org/10.1007/978-3-030-29996-5_60

Application of Digital Tools, Data Analytics and Machine Learning in Internal Audit Jelena Popara1,2

, Milena Savkovic2(B) , Danijela Ciric Lalic2 and Bojan Lalic2

,

1 NIS j.s.c, 21000 Novi Sad, Serbia 2 Faculty of Technical Sciences, University of Novi Sad, 21000 Novi Sad, Serbia

[email protected]

Abstract. The Internal Audit seeks to respond to the challenges of the modern business environment, with a special focus on working with large databases so as to provide an additional value to their organizations through intelligent testing, root cause and trend analysis, summarizing the number and value of exceptions and providing the starting point for detection of unusual trends. This article analyse different digital solutions used in the oil and gas industry, with a special focus on the comparative analysis of digital tools, advanced analytics and machine learning in different stages of internal audit activities for the purpose of performing trend analysis and testing internal controls. Proposals for further improvements of the existing internal audit practice is to reach the level of predictive analytics and data-driven decision-making. Theoretical and practical analysis of the current level of development and use of digital tools by the internal audit in the oil industry are used a basis for comparative analysis of different IT solutions in working with large databases, as well as the competencies internal auditors will require in order to recognize new risks, technologies, innovations and use their knowledge and experience to help the learning and improvement processes in the organization, encourage development of innovative business practices and processes that will help the organization reach its strategic business goals. The originality of this article lies in its focus on the use of digital solutions, advanced analytics and machine learning in different stages of internal audit activities within the oil and gas industry. Keywords: Risk management · Digital tools · Data analytics · Data-driven decision-making · Internal audit

1 Introduction One of the key responsibilities of the Internal Audit is to provide an independent and objective assurance to the Board of Directors, i.e. the Board of Directors Audit Committee, that the company risks have been identified and are duly managed in the appropriate way [1]. In order to fulfil the responsibility, the Internal Audit requires competent auditors, tools and techniques accommodated to the organization which may help identify, organize and present such information [2]. © IFIP International Federation for Information Processing 2023 Published by Springer Nature Switzerland AG 2023 E. Alfnes et al. (Eds.): APMS 2023, IFIP AICT 689, pp. 357–371, 2023. https://doi.org/10.1007/978-3-031-43662-8_26

358

J. Popara et al.

The Internal Audit seeks to respond to the challenges of the modern business environment, with a special focus on working with large databases so as to provide an additional value to their organizations through intelligent testing, root cause and trend analysis, summarizing the number and value of exceptions and providing the starting point for detection of unusual trends [3]. This article will present examples of different digital solutions used in the oil and gas industry (based on the example of the Serbian company NIS j.s.c.), with a special focus on the comparative analysis of digital tools, advanced analytics and machine learning in different stages of internal audit for the purpose of performing an automated analysis of trends and testing internal controls. The collected data were taken over from internal auditors and from digital technology and innovation experts in the oil and gas industry. Proposals for further improvements of the existing internal audit practice is to reach the level of predictive analytics and data-driven decision-making. This article is divided into four chapters. The introductory section outlines the purpose of the article, while the theoretical background delves into the literature on internal audit, capability models, competencies, digital tools, data analytics, and data science. The third section provides an overview of the selected digital tools and presents a comparative analysis between them. Lastly, the article is concluded with a summary and suggestions for future research.

2 Theoretical Background 2.1 Understanding Internal Audit and the Capability Models Internal auditing is an independent, objective assurance and consulting activity designed to add value and improve an organization’s operations. It helps an organization accomplish its objectives by bringing a systematic, disciplined approach to evaluate and improve the effectiveness of risk management, control, and governance processes [4]. Internal audit is performed in different legal and cultural contexts in organisations differing in their purpose, size and complexity [5]. Although these differences have an impact on different internal audit practices in specific industries, compliance with the Standards is required in order to discharge the obligations of internal auditors and to carry out internal audit activities [1, 6, 7]. Internal auditing is performed by professionals with an in-depth understanding of the organisation of the business culture, systems and processes. Internal auditors are expected to observe the International Standards for the Professional Practice of Internal Auditing (hereinafter: the Standards) and the Code of Ethics [1]. The purpose, authority, and responsibility of the internal audit activity must be formally defined in an Internal Audit Charter, consistent with the Mission of the Internal Audit and Mandatory Elements of the Framework (including the Core Principles for the Profession of Internal Auditing), the Code of Ethics, and the Standards and the Definition of Internal Auditing. In 2018, the Association of Internal Auditors of Netherlands drafted the Internal Audit Capability Model IA-CM as a framework for strengthening or improving the function of internal audit through a sequence of evolutionary steps. Each of the steps

Application of Digital Tools, Data Analytics and Machine Learning

359

is organised in five progressive maturity levels. Improvements in the processes and practices at every stage provide the basis for progressing further to the next maturity level. The Internal Audit Capability Model consists of five maturity levels, namely Initial, Infrastructural, Integrated, Managed, and Optimized, which serve as benchmarks for assessing and improving the effectiveness and efficiency of internal audit activities within an organisation [8]. At the Initial level, internal auditing is ad hoc or unstructured, only few processes are identified and practices are carried out inconsistently. Only isolated individual audits and/or document and transaction reviews can be implemented. “The three lines of defence” have not been established and there are no consulting services provided by the Internal Audit [9]. The audit is probably limited to a transaction audit; i.e. inspection of the regularity and accuracy of individual economic transactions or some basic compliance audits. The infrastructure for the internal audit activity is not established and the auditors are probably a part of a larger organisational unit. At this level, internal auditing must rely on individual efforts or personal skills of the auditor performing auditing and their personal objectivity. At the Infrastructural level, the Internal Audit performs primarily a traditional compliance audit or, in other words, compliance audits observing particular areas, processes or systems of policies, plans, procedures, laws, regulations, contracts or other requirements. This could include financial audits, as well as IT system and process audits assessing whether an appropriate internal control framework is in place and whether it is functional. The internal audit activity has started identifying and recruiting the people with the required competency and the relevant skills for performing the audit engagement. The internal audit is developed into an activity with “an added value” helping the organisation to manage its risks and use the opportunities for improvement [8, 9]. The Internal Audit also considers other subject matters, including the strategy and informal controls, highlighting the importance of the Integrated level. The internal audit services have become more diverse in order to support the needs of the organisation’s management. The use of data analytics is organised, the expertise is available in the Internal Audit or may be easily obtained, while the tools for data analysis are available. As appropriate, consultancy services are also undertaken within the internal audit activity in order to provide the guidelines and advice to the management. At the Managed level, the internal audit activity functions as part of the organisation management and of the risk management [10, 11]. The audit director is in charge of providing formal and informal counsel on strategic issues and of having an impact on the executive board and the audit commission and/or the supervisory board. The internal audit activity has integrated its use of quantitative and qualitative data and information in order to help it achieve its strategic objectives and continuously improve its performance. The activity of internal audit functions as a well-managed business unit and has a long-term vision and plan in line with the organisation’s guidelines. At level 5 – Optimized, the focus is on learning to continuously improve in order to improve the capability. The internal audit activity at level 5 is characterized as a learning organization with continuous improvement of the processes and innovations. It follows the volatile external environment and uses the information inside and outside the

360

J. Popara et al.

organization so as to improve its approach to the assessment of the governance practice, risk management and controls (Fig. 1).

Fig. 1. The five levels of the Internal Audit Capability Model [9]

2.2 Internal Auditor Competencies Over the last decades, people and technology have become very much related, as never before. The appearance of artificial intelligence has accelerated the processes, offering previously impossible opportunities. Data have become readily available, robots are used more than ever before, and all this leads to further technological development that changes operations, creates new industries and changes the labor market [12]. Internal auditor competencies represent a set of knowledge, skills, capabilities and behaviors the employees engaging in internal auditing should have and apply in their work so as to perform it efficiently [13]. The predominant position in the literature is that these are knowledge, skills, attitudes, values, motives, traits and capabilities. To understand the term competencies, the iceberg model is often used [14]. To become an authorized internal auditor and maintain professional competence, individuals must acquire functional competencies in internal auditing through training and regular professional development, which includes gaining knowledge and skills in international standards, legal frameworks in audit and accounting, industry-specific knowledge, internal auditor code of ethics, computer-assisted audit tools and techniques, governance, risk and internal control management, critical thinking, and communication [15]. Specific skills and training programs necessary for developing these competencies can be found in trainings in programming (Python and R), SQL, advanced finance modeling machine learning. Programs in computational finance, statistics, digital transformation, database management and other similar topics would offer practical guidance to internal auditors seeking to enhance their capabilities in a digitally-driven landscape. The traditional audit model enables auditors to arrive at conclusions based on the limited sample of the population, instead of analysing all available or a large sample of

Application of Digital Tools, Data Analytics and Machine Learning

361

data [13–15]. Digital tools, in the way it is usually used, is a practice of analysing large quantities of data in search of anomalies. A well-conceived audit supported with digital tools will not be a sample, but a comprehensive review of all transactions. Another advantage of an audit supported with digital tools is in the fact that it enables auditors to test specific risks. Digital tools analyses are limited only to the data saved in the files in line with the systematic patterns. Many data have never been documented in this way. In addition, saved data often contain deficiencies, they are badly classified, not readily available and in practice it is very difficult to confirm their integrity. In the existing practice, the use of digital tools represents an addition to internal audit tools and techniques. In some business processes, an audit supported with digital tools is completely impossible to perform. 2.3 Digital Tools, Data Analytics and Data Science The use of the existing technologies for the purpose of improving business processes represents the definition of digitalization [16, 17]. Digitalization is the use of the existing technologies and information for improving or replacing business processes, making profit and creating a digital operation environment where information has the central role [18]. Other definitions of digitalization connect it with Industry 4.0 which refers to intelligent networking of industrial machines and processes aided by information and communication technologies [19]. There are many ways in which companies may use intelligent networking. Digital transformation is not IT sector, but a process including business as a whole [20]. Additionally, it is also not only the use of digital tools and technologies, but the method in which the tools and technologies change modern operations, the work of employees, cooperation with the partners and user experience itself [10]. Digital transformation represents the integration of digital technologies and business models through applications such as big data, internet of things (IoT), artificial intelligence, for the purpose of increasing the value of the company and changing the operational scheme [17, 21–23]. There are a lot of tools for obtained data transformation which can be used to predict and discard patterns and new values. The greatest challenge is precisely how to turn the huge quantity of data into valuable information. The next figure shows interrelation of digital tools and their necessity as the basis for further development of data analytics and data science (Fig. 2). Data analytics may be defined as the process of analysing raw data in order to identify trends and answers to the questions. Advanced analytics deals with the ‘What if?’ question. This part of data science uses advanced tools for data extraction, making predictions and identifying trends. These tools include classical statistics, as well as machine learning. The job of a data analyst includes working with data in the entire data analytics chain. This includes working with data in different ways. The primary steps in the process of data analytics are data collection, data management, statistical analysis and data presentation. Data collection is the essential process for many tasks of data analytics. It means extraction of data from the unstructured data sources. Statistics and machine

362

J. Popara et al.

Fig. 2. Graphical representation of interrelation of digital tools, data analytics, data science, machine learning and artificial intelligence including IT solutions that were analysed

learning technique are also used to analyse data. Large datasets are used to create statistical models which identify data trends. These models may then be applied to new data in order to make predictions and informed decisions. Statistical programming languages such as R or Python are essential for this process. In addition, open source libraries and packages such as TensorFlow enable advanced analysis. By 2025, the total amount of available data will rise to 175 zettabytes (a zettabyte is a trillion gigabytes). Data are also called ‘the oil of the 21st century’. Data science may be thought of as having a five stage life cycle that is consisted of collection, maintenance, process, analysis, communication. There are some general skills that need to be acquired to create successful data science professionals. They include the skills of: – Programming – use of languages such as Python and R. – Database management – SQL learning and application in communication with databases. – Statistics – have a method for analysing data in order to solve problems[24]. Machine learning is defined as the use of algorithms to make decisions by way of generalisation (or by finding the patterns) in the given database. Machine learning is showing better results than the standard statistical approach when it comes to a large number of large dimension variables and when their interrelations are non-linear [25]. In the oil and gas industry, based on the application of advanced analytics and machine learning for the purpose of reducing operating expenses and risks, the so-called digital twin solutions are being developed. A digital twin is a virtual representation of a structure or system including its life cycle, it is updated from the data in real time and uses stimulation, machine learning and reasoning to help in the decision-making process.

Application of Digital Tools, Data Analytics and Machine Learning

363

Machine learning together with deep learning form artificial intelligence. Both deep learning and machine learning are sub-fields of artificial intelligence, and deep learning is actually a sub-field of machine learning. AI is one of the newest fields in science and engineering. AI currently encompasses a huge variety of subfields, ranging from the general (learning and perception) to the specific, such as playing chess, proving mathematical theorems, writing poetry, driving a car on a crowded street, and diagnosing diseases. AI is relevant to any intellectual task; it is truly a universal field [21]. There are numerous, real-world applications of AI systems today. Below are some of the most common examples: – Speech recognition: It is also known as automatic speech recognition (ASR), computer speech recognition, or speech-to-text, and it is a capability which uses natural language processing (NLP) to process human speech into a written format. Many mobile devices incorporate speech recognition into their systems to conduct voice search - e.g. Siri or provide more accessibility around texting. – Automated stock trading: Designed to optimize stock portfolios, AI-driven highfrequency trading platforms make thousands or even millions of trades per day without human intervention. – Detecting frauds: Banks and other financial institutions can use machine learning to detect suspicious transactions. Supervised learning can train a model using information about known fraudulent transactions. Anomaly detection can identify transactions that appear atypical and require further investigation [26].

3 Exploring and Evaluating Digital Tools for Internal Audit: A Comparative Analysis of Key Solutions For the research purposes, five IT solutions were selected that are currently in the development or implementation stage in the implementation of internal audit procedures in the oil and gas industry. Considering that oil and gas industry is in top 10 most digitalized industries, following IT solutions such as TeamMate, Jira, Process mining, OCR (Optical character recognition), as well as Oracle BI dashboard solutions. In addition, 2 IT solutions used in the oil and gas industry and developed for business needs were analysed, e.g. for the purposes of predicting equipment downtime, improving the efficiency of business processes, etc. They are presented in order to understand the industry development trends and future competencies that the project managers, employees, and internal auditors auditing business processes will be required to have to be able to perform on the Managed or Optimized maturity level as per Internal Audit Capability Model. The observation period covers the application and development of IT solutions in the previous two years, starting from 2019, and represents the solutions that are currently being used and are still being developed. The collected data were taken over from internal auditors and from digital technology and innovation experts in the oil and gas industry. 3.1 IT Solution – TeamMate TeamMate is an audit management system, designed to bring efficiency and consistency to the entire audit process. It facilitates the audit process from risk assessment to reporting, i.e. through all stages of audit engagement. The TeamMate Suite was designed with

364

J. Popara et al.

the idea of flexibility, the system can be applied to many different types of assignments and audit engagements. The TeamMate is a fully, comprehensive, dedicated internal audit management solution designed to help auditors in managing all aspects of the process (Fig. 3). – TeamRisk is an advanced risk assessment system that enables internal audit departments to develop a risk-based audit plan. – The TeamSchedule module is a very flexible tool used to manage resources, projects and budget. – The TeamTEC module provides the ability to record, track and report on audit time and resource costs. – TeamEWP is used for all stages of the audit engagement - preparation of working paper, review, report generation, and storage - in electronic format. A TeamEWP project file is available to multiple users in different locations, helping teams work together even when working remotely. – TeamCentral is used for software package management and allows access to key historical data related to projects and issues.

Fig. 3. Modular structure of the TeamMate Audit solution

3.2 IT Solution – Jira Jira is an issue tracking product that allows bug tracking and agile project management. Jira is a modern, state-of-the-art tool that enables process management through a highly customizable flow - from start to finish. It allows the work to be divided into smaller units/parts that can be assigned to responsible persons and to monitor the progress until the work is completed. With additional information, such as status reports, priorities, comments, attachments, and functionalities such as availability of real-time changes, automatic notification system, reporting and visualization, this tool provides accurate and timely information to perform the tasks.

Application of Digital Tools, Data Analytics and Machine Learning

365

3.3 IT Solution – Process Mining Process mining tool is an intellectual data analysis solution, which enables to build business process models on the basis of data, directly obtained from information systems in which processes are performed (Fig. 4).

Fig. 4. Visual representation of internal controls of business processes before and after using the Process mining IT solution- reaching the target design of internal controls

The Process Mining tool is currently in the phase of testing and training of internal auditors of NIS j.s.c. company for its use in work during the audit engagement. The current practice shows that the analysis and transformation of data with verification of completeness and accuracy require between 4 and 5 weeks on average. The whole Process mining project per individual business process lasts between 10 and 12 weeks, depending on the complexity of the process itself. Further training and practice with this IT solution need to be continued. 3.4 IT Solution – Oracle BI Dashboard Oracle BI Interactive Dashboards provide any knowledge worker with intuitive, interactive access to information that is actionable and dynamically personalized based on the individual’s role and identity. In the Oracle BI Intelligence Dashboards environment, the end user is working with live reports, prompts, charts, tables, pivot tables, graphics, and tickers in a pure Web architecture. The user has full capability for drilling, navigating, modifying, and interacting with these reports. Oracle BI Intelligence Dashboards can also aggregate content from a wide variety of other sources, including the Internet, shared file servers, and document repositories. 3.5 IT Solution – Optical Character Recognition (OCR) Optical Character Recognition is the process that converts an image of text into a machine-readable text format. For example, if you scan an invoice, the computer saves the scan as an image file. You cannot use a text editor to edit, search, or count the words

366

J. Popara et al.

in the image file. However, you can use OCR to convert the image into a text document with its contents stored as text data. OCR can transform a scanned PDF file into an editable and searchable text-based document. In this way, the time required for the analysis of data contained in invoices, contracts, questionnaires or other types of documents required for the tests of details in the audit process is significantly reduced. In data extraction software, documents pass through four main stages (Import, Recognition, Verification, Export) and as a result they are classified, recognized, verified and transferred into reliable, accurate, searchable and highly structured electronic data that can drive business processes across a department or enterprise. The application of the OCR solution for audit purposes, i.e. tests of details is currently limited due to differences in the language and structure of invoices. The success of using this solution, i.e. extracting the data needed for audit tests, is 20–25% on average. Other imported invoices were not recognized by the OCR solution and the text images were not converted into text data. A new iteration of testing the solution is underway with the aim of improving the performance of this solution and achieving a data conversion success rate of at least 90%. 3.6 IT Solution – PyVision Risk Detection and Predictive Diagnostics Deep learning examples in oil industry are helping to improve worker safety around heavy machinery by automatically detecting when people or objects are within an unsafe distance of machines. There are different categories of “Computer Vision” function, such as, image processing including image recognition, face recognition, optical character recognition. The examples of currently used Computer Vision applications are: – A smart video surveillance system that helps detect fires; – Automated controls to prevent fires. The development and implementation of previously mentioned functionalities of IT solutions by business process owners in companies creates a need for new knowledge and internal auditors who, in the planning and fieldwork stage, must understand the available data, and are certainly a potential that will be further developed in the coming period [27]. 3.7 IT Solution – Digital Twins Digital twin technology is intended to inform users about their operations and suggest measures to avoid any downtime of production processes. As the operational life continues, the digital twin is updated automatically, in real-time, with current data, work records, and engineering information to optimize maintenance and operational activities. Using this information, engineers, managers, and operators can easily search the asset tags to access critical up-to-date engineering and work information and find the health of a particular asset. With the defined functionalities, operational and asset issues are flagged and addressed early on, and the workflow becomes preventative, instead of reactive.

Application of Digital Tools, Data Analytics and Machine Learning

367

With a digital twin, process simulation can also be performed to optimize operating models based on their physical properties and thermodynamic laws. In order to improve performance and increase profitability, three approaches are applied: Stable state, Dynamic modelling and Predictive analytics for monitoring the equipment conditions. 3.8 Results of the Comparative Analysis The market for digital tools for required internal audits is developed and is expanding globally [23]. There are many solutions that can be adapted to the needs of internal auditing. The research analyzed the digital solutions that were tested in the practice in the oil and gas industry and digital solutions used by internal auditors in different stages. Table 1. Comparative analysis of the application of digital tools in the internal audit management process. Internal Audit phases Digital tools Business process quality assurance Project management systems Data analycs Data analycs Machine learning AI based on deep learning AI based on machine learning

IT soluon TeamMate Jira Process mining Oracle BI OCR PyVision Digital twins

Planning phase Fieldwork phase Reporng phase a a a a a a a a a r r a a a a a a a a a a

Monitoring of recommendaons a a r

a r r r

Today’s level of internal audit development in NIS j.s.c. uses the TeamMate solution in the IA management process, and in this way the quality of the business process is ensured. Furthermore, the Jira IT solution, which is essentially a project management system, is also internally adapted and used in all phases of the audit engagement process, communication with the business regarding the status of the delivery of the required documentation and rapid data exchange. The advanced data analysis is provided by an in-house Process Mining tool used in the planning, fieldwork and reporting stages, and represents an advanced step compared to data analysis within the MS Office package. Process Mining is primarily used to improve business processes by increasing the efficiency of business processes, eliminating process “bottlenecks”, identifying gaps in automated controls in the business process. Process Mining contributes to the quality of audit reports and the analysis results form. The initial steps in working with machine learning-supported solutions produced limited results at this stage of using the OCR solution, which finds its application in the planning, fieldwork and reporting phase, reducing the auditor’s work time on testing the details of primary documentation. The application of the OCR solution is in the field work stage and is currently limited due to differences in the language and structure of the invoices. The success of using this solution, i.e. extracting the data needed for audit tests, is 20–25% on average. A new iteration of testing the solution is underway with the aim of improving the performance of this solution.

368

J. Popara et al.

The PyVision solutions and digital twins are IT solutions developed by business for the preventive management of business processes, reporting on key operational data in order to minimize non-productive time as well as the risk of equipment downtime. New technologies that support business processes must also be subject to detailed testing, and for this reason, internal auditors collect this information in the audit planning stage, familiarize themselves with the functionalities of IT solutions, identify key controls, including general IT controls that are tested in the field work stage. One integrated solution fitting all internal audit phases and activities can be one of the future developments but should not be considered as the only possible stream. Specific digital tools that are aimed to identify data “gaps” that require further automation are valuable for further digitalization process, enabling complete and accurate registers of business transactions. In this way, internal audit is generating digital solutions valuable for its business, creates innovations and improves governance practice, risk management and controls, pivoting to meet Managed or Optimized maturity level as per Internal Audit Capability Model. By combining data analytics, machine learning, big data and software applications, the power of data is used to create business value for oil and gas companies, which are leaders in the development of digital tools as a whole. The use and further improvement of digital tools, together with machine learning technologies will radically change the way audits are carried out in the future, and the following conclusions and proposals for further improvement of the existing practice can be reached: – – – – –

Work on the population sample will become obsolete; The assessment of internal controls will be proactive; Additional data will be used as audit evidence; Auditors will become data scientists; Auditors will become strategic advisors.

4 Conclusion In a globally connected world, there is a growing need for internal audit functions that provide added value to their organizations. Shareholders expect more from their internal audit functions. They want their assurance that controls are working properly, to advise on changes and operational issues, and to predict and provide insights into risks to the organization [28]. Process automation is the initial level and is a basis required to switch from paper documents to electronic data. Digital tools are developed on their basis and help the architecture of the processes themselves, and only established and developed digital tools can further improve processes, develop algorithms, machine learning and artificial intelligence. The Project Management Institute (PMI) has identified key characteristics that are essential for project managers in the digital age of business, which can also be applied to audit engagement managers. These include creativity and innovative thinking, expertise in data science, advanced knowledge of security and privacy, legal acumen, the ability to solve complex problems, effective team leadership and communication skills, and the capacity to make informed decisions based on data analysis [29].

Application of Digital Tools, Data Analytics and Machine Learning

369

Managerial creativity will be the key. New technologies, products and ways of working have opened great space for creativity [30, 31]. Although robots will help in certain tasks, they will not be capable of creating and designing innovations. Namely, this research is strictly limited to digital solutions used by internal audits in oil and gas industry. The content analyzes current digital solutions based on the sample of solutions that are applicable in current internal audit practices. Findings presented in this study can assist future research in this or related research areas in order to overcome these limitations. The next generation of internal auditors must recognize the challenges of new risks, technologies, innovations and be able to help preserve processes, assets and associated business development, implement new methods of creating and delivering value. Internal auditors have a broad picture of the organization and their knowledge and experience contribute to organizational learning, improvement and encourage the development of innovative business practices and processes that will help the organization achieve its strategic business goals.

References 1. IIA, T.: The role of internal auditing in enterprise-wide risk management (2009) 2. Tušek, B., Saˇcer, I.M., Halar, P.: The internal audit function as an effective tool for increasing business success in digital economy era. Presented at the (2018) 3. Lois, P., Drogalas, G., Karagiorgos, A., Tsikalakis, K.: Internal audits in the digital era: opportunities risks and challenges. EuroMed J. Bus. 15, 205–217 (2020). https://doi.org/10. 1108/EMJB-07-2019-0097 4. Betti, N., Sarens, G.: Understanding the internal audit function in a digitalised business environment. J. Account. Organ. Chang. 17, 197–216 (2021). https://doi.org/10.1108/JAOC11-2019-0114 5. Akhmedjanov, K.B.: Internal Audit and its Development Prospects. Akhmedjanov, K. B. “Internal audit and its development prospect. Int. J. Econ. Commerce Manage. 4, 555–560 (2016) 6. Pizzi, S., Venturelli, A., Variale, M., Macario, G.P.: Assessing the impacts of digital transformation on internal auditing: A bibliometric analysis. Technol Soc. 67 (2021). https://doi.org/ 10.1016/j.techsoc.2021.101738 7. Abbott, L.J., Parker, S., Peters, G.F.: Internal audit assistance and external audit timeliness. Auditing. 31, 3–20 (2012). https://doi.org/10.2308/ajpt-10296 8. Ain Zakiah Mohd Yusof, N., Haron, H., Ismail, I., Hock Chye, O.: Independence of Internal Audit Unit Influence the Internal Audit Capability of Malaysian Public Sector Organizations. KnE Social Sciences. (2019). https://doi.org/10.18502/kss.v3i22.5122 9. Janse van Rensburg, J.O., Coetzee, P.: Internal audit public sector capability: a case study. J Public Aff. 16, 181–191 (2016). https://doi.org/10.1002/pa.1574 10. Yucel, S.: Estimating the benefits, drawbacks and risk of digital transformation strategy. In: Proceedings - 2018 International Conference on Computational Science and Computational Intelligence, CSCI 2018, pp. 233–238. Institute of Electrical and Electronics Engineers Inc. (2018) 11. Rahi, K., Bourgault, M., Preece, C.: Risk and vulnerability management, project agility and resilience: a comparative analysis. International Journal of Information Systems and Project Management. 9, 5–21 (2021). https://doi.org/10.12821/ijispm090401

370

J. Popara et al.

12. Betti, N., Sarens, G., Poncin, I.: Effects of digitalisation of organisations on internal audit activities and practices. Manag. Audit. J. 36, 872–888 (2021). https://doi.org/10.1108/MAJ08-2020-2792 13. Kabuye, F., Nkundabanyanga, S.K., Opiso, J., Nakabuye, Z.: Internal audit organisational status, competencies, activities and fraud management in the financial services sector. Manag. Audit. J. 32, 924–944 (2017). https://doi.org/10.1108/MAJ-09-2016-1452 14. Meinlschmidt, J., Schleper, M.C., Foerstl, K.: Tackling the sustainability iceberg: a transaction cost economics approach to lower tier sustainability management. Int. J. Oper. Prod. Manag. 38, 1888–1914 (2018). https://doi.org/10.1108/IJOPM-03-2017-0141 15. Setyaningrum, D., Kuntadi, C.: The effects of competence, independence, audit work, and communication on the effectiveness of internal audit. J. Econ. Bus. Accountancy Ventura. 22 (2019). https://doi.org/10.14414/jebav.v22i1.879 16. Sofic, A., Rakic, S., Pezzotta, G., Markoski, B., Arioli, V., Marjanovic, U.: Smart and resilient transformation of manufacturing firms. Processes. 10 (2022). https://doi.org/10.3390/pr1012 2674 17. Rakic, S., Visnjic, I., Gaiardelli, P., Romero, D., Marjanovic, U.: Transformation of manufacturing firms: towards digital servitization. In: Dolgui, A., Bernard, A., Lemoine, D., von Cieminski, G., Romero, D. (eds.) APMS 2021. IFIP, vol. 631, pp. 153–161. Springer, Cham (2021). https://doi.org/10.1007/978-3-030-85902-2_17 18. Jankovic-Zugic, A., Medic, N., Pavlovic, M., Todorovic, T., Rakic, S.: Servitization 4.0 as a Trigger for Sustainable Business: Evidence from Automotive Digital Supply Chain. Sustainability (Switzerland). 15 (2023). https://doi.org/10.3390/su15032217 19. Ciric, D., Gracanin, D., Vuˇckovi´c, T., Lalic, B., Simeunovic, N.: Digital Transformation in the Engineering Research Area: Scientific Performance and Strategic Themes. Presented at the April (2022) 20. Rakic, S., Pero, M., Sianesi, A., Marjanovic, U.: Digital servitization and firm performance: technology intensity approach. Eng. Econ. 33, 398–413 (2022). https://doi.org/10.5755/j01. ee.33.4.29649 21. Jones, M.D., Hutcheson, S., Camba, J.D.: Past, present, and future barriers to digital transformation in manufacturing: a review. J. Manuf. Syst. 60, 936–948 (2021). https://doi.org/10. 1016/j.jmsy.2021.03.006 22. Jafari-Sadeghi, V., Garcia-Perez, A., Candelo, E., Couturier, J.: Exploring the impact of digital transformation on technology entrepreneurship and technological market expansion: the role of technology readiness, exploration and exploitation. J. Bus. Res. 124, 100–111 (2021). https://doi.org/10.1016/j.jbusres.2020.11.020 23. Tran, T., Tran, T., Hai, N., Tinh, N.: Digital transformation in oil and gas companies - A case study of Bien Dong POC. Petrovietnam J. 10, 67–78 (2020). https://doi.org/10.47800/PVJ. 2020.10-07 24. Barlette, Y., Baillette, P.: Big data analytics in turbulent contexts: towards organizational change for enhanced agility. Prod. Plann. Control. 33, 105–122 (2022). https://doi.org/10. 1080/09537287.2020.1810755 25. Babenko, V., Panchyshyn, A., Zomchak, L., Nehrey, M., Artym-Drohomyretska, Z., Lahotskyi, T.: Classical machine learning methods in economics research: macro and micro level examples. Wseas Trans. Bus. Econ. 18, 209–217 (2021). https://doi.org/10.37394/23207. 2021.18.22 26. Al-Mohammedi, Y.A.: The role of digitalization in developing internal audit practices in an IT environment. In: Proceedings - 2020 2nd Annual International Conference on Information and Sciences, AiCIS 2020, pp. 230–236. Institute of Electrical and Electronics Engineers Inc. (2020) 27. Applications of Machine Learning in Oil and Gas (2023) ij

Application of Digital Tools, Data Analytics and Machine Learning

371

28. Dakovic, M., Lalic, B., Delic, M., Tasic, N., Ciric, D.: Systematic mitigation of model sensitivity in the initiation phase of energy projects. Adv. Prod. Eng. Manage. 15, 217–232 (2020). https://doi.org/10.14743/apem2020.2.360 29. Project Management Institute: The standard for project management and a guide to the project management body of knowledge (PMBOK guide). Project Management Institute (2021) 30. Ciric, D., Lalic, B., Savkovic, M., Marjanovic, U., Stefanovic, D.: The Concept of Gamification in Manufacturing Research: Science Mapping Approach Using Bibliometrix R-Tool. In: 26th International Conference on Production Research-ICPR. Tunghai University, Taichung, Taiwan (2021) 31. Ciric, D., Lalic, B., Marjanovic, U., Savkovic, M., Rakic, S.: A bibliometric analysis approach to review mass customization scientific production. In: Dolgui, A., Bernard, A., Lemoine, D., von Cieminski, G., Romero, D. (eds) APMS 2021. IFIP, vol. 634, pp. 328–338. Springer, Cham (2021). https://doi.org/10.1007/978-3-030-85914-5_35

Consumer Engagement in the Design of PLM Systems: A Review of Best Practices Uchechukwu Nwogu(B)

and Richard Evans

Faculty of Computer Science, Dalhousie University, Halifax, NS B3H 4R2, Canada [email protected]

Abstract. Digitization has disrupted all aspects of Product Development (PD) from idea conceptualization through to the retirement or recycling of a product. With advancements in collaborative technologies, product development teams must seek competitive advantage from their use but, at the same time, improve their understanding about how they can best be integrated into current workflows and processes. Product Lifecycle Management (PLM) systems are designed to assist firms in the management of the entire PD lifecycle. They use advanced technologies to bring together all stakeholders involved in the PD lifecycle with the aim of satisfying the complex needs and requirements of consumers. Using a mix of literature review and content analysis tools, this paper presents a critical review of best practices for engaging consumers in the design of PLM systems. Specifically, we examine current levels of consumer engagement, benefits of consumer engagement, the effective integration of consumers, and best practices for consumer engagement in the design of PLM systems. Keywords: Consumer Engagement · Social Product Development · Product Lifecycle Management · Software Development · Co-Creation

1 Introduction Innovative product development is crucial for the economic success of modern firms. It requires firms to complete a range of knowledge-intensive and complex activities from idea conceptualization and market analysis through to the product being designed, manufactured (or coded), and launched into an identified market. This process is known as the PD lifecycle which aims to bring a new product (physical or digital) to market that meets or exceeds the needs and requirements of consumers. By following a carefully constructed PD Lifecycle, firms can reduce their likelihood of product failure, ensuring that their product stands the best possible chance of marketplace success. Product development activities, however, involve intricate collaboration between a range of internal and external actors, including customers, end-users, and innovation communities (external), and design, manufacturing, marketing, and sales teams (internal); each of these actors have their own set of requirements and these must be met, integrated, and managed to ensure a successful product launch. © IFIP International Federation for Information Processing 2023 Published by Springer Nature Switzerland AG 2023 E. Alfnes et al. (Eds.): APMS 2023, IFIP AICT 689, pp. 372–385, 2023. https://doi.org/10.1007/978-3-031-43662-8_27

Consumer Engagement in the Design of PLM Systems

373

Throughout the PD lifecycle, it is critical that firms develop an in-depth appreciation of their consumers whether they are the direct customer or the end-user of the product. By developing empathy with consumers and an understanding of their habits, needs, pain points, and desires, product designers can address the genuine problems experienced by them. This process, however, does not end simply when the product is launched and, therefore, firms must continuously refine their products using appropriate user-centered design methods to ensure seamless updates to product features. The PD lifecycle traditionally includes four phases: Introduction, Growth, Maturity, and Decline [1], as illustrated in Fig. 1. During the Introduction phase, the sales of a product are often slow and increase gradually, leading to minimal or no profit due to the uniqueness of the product developed. In contrast, during the growth phase, sales of the product typically accelerate as consumer and brand awareness increases, resulting in higher profits for the firm. During the maturity phase, sales continue at a steady pace and profits peak as consumers have already purchased the product and have developed an affinity with the brand. Eventually, however, during the decline phase, sales reduce as consumers lose interest or begin to identify alternatives on the market, resulting in a decrease of sales [2]. It should be noted here that the length of each phase can vary depending on the product and market conditions. Some products may experience short introduction phases and long maturity phases, while others may have a longer growth phase and a shorter decline phase. Firms must, therefore, monitor their products and adjust their strategies accordingly to maximize profits and extend their product’s life.

Fig. 1. Traditional Product Development Lifecycle.

Historically, firms have limited consumer involvement in the PD lifecycle to the end of production stage i.e., once the product has been launched in the market. However, with advancements in collaborative technologies, such as enterprise social media (e.g., Salesforce), product innovation has become increasingly consumer-driven with consumers demanding the ability to share their experiences and opinions with others and form greater connections with brands. Nowadays, consumers are evermore connected, informed, and empowered, and seek greater engagement in the PD lifecycle [3]. This, however, goes against the traditional practice of firms deciding which products should be launched into a market and what value should be created. Consumers also seek greater personalization of products with such customizability offering advantages to both firms and consumers. Through collaboration, firms obtain value by capturing the preferences of consumers while gaining insights into their behaviors. Conversely, consumers develop

374

U. Nwogu and R. Evans

greater understanding of the future products being developed by firms, ensuring that they receive enhanced value when the product is launched. During the Introduction phase, most firms focus on making consumers aware of the new product through advertising and marketing campaigns, while in the Growth phase, consumers accept the product and submit feedback on how well the product is meeting their needs and how it can be improved; often, this is when firms choose to release a Version 2.0 of the product (i.e., an updated version). In the Maturity phase, consumers should have built affinity with the brand and developed loyalty towards the firm and its products. At this point, firms can choose whether to further engage consumers through direct promotions and other marketing incentives; in so doing, they may be able to ‘fight off’ competition from other firms in intense markets. Finally, during the Decline phase, consumers become less interested in the product and look for alternatives that are provided by competitors. At this point, sales of the product often decline, and the product ultimately is retired unless it can be redesigned to remain relevant. Such interactions are required to ensure successful product development activities. Consumer participation provides a unique opportunity for firms and can help develop a competitive advantage over competitors. It is also crucial for improving decision making, increasing stakeholder satisfaction, helping identify early issues with products, and assisting with product iterations. To aid the management of the PD lifecycle, many firms now use PLM systems as part of their PD lifecycle which drives digitization through the factory all the way to the consumer. PLM, as a concept, refers to the process of data management that addresses the entire product lifecycle from ideation through to the retirement or recycling of a product [4]. With developments in web-based collaborative Table 1. PD Lifecycle adopted by most providers of PLM Systems. Phase

Description

Concept and Design

The Concept and Design phase or “Ideation” phase is where a product’s requirements are defined based on factors, including competitor analysis, gaps in the market, or customer needs

Development

In the Development phase, the detailed design of the product is created, along with any necessary tool designs. This phase includes validation and analysis of the planned product, as well as prototype development and piloting. This generates vital feedback on how the product is used and what further refinements are required

Production and Launch

Feedback from the pilot activities is used to adjust the product’s design and other components to produce a market-ready version. The production of the new product is scaled, which is followed by the launch and distribution of the product to the market

Service and Support

Following the launch of the product, the firm offers servicing and support to customers and end-users

Retirement

At the end of the product’s lifecycle, its withdrawal from the market must be managed – along with any retrials or absorption into new concept ideas

Consumer Engagement in the Design of PLM Systems

375

technologies, PLM systems have been extended to involve all actors involved in the PD lifecycle to help them collaborate throughout the PD lifecycle. In most modern PLM systems, the PD lifecycle is split into five phases (see Table 1). In the design of PLM systems, it is crucial that firms engage internal and external actors to develop a system that accommodates all stakeholders involved and integrates their needs and requirements. PLM systems facilitate the capture and dissemination of information throughout the PD lifecycle by providing an artefact-centric approach to information management [5] and help firms to capture, manage, and preserve the created information for their entire product portfolio rather than a single project or product [6]. Internal actors, such as engineers and designers, can provide valuable input on the specific features and functionality required for their product, while external actors can provide insights into the needs and preferences of the market when engaging with such systems. Similarly, online innovation communities can contribute to idea generation activities through initiatives, such as crowdsourcing. Such engagement activities can lead to a more successful product launch and deployment of the product. PLM systems have evolved as digital technology has advanced, expanding across the entire PD lifecycle. Traditional PLM has relied on document-centered systems, but modern use favors data sharing and exchange to store information as metadata or partial models in a database. The critical challenge here is to maintain the stored data as the sole source of truth for the product throughout its lifecycle. The emergence of cloud-based data management systems has also led to the prioritization of information-centered management [7–9]. For example, Oracle Cloud, offered as a Software-as-a-Service (SaaS) product, scales to meet the requirements of firms as it has the capability to extend well beyond product innovation to support the digital transformation needs of firms across industries and multiple enterprise software domains [8]. Another approach is to collaborate with other technology firms to integrate PLM systems with other software; for example, Siemens PLM has partnered with Microsoft to integrate their Teamcenter PLM system with Microsoft Teams [10], which allows users to collaborate on PD projects in real-time. More recently, PLM system developers have begun to use social media and online communities to engage external actors. For example, Autodesk has created an online community where users can share ideas and provide feedback on their PLM software, called Fusion Lifecycle [11]. Similarly, PTC Windchill has developed an online community where users can post questions, share ideas, and provide feedback. Finally, PLM systems are beginning to integrate advanced technologies, such as Artificial Intelligence (AI), to gain greater insights into how their PLM systems are being used and to identify areas for improvement; for example, SAP PLM has created an AI-powered system that analyzes user behavior and provides personalized recommendations for optimizing their PLM system [12]. Although much research exists on how internal and external actors are involved in the PD lifecycle, limited research exists on how consumers are involved in the design of PLM systems. The aim of this research, therefore, is to review and synthesize the existing literature on identified best practices of how consumers are involved in the design of PLM systems. The research questions addressed in this paper are as follows: 1. What is the current level of consumer engagement in the design of PLM systems? 2. What are the benefits of consumer engagement in the design process?

376

U. Nwogu and R. Evans

3. How can consumers be effectively integrated into the design of PLM systems? 4. What best practices currently exist for the involvement of consumers in the design of PLM systems? The remainder of this paper is structured as follows: in Sect. 2, the method used in this study is introduced, including an overview of the search strategy adopted to identify relevant literature. In Sect. 3, a review of literature on consumer engagement in product development is provided; this is given with the aim of answering the proposed research questions. Section 4 identifies the current best practices for consumer engagement in the design of PLM systems. Finally, Sect. 5 provides conclusions and suggests future research directions.

2 Method To develop a better understanding of how PLM systems are currently designed, and how consumers are involved in this process, this study used a mix of literature review and content analysis tools to analyze the content of related papers. Papers related to consumer engagement in the design and development of PLM systems were collected using relevant scientific citation databases, including Web of Science Core Collection database. A search for all relevant literature was performed in February 2023. The topic search string was composed of four parts. The first part focused on consumer engagement in the design and/or development process using: “user” OR “consumer” OR “customer” OR “community”. The second focused on consumer engagement methods using: “social product development” OR “open innovation” OR “crowdsourcing” OR “co-creation” OR “collaboration” OR “participation”. The third part focused on the system being designed and developed using: “system” OR “PLM” OR “tool” OR “software” OR “SaaS”. Finally, the fourth part focused on the development stage for consumer engagement using: “design” OR “development” OR “conceptualization” OR “ideation” OR “manufacturing” OR “production”. The total number of papers in all years from the search query was 88 articles. To identify the most relevant papers, we refined the papers by excluding editorial material (1), news item (1), and books (1). Then, we performed a more detailed control of titles, keywords, and abstracts. After fine-tuning, 85 papers were analyzed. Following examination of all records obtained by the search, we reviewed all titles, abstracts, and keywords to identify if the paper included discussions on best practices or offered advice for engaging consumers in the design or development of PLM systems. If the paper included commentary on best practices, it was marked as “1” and if not “0”. In total, 21 papers included discussions on best practices.

3 Literature Review PLM has emerged as a critical tool in modern firms for managing product development activities. The survival of any firm lies in their ability to make crucial decisions in managing their products throughout their PD lifecycle. PLM systems are designed to improve a firm’s capability to manage its PD operations, facilitate teamwork across

Consumer Engagement in the Design of PLM Systems

377

various business units and functions, as well as their interaction between different firms, including suppliers and external actors. The review of existing literature performed in this study highlighted the importance of PLM systems and their impact on product development. One of the key themes in the existing literature is the importance of consumer engagement in the design and development of PLM systems process. 3.1 What is the Current Level of Consumer Engagement in the Design of PLM Systems? According to the Digital Transformation of Product Design report, published by Oracle in 2020 [13], 2 out of 3 PD teams lack robust social tools for collaboration. Products are no longer developed in isolation but with collaborative teams using internal and external actors. To cater to the needs of a wide range of stakeholders, modern PD teams should consider input not only from within their own firm but also beyond it. Surprisingly, the report revealed that of 358 product design professionals surveyed, 22% of PD teams do not use any social tools, nor are they exploring the possibility of using them. Another 11% of those surveyed are considering their use, but do not currently employ them. Out of all the teams surveyed, only 16% possess the ability to easily gather feedback from all stakeholders, whereas 51% have limited (33%) or strong (18%) internal collaboration tools but lack a mechanism to obtain external feedback. The level at which consumers are involved in design activities is critical to the successful development and deployment of PLM systems. Consumer engagement encompasses different levels of involvement, leading to the extent to which the consumer exerts some influence during the development process [14]. Zhang and Dong [15] explored how users are involved in design practice and summarized four models of user engagement in the design and development process, namely: designers representing users, designers consulting users, users participating in the design process, and users as designers. The authors suggested that consumer engagement is essential for the success of design projects, as it allows designers to understand user needs and preferences and create products that better meet those needs. The authors further described the key features and benefits of each of their models, as well as the challenges associated. 3.2 What Are the Benefits of Consumer Engagement in the Design Process? Consumer engagement in the design of PLM systems can have a positive effect on the systems’ effectiveness. A study by Cantamessa, Montagna and Neirotti [16] found that PLM systems that involved users in the design process led to more efficient product data retrieval and a tighter workflow logic in the new PD process. The authors posited that consumer engagement improves the workflow logic, product carryover, and data retrieval, leading to increased productivity for users. This can help reduce the time it takes to bring a new product to market. He and King [17] conducted a review of 82 empirical studies and found that consumer participation in the development of information systems was minimally-to-moderately beneficial to developmental outcomes, with a stronger effect being experienced on the attitudinal/behavioral outcomes than on productivity outcomes. Similarly, Kujala [18] reviewed the benefits and challenges of consumer engagement in the development process and found that user involvement generally has a

378

U. Nwogu and R. Evans

positive effect, especially on user satisfaction and requirements capture. These provisions enable users to provide insights into new and emerging trends, which allow for the creation of innovative systems which consider the diversity of users. Involving consumers in the design of PLM systems can result in numerous benefits, such as improved system usability, increased user satisfaction, cost savings, and enhanced system adoption. Thus, developers of PLM systems should prioritize engagement with consumers in the design process to maximize the successful deployment of their systems. While involving consumers in the design process of PLM systems can lead to more effective and robust systems, the effectiveness of the system will also depend on other factors, including system design, organizational factors (e.g., employee resistance, cost of adoption, replication of global solution in a local scale), and industry characteristics (e.g., governance and policy issues, information handling, intellectual property issues etc.). 3.3 How Can Consumers Be Effectively Integrated into the Design of PLM Systems? Consumer engagement is crucial in the design of PLM systems as it ensures that the system aligns with users’ needs, requirements, and preferences. Several strategies exist that can be employed to effectively integrate consumers in the design of PLM systems. Damodaran [19] provided practical guidance for consumer involvement in the systems design process, emphasizing the importance of a clear differentiation between roles, an infrastructure to support consumer engagement, and avoiding common pitfalls in quality assurance procedures. This will lead to consumers embracing the PLM system and increase the overall adoption rate. Prahalad and Ramaswamy [3] suggested co-creation experiences as the next frontier in value creation in the process of integrating consumers into the design of PLM systems, however the authors’ study could be considered somewhat dated. This approach involves working collaboratively with consumers through various ethnographic methods, such as workshops, focus groups, and participatory design sessions. The authors suggested that by involving consumers in the value creation process, firms can create more personalized and meaningful experiences that meet consumer needs. The study also identified three key principles of co-creation experiences, namely: engagement, personalization, and connectivity, suggesting that by focusing on these principles, firms can create experiences that are both valuable to users and sustainable for the firm. Another approach to effectively integrating consumers in the design of PLM systems is to use agile development methodologies. Humpert et al. [20] recommended the use of the agile product development process as an impression of the consumer. The study reported that agile product development helps reduce the PD timeframe, which results in a shorter time-to-market which keeps the firm competitive. This method ensures consistent interaction between the consumers and developers leading to a high number of text responses which can be integrated as part of the requirements management process. User-centered design is another approach proposed by researchers [21], which is based on the integration of tools in the search for feasible solutions focused on meeting user requirements. By facilitating communication among members of the design team, the proposed design process resulted in improvements in four key areas. These included

Consumer Engagement in the Design of PLM Systems

379

a clearer identification of design demands and user needs, incorporation of these needs into the design process, reduction in development time, and generation of feasible and innovative solutions that prioritized functionality and innovation. Other novel methods for consumer engagement included the use of 3D printing for creating collaborative mock-ups and prototypes which were found to enhance the visualization of technical solutions and streamlined the process of refining the product concept. It also allowed for greater stakeholder participation, contributing to a usercentered design approach. The testing approach will play a critical role in evaluating product functionality and identifying opportunities for future enhancements. Further, the use of social media and crowdsourcing can be leveraged to engage a broader user base in the design process. Such platforms allow users to provide feedback, suggestions, and ideas leading to the development of more innovative PLM systems. Integrating end-user involvement in the design of PLM systems is critical to ensure the development of user-centric, effective, and innovative PLM systems. Participatory design, agile methodologies, social media and crowdsourcing platforms, and effective communication channels, can be leveraged to effectively integrate consumers. 3.4 What Best Practices Currently Exist for the Involvement of Consumers in the De-sign of PLM Systems? Studies that provide insights or guidelines for how to successfully engage consumers in the design process of PLM systems are scarce. The MEMORAe project is one such practice that allows stakeholders of a design team, regardless of where they are located, to collaboratively produce a shared understanding through the collaborative development of ontologies [22]. Ontologies represent a shared world view and can support communication and knowledge sharing through a community of practice. Stakeholders involved in the design process can express what they think about given subjects and create their own subjects. This ensures that everyone involved in the PD lifecycle has access to every element correlated with the product throughout its lifecycle. Arduin et al. [23] identified challenges, however, such as the complexity of PLM systems, and the need for crossfunctional collaboration, as important considerations for firms designing and developing PLM systems. Additionally, the authors’ emphasized the importance of incorporating social and technical factors in knowledge sharing frameworks as being particularly relevant, as effective knowledge sharing requires not only technological solutions, but also a supportive organizational culture. The task of engaging consumers in a value co-creation process is highly challenging for firms, especially in the online context, as competitive environments continue to become more complex. Leclercq, Poncin and Hammedi [24] suggested the practice of gamification as an approach to engaging consumers in the design process. This practice involves the integration of game mechanics into non-game contexts, aimed at encouraging consumers to be involved in activities of the firm in a sustainable and inclusive manner. The authors presented an overview of value co-creation, describing its potential benefits and challenges with using gamification. In the authors’ study, gamification in the design process was found to increase consumer engagement, motivation, and enjoyment, and ultimately improved the quality and quantity of contributions made by consumers. Several factors were identified to have contributed to the approach. Among them was

380

U. Nwogu and R. Evans

the use of clear and meaningful game mechanics, the provision of regular feedback and rewards, and the alignment of the gamification approach with consumers’ motivations and goals. Gamification was suggested to be considered by firms when designing new product development platforms to enhance engagement and value co-creation outcomes. The authors recommended that firms should carefully design and test their gamification approach to ensure that it is aligned with their consumers’ needs and motivations, and that they regularly evaluate and adapt the approach to optimize engagement and value co-creation outcomes. According to Weinert et al. [25], co-creation techniques offer a promising avenue for enabling employees to independently design materials tailored to their specific work contexts. To create a co-creation system that supports employees throughout the process of creating learning materials related to work processes, the authors conducted focus groups in vocational training schools and manufacturing settings, gathering requirements from co-creation concepts and practices. This led to the creation of four design principles for a collaborative learning and qualification system for use in the manufacturing industry. The authors used a mixed methods approach to validate their principles and features, demonstrating the success of the resulting artifact. Overall, the study provided useful insights into how a co-creation system can be designed to support learners in creating learning material, while considering the cognitive load involved. Ringen, Holtskog and Welo [26] proposed a value-driven co-creation framework that helps firms to identify and address the needs of their consumers through collaboration. The authors’ framework consisted of four stages, namely: ideation, co-creation, realization, and value delivery. During the ideation stage, firms identify consumer needs and preferences. In the co-creation stage, the firm works with consumers to generate and develop ideas. In the realization stage, the firm creates a prototype and tests it with the consumer and, finally, in the value delivery stage, the firm delivers the final product or service to the consumer. The proposed framework supports firms in improving their understanding of consumers’ needs and preferences and provides well-structured guidelines for each stage of the co-creation process. Additionally, the importance of collaboration and communication between the firm and consumers was emphasized. Ringen, Paalsrud and Lodgaard [27] explored the concept of interorganizational learning in manufacturing networks. The authors presented interorganizational learning as a key factor in the success of manufacturing networks, as it allows firms to share knowledge, skills, and resources with other firms. Four key processes of interorganizational learning were identified in the study, namely: acquisition, assimilation, transformation, and exploitation. Firms can improve their ability to learn from each other and work collaboratively in manufacturing networks where these processes are understood and internalized. Additionally, the authors emphasized the importance of communication and trust in interorganizational learning which is particularly relevant in the context of designing PLM systems where stakeholders and firms involved in the design ecosystem must work together to achieve the common goals of the system. Katona [28] explored the concept of user participation in design activities, focusing on the idea of Democratic Product Design (DPD), where firms allow consumers to vote on new product designs through social media platforms. Using a linear city model, the author analyzed the strategic implications of DPD in monopoly and duopoly settings,

Consumer Engagement in the Design of PLM Systems

381

especially how it can be used as a differentiation strategy for firms. The author posited that consumer engagement can lead to increased product innovation and differentiation, as well as increased consumer satisfaction and loyalty towards the brand. The study highlighted several case studies of firms that have successfully implemented democratic design, including LEGO and Threadless. Another crucial practice for user-involvement in the design process is the management of consumer preferences. Kpamma et al. [29] explored the application of Choosing by Advantages (CBA) decision system at the ideation stage of the design process. The authors conducted a workshop involving the design team, user-group, and firm management, to decide on the options using the CBA approach. The findings of the study illustrated that the CBA decision aid is a transparent and participative decision system, capable of creating and sustaining a collaborative atmosphere of trust and respect in the involvement of consumers to make design choices. The study also suggested that the observed attributes of the CBA decision system made it worthwhile to incorporate it in a user-involvement framework for the design process, as the tool can foster an atmosphere of collaboration between designers and consumers to arrive at design choices acceptable to all stakeholders.

4 Best Practices for Consumer Engagement in the Design of PLM Systems Following the review of currently available literature, it was possible to identify the best practices proposed for consumer engagement in the design of PLM systems. These included six key practices, which are described in Table 2. Table 2. Best Practices for Consumer Engagement in the Design of PLM Systems Best Practice

Description

Early and Continuous Involvement

Schemmer et al. [30] reported that involving consumers early in the design process is suitable for systems for process planning, as it has proven to detect usability conflicts early, and enables the identification of additional feature requests. The authors suggested that participatory design is a key enabler to develop accepted, usable, and useful interfaces. This will provide a platform for stakeholders in the PLM ecosystem to easily make their contributions and observations known during the design process. Consumers should be involved from the very beginning and throughout the entire PD lifecycle to ensure that their needs and requirements are met

User-centered Design The design process should be centered around the needs and requirements of consumers [15]. This involves conducting user research and usability testing to understand their unique preferences and challenges. In production Systems, Schemmer et al. [30] reported that user-centered design can increase ergonomics, usability, and user experience of process planning applications

(continued)

382

U. Nwogu and R. Evans Table 2. (continued)

Best Practice

Description

Collaboration

Collaboration between users, developers and other stakeholders involved in the design process is essential for the successful design and deployment of PLM systems. This involves open communication and regular feedback sessions. Kang [31] noted that character, choice, and customization, are critical elements that influence consumer commitment to co-creation and collaboration in design. The platform for collaboration must be designed in such a way that consumers understand what the process is all about to keep them focused on the project [32]

Iterative Design

The design of PLM systems should be iterative, with regular testing and feedback sessions with consumers [28]. This allows for incremental improvements to be made and ensures that the final PLM system meets the needs of all stakeholders

Training and Support Consumers should receive adequate training and support to ensure that they can effectively use the PLM system that is created. Jokela [33] identified workshops as an effective tool for training consumers in using systems. Such workshops bring all the stakeholders in the design ecosystem together, helping designers to understand the needs of consumers and what kind of data is expected from the user. An ontology should be created to support communication and knowledge sharing among all participants [22, 23]. The output of such training sessions is to make the user understand the system properly and to incorporate the consumers’ requirements into the development of the PLM system Usability Testing

The PLM system should be tested for usability to ensure that it is intuitive and easy to use. Gulliksen et al. [34] suggested the availability of experts in the field and involving them in the project early and continuously. The authors also suggested that usability designers must be given full authority to decide on matters pertaining to the usability of the PLM system and its future use situation

5 Conclusion This paper provides a critical review of best practices related to the engagement of consumers in the design of PLM systems. Designing PLM systems can be a complex task that requires extensive knowledge sharing and collaboration with all stakeholders involved in the design ecosystem. Involving consumers in the design process is critical to the success of PLM systems as it guarantees that the needs and requirements of consumers are integrated in the launched product. The best practices outlined in this paper aim to serve as a guide in the development of future PLM systems. By following the suggested best practices, developers of PLM systems can create systems that meet the needs of consumers in a usable way. Of course, the framework proposed in this paper has some limitations as it has not been validated in a real-life design environment. Further work must, therefore, include an empirical investigation involving consumers in the design of PLM systems, especially with respect to the design of user interfaces.

Consumer Engagement in the Design of PLM Systems

383

Future work will include examining how the user interfaces and experiences of PLM systems can be enhanced through the engagement of consumers in their design process. Empirical study with software development companies in Nova Scotia, Canada, will also be conducted to confirm that the best practices reported in this paper can increase the value offered by PLM systems when consumers are engaged. Such research will require data collection from consumers using qualitative methods, including surveys, usability testing, and case study investigations with partner firms. By employing a mixed methods approach, the effectiveness of consumer engagement practices can be assessed, providing valuable insights for system refinement. Funding. This paper is based on work funded by the Natural Sciences and Engineering Research Council (NSERC) of Canada under Grant No. RGPIN-2022-05008.

References 1. Chen, B.-C.: Study on establishment of product life cycle model of cultural creative product industrialization. In: International Conference on Applied System Innovation, Japan, pp. 1259–1262. IEEE (2017) 2. Lin, J.-S., Chang, J.-H., Yang, I.-Y.: An adjustment method for material inventory control decision considered in product life cycle. In: International Conference on Industrial Engineering and Applications, Singapore, pp. 310–313. IEEE (2018) 3. Prahalad, C., Ramaswamy, V.: Co-creation experiences: the next practice in value creation. J. Interact. Mark. 18(3), 5–14 (2004) 4. Hayat, M., Winkler, H.: From traditional product lifecycle management systems to blockchain-based platforms. Logistics 6(3), 1–14 (2022) 5. Jones, D.E., et al.: The design of visual information objects in three-dimensional virtual environments for engineering information navigation. In: International Design Conference, Croatia, pp. 1583–1594. The Design Society (2018) 6. Aram, S., Eastman, C.: Integration of PLM solutions and BIM systems for the AEC industry. In: International Symposium on Automation and Robotics in Construction, pp. 1046–1055. The International Association for Automation and Robotics in Construction, Canada (2013) 7. Adamenko, D., Kunnen, S., Nagarajah, A.: Digital twin and product lifecycle management: what is the difference? In: Nyffenegger, F., Ríos, J., Rivest, L., Bouras, A. (eds.) PLM 2020. IAICT, vol. 594, pp. 150–162. Springer, Cham (2020). https://doi.org/10.1007/978-3-03062807-9_13 8. Arbor, A.: Next Generation – Digitally Connected PLM. CIM Data, Michigan (2019) 9. CFI Education Inc.: Product Lifecycle Management. https://corporatefinanceinstitute.com/ resources/management/product-life-cycle/. Accessed 23 Mar 2023 10. To-Increase B.V.: Siemens Teamcenter PLM integration. https://appsource.microsoft.com/ en-cy/product/dynamics-365-for-operations/to-increasebv.394060bdf-dqb5-46a0-b02d-c3c 3kkdfoc80fsjs056?tab=Overview&exp=ubp8. Accessed 12 Apr 2023 11. Autodesk, Autodesk Community. https://www.autodesk.ca/en/community. Accessed 13 Apr 2023 12. SAP, What is Product Lifecycle Management (PLM)? https://www.sap.com/insights/whatis-product-lifecycle-management.html. Accessed 8 Apr 2023 13. Tara, R.: The Digital Transformation of Product Design. https://www.engineering.com/story/ research-report---the-digital-transformation-of-product-design-ydxqj. Accessed 27 Mar 2023 14. Fischer, B., Peine, A., Ostlund, B.: The importance of user involvement: a systematic review of involving older users in technology design. Gerontologist 60(7), 513–523 (2020)

384

U. Nwogu and R. Evans

15. Zhang, B., Dong, H.: User involvement in design: the four models. In: Zhou, J., Salvendy, G. (eds.) ITAP 2016. LNCS, vol. 9754, pp. 141–152. Springer, Cham (2016). https://doi.org/10. 1007/978-3-319-39943-0_14 16. Cantamessa, M., Montagna, F., Neirotti, P.: Understanding the organizational impact of PLM systems: evidence from an aerospace company. Int. J. Oper. Prod. Manag. 32(2), 191–215 (2012) 17. He, J., King, W.R.: The role of user participation in information systems development: implications from a meta-analysis. J. Manag. Inf. Syst. 25(1), 301–331 (2008) 18. Kujala, S.: User involvement: a review of the benefits and challenges. Behav. Inf. Technol. 22(1), 1–16 (2003) 19. Damodaran, L.: User involvement in the systems design process – a practical guide for users. Beh. Inf. Technol. 15(6), 363–377 (1996) 20. Humpert, L., Röhm, B., Anacker, H., Dumitrescu, R., Anderl, R.: Method for direct end customer integration into the agile product development. Procedia CIRP 109, 215–220 (2022) 21. Santos, A.V.F., Licursi, L., Amaral, M.F., Cavalcanti, A., Silveira, Z.C.: User-centered design of a customized assistive device to support feeding. Procedia CIRP 84, 743–748 (2019) 22. Abel, M.-H.: Competencies management and learning organizational memory. J. Knowl. Manag. 12(6), 15–30 (2008) 23. Arduin, P.-P., Le Duigou, J., Abel, M.-H., Eynard, B.: Knowledge sharing in design based on product lifecycle management system. In: Chakrabarti, A. (ed.) ICoRD’15 – Research into Design Across Boundaries Volume 2. SIST, vol. 35, pp. 507–517. Springer, New Delhi (2015). https://doi.org/10.1007/978-81-322-2229-3_43 24. Leclercq, T., Poncin, I., Hammedi, W.: The engagement process during value co-creation: gamification in new product development platforms. Int. J. Electron. Commer. 21(4), 454–488 (2017) 25. Weinert, T., Billert, M., Thiel de Gafenco, M., Janson, A., Leimeister, J.M.: Designing a co-creation system for the development of work-process-related learning material in manufacturing. Computer Supported Cooperative Work (CSCW) 32(1), 5–53 (2022) 26. Ringen, G., Holtskog, H., Welo, T.: A Novel Value Driven Co-creation Framework. In: Lalic, B., Majstorovic, V., Marjanovic, U., von Cieminski, G., Romero, D. (eds.) APMS 2020. IAICT, vol. 591, pp. 663–671. Springer, Cham (2020). https://doi.org/10.1007/978-3-03057993-7_75 27. Ringen, G., Paalsrud, F., Lodgaard, E.: Interorganizational learning in manufacturing networks. In: Lalic, B., Majstorovic, V., Marjanovic, U., von Cieminski, G., Romero, D. (eds.) APMS 2020. IAICT, vol. 591, pp. 680–686. Springer, Cham (2020). https://doi.org/10.1007/ 978-3-030-57993-7_77 28. Katona, Z.: Democracy in product design: consumer participation and differentiation strategies. Quant. Mark. Econ. 13(4), 359–394 (2015). https://doi.org/10.1007/s11129-0159164-z 29. Kpamma, Z.E., Adinyira, E., Ayarkwa, J., Adjei-Kumi, T.: Application of the CBA decision system to manage user preferences in the design process. J. Professional Issues Eng. Educ. Pract. 142(1), 05015004 (2016) 30. Schemmer, T., Brauner, P., Schaar, A.K., Ziefle, M., Brillowski, F.: User-centred design of a process-recommender system for fibre-reinforced polymer production. In: Yamamoto, S., Mori, H. (eds.) HCII 2020. LNCS, vol. 12185, pp. 111–127. Springer, Cham (2020). https:// doi.org/10.1007/978-3-030-50017-7_8 31. Kang, J.-Y.M.: Customer interface design for customer co-creation in the social era. Comput. Hum. Behav. 73, 554–567 (2017) 32. Norman, D., Euchner, J.: Design for use. Res.-Technol. Manag. 59(1), 15–20 (2016)

Consumer Engagement in the Design of PLM Systems

385

33. Jokela, T.: Guiding designers to the world of usability: determining usability requirements through teamwork. In: Seffah, A., Gulliksen, J., Desmarais, M.C. (eds.) Human-Centered Software Engineering – Integrating Usability in the Software Development Lifecycle. HumanComputer Interaction Series, vol. 8, pp 127–145. Springer, Dordrecht (2005). https://doi.org/ 10.1007/1-4020-4113-6_8 34. Gulliksen, J., Goransson, B., Boivie, I., Persson, J., Blomkvist, S., Cajander, A.: Key principles for user-centred systems design. Behav. Inf. Technol. 22(6), 397–409 (2010)

A Distributed Ledger Technology Solution for Connecting E-mobility Partners Radu Ungureanu1(B) , Selver Softic2 , Emil St. Chifu1 , and Ioan Turcin2 1 Technical University of Cluj-Napoca, Memorandumului 28, 400114 Cluj-Napoca, Romania

[email protected] 2 CAMPUS 02, University of Applied Sciences, Körblergasse 126, 8010 Graz, Austria

Abstract. The work is focused on the e-mobility industry and proposes an innovative solution for handling roaming connections in the context of electric vehicles charging. We address the privacy and performance issues of the roaming hub approach for enabling roaming by designing a peer-to-peer platform for connecting e-mobility partners. We attempt to achieve interoperability by using the current industry standards as a foundation to our approach. The idea behind this solution is to record and verify transactions between the charging point operators (CPOs) and the e-mobility service providers (EMSPs) on a permissioned distributed ledger to increase the trust between the two parties. The agreements between the partners are signed and enforced via smart contracts. Keywords: e-mobility · EV charging · Roaming · Distributed ledger technology · Smart contracts

1 Introduction The e-mobility industry has been on a raising trend over the past few years. In 2010 the total number of registered electric vehicles in the European Union was 600, while in 2021 there were over 1,729,000 such vehicles registered [1]. Moreover, the European Parliament and European Council signed an agreement that states that the sale of vehicles powered by internal combustion engines will be forbidden starting with the year 2035 and only electric vehicles could be registered from that point onwards to reduce carbon dioxide emissions [2]. The widespread adoption of this technology brings with it a series of challenges, one of those being the charging process. The engines of electric vehicles are powered using the energy stored in a battery attached to the car. The charging process of an electric vehicle is therefore the process of filling up its battery. An electric vehicle user has several options in terms of locations when they want to perform the charging process, such as: on their premises, in a private location (e.g. workplace, hotel, restaurant), or at a public charge station. The current work is focused on the functioning of public charge stations that can be accessed by customers based on an agreement. If we consider the energy flow from the source to the end customer, which is the EV user in our case, © IFIP International Federation for Information Processing 2023 Published by Springer Nature Switzerland AG 2023 E. Alfnes et al. (Eds.): APMS 2023, IFIP AICT 689, pp. 386–398, 2023. https://doi.org/10.1007/978-3-031-43662-8_28

A Distributed Ledger Technology Solution for Connecting E-mobility Partners

387

there may be multiple organizations involved. An organization that owns and operates charging stations takes the role of a charge point operator (CPO) [3]. The CPO role is responsible for facilitating the physical charging process on their premises. Therefore, they must supply the energy and offer a physical interface for the electric vehicle. In general, the charging stations managed and maintained by the CPO are distributed at different geographical locations and need to be in permanent connection with the central system of the CPO. The operator must collect different sorts of live data from the stations like the availability of connectors or information about the charging sessions. The EV user has two options for accessing the charging stations offered by a CPO: ad-hoc or contract-based [4]. The ad-hoc method, where the user must use a bank card to initiate the process, has a series of disadvantages: it may add extra processing fees, reluctance from the user side on providing their cards without knowing beforehand the sum that will be debited, and increased implementation and maintenance effort for the charge point operator. The other option, contract-based, implies that the electric vehicle user signs an agreement beforehand with an organization that has the role of an e-mobility service provider (EMSP) [5]. The EMSP has a series of public or private contract offers that potential customers can sign. These offers would stipulate a certain tariff that would be applied when the customer performs the charging process. After a customer completes the registration process at an EMSP they will be provided with a method through which they can authenticate themselves at a charge station. There are several methods through which the users can perform this: RFID cards, tokens stored on mobile-based applications, or via a plug and charge protocol (ISO 15118) which reads the identity of the user when the physical connector is plugged into the car. These tokens are read by the charging stations and then sent further for verification to check whether the user is allowed to initiate the charging process. In the e-mobility industry, an organization could take on both the charging point operator and e-mobility service provider roles. This is also the most common scenario when using a public charging infrastructure, on a day-to-day basis. However, there may arise situations in which the user is not able to use the charge points of the EMSP he has signed a contract with due to various reasons like lack of charge points in the desired range or when all the charging poles are already in use. The only way through which the users could overcome these problems by themselves is by signing contracts with multiple EMSPs. However, managing multiple contracts, charging cards and mobile applications could be too much of an inconvenience for many individuals. The solution for this problem is roaming, where an entity that has the EMSP role signs an agreement with an entity that has the CPO role. The agreement would then stimulate the fact that the clients of the EMSP can use the charging infrastructure of the CPO under certain conditions. As the report [6] issued by the European Court of Auditors specifies, currently at the level of the European Union there is no standard for achieving interoperability between the different actors in the roaming process and there are currently multiple approaches from different companies for addressing this topic. The goal would be to have a single standard that is implemented and used by most of the providers, which would lead to larger charging networks and increased availability for the EV users. Our solution seeks to simplify and improve the efficiency of EV roaming from a technical perspective.

388

R. Ungureanu et al.

2 Current State of the Industry According to the authors of [7], there are two ways through which two partners could establish a roaming connection: directly, by establishing a peer-to-peer connection, or indirectly, via a roaming hub. As [8] specifies, the roaming platform must satisfy a series of requirements such as interoperability, data consistency, reliability, and data security (Fig. 1).

Fig. 1. Central hub in the context of EV roaming.

Roaming hubs are centralized entities that act as an intermediary for every message sent between the partners. This means that all requests are sent to the roaming hub which will in turn route them to the appropriate partner. Some examples of roaming-hub platforms that are currently active in the EU are e-clearing.net, Hubject, and Gireve. From a partner’s perspective, the biggest advantage of this technique is that you will only have to implement a single communication interface and connect to a single entity, the hub, to establish one or multiple roaming connections. The hub then must keep track of all the agreements signed between the partners and do the routing of a request only if there is a connection between the two sides. However, this method comes with a series of downsides due to the third party involved in the transactions: performance overhead, extra processing fees, and reduced privacy. The performance overhead comes from the fact that there will always be an extra hop in communication. Another disadvantage is the ever-increasing amount of data on the hub’s side, which may then slow down actions like database operations and lead to longer response times. In this case, the partners have no choice, but to rely on the optimization methods applied by the hub. By establishing a direct connection with the roaming partners, one could eliminate all third-party-related problems. On the other hand, if the peer-to-peer connection is established in a non-regulated environment then certain problems could appear such as different interfaces for different partners or a lack of trust between the transaction’s parties. These problems can be solved by creating a peer-to-peer network with a set of

A Distributed Ledger Technology Solution for Connecting E-mobility Partners

389

pre-defined rules stipulated within smart contracts that makes use of distributed ledger technology (DLT) capabilities. DLTs gained widespread attention in 2008 with Satoshi’s introduction of the Bitcoin blockchain network [9]. The main idea behind the Bitcoin concept was to provide a platform on which funds can be transferred directly between peers without the intervention of a third party, such as a bank. According to the author, the conventional means of electronic money transfer has a major drawback in the trust model. The participants of a transaction must always rely on the same third-party to mediate any conflicts. They introduced a peer-to-peer transaction model that relies on public key cryptography. The transactions would then be included in blocks, using the Proof of Work consensus algorithm, and then stored on a ledger that is distributed across all nodes of the network. Smart contracts were first introduced by Nick Szabo in 1996 [10]. The goal of a smart contract is to transpose the terms of an agreement within a structure that can be executed by a computer and therefore automating the validation process. Bitcoin could also execute smart contracts, but they were specified in a stack-based language whose set of instructions did not allow the modeling of complex business logic. The first introduction of a Turing-complete language for specifying smart contracts within a DLT network was in 2014 with Ethereum’s release [11]. The bytecode of an Ethereum smart contract is executed by the Ethereum Virtual Machine, which runs on top of the network nodes. With the aid of smart contracts running on top of blockchain-based solutions, organizations can achieve increased trust and transparency in a B2B environment. This fact is highlighted in the following studies [12, 13]. Regarding the usage of distributed ledgers in the context of e-mobility, we have a series of works around this topic. The authors of [14] describe how the transaction corresponding to a charging session can be processed on a distributed ledger and its corresponding steps: authorization, energy transfer, and payment. However, the authors discuss the solution in the context of public blockchains and they do not address the confidentiality of the data and of the transaction, which must be considered in an enterprise context where you store and process the data of end customers. The work [15] presents an entire architecture for billing roaming EV users. It describes all the components, including energy flow from the grid to the electric vehicle. On the other hand, they vaguely describe how can the transactions be included on the ledger and how will the partners reach a consensus.

3 Peer-to-Peer Roaming via DLT 3.1 Context We address the problems of the central hubs and classical peer-to-peer connections by introducing a regulated peer-to-peer network. All transactions will be stored on a permissioned distributed ledger. The network contains two types of nodes: EMSP and CPO. Each node will be responsible for maintaining the list of partners that it is connected to. The trust between the partners will be increased through smart contracts. These will be written in a Turing-complete programming language and can impose the settlement of transactions in an automated manner.

390

R. Ungureanu et al.

As [16] specifies, in the e-mobility industry there are currently three tiers of communications (Fig. 2).

Fig. 2. Tiers of communication in EV roaming as per [16].

The first tier represents the communication between the electric vehicle and the charging station. Currently, most charging stations allow only energy transfer through the cable. However, there are currently protocol standards at this level, such as ISO/IEC 15118, which allow the vehicle to store information and transfer it to the charging station. The plug and charge specification requires the standard to be implemented physically on the vehicle. The authentication token of the user can therefore be stored within the vehicle and when the connector is plugged in, the authorization is done automatically, without having to provide an RFID card or a token stored on a mobile-based application. On the second tier, there is communication between the charging station and the central system of the operator. This is usually done via the Open Charge Point Protocol (OCPP). The protocol allows live updates regarding the performed charging sessions and the state of the charging poles within the station. If the EV user is a direct client of the operator, then the first two tiers can satisfy the entire process. However, if roaming is involved then a third tier must be introduced. This tier contains the communication between the system of the operator that facilitates the charging session and the system of the roaming partner to which the EV user is subscribed. Various protocols can be used to facilitate the communication at this level, the most popular being: Open Charge Point Interface (OCPI), Open Intercharge Protocol (OICP), Open Clearing House Protocol (OCHP) or eMIP (eMobility Inter-operation Protocol). All these protocols are developed by different entities and have slightly different features. Our solution is focused on the third tier of communication, but it also involves the EV user, which is placed on the first tier.

A Distributed Ledger Technology Solution for Connecting E-mobility Partners

391

3.2 Network Architecture The system runs as a permissioned DLT network running on a series of nodes with the incentive of enabling roaming functionalities on their side. A node can act as both a CPO and an EMSP. Each party can sign agreements with other peers, the terms of the agreement being stored on the ledger and then ensured via smart contracts. Since this is a permissioned network, but all members are equal in terms of rights, a consortium must be defined for the governance of the network and the establishment of the membership criteria. Even though the charge point and the EV user are not part of the DLT network itself, they are part of the system since they are the main actors in the charging process (Fig. 3).

Fig. 3. DLT for roaming conceptual architecture.

3.3 System Capabilities Privacy. One of the most important aspects for a business in the e-mobility industry is keeping its data private. With a public blockchain solution like Bitcoin [9] or Ethereum [10], where everyone can see the transactions from the DLT and it is therefore impossible to satisfy the privacy requirements. However, if a permissioned ledger is chosen then

392

R. Ungureanu et al.

the network members can choose the visibility level of the transactions and can make it available only to a subset of peers that are interested in the outcome of the transaction. By having a peer-to-peer solution where we do not have to rely on the forwarding capabilities of a roaming hub we also eliminate the possible storing issues that this approach may bring. An entity may be reluctant to store data that is not supposed to be public on a system that is not under its direct control. Increased Efficiency. Unlike the hub scenario, where two partners that enabled roaming functionalities between their systems must first send the data to a third party which forwards them to the sender, the communication in our system will be direct. The two partners can send messages to one another without performing an extra hop and therefore reduce the complexity of communication. This also eliminates the scalability issues that a central hub would face and which may lead to performance issues. Static data such as information about charging stations or authentication tokens grows exponentially since it must be stored and processed for each partner that joins the platform. In the peer-topeer scenario each partner is responsible for storing and maintaining its own data the growth is usually linear. Standardization. Many of the downsides of a central hub can be eliminated by connecting directly to a partner. However, if there is no universal approach among the industry partners then it would be very hard to keep and maintain such systems. Each entity could have its preferences in terms of interfaces and handling all these would increase the complexity of the edge services. In our system, we propose a standard way of interfacing where the agreements would become eligible immediately after the execution of a smart contract. Automated Billing. One of the advantages of distributed ledger technologies is their ability to handle digital assets. In the usual roaming scenario, a charge point operator allows the customers of an e-mobility service provider to charge their vehicles on their infrastructure. The EMSP must then pay a fee to the CPO accounting for the access and the energy consumed by its customer at the charging site. These fees could be settled immediately in the form of digital assets in a secure and private manner.

4 Design Choices 4.1 DLT Platform As we have previously mentioned, public blockchain networks cannot be used as a foundation for our system due to privacy issues. However, several permissioned DLT platforms can be taken into consideration to form the base of such a solution: Quorum [17], Hyperledger [18], or Corda [19]. We analyzed the features and limitations of these three platforms and decided that a good foundation for the current state of our solution is represented by Hyperledger. The main argument is that this is a platform that possesses a high degree of flexibility in terms of technology. This allows us to design and implement a solution that fits the standards that are already present in the industry, facilitating the adoption of our solution by the existing players. Because this is a permissioned network, the participants must be identified. The network can be managed as a consortium led

A Distributed Ledger Technology Solution for Connecting E-mobility Partners

393

by the participating organizations. To adopt this solution, a party will have to deploy a physical node on the network. The partner will then be able to connect to the other parties by setting up roaming agreements. Once a roaming agreement is signed the EV users can access the infrastructure of their provider’s partners. 4.2 Roaming Protocols Since the connection between the partners is a direct one, an appropriate interface of communication must be defined. Regarding this, there are two approaches: either to define a new way of interfacing or to make use of an existing e-roaming protocol. For the current model, we propose a solution based on one of the protocols that are already being used in the e-mobility industry. This would make the approach easier to adopt by the current market players. In [20] the authors present a comparative analysis of the most popular protocols used across the European Union countries. Out of the analyzed protocols only OCPI and OCHP-direct support peer-to-peer communication. However, OCHP-direct is rather a subset of the OCHP protocol which was originally designed to work along with roaming hubs. Another minus of OCHP-direct is that it currently has only a SOAP-based interface. On the other hand, all OCPI’s operations are modeled to work for both peer-to-peer and centralized messages. Moreover, OCPI has a RESTful interface which makes it a better solution for our current design (Table 1). Table 1. Comparative analysis of e-roaming protocols [20]. Protocol

Owner

Open Source

Peer-to-peer

OCPI (2.2)

NKL

Yes

Yes

OCHP-direct (0.2)

e-clearing.net

Yes

No

eMIP (0.7.4)

Gireve

No

No

OICP (2.2)

Hubject

No

No

5 Transaction Model 5.1 Proposed Interaction Roaming capabilities enable the exchange of multiple data categories between the partners, including charge point locations or live information about sessions and, most important, billing data for the charging sessions performed by the other partner’s users. The billing information is created by the operator and stored in an object called charge data record (CDR). Our goal is to create a transaction model in which we can verify the validity of CDRs by recording timestamps from both the operator and the EV user and feeding those inputs to a smart contract. The OCPI protocol specification [21] includes a “Commands” module through which EV users can send requests to the operator via the mobile application of the provider.

394

R. Ungureanu et al.

Fig. 4. Sequence diagram for the start session scenario.

These commands are then forwarded to the operator which in turn routes the requests to the charging stations, via an internal call within their organization. Among these commands, we have an operation for starting a charging session, called “START_SESSION” and an analogous operation for stopping the session, named “STOP_SESSION”. In the proposed workflow the EV user who is subscribed to the EMSP would authenticate at the charging station of the CPO using their token. They would then proceed to plug the connector to their vehicle and start the charging process by sending a “START_SESSION” command from the mobile application. The provider then forwards the command to the operator. The operator then triggers the charging session remotely on their charging station and sends an asynchronous response to the EMSP for the initial “START_SESSION” command. The EV user can get live updates about the charging process on their mobile application via the OCPI “Sessions” module. To stop the charging session, they would again have to use the mobile application and the provider will again do the forwarding. The provider will record the timestamps when they received the “START_SESSION” and “STOP_SESSION” commands from the EV user. A charge data record, which acts

A Distributed Ledger Technology Solution for Connecting E-mobility Partners

395

as a bill for the charging session, includes the start and end timestamps of the session, but they are created by the operator. With our design, we want to ensure the operator’s timestamps and the provider’s timestamps correspond and that the tariffing rules were properly applied. If notable differences appear this means that the operator could be altering the timestamps of the charging session to obtain larger profits and take advantage of the provider’s clients. In the “START_SESSION” scenario (see Fig. 4), the EMSP is recording multiple timestamps. The first one, tSENT , is stored when the EMSP forwards the “START_SESSION” command to the CPO. When the EMSP receives the synchronous response from the CPO it will record the tSTART_SESSION_RECEIVED timestamp, being able to see the response time of the operator. After the operator receives a positive response from the charging station for initiating the session, it will asynchronously send the EMSP an “ACCEPTED” response. The EMSP again records the timestamp tACCEPTED , which is used for both verifying the start of the charging session and for computing the response time of the station. The CPO gets continuous updates from its charging stations regarding the status of the charging sessions. Therefore, the CPO will construct a “Session” object following the OCPI specification and forward it to the EMSP. The EMSP will then extract the tSESSION_START , which will be sent to a smart contract for verification, together with the tACCEPTED timestamp. According to the workflow, the session is started after the “ACCEPTED” response is recorded. If this condition is not satisfied, then the details of the charging session may be altered. This verification allows us to check whether the start of the session timestamp is correct, the end of the session timestamp can be verified similarly if the “STOP_SESSION” is sent by the user. To compensate for any possible network delays, the provider and operator could agree on specific thresholds when signing their agreements and integrate it into the smart contract. The violation of the threshold would allow the provider to reject a certain CDR. Moreover, when the EMSP is receiving updates from the CPO, it is recording the timestamps and can compute the response times of both the CPO central system and the charging stations. A service level agreement will then be introduced between the CPO and the EMSP to define the acceptable response times. Other plausibility checks can be integrated within these smart contracts, for instance analyzing the consumed energy in the specified time frame, since the provider already knows the power capabilities of the charging station. If a CDR is rejected then the manual intervention of the provider and operator is required, via a separate smart contract to establish the sum that must be paid. The platform can also keep track of the violations of the agreement so that multiple breaches would eventually generate penalties. 5.2 Consensus Since this is a peer-to-peer platform, the transactions that are generated by the smart contract executions must be validated within the network. The goal of this architecture is to facilitate a platform for multiple partners where they could fluently achieve their business goal. Therefore, for validating the transactions, the two participating peers could implement a rotation mechanism and always choose a random notary party from a pool of partner nodes. This pool must be agreed upon by both parties and will preferably be composed of nodes that have a roaming connection with both peers. Our choice of

396

R. Ungureanu et al.

technology in terms of the DLT, Hyperledger Fabric, supports the notary concept in terms of endorsement policies [22]. The incentive for a node to act as a notary for two of its partners is that when they will be part of a transaction, they will also need a validator to have their changes published on the ledger. Moreover, a rewarding mechanism can be defined for the notaries by adding an additional smart contract to provide an additional incentive for the endorsers.

6 Conclusion The e-mobility industry is still evolving and currently, from a technical perspective, we have multiple approaches that are trying to become predominant in this business market. According to the current trends, the user base of e-mobility organizations will keep increasing over the next few years and it is hence necessary to have a common view regarding the processes involving roaming. With our work, we attempt to provide an innovative way of connecting the e-mobility partners by building upon the existing foundations. We analyzed the current industry standards: the centralized roaming hub and the peer-to-peer approach. The downsides of hubs, which include increased trust and limited privacy, made us look into the peer-to-peer approach. However, in direct communication, without a concrete approach and a regulated environment, the peers may be misusing the protocols or attempting to violate the conditions of an agreement. Concerning this issue, we introduced a permissioned network based on distributed ledger technologies. By using a permissioned approach for joining the network, the data of an organization can be kept private and the information can be shared with the other peers on a needto-know basis. Therefore, objects like the charging sessions or the charge detail records will only be shared by the parties that take part in the transaction. For the communication between the partners, we proposed the implementation of a current standard, the Open Charge Point Interface (OCPI) protocol, since the businessspecific objects are already designed and supports peer-to-peer communication. Besides, the organizations that already use this protocol would be able to migrate to our solution with reduced effort. Our analysis led us to the conclusion that DLTs can provide benefits by automating processes with the use of smart contracts. In a traditional peer-to-peer interaction, without the use of DLTs, the settlement of conflicts can prove to be a cumbersome process. On the other hand, smart contracts verify and impose possible penalties automatically. Billing between the partners can be realized directly on the network by using digital assets. For the exchange of session data between an EMSP and a CPO, we introduced a plausibility check that can be performed within smart contracts. The plausibility check relies on recording and verifying timestamps of the commands received by the EMSP from the EV user through a mobile app and the responses of the CPO to these commands. The next steps for our work would be to implement a proof of concept based on the current design, using Hyperledger Fabric as a DLT and the OCPI protocol for modeling the communication. Currently we developed the interaction based on the existing industry standards. However, since the e-mobility industry is still evolving, our platform is flexible to changes and to the integration of other approaches. New standards, like

A Distributed Ledger Technology Solution for Connecting E-mobility Partners

397

the plug and charge (ISO 15118) will probably modify current processes. This could simplify the communication on the EV user side and might allow the vehicle to communicate directly with the charging station or even with the EMSP, without having to rely on the mobile app for asserting the validity of the charging session timestamps.

References 1. European Environment Agency: New registrations of electric vehicles in Europe (2022). https://www.eea.europa.eu/ims/new-registrations-of-electric-vehicles/. Accessed 15 Apr 2023 2. European Commission: Zero emission vehicles: first ‘Fit for 55’ deal will end the sale of new CO2 emitting cars in Europe by 2035 (2022). https://ec.europa.eu/commission/presscorner/ detail/en/ip_22_6462/. Accessed 15 Apr 2023 3. Greenflux blog. https://www.greenflux.com/what-does-a-charge-point-operator-do/. Accessed 15 Apr 2023 4. The AA: Charging around Europe in an Electric Vehicle (2018). https://www.theaa.com/eur opean-breakdown-cover/driving-in-europe/charging-around-europe-in-an-electric-vehicle. Accessed 15 Apr 2023 5. Parker, M.: What is EV roaming and how it is developing (2021). https://monta.com/uk/blog/ what-is-ev-roaming-how-its-developing/. Accessed 15 Apr 2023 6. European Court of Auditors: Infrastructure for charging electric vehicles: more charging stations but uneven deployment makes travel across the EU complicated (2021). https://www.eca.europa.eu/Lists/ECADocuments/SR21_05/SR_Electrical_c harging_infrastructure_EN.pdf. Accessed 15 Apr 2023 7. Ferwerda, R., Bayings, M., Kam, M., Bekkers, R.: Advancing e-roaming in Europe: towards a single “Language” for the European charging infrastructure. World Electric Veh. J. 9, 50 (2018). https://doi.org/10.3390/wevj9040050 8. Noyen, K., Baumann, M., Michahelles, F.: Electric mobility roaming for extending range limitations. In: 2013 International Conference on Mobile Business (2013) 9. Nakamoto, S.: Bitcoin: A Peer-to-Peer Electronic Cash System (2008) 10. Szabo, N.: Smart Contracts: Building Blocks for Digital Markets (1996) 11. Buterin, V.: Ethereum Whitepaper (2013) 12. Desai, S., Deng, Q., Wellsandt, S., Thoben, KD.: An architecture of IoT-based product tracking with blockchain in multi-sided B2B platform. In: Ameri, F., Stecke, K., von Cieminski, G., Kiritsis, D. (eds.) APMS 2019. IFIP, vol. 566, pp. 458–465. Springer, Cham (2019). https:// doi.org/10.1007/978-3-030-30000-5_57 13. Chiacchio, F., D’Urso, D., Compagno, L., Chiarenza, M., Velardita, L.: Towards a blockchain based traceability process: a case study from pharma industry. In: Ameri, F., Stecke, K., von Cieminski, G., Kiritsis, D. (eds.) APMS 2019. IFIP, vol. 566, pp. 451–457. Springer, Cham (2019). https://doi.org/10.1007/978-3-030-30000-5_56 14. Kirpes, B., Becker, C.: Processing electric vehicle charging transactions in a blockchain-based information system (2018) 15. Yaqub, R., Ali, H., Asar, A.: AI and blockchain integrated billing architecture for charging the roaming electric vehicles. IoT. 1, 382–397 (2020). https://doi.org/10.3390/iot1020022 16. Buamod, I., Abdelmoghith, E., Mouftah, H.T.: A review of OSI-based charging standards and eMobility open protocols. 2015 6th International Conference on the Network of the Future (NOF) (2015). https://doi.org/10.1109/nof.2015.7333288 17. Quorum Whitepaper. https://github.com/ConsenSys/quorum/blob/master/docs/Quorum% 20Whitepaper%20v0.2.pdf. Accessed 15 Apr 2023

398

R. Ungureanu et al.

18. Hyperledger Fabric: Open, Proven, Enterprise-grade DLT. https://www.hyperledger.org/wpcontent/uploads/2020/03/hyperledger_fabric_whitepaper.pdf. Accessed 15 Apr 2023 19. Brown, R., Carlyle, J., Grigg, I., Hearn, M.: Corda: An Introduction (2016). https://doi.org/ 10.13140/RG.2.2.30487.37284 20. van der Kam, M., Bekkers, R.: Comparative analysis of standardized protocols for EV roaming. Report D6.1 for the evRoaming4EU project (2020). https://doi.org/10.13140/RG.2.2. 19988.17283 21. Open Charge Point Interface 2.2 Documentation. https://evroaming.org/app/uploads/2020/ 06/OCPI-2.2-d2.pdf. Accessed 15 Apr 2023 22. Hyperledger Fabric Documentation: Endorsement Policies. https://hyperledger-fabric.readth edocs.io/en/latest/endorsement-policies.html. Accessed 15 Apr 2023

Managing Digitalization of Production Systems

Leveraging Advanced Digital Technology Practices to Enhance Information Quality in Low-Volume Product Introduction and Manufacturing Siavash Javadi1,2(B) and Koteshwar Chirumalla2 1 Scania AB, 151 87 Södertälje, Sweden

[email protected]

2 Mälardalen University, Hamngatan 15, 631 05 Eskilstuna, Sweden

[email protected] Abstract. This study aims to examine the potential of advanced digital technology practices to enhance information quality during low-volume product introduction and manufacturing. The successful introduction of products is vital for the competitiveness and prosperity of manufacturing companies. However, low-volume products have distinct characteristics that affect the introduction process, and one significant challenge is the presence of poor information quality. This adversely impacts the introduction process and can lead to disruptions in new product production. Consequently, it is crucial to identify alternative approaches that align with the specific characteristics of such industries to enhance information quality. This research conducted a longitudinal study within a large manufacturing company specializing in heavy commercial vehicles. Two product introduction projects were examined to investigate how the adoption of advanced digital technology practices, including interactive digital design reviews, digital test assemblies, virtual builds, and digital clinics, has influenced the quality of information exchanged between the research and development (R&D) department and the production team throughout the product introduction process. The study begins by highlighting the primary challenges associated with the product introduction process within the case company. Subsequently, it explores the impacts of implementing digital technology practices on various criteria related to information content quality, such as completeness, accuracy, clarity, consistency, correctness, and timeliness. These criteria are discussed in detail, and their impact is supported by examples from the two studied product introduction projects. Keywords: Low-volume manufacturing · product industrialization · digital transformation · Industry 4.0 · smart production

1 Introduction In today’s highly competitive global market, manufacturing companies are under pressure to develop high-quality products rapidly, at lower costs, and within shorter timeframes. These companies have significantly accelerated their product development © IFIP International Federation for Information Processing 2023 Published by Springer Nature Switzerland AG 2023 E. Alfnes et al. (Eds.): APMS 2023, IFIP AICT 689, pp. 401–416, 2023. https://doi.org/10.1007/978-3-031-43662-8_29

402

S. Javadi and K. Chirumalla

cycles, achieving two to three times faster performance than a decade ago. Such accelerated processes have important managerial implications for production systems and their processes [1], particularly during the product introduction or industrialization process [2, 3]. The product introduction process involves transitioning from engineering design to production, encompassing activities necessary to make the product manufacturable and prepare for production [4]. It refers to the actions taken to enable the planned volume production of the product for customers [5]. Efficient management of the product introduction process can lead to shorter time-to-volume and cost-effective production systems with fewer disruptions [6, 7]. The significance of effective product introduction management is even more pronounced in low-volume manufacturing, where unique challenges arise. These challenges include limited opportunities for testing and refining the fit between products and production systems, the absence of a conventional ramp-up process, and a strong emphasis on product functionality due to inherent complexity [8–11]. In this context, information quality becomes a critical determinant of success in the product introduction process [12, 13]. Effective coordination between research and development (R&D) and production relies heavily on continuous information sharing [12–15]. Previous research indicates that interface problems arise from poor information quality [6, 16, 17] and the lack of feedback loops between R&D and production [2, 18]. Advanced digital technologies, encompassing Industry 4.0 (I4.0) technologies, offer new possibilities and capabilities for manufacturing companies [19]. I4.0 encompasses a wide range of concepts and technological advancements, including the Internet of Things (IoT), artificial intelligence (AI), cloud computing, big data technologies, blockchain, augmented reality, automation, advanced robots, additive manufacturing, and simulation [19, 20]. By integrating the physical and digital realms, these technologies have the potential to revolutionize business operations. Previous research has extensively explored the role of advanced digital technologies in the context of new product development [21–24] and manufacturing [25, 26] from multiple perspectives. These studies have investigated their influence on processes and linking capabilities across various application areas of manufacturing [21]. Additionally, the significance of virtual manufacturing [27] and virtual reality [28, 29] in enhancing product and process development has been examined. Virtual manufacturing enables interactive simulations of manufacturing processes, such as virtual prototyping [30], virtual machining, virtual inspection, virtual assembly, and virtual operational systems [27]. On the other hand, virtual reality is actively employed in diverse industries to support decision-making, foster innovation in product design and manufacturing, and evaluate factors like visibility, ergonomics, aesthetic quality, abstract data visualization, and interdisciplinary communication [28]. However, existing literature on Industry 4.0 (I4.0) technologies primarily focuses on product development and manufacturing in high-volume manufacturing settings. Limited research specifically concentrates on the product introduction process in low-volume manufacturing. Furthermore, the role of digital technologies in enhancing information quality during low-volume product introduction and manufacturing requires further exploration. Therefore, this study aims to address this gap and examine the potential

Leveraging Advanced Digital Technology Practices to Enhance Information Quality

403

of advanced digital technology practices to enhance information quality during lowvolume product introduction and manufacturing. The research is based on a longitudinal study conducted within a large manufacturing company specializing in heavy commercial vehicles. Two low-volume product introduction projects are examined to explore how the adoption of advanced digital technology practices has influenced information quality.

2 Theoretical Background 2.1 Low-Volume Manufacturing Low-volume manufacturing is distinguished by its yearly production volumes typically ranging from 20 to 500 products, along with the high variety, complexity, and customizability of its product offerings to avoid substantial investment costs [31]. Consequently, the existing production systems are designed to accommodate the production of new products without requiring significant modifications [9]. The design of production systems in low-volume manufacturing prioritizes flexibility to cater to diverse product requirements. This flexibility is achieved through the use of universal production equipment, highly skilled workers, limited automation, and shared production resources across different products [9, 32]. However, these characteristics also result in limited opportunities to thoroughly test and refine both the product and the production system. The scarcity of prototypes, constraints in pre-series productions, and impracticality of conventional production ramp-up make it challenging to identify and address disturbances at the start of production in low-volume manufacturing [33]. In essence, the traditional approach of relying on test and refinement during later phases of the product introduction process is not feasible in the low-volume manufacturing context. 2.2 Information Quality Information quality is commonly defined as the suitability of information for its intended use [34, 35]. Several information quality frameworks have been proposed by researchers, considering different criteria, contexts, and purposes of use [34–37]. These frameworks have been applied in various manufacturing industry contexts, such as supply chain management [e.g., 38] and production planning [e.g., 41]. Studies by Eppler [40] and Knight & Burn [41] have compared different aspects of these frameworks. Eppler [40] identified certain limitations in existing frameworks, including their context-specific nature, lack of consideration for conflicts between different quality criteria, absence of measurement methods, lack of implementation tools, and insufficient foundation in theory or practice. To address these limitations, Eppler [40] proposed a generic information quality framework for knowledge-intensive processes, drawing from both literature and practical insights. This framework categorizes sixteen information quality criteria into two main categories: content quality and media quality. Various studies have defined specific information content quality criteria [35, 41, 42]. By comparing the frameworks proposed by Wang and Strong [35], Eppler [40], and English [41], a synthesized information content quality criteria framework was developed for analyzing information content quality within the low-volume product introduction and manufacturing in this study.

404

S. Javadi and K. Chirumalla

Figure 1 depicts the information content quality framework used in this paper. Based on Eppler’s [4] framework, the criteria are divided into two levels: (1) relevant information, which addresses the quality of information in relation to its intended community (i.e., the users), and (2) sound information, which encompasses criteria describing the inherent characteristics of information.

Complete

Sound information

Concise

Accurate Clear Applicable

Consistent

Correct

Timely

Content Quality

Relevant information

Fig. 1. The applied information content quality framework adapted from Eppler [41]

2.3 Advanced Digital Technologies For Product Introduction Advanced digital technologies, often associated with Industry 4.0, possess various capabilities such as real-time capability, intelligence, interoperability, virtualization, decentralization, connectivity, service orientation, modularity, and analytical capabilities [20, 21]. These technologies collectively support the need for increased productivity, improved flexibility and resilience, and reduced manufacturing and operations management costs [43]. In the context of Industry 4.0, integration and virtualization of manufacturing design and production processes are enabled through the use of information and the internet, resulting in the creation of smart products [43]. The role of Industry 4.0 technologies in improving and facilitating key activities in new product development and introduction has been highlighted in several studies. For example, the implementation of Industry 4.0 tools has the potential to shorten certain stages of the new product development process by more than 25 percent [23]. Additionally, augmented/virtual reality tools, simulations, and virtual manufacturing have shown applicability in product development and introduction. Researchers have proposed the concept of smart virtual product development, which leverages Industry 4.0 technologies, to enhance product manufacturing [44]. Smart virtual product development aids in decision making by utilizing explicit knowledge of formal decision events or experiences associated with various activities throughout the product development process, including design, manufacturing, and product inspection [44]. The contribution of Industry 4.0 technologies to important aspects of product introduction, such as design for assembly (DFA), has been explored by Bharath and Patil [27]. They introduced the concept of virtual manufacturing, which involves executing manufacturing processes in both computer simulations and the real world. Virtual models enable the prediction of potential problems concerning product functionality and manufacturability before actual production takes place.

Leveraging Advanced Digital Technology Practices to Enhance Information Quality

405

During product introduction, an important activity is to test and refine the product and production system, ensuring their compatibility [30]. A review on virtual reality applications in design highlighted that these applications not only assist with traditional design activities like 3D modeling, virtual prototyping, and product evaluation, but also enable co-design and design education beyond the early design stages [29]. Virtual reality can support early design phases, and there is a growing interest in using it for products that have emotional dimensions beyond functionality, with a specific focus on virtual assemblies and prototypes. Furthermore, the role of advanced digital technologies in the development of new production processes during the introduction of new products has been investigated. Digitally-enabled process innovation involves the use of new digital technologies to support the development and implementation process of completely or significantly new production methods, procedures, or techniques by leveraging organizational resources, structures, infrastructure, and culture [21].

3 Research Method The research method employed in this study is a longitudinal investigation of the product introduction practice within a company specializing in heavy commercial vehicles. The study specifically focuses on two product introductions to illustrate the impact of digital technology practices on the information content quality during the process. By adopting a longitudinal approach, the study aims to explore and comprehend how the adoption of advanced digital technologies influences the quality of information throughout the entire duration of the projects, from initiation to completion and handover to production. The empirical findings are derived from direct involvement in the product introduction activities as a process engineer and project manager over a span of four years. More than 20 product development projects with varying scopes were closely examined to gather data. Additionally, two specific product introductions were closely monitored as product introduction project manager to gain a deeper understanding of how the adoption of digital technology practices influenced the quality of information exchanged between the R&D and the production department during the product introduction process. These specific cases offer concrete examples of the effects of digital technology practices on information content quality. Data collection was carried out through active participation in project meetings and activities, review of project documents, analysis of meeting notes, and analysis of email conversations and discussions with project participants. The selected projects were focused on low-volume products within the case company, as these typically present greater challenges in terms of testing and verification. In the subsequent section, the two projects, referred to as Case A and Case B, are briefly described to provide context and specific illustrations.

4 Case Description The case company is headquartered in Sweden and has R&D facilities located there as well. Additionally, it operates six final assembly production units in six different countries. The company’s product range caters to various global markets and serves diverse

406

S. Javadi and K. Chirumalla

purposes such as mining, construction, transportation and distribution of goods, military applications, and urban and sanitary services. This entails the production of both highvolume products for general markets and customized, low-volume products tailored to specific market and customer needs. The company follows a stage-gate product development and introduction process for all its product development projects. These projects are typically managed through cross-functional teams composed of representatives from project management, R&D, production, service, and marketing. The product introduction department, represented by a project manager, oversees the production aspect of the projects, ensuring the smooth introduction of new products across all six production units within the planned timeline and desired quality. To achieve this, the product introduction department is involved in the projects from the early stages, identifying necessary changes in the production system and planning and developing the required deliveries. These deliveries encompass production tools, processes, assembly sequences, logistics flows, part and component packaging, as well as new work instructions. To ensure successful deliveries, a cross-functional group within the product introduction department collaborates, including process engineers, logistics planners, and preparators. Continuous communication is maintained between the product introduction department, the product development team, and the production units to ensure that production inputs and requirements are adequately addressed in the product development projects, and that production deliveries are of high quality and aligned with the project schedule. The company’s product introduction process comprises four main phases: configuration and target setting, development, test and verification, and termination. The configuration and target setting phase involves assessing the feasibility of producing new products, evaluating required investments, and estimating additional assembly times and costs. The development phase encompasses activities related to the development of new tools and production processes, including assembly sequences and instructions. The test and verification phase begins with the start of production and involves the production of pre-series products on the verification line to test, verify, and refine the production deliveries. The termination phase begins with the commencement of commercial production, during which the product is handed over to the production units and closely monitored to identify and rectify any remaining deviations. In Case A, the objective of the product introduction was to develop a new product variant with additional compressed natural gas (CNG) cylinders behind the drivers’ cabin to meet the requirements of a specific market demanding longer-range trucks. The product was planned for production in two units across two continents, with an estimated annual production volume of 70 vehicles. The project spanned 20 months from initiation to handover to production in the period of 2021–2022. Due to the complexity of the installation and the need for extra safety measures in gas installations, two verification trucks were planned for the project, despite the low product variant volume. In Case B, the aim was to introduce a new, larger battery box to meet the needs of a customer and purpose that required higher-capacity batteries. The product was planned to be introduced in all production units, with an estimated annual production volume of 200 vehicles. The product introduction process was monitored over a 12-month period,

Leveraging Advanced Digital Technology Practices to Enhance Information Quality

407

from initiation to handover to production. Unlike the first case, no verification trucks were required for this product.

5 Empirical Findings 5.1 Key Challenges in Product Introduction Process in the Case Company The case company has faced intense market competition, the emergence of electric vehicles, and stricter environmental regulations, which have compelled them to introduce more products and variants within shorter timeframes. These time constraints pose challenges in thoroughly testing, verifying, and refining products and production systems, including physical prototypes for design evaluations and testing manufacturing processes and tools. To overcome these challenges, the case company implemented front-end loading strategies to ensure the delivery of high-quality information to stakeholders involved in product introduction, especially the production department. Front-end loading was particularly crucial for low-volume products. Unlike high-volume products that undergo extensive testing and verification with multiple prototypes during a 3–6-month pre-series production phase, low-volume products often have limited or no prototypes and a shorter pre-series production period. This lack of testing and verification opportunities for both the product and production process poses even greater challenges in providing highquality information for low-volume products. To address these issues, the case company leveraged digital technology practices as a key tool to improve front-end loading and enhance the quality of the provided information in product development and introduction projects. The following sections will discuss the specific practices adopted and their impact on the quality of information content. 5.2 Advanced Digital Technology Practices in Product Introduction The case company has implemented several practices leveraging advanced digital technologies in product introduction. These practices include interactive digital design reviews, digital test assemblies, virtual builds, and digital clinics. Interactive Digital Design Reviews. Interactive digital design reviews have evolved from a routine practice of presenting early product concepts by R&D. Now, production is also involved in the process, enabling the simulation of human and production tool interactions with the products within the same environment. Product introduction process engineers have been equipped with updated competencies in CAD modeling and simulation. This empowers them to simulate production processes using 3D models of production tools and digital mannequins. This enhanced practice has proven instrumental in identifying, filtering out, or improving concepts that are incompatible with the production system or require significant investment in new production tools during the initial stages of product development. Digital Test Assemblies. Digital test assemblies have become increasingly prevalent in the introduction of new products, replacing physical tests to a large extent. These

408

S. Javadi and K. Chirumalla

assemblies are crucial for testing various aspects of assembling new installations on the product. The objectives of these tests include: (1) verifying the feasibility of the assembly sequence and identifying any potential issues such as collisions between the new installation and its surroundings during assembly, (2) identifying potential ergonomic and health risks for operators through the use of digital manikins, and (3) determining the need for new production tools or verifying the functionality of existing tools for new installations. A key factor enabling the transition to digital tests is the utilization of new CAD software capable of simulating the behavior of soft materials like plastic pipes and cables. These materials have been a significant source of uncertainties in digital tests. By relying primarily on digital test assemblies, several benefits have been realized: (1) significant reduction in costs and lead time associated with test assemblies, (2) the ability to initiate testing of new installations in earlier phases of product introduction and conduct multiple test and modification cycles when necessary, and (3) the capacity to test all required product variants without incurring additional costs and time in the project. Virtual Builds. Virtual builds involve a step-by-step animated assembly of the entire vehicle, aiming to validate the assembly sequence for the complete truck and detect any potential collisions between parts and installations. These virtual builds are conducted in response to requests from either the R&D department or the product introduction team, particularly in larger projects that require testing and verifying more extensive product changes and covering a wider range of product variants. The use of virtual builds enables the case company to achieve two main objectives: (1) testing and verifying the production sequence for all necessary product variants, and (2) obtaining an early overview of the production sequence for new installations and identifying the required changes in the assembly sequences of existing installations. Digital Clinics. Digital clinics are organized to introduce new production tools to the production units, aiming to simulate their functions in a 3D environment with accurate parts and relevant surroundings, including vehicle installations and digital mannequins. These digital clinics have replaced the traditional physical clinics, which used to rely on physical prototypes and the physical presence of engineers and operators from the production units. Comparing to the physical clinics, digital clinics have the following benefits: 1) Early feedback from production units: The new tools are presented to the production units approximately six to three months before the start of pre-series production. This allows the production units to provide direct feedback on the initial concepts for improvements and modifications. By obtaining early feedback, the need for manufacturing and modifying physical prototypes is reduced, saving costs and time. Moreover, it helps avoid last-minute changes in tool design that were previously common due to delayed feedback from the production units caused by tight project schedules. 2) Testing more variants and assembly scenarios: Digital clinics provide the opportunity to test the new tools on different product variants and various assembly scenarios. 3) Effective involvement of production units: Digital clinics foster increased engagement from the production units in the tool development process. With digital meetings, more people can participate without the need for travel. Additionally, the availability

Leveraging Advanced Digital Technology Practices to Enhance Information Quality

409

of digital materials after the clinic facilitates easy transfer of information to their respective organizations and allows production unit engineers to gather input more efficiently. 5.3 Improving Information Content Quality Through Advanced Digital Technology Practices The impact of adopting digital technology practices on information content quality criteria is discussed in the following sections, focusing on examples from two product introduction projects studied. By incorporating design reviews and digital test assemblies in iterative cycles, there was a continuous exchange of information between product design and production, ensuring that details were complete, correct, and clear. In Case A, early design reviews and digital test assemblies highlighted ergonomic issues and challenges in accessing the mechanical joints between the gas cylinders package and the chassis frame. Through a digital manikin sight test, it was discovered that operators couldn’t see the holes for inserting vertical screws under the gas cylinders, making it difficult to mount the components (Fig. 2). Additionally, the space between the gas cylinders and the vertical screws was insufficient for using a standard tightening tool. As a result, the design was modified to replace the vertical screws with horizontal ones, improving visibility and accessibility.

Fig. 2. Simulating operators access and sight to the mechanical joints in Case A.

Also, the continuous digital communication between R&D and production significantly improved the timeliness of information, ensuring that production teams were kept updated with the latest design concepts. In Case B, weekly mini design reviews were conducted to discuss the most recent design changes and ideas. Subsequently, on a monthly basis, the relevant production units were provided with this information to gather their input and feedback. This approach enhanced the consistency of the information throughout the process. By digitally transforming the information about the latest design changes from R&D to production units, with consistent details and content, the likelihood of different interpretations and contradictions in the information

410

S. Javadi and K. Chirumalla

was minimized. This prevented misunderstandings or lack of details that often lead to inconsistencies. In Case B, the digital test assemblies focused on examining the behavior of cables in various routing and connection scenarios (Fig. 3). This new capability in the 3D environment enabled R&D and production teams to obtain complete, correct, accurate, and clear information about the optimal assembly scenarios at an early stage of the project. Previously, such insights were only attainable at later stages with higher design maturity and the availability of physical prototypes. However, it is important to note that one aspect of information accuracy that still requires improvement in these digital practices is the simulation of friction forces. This limitation became apparent in Case B when simulating the connection of cable connectors to the battery box.

Fig. 3. Change of design from vertical mechanical joints (left) to horizontal ones (right) after identification of design for assembly issues.

In both cases, the utilization of virtual builds played a crucial role in defining and refining assembly sequences for new installations. Specifically, in Case B, a virtual build successfully identified a collision between the exhaust assembly and the new line carriers in a particular product variant, attributable to a lower battery box height. This discovery prompted necessary design modifications (Fig. 4).

Fig. 4. Simulating and analyzing behavior of the cables in different routing scenarios in Case B with a new 3d modelling software.

Virtual builds proved to be invaluable by offering clear, complete, and correct information pertaining to production possibilities, limitations, and the feasibility of various

Leveraging Advanced Digital Technology Practices to Enhance Information Quality

411

assembly scenarios. Furthermore, they significantly contributed to the timeliness of information dissemination, as the collision between the exhaust assembly and the new line carriers was detected prior to their installation in the factory (Fig. 5). Ultimately, virtual builds seamlessly integrated the most up-to-date information from both projects into the production sequence, ensuring the smooth execution of assembly processes.

Fig. 5. Detection of a collision between the exhaust and the new line carrier in Case B

Later in the product A introduction process, a digital clinic was conducted to showcase the new lifting tool designed for the gas cylinder package. This digital clinic offered clear, complete and correct information of the tool’s functionality by simulating it within a 3D environment (Fig. 6). By leveraging the digital clinic, the production units were able to identify potential risks and areas for improvement at an early stage, even before finalizing the tool concept with the supplier. One crucial issue identified during the digital clinic was the accessibility of the release handle of the tool. To mitigate safety and ergonomic concerns, the handle was relocated to the outer side of the truck, ensuring easy access for the operator. Additionally, the use of digital clinics played a pivotal role in enhancing the timeliness and consistency of information related to the new lifting tool. All stakeholders, including R&D, product introduction, and the tool supplier, had access to the same up-to-date information about the tool’s design approximately eight

Fig. 6. Changing the release mechanism of the lifting tool (top right) after detection of safety and ergonomic issues with the first concept (top left).

412

S. Javadi and K. Chirumalla

months prior to the start of production. This facilitated the avoidance of information gaps between different functions and minimized the risk of late detection of tool-related issues due to discrepancies in information from various sources, which was a common occurrence in traditional physical clinics. While the digital clinic provided a high level of accuracy in terms of simulating the tool’s function, it does have limitations when it comes to replicating the effects of gravity on the movement of the installation during the lifting process. The Table 1 summarises the influences of the advanced digital technology practices on different aspects of information content quality in low-volume manufacturing. Table 1. Influences of the implemented advanced digital technologies on information quality Digital practices

Information content quality criteria Complete, correct Accurate & clear

Consistent

Timely

Interactive design reviews

Improved information about production possibilities and limitations in early phases

Ability to simulate realistic information about ergonomics and DFA aspects in early phases of product development

Access to the same information about product concept for R&D and PI

Production has instant access to the latest design concepts

Virtual builds

Providing information about the sequence of all necessary product variants

Providing realistic information about complete production sequence

Same information about production sequence and possible issues is available for R&D & PI

Both PI and R&D has direct access to the latest production sequence, including future cross-project changes in sequence

Digital test assembly

Providing more complete, correct, and clear information by testing all product variants and all assembly scenarios

Simulating assembly of new installation with realistic information about ergonomic and DFA aspects. Still lacks ability to simulate aspects such as gravity and friction

Same information about product and production processes are accessible for R&D & PI

PI direct access to the latest design changes to investigate DFA/DFAM issues

(continued)

Leveraging Advanced Digital Technology Practices to Enhance Information Quality

413

Table 1. (continued) Digital practices

Digital tool clinic

Information content quality criteria Complete, correct Accurate & clear

Consistent

Timely

Complete, correct, and clear information about new production tools by presenting different scenarios and possibility of more tool design improvement cycles

Access to the same information about the new production tools for PI, tool suppliers, and production units

Production units have direct access to the latest production tools concepts

Realistic information about new tools’ function. Still lacks ability to simulate aspects such as gravity and friction

6 Discussion and Conclusion This study examined the potential of advanced digital technology practices [19, 20, 25] in enhancing information quality during low-volume product introduction and manufacturing. The findings highlight the significance of information content quality in successful product introductions [3, 4], particularly in the context of low-volume manufacturing with its unique challenges [8–11]. The research demonstrates that the adoption of advanced digital technology practices, including interactive digital design reviews, digital test assemblies, virtual builds, and digital clinics, positively impacted information content quality throughout the product introduction process. These practices effectively improved the completeness, accuracy, clarity, applicability, conciseness, consistency, correctness, and timeliness of exchanged information. In the existing literature, this study complements prior works on enhancing information content quality during product introductions, such as Zahay et al. [13], Frishammar et al. [15], Eppler [34], and English [42]. It is worth noting that while most studies on advanced digital technologies primarily focus on their application in specific manufacturing areas such as product design, maintenance, production planning, and quality inspection, this research showcases their utilization in addressing a long-standing engineering and management challenge—facilitating effective information exchange and reducing interface problems by enhancing the quality of exchanged information between R&D and production during new product introductions [1, 7, 12, 16, 17]. Furthermore, this study contributes to the knowledge base by exploring the application of virtual reality technologies [27] and the convergence of digital and physical spaces to enhance information quality in product introduction [28–30, 44].

414

S. Javadi and K. Chirumalla

The findings underscore the potential benefits of these practices in enhancing competitiveness, efficiency, and success rates for low-volume manufacturing. While this study focuses on a single company, future research involving multiple company cases can offer additional insights into the challenges and opportunities of implementing advanced digital technologies to improve information content quality. Additionally, comparative studies between high- and low-volume introductions can shed light on the limitations and opportunities of adopting such technologies in low-volume manufacturing contexts, enabling companies to achieve sustainable growth and excel in product introductions. Furthermore, future studies that explore information media quality can provide a more comprehensive understanding of how advanced digital technologies can enhance information quality for both researchers and practitioners.

References 1. Ettlie, J.E.: Product-process development integration in manufacturing. Manage. Sci. 41(7), 1224–1237 (1995) 2. Chirumalla, K., et al.: Exploring feedback loops in the industrialization process: a case study. Procedia Manuf. 25, 169–176 (2018) 3. Ruffles, P.C.: Improving the new product introduction process in manufacturing companies. Int. J. Mfg. Tech. Mgt. 1, 1–19 (2000) 4. Bellgran, M., Säfsten, K.: Production Development: Design and Operation of Production Systems. Springer, London (2010). https://doi.org/10.1007/978-1-84882-495-9 5. Lager, T.: Managing Process Innovation: from Idea Generation to Implementation. Imperial College Press, London (2011) 6. Fjallström, S., et al.: Information enabling production ramp-up. J. Mfg. Tech. and Mgt 20(2), 178–196 (2009) 7. Adler, P.: Interdepartmental interdependence and coordination: the case of the design/manufacturing interface. Organ. Sci. 6(2), 147–167 (1995) 8. Javadi, S., Bruch, J., Bellgran, M.: Characteristics of product introduction process in lowvolume manufacturing industries: a case study. J. Manuf. Technol. Manag. 27(4), 535–559 (2016) 9. Qudrat-Ullah, H., Seong, B.S., Mills, B.L.: Improving high variable-low volume operations: an exploration into the lean product development. Int. J. Technol. Manage. 57(1), 49–70 (2012) 10. Rahim, A.R.A., Baksh, M.S.N.: The need for a new product development framework for engineer-to-order products. Eur. J. Innov. Manag. 6(3), 182–196 (2003) 11. Surbier, L., Alpan, G., Blanco, E.: A comparative study on production ramp-up: state-of-theart and new challenges. Prod. Plan. Control 25(15), 1264–1286 (2014) 12. Bruch, J., Bellgran, M.: Characteristics affecting management of design information in the production system design process. Int. J. Prod. Res. 51(11), 3241–3251 (2013) 13. Zahay, D., Griffin, A., Fredericks, E.: Information use in new product development: an initial exploratory empirical investigation in the chemical industry. J. Prod. Innov. Manag. 28(4), 485–502 (2011) 14. Dröge, C., Jayaram, J., Vickery, S.K.: The ability to minimize the timing of new product development and introduction: an examination of antecedent factors in the North American automobile supplier industry. J. Prod. Innov. Manag. 17(1), 24–40 (2000) 15. Frishammar, J., Ylinenpää, H.: Managing information in new product development: a conceptual review, research propositions and tentative model. Int. J. Innov. Manag. 11(04), 441–467 (2007)

Leveraging Advanced Digital Technology Practices to Enhance Information Quality

415

16. Lakemond, N., Johansson, G., Magnusson, T., Safsten, K.: Interfaces between technology development, product development and production: critical factors and a conceptual model. Int. J. Technol. Intell. Planning 3(4), 317–330 (2007) 17. Chirumalla, K.: Managing Product Introduction Projects in Operations: Key challenges in heavy-duty vehicle industry. J. Modern Project Manag. 5(3) (2018) 18. Chirumalla, K.: Clarifying the feedback loop concept for innovation capability: a literature review. In: 28th The International Society for Professional Innovation Management (2017) 19. Rad, F.F., Oghazi, P., Palmie, M., Chirumalla, K., Pashkevich, N., Patel, P.C.: Industry 4.0 and supply chain performance: a systematic literature review of the benefits, challenges, and critical success factors of 11 core technologies. Ind. Mark. Manage. 105, 268–293 (2022) 20. Oztemel, E., Gursev, S.: Literature review of Industry 4.0 and related technologies. J. Intell. Manuf. 31, 127–182 (2020) 21. Chirumalla, K.: Building digitally-enabled process innovation in the process Industries: A dynamic capabilities approach. Technovation 105, 102256 (2021) 22. Santos, K., et al.: Opportunities assessment of product development process in industry 4.0. Procedia Manufact. 11, 1358–1365 (2017) 23. Urban, W., et al.: Application of industry 4.0 to the product development process in projecttype production. Energies 13(21), 5553 (2020) 24. Ahmed, M.B., et al.: smart virtual product development (SVPD) to enhance product manufacturing in industry 4.0. Procedia Comput. Sci. 159, 2232–2239 (2019) 25. Zheng, T., et al.: The applications of Industry 4.0 technologies in manufacturing context: a systematic literature review. Int. J. Prod. Res. 59(6), 1922–1954 (2021) 26. Oliveria, J., et al.: New product development in the context of industry 4.0: insights from the automotive components industry. In: IJCIEOM 2018: Industrial Engineering and Operations Management II, pp. 83–94 (2018) 27. Bharath, V.G., Patil, R.: Virtual manufacturing: a review. Int. J. Eng. Res. Technol. (2015) 28. Berg, L.P., Vance, J.M.: Industry use of virtual reality in product design and manufacturing: a survey. Virt. Real. 21, 1–17 (2017) 29. Berni, A., Borgianni, Y.: Applications of virtual reality in engineering and product design: why, what, how, when and where. Electronics 9(7), 1064 (2020) 30. Zorriassatine, F., et al.: A survey of virtual prototyping techniques for mechanical product development. Proc. Inst. Mech. Eng. Part B J. Eng. Manuf. 217(4), 513–530 (2003) 31. Jina, J., Bhattacharya, A.K., Walton, A.D.: Applying lean principles for high product variety and low volumes: some issues and propositions. Logist. Inf. Manag. 10, 5–13 (1997) 32. HILL, T.: Manufacturing Strategy: Text and Cases, Irwin/McGraw-Hill 33. Javadi, S., Bruch, J., Bellgran, M., Hallemark, P.: Challenges in the industrialization process of low-volume production systems. In: International Conference on Manufacturing Research 2013. Cranfield, United Kingdom: Cranfield University Press (2013) 34. Eppler, M.J.: Managing Information Quality: Increasing the Value of Information in Knowledge-intensive Products and Processes. Springer, Heidelberg (2006) 35. Wang, R.Y., Strong, D.M.: Beyond accuracy: what data quality means to data consumers. J. Manag. Inf. Syst. 12(4), 5–33 (1996) 36. Price, R., Shanks, G.: A semiotic information quality framework: development and comparative analysis. J. Inf. Technol. 20(2), 88–102 (2005) 37. Kahn, B.K., Strong, D.M., Wang, R.Y.: Information quality benchmarks: product and service performance. Commun. ACM 45(4), 184–192 (2002) 38. Li, S., Lin, B.: Accessing information sharing and information quality in supply chain management. Decis. Support Syst. 42(3), 1641–1656 (2006) 39. Gustavsson, M., Wänström, C.: Assessing information quality in manufacturing planning and control processes. Int. J. Qual. Reliabil. Manag. 26(4), 325–340 (2009)

416

S. Javadi and K. Chirumalla

40. Eppler, M.J.: A generic framework for information quality in knowledge-intensive processes. In: Pierce, E.M., Quality, I.C.O.I., Katz-Haas, R. (eds.) Proceedings of the Sixth International Conference on Information Quality, Cambridge, MA, USA, pp. 329–346. Massachusetts Institute of Technology (2001) 41. Knight, S.-A., Burn, J.M.: Developing a framework for assessing information quality on the World Wide Web. Inf. Sci. Int. J. Emerg. Transdisc. 8(5), 159–172 (2005) 42. English, L.P.: Information Quality Applied: Best Practices for Improving Business Information, Processes and Systems. Wiley Publishing, Indianapolis (2009) 43. Guo, D., et al.: Synchroperation in industry 4.0 manufacturing. Int. J. Prod. Econ. 238, 108171 (2021) 44. Anderl, R.: Industrie 4.0 - advanced engineering of smart products and smart production. In: 19th International Seminar on High Technology, October 9 2014, Piracicaba, Brasil (2014)

Evaluating Augmented Reality, Deep Learning and Paper-Based Assistance Systems in Industrial Manual Assembly Alexander Riedel(B)

, Johanna Gerlach, Maximilian Dietsch, Frank Engelmann, Nico Brehm, and Tobias Pfeifroth

Ernst-Abbe University of Applied Sciences, Jena, Germany [email protected]

Abstract. Augmented reality or deep learning-based methods are increasingly being investigated for use in manufacturing tasks. As manual assembly tasks become more complex and change rapidly, these new technologies offer additional training, cognitive load reduction and error prevention. To realize their potential in industrial applications, technologically simple implementations and real-world experiments are needed. This study aims to explore the potential of deep learningbased object recognition as a driver for modular assembly assistance systems. The proposed deep learning framework and a state-of-the-art augmented reality head-mounted display will be compared for their potential to prevent errors and reduce training time in a real-world assembly task. To ensure that the results are transferable to industrial applications, experiments are conducted on the assembly of a complex product over a full working day of 8 h with a subject size of n = 20. It is found that both the proposed deep learning framework and the augmented reality display drastically reduce the error rate for certain types of errors. Keywords: assembly assistance · deep learning · augmented reality · industry 5.0

1 Introduction Manufacturing companies are faced with a rapidly changing, customer-driven market. Customers expect continuous technological advances in their products, resulting in shorter time-to-market for new products [1]. These changes have led to an increasing number of product variants. In response to these market demands, manufacturing companies have turned to flexible and adaptable production lines that include manual labor, as human workers bring an unparalleled level of flexibility to the process [2]. Flexible production lines require rapidly changing processes to which workers must adapt, and the use of workers and tasks must be flexible enough to adapt to the changing demands of the production process. This can result in less training time per employee, which can be a major challenge for manufacturers [2]. Complex products with small batch sizes are best manufactured in a manual assembly environment for economic and © IFIP International Federation for Information Processing 2023 Published by Springer Nature Switzerland AG 2023 E. Alfnes et al. (Eds.): APMS 2023, IFIP AICT 689, pp. 417–431, 2023. https://doi.org/10.1007/978-3-031-43662-8_30

418

A. Riedel et al.

technological reasons [3]. However, in this type of production, product quality and assembly time are highly dependent on worker characteristics, such as their skills, working conditions, and experience. Flexibility constraints may hinder these desired characteristics, e.g., a short-term temporary worker cannot gain many years of experience. In addition, high product complexity and variance place higher cognitive demands on each worker, leading to a need for error-preventing assembly assistance systems [4]. These systems, also called cyber-physical-systems (CPS) [5], use sensors to collect data of the assembly process and provide visual instructions to the worker. Assembly assistance systems have shown to reduce the error rate and increase workers’ productivity [6, 7]. The trend of using augmented reality through head mounted displays (AR HMD) and artificial intelligence for human assistance has been growing steadily in recent years [8]. The effectiveness of assembly assistance systems can be evaluated by their ability to detect and prevent assembly errors, their time saving potential and their mental load reduction potential. In order to compare different cyber-physical assistance systems in the aforementioned areas, experiments must be conducted with the same assembly task under the same conditions. A realistic and complex manual assembly task must be performed to transfer the results to industrial assembly. In this paper, we show comparative experimental results of three types of assistance systems for manual assembly, each with a sample size of 20 people working for 8 h (see Table 1). The used assistance methods are a traditional paper-based manual, an augmented reality system (AR HMD) and a newly development deep learning-based assistance framework. The single components of the proposed deep learning-based framework can be used without object detection. The use of AI-driven object detection however enables to orchestrate a large number of individual assistance modules through knowledge of the assembly state and the occurrence of faults. The HoloLens AR HMD can be equipped with an AI object detection but was only used with its out-of-the-box proprietary software package which does not support object detection. Table 1. Comparing augmented reality head mounted display and deep learning assembly assistance systems against traditional paper-based assistance in error rate change and assembly time change. Paper-Based

Augmented Reality HMD

Deep Learning Framework

Error rate change

-58%

-60%

Assembly time change

-4%

-18%

Evaluating Augmented Reality, Deep Learning

419

To evaluate their error prevention, time saving and mental load reducing potential, we conducted experiments in which test subjects carried out the manual assembly of an explosion-proof tube lamp under all three assistance types. The assembly product was provided by the internationally operating company R. Stahl AG for this investigation. Especially in safety-critical areas such as explosion-proof product manufacturing, the risk of assembly errors must be reduced to near zero, which is why the error prevention potential is of special interest in this study. Several studies have compared deep learning-based [9–12] and augmented realitybased [13–16] assistance systems with traditional paper-based methods. However, most of them compare only one type of assistance (e.g., paper-based vs. AR, paper-based vs. AI), use only simple products (e.g. LEGO® [13, 17, 18]), have small test groups, and perform few consecutive assemblies. The impact of assembly support systems in industrial environments and the comparative performance of two innovative methods for real industrial assembly tasks remain partially unclear. To address this lack of knowledge about modern assembly assistance systems, this work aims to provide: – A quantitative analysis of the time and error reduction potential of paper-based, AR and deep learning-based assembly assistance with 20 subjects each. – Performed experiments over the course of an 8-h assembling a complex product and analyzing of the different types of errors committed. – Design-architectural details of a deep learning-based (AI) modular assistance system framework using object detection technologies. 1.1 Related Work The main contributions of this work are (1) to compare AR and AI assistance systems with traditional paper-based manual assistance, and (2) to propose a deep learning-based assistance framework for manual assembly tasks. Both are compared to the existing literature in the field. Augmented Reality in Assembly Assistance Chu et al. [16] evaluate the impact of a Microsoft HoloLens AR device on assembly time and error rate. They compare the results for the task of inserting a power cord and a wooden tenon into a given area using video instructions or an AR environment. It is found that both assembly time and error rate can be reduced using the HoloLens HMD compared to video instructions. However, the tasks evaluated are simple (one step per task) and fast (10–100 s) and therefore only transferrable to low-complexity industrial applications. Loch et al. [17] compare video assistance with on-screen AR support for assembling LEGO® models. Results show a 10-fold error reduction and nearly constant assembly time for the AR system compared to the video instruction. The subjects also reported that they found the AR system more comfortable to use. Comparing the HoloLens AR HMD and traditional paper-based instructions, Blattgerste et al. [13] found an increase in assembly time for the AR device and no difference in total errors for either. However, the AR assembly group made significantly fewer errors in finding the correct parts. The simplicity of the assembly task, building

420

A. Riedel et al.

a LEGO® wall in 32 steps from identically shaped pieces, could make it difficult to transfer the results to industrial use cases. In their paper [20], Chu et al. demonstrated the effect of a smartphone-based AR assistance system for assembling an architectural structure. Compared to a paper-based control group, it is shown that the AR system almost doubled the assembly time and halved the error rate. Users indicated that a hands-free AR method such as HMDs (e.g., Microsoft HoloLens) would provide faster assembly. The literature acknowledges that AR-based technologies can be important tools to reduce errors and decrease the user’s mental workload compared to other methods. In addition, previous studies do not fully agree on whether AR applications can improve operational efficiency or the quality of manual assembly of complex objects. This is mainly due to the simplicity of the assembly tasks studied, where no tools are used or only toy objects are assembled. Vision-Based Deep Learning in Assembly Assistance The application of deep learning-based object recognition in manual assembly has been extensively investigated for its feasibility and how the acquired information can be used in the assembly process. Use cases include, for example, counting assembly parts to support the handling process [9], detecting small electrical parts during assembly [21], or recognizing work steps [22]. These approaches highlight the purpose and importance of using advanced object recognition in manual assembly, but without implementing the method in a productive assembly assistance system or evaluating its utility. In a preliminary study, Riedel et al. [12] show the effects of the proposed deep learning-based assistance framework for a smaller test subject group and without the use of additional assistance modules such as a laser projector. Also, it is only compared to paper-based assistance lacking the performance compromise to an AR-HMD. A study conducted by Faccio et al. proposes a more advanced approach to error and progress monitoring that relies on vision [23]. Here, workers’ hand positions are compared to virtual, predefined three-dimensional control volumes to determine correct or incorrect hand positions. The assistance system provides visual feedback based on the hand reaching into the control volume. In a follow-up work, the authors conduct user studies to compare the assembly time with and without the assistance system [24]. The proposed system reduces the average assembly time by 22%. However, the error reduction rate by the system is not reported. The use of predefined control volumes is inflexible when different workers have different types of hand movements. Deeplearning object detection methods could provide a sufficient way to directly observe object status and potential errors. A case study by Paz Rocha et al. presents an image processing assembly assistance system based on hand motion detection over control areas [25]. The system follows a modular approach where the captured motion data is used to determine component status. Paz Rocha et al. focus on implementation details and data collection rather than error detection or assembly time reduction. The literature review highlights the current research situation, where deep learningbased object recognition is underrepresented in assembly assistance systems and its impact on key performance indicators such as error rate has hardly been investigated and compared with other state-of-the-art methods such as AR. The impact of AI-based

Evaluating Augmented Reality, Deep Learning

421

assistance systems on assembly time has not been extensively studied in the literature, while the impact of augmented reality HMDs shows ambivalent results.

2 Methods and Materials The assembly assistance experiments were conducted using the AR HMD Microsoft® HoloLens 2, which runs the proprietary Dynamics 365 Guides software to create and visualize assembly instructions. The deep learning-based framework solution used in the AI assisted assembly cohort relies on an image processing algorithm to determine the current assembly state and then trigger a pick-by-light module, an electric screwdriver, the appropriate work instructions, and a laser projector. 2.1 Augmented Reality Head Mounted Solution The Microsoft HoloLens assistance system is realized by visualizing the storing location of parts, the desired assembly location, and the action to be performed for each work step following the 2W1H (what, where, how) principle (see Fig. 1).

Fig. 1. HoloLens assembly assistance visualization. The left arrow shows the storing location of the part to be assembled, the white hologram shows the assembly location, the animated hands show the assembly action.

Upon completion, a virtual button can be clicked to visualize the instructions for the following steps. In addition, a written instruction and a static image are provided for each step to ensure that all three experimental groups receive the same information. The animated instructions are created using CAD models of the assembly components. To locate object positions, HoloLens uses fixed ArUco markers placed on the workbench. 2.2 Deep Learning-Based Assistance Framework The proposed deep learning-based framework is based on the use of an object detection algorithm that predicts the location and state of single assembly components. It is

422

A. Riedel et al.

originally introduced in [12] but extended by a laser projector module. Based on the predicted input information, the current work step can be derived statically or via a simple neural network. Using the current work step and information about the component’s status, any type of assistance module can be triggered. In the present configuration, a pick-by-light module, an electrical screwdriver, a display, and a laser projector are used within the assistance framework. The display module visualizes the current assembly instructions and indicates assembly errors. The laser projector module highlights the current assembly region of interest (Fig. 2). Object Detection

Laser Projector Module

Pick-byLight Module

Screwdriver Module

Display Module Additional Modules

Fig. 2. Schematic architecture of the proposed deep learning-based assembly assistance framework in the configuration used for this work.

The components status and position are acquired via the single shot object detection model YOLOv4 [26]. The architecture was chosen because it can deliver real-time results at a high accuracy on GPU edge devices. More current deep learning models like YOLOv8 or transformer-based models might give an increase in detection accuracy and inference speed. In total, 1512 training images containing 5485 object instances and 790 validation images containing 1468 object instances were used for model training. The training data consists of 41 individual object classes, so each class is represented by 134 training instances on average. The small training set size is reasonable due to the low intraclass and high inter-class variance of the assembly parts. The inference is performed on a Nvidia Jetson AGX Xavier, publishing the detected objects in real-time via MQTT messages to the connected assistance modules. The corresponding work step or assembly faults are derived based on the object detection results (Fig. 3).

3 Experimental Setup The experimental setup is based on the assembly of a complex tubular lamp with the support of either the AR HMD or the deep learning-based assistance framework and a control group using paper-based assistance. Three physically identical workstations are set up for the assembly experiments. The work instructions for all groups are identical in content.

Evaluating Augmented Reality, Deep Learning

423

Fig. 3. Spatial relation of the assistance modules used in the prototypical implementation of the proposed assistance framework.

The assembly process includes 34 individual work steps, including 21 steps of plugging parts, 5 cable connections, 3 screwing operations, and 2 functional tests. The assembly components are made of materials such as plastic, aluminium, steel, glass, and electrical components. Their size ranges from 1mm cable wire to a 50 cm pipe housing. The wide variety of assembly methods, materials, and component sizes ensures the industrial relevance and makes the results transferable to other production methods and assembly lines. In each of the three groups, 20 subjects assembled tube lamps over a period of 8 h. During this time, at least 25 tube lamps were mounted by each subject. After two and five hours of work, the test subjects took a break of 30 min. The test subjects had never assembled this kind of lamp before but may have had various general manual assembly skills. Besides from the work instructions, they received no additional information about the assembly process. The experimental phase was supervised by experienced engineers from the manufacturing department. During the assembly process, any assembly error that occurred was manually recorded by the supervisors. To minimize the evaluation

424

A. Riedel et al.

error between different supervisors, a supervision protocol was established on how to count errors and measure assembly times (Table 2). Table 2. Composition of the control and experimental groups in terms of age, gender, and physical characteristics. F – female, M – male, R – right-handed, L – left-handed. Paper-Based

Augmented Reality HMD

Deep Learning Framework

N

20

20

20

Age

25.95

24.5

25.75

Gender

40% F/60% M

40% F/60% M

40% F/60% M

Handedness

85% R/15% L

95% R/5% L

75% R/25% L

Eyeglasses

55%

50%

50%

To determine the perceived mental load, the NASA task load index (TLX) questionnaire was used [27]. It rates six dimensions of mental workload (mental demand, physical demand, temporal demand, effort, performance, frustration level) using a 20-point scale and is a common tool to be used for determining mental load during assembly [16, 17]. The NASA-TLX was measured every 5 assembly cycles. The time for each complete assembly cycle was measured manually. In the experimental groups, the duration of each individual work step was additionally recorded by the respective systems. Using this data, a training period and a routine period could be identified. 3.1 Error Assessment To determine the types of errors for each work task, an error evaluation was performed. As a basis for further analysis, the tasks are described using the hierarchical task analysis (HTA) method [26]. HTA identifies all possible interaction paths with a system by subdividing each task into subtasks until no further subdivisions are possible. Based on the possible assembly states, a human reliability analysis such as THERP (technique for human error-rate prediction) [27] is used to assign the states to relevant error categories. Since the objective of this work is a comparative error reduction study and not an error analysis, THERP was used to identify errors. Due to the high complexity of the assembly product, only safety and functionally relevant defect types are included for further consideration. In total, the study’s supervisory team identified 70 potential defect types. Errors made by the test subjects are counted as such if no error correction is made in the next work step. This approach corresponds to the real-life scenario targeted by the study. Following the guideline VDI 4006-2 [29], the identified errors were classified into the four categories described in Table 3. By using error categories, the results become interpretable on a higher level.

Evaluating Augmented Reality, Deep Learning

425

Table 3. Error categories according to VDI 40062, that appeared during the assembly process of the explosion-proof tube lamp. Type (Abbreviation)

Description

Example

Error by omission (OM) Some action was not carried out

- forgot to plug in two cable wires - forgot to perform functionality test

Execution error (EX)

- chose a wrong plastic cap - sets wrong torque value for screwing

Something is wrongly set or selected

Error by confusion (CE) Something is done in- - reverses positive and negative connections stead of something else - confuses two similar looking plastic parts Quantitative error (QE)

Something is too much - applies two insulating films instead of one or too little - only screws 7 of the 8 required screws

4 Results and Discussion 4.1 Error Reduction Potential To track the test subject’s training process over the course of the multiple assemblies, the errors were accumulated for every five assemblies. All three groups generally show a decline in mean assembly errors over the course of 25 assembly cycles (Fig. 4).

Fig. 4. Number of assembly errors encountered in assembly using one of the three assistance methods. The errors are accumulated for five assemblies.

Compared to the paper-based assembly, the deep learning framework assistance shows a total error reduction of 60% (11.05 errors vs. 4.4 errors). The AR HMD shows

426

A. Riedel et al.

a similar reduction rate of 58%. After an initial 10 assembly cycles, the errors start increasing for the AR HMD assembly group, however, on a low level of 0.5 to 1 errors per assembly. This effect might be due to a false sense of familiarity with the process when using the HoloLens instruction. The generally smaller error interquartile range of both the deep learning framework and AR HMD groups indicate a lower statistical dispersion of the number of errors. The lower spread over the number of test subjects signals a more stable assembly environment. The general error reduction effect provided by the assembly assistance systems brings a variety of positive effects in the industrial use. First, the required amount of supervision during the training period can be reduced. Second, the number of rejects due to errors is reduced. Third, the assembly time is reduced due to less rework caused by faulty assemblies. Since these positive effects are difficult to quantify, we assume a strong correlation of each factor with the number of total errors. To evaluate the assistance system’s abilities to prevent certain types or errors, the number of errors per category (Table 3) are analyzed. The results show that both the deep learning framework assistance and the AR HMD group produces fewer errors for each of the four error types (Fig. 5). The categorical errors are normed by their chance to occur.

Fig. 5. Normed assembly errors per category accumulated for 25 assemblies per test person. The four categories are errors by omission (OM), execution errors (EX), errors by confusion (CE) and quantitative errors (QE).

In particular, errors by omission (62% error reduction), execution errors (52% error reduction) and quantitative errors (72% error reduction) are effectively prevented by the deep learning framework assistance. Since it allows the execution of the following work step only if the current work step is visually detectable, errors by omission are actively prevented. Additionally, the pick-by-light system is likely to prevent several types of errors by reducing the cognitive load and the risk of assembling incorrect parts.

Evaluating Augmented Reality, Deep Learning

427

The qualitative analysis of the error prevention potential for each assistance module coincides with the quantitative results previously explained. The augmented reality HMD shows similar error prevention abilities although does performs worse in preventing execution errors. This might be due to the fact that the correct execution is not actively supervised by a camera but only passively demonstrated to the user. The strong potential of the AR HMD assistance to prevent errors by omission (70% reduction) is promising for the technology to be used in small scale part industrial assembly. 4.2 Influence on Assembly Time The evolution of the assembly time over several assembly cycles provides information about the duration of a training phase for the first cycles and a routine phase for the following assemblies. The paper-based control group took an average of 4 h 56 min to assemble 25 tube lamps. The AR HMD reduced the total assembly time by 4% (4 h 45 min) and the deep learning framework by 18% (4 h 4 min). The time differences are explained by the additional time to confirm a work step completion in the AR HMD whereas the deep learning framework automatically detects completion (Fig. 6).

Fig. 6. Evolution of assembly time for the first five assemblies and the mean time of five for the sixth to 25th assembly. The first five assemblies are referred to as the training period.

The mean assembly time for all groups changes by only 5% after the fifth assembly. The training phase is therefore defined as the first five completed assemblies.

428

A. Riedel et al.

4.3 Influence on Mental Workload The mean NASA-TLX value is calculated across all subjects and assembly cycles separately for each assistance group. The evaluation shows a mean NASA TLX of 6.19 for the control group, 5.23 for the group using the Deep Learning Framework and 7.04 for the group using the AR HMD. This represents a 14% increase in NASA-TLX when using the AR HMD and a 15% decrease when using the proposed deep learning framework compared to the control group. In terms of mental and physical effort, subjects reported higher scores when using the AR HMD compared to the other groups. The larger NASA-TLX value for AR HMD systems is a known effect that is a strong barrier to daily use in industry (Table 4). Table 4. Mean NASA-TLX value and respective standard deviation (SD) for the control and experimental groups. The percentage change is calculated compared to the paper-based group. Paper-Based

Augmented Reality HMD

Deep Learning Framework

NASA-TLX mean (SD)

6.19 (1.47)

7.04 (1.25)

5.23 (1.33)

NASA-TLX change

–

+14%

−15%

5 Conclusion In this work, two state-of-the-art concepts of manual assembly assistance were evaluated with respect to their effects on assembly time and assembly errors. Interestingly, the positive effects of augmented reality assistance on error reduction can be similarly achieved by an object recognition-based assistance system in the proposed configuration. To the best of our knowledge, this is the first study to investigate advanced worker assistance systems in which 20 test subjects manually assemble a complex product for a full workday (8 h). This makes the results particularly applicable to an industrial setting. Compared to a traditional paper-based method, we find that both an augmented reality head-mounted display (HoloLens 2) and a deep learning-based assistance framework show strong potential for reducing assembly errors (58% and 60%, respectively). However, the use of the AR HMD shows an increased mental workload during assembly and almost equal assembly time compared to the paper-based control group. In this study, the proprietary HoloLens 2 assistant guide creation tool was used, which may have had a negative impact on assembly time but is the intended operation tool. For the use of assistance systems in industrial manufacturing, the acceptance of the system by its users is crucial. The results of our qualitative experiments show that the test subjects had problems interacting with the HoloLens environment and therefore showed lower acceptance. Still its use is highly valuable for short-term and more complex tasks. The experiments demonstrate the previously researched fact that AR HMDs have issues in employee acceptance when used for 8 h, as they cause fatigue and disrupt the

Evaluating Augmented Reality, Deep Learning

429

workflow. Our results show that it is possible to achieve the positive effects on error reduction with additional reduced mental workload and assembly time reduction by using the proposed deep learning-based framework for assistance systems. The proposed framework would have had an even greater impact on assembly time if subjects did not have to wait for the object detection algorithm to detect the current assembly state. The timing should be considered as an important point in future implementations of the system. In the future work, reducing the training effort of a deep learning recognition system for manual assembly should be addressed. This could lead to non-invasive and costeffective advanced assistance systems capable of learning assembly processes in an unsupervised manner. In addition, future experiments should be conducted with test groups that reflect industrial reality. The unusual age and skill distribution limits the interpretability and transferability of this and most similar studies. To reflect reality even better, the effect of memory retention should also be evaluated by conducting experiments over multiple days. Acknowledgements. The authors would like to thank the Carl-Zeiss-Stiftung for their support of the Smart Assembly research project. We would also like to thank our industrial partners for providing assembly products and sharing knowledge.

References 1. Koren, Y.: The Global Manufacturing Revolution. Wiley, Hoboken (2010) 2. Zhou, F., Ji, Y., Jiao, R.J.: Affective and cognitive design for mass personalization: status and prospect. J. Intell. Manuf. 24(5), 1047–1069 (2013). https://doi.org/10.1007/s10845-0120673-2 3. Metzmacher, A.I., Hellebrandt, T., Ruessmann, M., Heine, I., Schmitt, R.H.: Aligning the social perspective with the technical vision of the smart factory. In: Schmitt, R., Schuh, G. (eds.) WGP 2018, pp. 715–729. Springer, Cham (2019). https://doi.org/10.1007/978-3-03003451-1_69 4. Merkel, L., Atug, J., Berger, C., Braunreuther, S., Reinhart, G.: Mass customization and paperless assembly in the learning factory for cyber-physical-production systems: learning module ‘from paperbased to paperless assembly’. In: 2018 IEEE 18th International Conference on Advanced Learning Technologies (ICALT), July 2018, pp. 270–271 (2018). https://doi.org/ 10.1109/ICALT.2018.00130 5. Villani, V., Sabattini, L., Czerniaki, J.N., Mertens, A., Vogel-Heuser, B., Fantuzzi, C.: Towards modern inclusive factories: a methodology for the development of smart adaptive humanmachine interfaces. In: 2017 22nd IEEE International Conference on Emerging Technologies and Factory Automation (ETFA), September 2017, pp. 1–7 (2017). https://doi.org/10.1109/ ETFA.2017.8247634 6. Quint, F., Loch, F., Orfgen, M., Zühlke, D.: A system architecture for assistance in manual tasks (2016). https://doi.org/10.3233/978-1-61499-690-3-43 7. Hinrichsen, S., Bendzioch, S.: How digital assistance systems improve work productivity in assembly, pp. 332–342 (2019) 8. de Souza Cardoso, L.F., Mariano, F.C.M.Q., Zorzal, E.R.: A survey of industrial augmented reality. Comput. Ind. Eng. 139, 106159 (2020). https://doi.org/10.1016/j.cie.2019.106159

430

A. Riedel et al.

9. Börold, A., Teucke, M., Rust, A., Freitag, M.: Deep learning-based object recognition for counting car components to support handling and packing processes in automotive supply chains. IFAC-PapersOnLine 53(2), 10645–10650 (2020). https://doi.org/10.1016/j.ifacol. 2020.12.2828 10. Ozdemir, R., Koc, M.: A quality control application on a smart factory prototype using deep learning methods. In: 2019 IEEE 14th International Conference on Computer Sciences and Information Technologies (CSIT), September 2019, pp. 46–49 (2019). https://doi.org/10. 1109/STC-CSIT.2019.8929734 11. Pierleoni, P., Belli, A., Palma, L., Palmucci, M., Sabbatini, L.: A machine vision system for manual assembly line monitoring. In: 2020 International Conference on Intelligent Engineering and Management (ICIEM), June 2020, pp. 33–38 (2020). https://doi.org/10.1109/ICIEM4 8762.2020.9160011 12. Riedel, A., et al.: A deep learning-based worker assistance system for error prevention: Case study in a real-world manual assembly. Adv. Prod. Eng. Manag. 16(4), 393–404 (2021). https://doi.org/10.14743/apem2021.4.408 13. Blattgerste, J., Strenge, B., Renner, P., Pfeiffer, T., Essig, K.: Comparing conventional and augmented reality instructions for manual assembly tasks. In: Proceedings of the 10th International Conference on PErvasive Technologies Related to Assistive Environments, June 2017, pp. 75–82 (2017). https://doi.org/10.1145/3056540.3056547 14. Hebenstreit, M., Spitzer, M., Eder, M., Ramsauer, C.: An Industry 4.0 production workplace enhanced by using mixed reality assembly instructions with Microsoft Hololens. Lecture Notes in Informatics (LNI), Proceedings - Series of the Gesellschaft fur Informatik (GI) (2020). https://doi.org/10.18420/muc2020-ws116-005 15. Lampen, E., Teuber, J., Gaisbauer, F., Bär, T., Pfeiffer, T., Wachsmuth, S.: Combining simulation and augmented reality methods for enhanced worker assistance in manual assembly. Procedia CIRP 81, 588–593 (2019). https://doi.org/10.1016/j.procir.2019.03.160 16. Chu, C.-H., Ko, C.-H.: An experimental study on augmented reality assisted manual assembly with occluded components. J. Manuf. Syst. 61, 685–695 (2021). https://doi.org/10.1016/j. jmsy.2021.04.003 17. Loch, F., Quint, F., Brishtel, I.: Comparing video and augmented reality assistance in manual assembly. In: 2016 12th International Conference on Intelligent Environments (IE), September 2016, pp. 147–150 (2016). https://doi.org/10.1109/IE.2016.31 18. Simonetto, M., Peron, M., Fragapane, G., Sgarbossa, F.: Digital assembly assistance system in industry 4.0 era: a case study with projected augmented reality. In: Wang, Y., Martinsen, K., Yu, T., Wang, K. (eds.) IWAMA 2020. LNEE, vol. 737, pp. 644–651. Springer, Singapore (2021). https://doi.org/10.1007/978-981-33-6318-2_80 19. Hoover, M., Miller, J., Gilbert, S., Winer, E.: Measuring the performance impact of using the microsoft hololens 1 to provide guided assembly work instructions. J. Comput. Inf. Sci. Eng. 20(6) (2020). https://doi.org/10.1115/1.4046006 20. Chu, C.-H., Liao, C.-J., Lin, S.-C.: Comparing augmented reality-assisted assembly functions—a case study on dougong structure. Appl. Sci. 10(10), 3383 (2020). https://doi.org/10. 3390/app10103383 21. Tavakoli, H., Walunj, S., Pahlevannejad, P., Plociennik, C., Ruskowski, M.: Small object detection for near real-time egocentric perception in a manual assembly scenario,” June 2021. 2106.06403 22. Eversberg, L., Grosenick, P., Meusel, M., Lambrecht, J.: An industrial assistance system with manual assembly step recognition in virtual reality. In: 2021 International Conference on Applied Artificial Intelligence (ICAPAI), May 2021, pp. 1–6 (2021). https://doi.org/10.1109/ ICAPAI49758.2021.9462061

Evaluating Augmented Reality, Deep Learning

431

23. Faccio, M., Ferrari, E., Galizia, F.G., Gamberi, M., Pilati, F.: Real-time assistance to manual assembly through depth camera and visual feedback. Procedia CIRP 81, 1254–1259 (2019). https://doi.org/10.1016/j.procir.2019.03.303 24. Pilati, F., Faccio, M., Gamberi, M., Regattieri, A.: Learning manual assembly through realtime motion capture for operator training with augmented reality. Procedia Manuf. 45, 189– 195 (2020). https://doi.org/10.1016/j.promfg.2020.04.093 25. Rocha, C.A.P., Rauch, E., Vaimel, T., Garcia, M.A.R., Vidoni, R.: Implementation of a visionbased worker assistance system in assembly: a case study. Procedia CIRP 96, 295–300 (2021). https://doi.org/10.1016/j.procir.2021.01.090 26. Bochkovskiy, A., Wang, C.-Y., Liao, H.-Y.M.: YOLOv4: optimal speed and accuracy of object detection. April 2020. http://arxiv.org/abs/2004.10934 27. Hart, S.G., Staveland, L.E.: Development of NASA-TLX (Task Load Index): results of empirical and theoretical research. In: Meshkati, P. (eds.) Human Mental Workload, vol. 52, P. A. Hancock and N. B. T.-A. North-Holland, 1988, pp. 139–183

Reinforcing the Closing of the Circular Economy Loop Through Artificial Intelligence and Robotics Waleska Sigüenza Tamayo1 , Naiara Uriarte-Gallastegi2 , Beñat Landeta-Manzano2 , and Germán Arana-Landin3(B) 1 Public Policy and Economic History Department, Faculty of Economic and Business,

University of the Basque Country, 01006 Vitoria, Spain [email protected] 2 Business Management Department, Faculty of Engineering, University of the Basque Country, 48013 Bilbao, Spain {naiara.uriarte,Benat.landeta}@ehu.eus 3 Business Management Department, Faculty of Engineering, University of the Basque Country, 20018 San Sebastián, Spain [email protected]

Abstract. The aim of this study is to analyse how Artificial Intelligence and Robotics contribute to the commonly employed strategies of Reuse, Recovery, and Recycling in the Circular Economy. Through a multi-case study of 12 technological development projects, we have examined the influence of these technologies on these circular economy variables. The analysis proceeded in two phases: firstly, we analysed the survey results, and secondly, we conducted personal interviews with technicians and managers, complemented by documentary analysis, to delve deeper into the research purpose. The findings indicate that, except for two cases, the relevance of these variables was not considered when integrating these technologies. However, across all the cases studied, we identified certain aspects that positively influence these variables, albeit with varying degrees of impact. In many instances, these influences were relatively low. Among the three variables analysed, Reuse exhibited the greatest influence, with both technologies demonstrating a medium to high degree of impact. In comparison, the influence of artificial intelligence on the variables of Recovery and Recycling was moderate to low, while Robotics exerted a low influence. These results emphasize the need for further research into the new capabilities offered by these technologies, enabling companies to increase the utilization of secondary materials. Keywords: Artificial Intelligence · Circular Economy · Robotics · Reuse · Recovery · Recycling

© IFIP International Federation for Information Processing 2023 Published by Springer Nature Switzerland AG 2023 E. Alfnes et al. (Eds.): APMS 2023, IFIP AICT 689, pp. 432–443, 2023. https://doi.org/10.1007/978-3-031-43662-8_31

Reinforcing the Closing of the Circular Economy Loop

433

1 Introduction Robotics (RBs) and Artificial Intelligence (AI) are two of the technologies associated with the technological revolution of Industry 4.0 (I4.0) that causes a transformative effect on research and development in the business world [1, 2]. Within this technological development, advances in AI and RBs are not only potentially beneficial due to their economical manufacturing advantages, but also because they can be used to drive the transition to the Circular Economy (CE) [3, 4]. To advance in this transition, organizations should work on maximizing the use of all available resources [5]. To this end, the need for products, materials, and services to be maintained for as long as possible in the circle through the principles of CEs is pointed out [6]. To achieve this fact, it is necessary to reinforce the focus on the variables of the CE with the objective of closing the loop. Taking into consideration, the work of the Ellen Macarthur Foundation, CE indicators could be divided in three different groups depending on the focus of the improvement [7]: Smarter product use and manufacture: • Refuse: Make product redundant by abandoning its function or by offering the same function with a radically different product. • Rethink: Make product use more intensive. • Reduce: Increase efficiency in product manufacture or use by consuming fewer natural resources and materials. Extend lifespan of product and its parts: • Reuse: Reuse by another consumer of discarded product, which is still in good condition and fulfils its original function. • Repair: Repair and maintenance of defective product so it can be used with its original function. • Refurbish: Repair and maintenance of defective product so it can be used with its original function. • Remanufacture: Use parts of discarded products in a new product with the same function. • Repurpose: Use discarded products its parts in a new product with a different function. Useful application of materials: • Recycle: Process materials to obtain the same (high grade) or lower (low grade) quality. • Recover: Incineration of material with energy recovery. In addition to these indicators, the Ellen Macarthur foundation published the famous Butterfly Diagram [7]. In the right side, the technical cycle focuses on non-biodegradable materials. Based on this technical cycle, in the literature some authors [1] summarize three central variables to analyse the influences of the I.40 technologies on the CE, as it can be seen in the right side of Fig. 1. In this case, in the literature, these variables are defined in the following way: • Reuse: Focuses on reusing products or components in the same type of product previously discarded. Always, if it maintains conformity to the original functions.

434

W. Sigüenza Tamayo et al.

Fig. 1. Influence of IA and RB on Reuse, Recovery and Recycling variables.

• Recovery: Through repair, reconditioning, remanufacturing and reuse of components, the useful life is extended. • Recycling: The objective is to replace the consumption of natural resources through transformation processes that allow to obtain a high degree of quality of the materials. Considering the CE indicators, some authors point out that I40Ts should play an important role. They emphasize that AI can contribute to CE: enhancing the improvement of production efficiency and resource recovery [7]. However, other authors highlighted, how poor AI training can generate serious problems in countless areas, such as decision making, and detracting competitiveness to companies that can influence in these variables [8, 9]. On the other hand, with respect to robotics, the ability to increase the reuse, recovery, and recycling of robotic-based sorting systems in waste management is highlighted [10]. Moreover, they are increasingly used in sorting lines and can improve these 3Rs by performing heavy physical, sorting work, or working in areas not suitable for human intervention [11–14]. However, it is also found that the manufacture of these robots generates significant resource consumption and generates new waste that cannot be avoided [13, 14]. Considering these elements, the purpose of this study is to deepen the contributions of Artificial Intelligence and Robotics on Reuse, Recovery and Recycling variables. With these objectives in mind, after this introduction, the academic literature on the subject is analysed. In the third section, the methodology developed in the research is described; and in the fourth section, the results obtained are shown. The fifth section presents the discussions and to finish before the references, the conclusions obtained in the research, as well as the limitations and future lines of work are presented.

Reinforcing the Closing of the Circular Economy Loop

435

2 Literature Review In the academic literature, it is pointed out that I4.0 technologies offer a wide range of possibilities that can help companies to improve their circular performance [15]. Specifically, in vertical e-commerce, AI contributes to the management of goods [16] that positively affects the use [17] and fosters the knowledge perspectives needed to achieve sustainable development goals, taking into account the processes to extend their useful life [18]. AI, through algorithms and the use of smart meters, contributes to the change of system dynamics developing innovative circular business models [19]. They allow analysing variables such as: product age, wear and tear, recovery, product recycling and market conditions [20]. In this line, the Ellen MacArthur Foundation Report states that AI and robotics optimizes the circular infrastructure, improving the processes of sorting, reuse, disassembly, re-manufacturing of components and recycling of materials [21, 22]. On the other hand, other authors add that AI drives the “collaborative economy” and the sharing of products and services, thus improving the reuse of materials and enhancing sustainability [23]. However, there are studies that point out that AI does not represent sufficiently solid evidence in relation to reuse, recovery and recycling [1]. Regarding the influence of robotics, some studies highlight its ability to influence the variables reuse, recovery and recycling. Specifically, robotics-based sorting systems in waste management can perform sorting operations by performing heavy physical sorting work or working in areas not suitable for human intervention (noise, dust, dirt, contaminants, etc.) [11]. In addition, their application in maintenance operations is also observed to extend the useful life of products and components, and to develop waste recovery operations [12, 24, 25]. Taking these aspects into account, three research questions are proposed, which can be subdivided into two sub-questions (see Fig. 2): • RQ1: How do I40Ts influence the improvement of the reuse variable? • RQ1.1: How does AI influence the improvement of the reuse variable? • RQ1.2: How does RB influence the improvement of the reuse variable? • RQ2: How do I40Ts influence the improvement of the recovery variable? • RQ2.1: How does IA influence the improvement of the recovery variable? • RQ2.2: How does RB influence the improvement of the recovery variable? • RQ3: How do I40Ts influence the improvement of the recycling variable? • RQ3.1: How does IA influence the improvement of the recycling variable? • RQ3.2: How does RB influence the improvement of the recycling variable?

436

W. Sigüenza Tamayo et al.

Fig. 2. Graphical summary of the research questions

3 Methodology The research process is summarizing in Fig. 3. As it can be seen, first, the aims, objectives and research questions were defined. These allowed us to focus the literature review and achieve a greater level of detail in their definition. In order to achieve the objectives set, a multi-case study methodology was selected as it was found to be adequately adapted to the object of the research and allowed us to obtain a greater level of understanding and detail of the topic studied [26, 27]. The case selection process was developed among the innovative AI and Robotics technology projects submitted in the first 4 editions of the BIND 4.0 program. The objective of this program is to accelerate the implementation of technological projects by increasing public-private collaboration in the Basque Country, Spain. Specifically, 12 projects in which IA and/or RBs were applied were analysed, after conducting a preliminary survey among 120 project managers that allowed us to determine the degree to which the projects were suitable for the object of study (see Table 1) [28]. In order to identify common patterns of behaviour and to achieve the objectives mapped, a protocol was designed to examine, categorize, tabulate and review the evidences gathered by the survey, interviews with managers and technicians and documentary analysis, as it is shown in Fig. 2. This unique research protocol, which integrated the construction of the theoretical framework, the study of the chain of evidence through multiple sources of evidence, and an analysis of the research data from different perspectives [29], allowed us to confirm the validity of the research and its reliability (see

Reinforcing the Closing of the Circular Economy Loop

437

Table 2). Once the reports of the individual cases were completed, we proceeded to carry out the cross analysis that allowed us to write the final report of the investigation. In addition, following the guidelines set by these authors, a scale of 5 levels of graduation from o no influence on very high influence was defined + + + +. Table 1. Description of the projects. Project

Technology

Application

A

AI/RB

Human Machine Communication

B

AI/RB

Processing solutions to improve safety

C

AI/RB

Food distribution of fresh produce

D

AI

Predictive maintenance with machine learning

E

AI

Supply chain management

F

AI

Digital manufacturing solutions

G

AI

Image analysis and diagnostics

H

AI

Data and text processing management

I

AI

Management and organizational support

J

RB

Food warehouse management

K

RB

Waste sorting in hazardous environments

L

RB

Tool management in machining workplace

Table 2. Measures taken to ensure the validity and reliability of the case study. Design

Data collection

Data analysis

Case study protocols

Provide the script to all interviewees

Conducting several pilot studies

Develop a case study database

Third-party review of the processes followed in the research

Internal validity

Establish the theoretical framework prior to data analysis

Records of factors that could serve as alternative explanations

Pattern matching

Construct validity

Use of multiple sources of evidence

Peer review of transcripts and drafts

Establishment of a chain of evidence Review the draft case study report

External validity

Justification for the selection • N / A of case studies Defining the scope and limits at the research design stage

Reliability

Source: adapted from [27, 30]

Use of triangulation techniques

• N/A

438

W. Sigüenza Tamayo et al.

Fig. 3. Research process carried out.

4 Results In the selected AI projects, there are evidence of contributions in relation to reuse, recovery and recycling. However, the degree of contribution is highly variable. It has been detected that the use of AI increases the efficiency and accuracy of the recycling process (see Table 3). Furthermore, in an application, analysing large amounts of data has contributed to finding patterns in objects that allow the life of products and components to be extended through reuse. In another application, very positive results have been obtained classifying objects integrated with artificial vision. It allows separating those that cannot be reused and separates them for recycling. In this way, it increases efficiency and optimizes the reuse process, reduces waste, and facilitates recycling. On the other hand, the use of modelling and simulation techniques, predicting how products will behave under different conditions of use, makes it possible to extend the life of products and components. Specifically, it improves decision-making regarding the selection of materials in production processes, improving subsequent efficiency in the reuse, recovery, and recycling in different phases. At a lower level of evidence, improvements have been proven through advanced solutions for monitoring. Specifically, thanks to their automated sorting and separation capabilities, prediction, and optimization of the recycling process. In addition, by monitoring and tracking recycled materials in the supply chain, it is possible to achieve

Reinforcing the Closing of the Circular Economy Loop

439

greater assurance of compliance with the requirements of these materials, as well as environmental standards and regulations. In most of the selected cases, RB is applied together with other types of technological solutions such as the Internet of things or artificial vision, among others. Managers highlight applications aimed at separating and sorting components to increase recovery and, in some cases, to prepare waste in a way that facilitates recycling. In this type of solution, the integration of robotics and mechanical AI has yielded good results. In addition, the managers interviewed emphasize that RBs become more necessary when working in hazardous environments and when dealing with repetitive applications. Specifically, in half of the projects analysed, it is highlighted that a significant reduction of waste has been achieved, obtaining a better use of the resources used. In addition, the technicians have integrated elements in the robots such as dust and toxic fume extractors by means of sprinkling and filtering systems to improve occupational safety in waste recovery plants. Table 3. Influence of AI and RBs on Reuse, Recovery and Recycling. Technology

Project

Reuse

Recovery

Recycling

BOTH

A

+

+

+

B

+++

o

o

C

++++

o

o

D

o

+++

o

E

+++

++

++

F

+

+

+

G

+++

+++

+++

H

o

+

++

I

++++

++

++++

J

+

+

+

K

++

++

++

L

+

+

+

AI

RB

Note:o no influence, + low influence (less than 1%), + + medium influence (1% or more and less than 5%), + + high influence (5% or more and less than 20) and + + + + very high influence (20% or more)

5 Discussion In some studies, in the academic literature, AI and Robotics are considered to be among the technologies belonging to I4.0 that have the greatest impact on CE [1, 2]. In this research, it has been possible to contrast some previous studies on how AI can have a positive influence on the 3 activities reuse, recovery, recycling that are needed to close the loop of the CE [13].

440

W. Sigüenza Tamayo et al.

However, neither AI nor RB influence all indicators in the same way and/or to the same extent. This research has confirmed that their contribution varies greatly depending on the type of application and on how the technology is adapted to the specific case [30]. The improvement of these indicators in the adoption of these technologies has not been a priority, except in the case of a maintenance company that has adopted AI and a waste management company that integrated RB. Despite this, in the selected cases, there is evidence of a positive influence on the three variables. Improvements in recovery, recycling and reuse have been observed by analysing data and obtaining detailed information in real time in the case of AI and by detecting and classifying items in the case of RB. Some of these evidence had already been collected in other studies in the literature. For example, smart containers equipped with sensors for material detection or level measurement and waste collection the reuse of materials is improved [13]. Other example, in greenhouse cultivation and control of greenhouse factors controlling the temperature range AI contributes to water recovery [2, 31]. Finally, in existing oil infrastructures it is highlighted that improved planning and optimization of drilling facilitate efficient hydrocarbon recovery [7]. In addition, the results reported by some authors about the capacity of the contribution of RB to these variables in areas such as waste management, process management or supply chain are confirmed [13, 24, 25]. When measuring the degree of influence of AI and RB on the variables Reuse, Recovery and Recycling, the projects selected show that both technologies have a positive influence on the three variables (see Fig. 4). Specifically in relation to the RQs, the following results have been obtained in relation to the evidence obtained:

Fig. 4. Influence of IA and RB on Reuse, Recovery and Recycling variables. Note: Note: o no influence, + low influence (less than 1%), + + medium influence (1% or more and less than 5%), + + high influence (5% or more and less than 20) and + + + + very high influence (20% or more).

Reinforcing the Closing of the Circular Economy Loop

441

• With respect to RQ1, it is found that there is a level of influence on the reuse variable that can be considered medium-high. This level of influence is similar in both cases for AI technology (RQ1.1) and RB (RQ1.2). • The influence exerted on the Recovery variable (RQ2) is lower. A medium-low level of influence has been evidenced in the AI projects (RQ2.1) and low in relation to RB technology (RQ2.2). • Finally, in relation to the Recycling variable (RQ3), it was found that, despite being positive, it is very variable. Specifically, in relation to AI technology the level of influence was medium-low (RQ3.1) and in relation to RB technology it was low (RQ3.2).

6 Conclusions The technological revolution offers a multitude of opportunities and risks in the three pillars of sustainability (economic, social, and environmental) [32]. The increase in resource consumption requires moving towards a paradigm shift towards a more circular society. It becomes necessary to extend the life of the materials used, minimize the consumption of primary resources, and develop new processes that allow an increasing proportion of secondary materials to be integrated into the loop. To this end, the variables Reuse, Recovery and Recycling become necessary. The activities oriented to improve these variables must take advantage of new technologies, such as AI and RB. However, both business and academia have focused their attention on the influence of AI and RB on economic competitiveness and, although with some delay, issues related to the other two pillars of CE, the social and environmental part, are being integrated. In the literature, moreover, there are varied opinions on their contribution to the improvement of these three variables. For this reason, we have analysed real cases of highly innovative AI and RB projects in which it has been found that both AI and RB have the capacity to support the transition process and may have a potential contribution in these 3 variables. In general, they have a positive influence that could be greater if these variables were integrated within the objectives of the technological adoption processes. In relation to the limitations, we highlight those derived from the methodology used. This technology makes it possible to reach a greater degree of depth on the subject with a higher degree of detail, but it has problems in generalizing the results. For this reason, to improve the level of knowledge, it would be of great interest to carry out quantitative studies at a global level, by sector and by type of technological application, as the use of these technologies become more widespread. Acknowledgements. This study has been funded by the Basque Autonomous Government (Research Group GIC-IT1691–22), the Euskampus Foundation within the projects ZIRBOTICS and SOFIA (Euskampus Missions 1.0 and 2.0), the MCIN/AEI/10. 13039/501100011033 with the project PID2020-113650RB-I00 and FEDER “Una manera de hacer Europa” / European Union “NextGenerationEU”/PRTR, the University of the Basque Country (Research Group UPV/EHU 21/005) and the Ministry of Universities and the European Union through the European UnionNext Generation EU fund (EU-RECUALI 21/11, 21/18 and 22/03). We are grateful for the technical

442

W. Sigüenza Tamayo et al.

and human support provided by the CDRE (Universite de Pau et des Pays de l’Adour) and the Environmental Sustainability and Health Institute (Technological University of Dublin).

References 1. Laskurain-Iturbe, I., Arana-Landín, G., Landeta-Manzano, B., and Uriarte-Gallastegi, N.: Exploring the influence of industry 4.0 technologies on the circular economy. J. Cleaner Prod. 321, 128944 (2021) 2. Zhou, Y., Xia, Q., Zhang, Z., Quan, M., Li, H.: Artificial intelligence and machine learning for the green development of agriculture in the emerging manufacturing industry in the IoT platform. Acta Agriculturae Scandinavica, Section B—Soil & Plant Sci. 72(1), 284–299 (2022) 3. De Sousa Jabbour, A.B., Jabbour, C.J.C., Godinho Filho, M., Roubaud, D.: Industry 4.0 and the circular economy: a proposed research agenda and original roadmap for sustainable operations. Ann. Oper. Res. 270(1–2), 273–286 (2018) 4. Buenrostro Mercado, E.: Proposal for the incorporation of technologies associated with industry 4.0 in Mexican SMEs. Entreciencias: diálogos en la sociedad del conocimiento 10, no. 24 (2022). Greenstein, S. Preserving the rule of law in the era of artificial intelligence (AI). Artif. Intell. Law, 30(3), 291–323 (2022) 5. Geissdoerfer, M., Vladimirova, D., Evans, S.: Sustainable business model innovation: a review. J. Clean. Prod. 198, 401–416 (2018) 6. Uriarte-Gallastegi, N., Landeta-Manzano, B., Arana-Landín, G., Laskurain-Iturbe, I.: How do technologies based on cyber-physical systems affect the environmental performance of products? a comparative study of manufacturers’ and customers’ perspectives. Sustainability 14(20), 13437 (2022) 7. Ellen MacArthur Foundation. What is the circular economy? (2023): https://ellenmacarthurf oundation.org/topics/circular-economy-introduction/overview 8. Solanki, P., Baldaniya, D., Jogani, D., Chaudhary, B., Shah, M., Kshirsagar, A.: Artificial intelligence: new age of transformation in petroleum upstream. Petrol. Res. 7(1), 106–114 (2022) 9. McGovern, A., Ebert-Uphoff, I., Gagne, D.J., Bostrom, A.: Why we need to focus on developing ethical, responsible, and trustworthy artificial intelligence approaches for environmental science. Environ. Data Sci. 1, e6 (2022) 10. Rana, N.P., Chatterjee, S., Dwivedi, Y.K., Akter, S.: Understanding dark side of artificial intelligence (AI) integrated business analytics: assessing firm’s operational inefficiency and competitiveness. Eur. J. Inf. Syst. 31(3), 364–387 (2022) 11. Kurniawan, T.A., Meidiana, C., Othman, M.H.D., Goh, H.H., Chew, K.W.: Strengthening waste recycling industry in Malang (Indonesia): Lessons from waste management in the era of Industry 4.0. J. Cleaner Prod. 382, 135296 (2023) 12. Herterich, M.M., Uebernickel, F., Brenner, W.: The impact of cyber-physical systems on industrial services in manufacturing. Procedia Cirp 30, 323–328 (2015) 13. Caggiano, A.: Cloud-based manufacturing process monitoring for smart diagnosis services. Int. J. Comput. Integr. Manuf. 31(7), 612–623 (2018) 14. Sarc, R., Curtis, A., Kandlbauer, L., Khodier, K., Lorber, K.E., Pomberger, R.: Digitalisation and intelligent robotics in value chain of circular economy oriented waste management–a review. Waste Manage. 95, 476–492 (2019) 15. Shreyas Madhav, A.V., Rajaraman, R., Harini, S., Kiliroor, C.C.: Application of artificial intelligence to enhance collection of E-waste: a potential solution for household WEEE collection and segregation in India. Waste Manage. Res. 40(7), 1047–1053 (2022)

Reinforcing the Closing of the Circular Economy Loop

443

16. Rosa, P., Sassanelli, C., Urbinati, A., Chiaroni, D., Terzi, S.: Assessing relations between circular economy and industry 4.0: a systematic literature review. Int. J. Prod. Res. 58(6), 1662–1687 (2020) 17. Rangaswamy, E., Nawaz, N., Changzhuang, Z.: The impact of digital technology on changing consumer behaviours with special reference to the home furnishing sector in Singapore. Human. Social Sci. Commun. 9(1), 1–10 (2022) 18. Teng, Z., He, Y., Renhong, W.: E-commerce: does sustainable logistics development matter? Sustainability 15(1), 579 (2023) 19. Hsu, C.-C.: The role of the core competence and core resource features of a sharing economy on the achievement of SDGs 2030. J. Innov. Knowl. 8(1), 100283 (2023) 20. Koraganji, D.V., Garimella, R., Kandra, P.: Current trends and future challenges in smart waste management in smart cities. In: Advanced Organic Waste Management, pp. 395–406. Elsevier (2022) 21. Roberts, H., et al.: Artificial intelligence in support of the circular economy: ethical considerations and a path forward. AI & SOCIETY 1–14 (2022) 22. Ellen MacArthur Foundation. Artificial intelligence and the circular economy - Al as a too/ to accelerate the transition, 2019. https://ellenmacarthurfoundation.org/artificial-intelligenceand-the-circular-economy. Accessed 1 Dec 2022 23. Wu, Z., Zhou, W., Aisi, Y.: Analysis of a legal regulation approach and strategy of a sharing economy based on technological change and sustainable development. Sustainability 15(2), 1056 (2023) 24. Yu, K.H., Zhang, Y., Li, D., Montenegro-Marin, C.E., Kumar, P.M.: Environmental planning based on reduce, reuse, recycle and recover using artificial intelligence. Environ. Impact Assess. Rev. 86, 106492 (2021) 25. Álvarez-de-los-Mozos, E., Rentería-Bilbao, A., Díaz-Martín, F.: WEEE recycling and circular economy assisted by collaborative robots. Appl. Sci. 10(14), 4800 (2020) 26. Daneshmand, M., Noroozi, F., Corneanu, C., Mafakheri, F., Fiorini, P.: Industry 4.0 and prospects of circular economy: a survey of robotic assembly and disassembly. Int. J. Adv. Manufact. Technol. 124(9), 2973–3000 (2023) 27. Yin, R.K.: Case study research and applications: Design and methods. Sage Books, (2018) 28. Verleye, K.: Designing, writing-up and reviewing case study research: an equifinality perspective. J. Serv. Manag. 30(5), 549–576 (2019) 29. Leino, M., Pekkarinen, J., Soukka, R.: The role of laser additive manufacturing methods of metals in repair, refurbishment and remanufacturing–enabling circular economy. Phys. Procedia 83, 752–760 (2016) 30. Miles, Matthew B., Michael Huberman, A., Johnny, S.: A methods sourcebook. Qualitative data analysis (2014) 31. Chauhan, C., Parida, V., Dhir, A.: Linking circular economy and digitalisation technologies: a systematic literature review of past achievements and future promises. Technol. Forecast. Soc. Chang. 177, 121508 (2022) 32. Arshad, J., et al.: Intelligent greenhouse monitoring and control scheme: An arrangement of Sensors Raspberry Pi based Embedded System and IoT platform. Indian J. Sci. Technol. 13(27), 2811–2822 (2020) 33. Durst, S., Davila, A., Foli, S., Kraus, S., Cheng, C.F.: Antecedents of technological readiness in times of crises: a comparison between before and during COVID-19. Technol. Society 72 102195 (2023)

A New Generation? A Discussion on Deep Generative Models in Supply Chains Eduardo e Oliveira1,2(B) and Teresa Pereira1,3 1

Instituto de Ciˆencia e Inova¸ca ˜o em Engenharia Mecˆ anica e Engenharia Industrial (INEGI), Associate Laboratory for Energy, Transports and Aerospace (LAETA) Rua Dr. Roberto Frias 400, 4200-465 Porto, Portugal 2 Faculdade de Engenharia da Universidade do Porto (FEUP) - Rua Dr. Roberto Frias 400, 4200-465 Porto, Portugal [email protected] 3 Instituto Superior de Engenharia do Porto (ISEP) - Instituto Polit´ecnico do Porto (IPP) - R. Dr. Ant´ onio Bernardino de Almeida 431, 4249-015 Porto, Portugal Abstract. With the advent of Chat-GPT, Artificial Intelligence (AI) became one of the most discussed technological developments of today. Although there are not yet many studies discussing the application of Chat-GPT and other Large Language Models (LLM) to the Supply Chain (SC), several works have used a larger family of techniques called Deep Generative Models (DGM). DGM are deep learning networks that aim to discover complex and high-dimensional probability distributions from, instead of just discriminating between the samples. In this paper, we briefly explain the recent developments in DGM (especially Energybased models, Variational Autoencoders, Generative Adversarial Networks, and Autoregressive models), review previous applications of these techniques to the SC, and discuss research and application opportunities. Keywords: Supply Chain · Artificial Intelligence Deep Learning · Deep Generative Models

1

· Machine Learning ·

Introduction

In the field of Artificial Intelligence (AI), tools like Chat-GPT (OpenAI, 2022) and DALL-E2 (Open AI, 2022), that are capable of generating content (whether text or image) from a prompt, are currently taking the world by storm, getting the most attention of mainstream media any AI topic has gotten in recent years. These tools, when text-based, are called Large Language Models (LLM), and are very complex AI models trained on a colossal amount of data, but which can be used in a reasonable amount of time to answer questions or requests by its users. Given how recently these models have achieved a ready-to-apply state and notoriety, there are little to no known applications to the area of Supply Chains (SC). However, LLM are part of a larger family of AI and Machine learning (ML) c IFIP International Federation for Information Processing 2023  Published by Springer Nature Switzerland AG 2023 E. Alfnes et al. (Eds.): APMS 2023, IFIP AICT 689, pp. 444–457, 2023. https://doi.org/10.1007/978-3-031-43662-8_32

Deep Generative Models in Supply Chains

445

models that focus on using deep learning to understand the data they analyze called Deep Generative Models (DGM), and there have been some applications of DGM to SC. In terms of algorithms, AI/ML can be divided in two categories: discriminative and generative algorithms. Most algorithms are of the discriminative category, and their focus is on distinguishing the instances according to their class or similitude to other instances (Murphy, 2012). In contrast, generative algorithms focus on finding the probability distributions that underlie the data, in other words they attempt to determine how each instance is generated. DGM are neural network models with many layers, belonging to this last category, aiming at determining highly-dimensional probability distributions (Ruthotto and Haber, 2021), usually without any form of labeled data or supervision (Bond-Taylor et al., 2022). This last characteristic of not requiring labeled data is one of their main advantages. Despite the exponential growth in the production and availability of data, most of it is unlabeled and even unstructured, which makes it difficult to apply supervised algorithms to analyze it. Most DGM belong to the selfsupervised type of algorithms (Bond-Taylor et al., 2022), which aim at learning intermediate representations (e.g., pseudo-labels) that can be used for downstream tasks (e.g., classification) without supervision (Jing and Tian, 2021). Both the self-supervised and generative nature of DGM grant them considerable flexibility when performing data analysis and generation tasks, ranging from clustering similar groups of data, classification/regression, or generating text or images to answer questions. The deep learning component ensures great performance on very complex tasks, from high-dimensional functions, to language processing and computer vision. However, the complexity of DGM brings some challenges. The first one has to do with the size of the model. Their high performance in complex tasks requires a very large model, which has many hyperparameters that require finetuning (Ruthotto and Haber, 2021). Optimization would be ideal, but the sheer amount of these hyperparameters makes this task extremely difficult and inefficient. These hyperparameters include the network’s design and structure, the objective-function to use, regularization, and other parameters that influence training (e.g., learning rates) (Ruthotto and Haber, 2021). The large size also means a large volumes of data are required for training and optimization (BondTaylor et al., 2022). As most of these models try to recreate the data in a latent, unknown mathematical space, it is difficult to estimate the similarity between instances in the real space and in the latent one. As such, many times proxy objective-functions are used, which may not lead to the best results. Despite these general features and challenges of DGM, several types and architectures exist. In the literature, four types of DGM can be identified: Energy-based Models (EBM), Generative Adversarial Networks (GAN), Variational Autoencoders (VAE), and Autoregressive models. The aim of this paper and main contributions are to provide: 1) an overview of the most relevant DGM techniques; 2) a review of the existing applications of these techniques to SC in the literature; and 3) initiate a discussion on the

446

E. Oliveira and T. Pereira

current state of application of DGM to SC, and future research avenues and possible applications. This paper is structured as follows. After this introduction, Sect. 2 presents and overview of the most used DGM techniques, divided by each of the above mentioned four types. In Sect. 3, we review a total of 12 papers, which apply of DGM to SC. Afterwards, in Sect. 4, based on the papers presented in the previous section, we discuss the current state of the application of DGM techniques to SC, as well as the possibilities in terms of future research and applications. Finally, in Sect. 5 we present a brief summary of the paper and our main conclusions.

2

Deep Generative Models

In this section we present the basic theory of the four types of DGM, and the advantages and disadvantages of each type. 2.1

Energy-based Models

Energy-based models (EBMs) are based on the concept of an energy function, which assigns a scalar value (or energy) to each possible configuration of the model (LeCun et al., 2006). In an EBM, the goal is to learn the energy function such that it assigns lower energy values to configurations that are more likely to occur in the training data, and higher energy values to configurations that are less likely to occur (BondTaylor et al., 2022). This is accomplished by minimizing a loss function that measures the discrepancy between the energy function and the true distribution of the data. One way to train an EBM is to use a Markov chain Monte Carlo (MCMC) method to generate samples from the model distribution and estimate the gradient of the loss function. However, MCMC methods such as random walk and Gibbs sampling have long training times when applied to high dimensional data, which makes them inefficient. Recently, it has been advocated the use of stochastic gradient Langevin dynamics (Bond-Taylor et al., 2022). However, this still requires a large amount of steps. A possible solution is to use a contrastive divergence algorithm, which involves repeatedly sampling from the model and updating the energy function to make the model better at generating realistic samples (Bond-Taylor et al., 2022). EBMs have several advantages, such as lower amount of parameters, flexibility of generation, allowing for composition of several energy functions (Du and Mordatch, 2019). Some disadvantages are long sampling processes, and complex evaluation and optimization (Duvenaud et al., 2020). The two most used EBM are Boltzmann machines (both original (Hinton and Sejnowski, 1983) and restricted versions (Hinton, 2002)), and diffusion models (Alain et al., 2016; Bengio et al., 2013; Ho et al., 2020; Sohl-Dickstein et al., 2015).

Deep Generative Models in Supply Chains

447

One of the earliest EBM were Boltzmann machines. A Boltzmann machine is a fully connected network of binary neurons that are turned on with probability determined by a weighted sum of their inputs (Bond-Taylor et al., 2022). These can be considered an energy function that is defined over a set of binary variables. The neurons can be divided those which are set as inputs to the model, called visible, and all the others neurons, called hidden. The energy function is used to assign an energy value to each possible configuration of the binary variables, and the probability distribution of the Boltzmann machine is given by the Boltzmann distribution. However, their fully connected nature makes them hard to train. Given this limitation, Hinton (2002) proposed Restricted Boltzmann Machines (RBM), which removes connections between units in the same group, allowing exact calculation of hidden units (Bond-Taylor et al., 2022). Most recent progresses in terms of EBM have been done using diffusion models, namely in image synthesis, video generation, and molecule design (Yang et al., 2013). They work by gradually corrupt the data by adding noise over a certain number of steps. This is called the forward process, and can be defined as a discrete Markov chain (Bond-Taylor et al., 2022). By learning a mapping between an input and a noise distribution, one can repeatedly apply it to the noise distribution to produce samples that are increasingly similar to the input. 2.2

Generative Adversarial Networks

Instead of using Markov chains as in EBM, Generative Adversarial Networks (GAN) (Goodfellow et al., 2014) use a pair of networks to discover the desired probability distribution (Bond-Taylor et al., 2022). One of the networks is a generator that tries to create false data as similar to the true data as possible, while the other one is a discriminator that estimates the probability that a sample comes from the real distribution. The training process of these two networks occurs in tandem, with the objective of the discriminator being the capacity to distinguish fake data from the real one, while the generator aims to minimise this capacity. This type of training is called adversarial, as both networks are trying to decrease each others’ probability of success. After training, each of the networks can be used separately, whether the objective is to generate new data or classify it, and they may be used in conjunction with other methods for increasing the array of tasks they can perform. The above mentioned flexibility is one of the advantages of GAN. They also have strong feature extraction and sample generation power (Ren et al., 2022), not requiring labeled data. Because of this, GAN have been used for data augmentation and generation (Brownlee, 2019), as is exemplified in Deeluea et al (2022). Another advantage of GAN is the capacity to use transfer learning to speed up the training process by only requiring fine-tuning, as can be seen in Fu and Lee (2021). GAN models also do not require assumptions on the probability distribution (Zhao and You, 2020). However, GAN possess some unique challenges. The adversarial nature of GAN makes them complex to train, as the equilibrium between the generator and discriminator’s objectives is hard to achieve as non-cooperation cannot guarantee convergence (Bond-Taylor et al.,

448

E. Oliveira and T. Pereira

2022). Another issue is mode collapse, where the generator produces only a limited set of samples, getting stuck on local optima, and fails to capture the full distribution of the training data (Thanh-Tung and Tran, 2020). Several network designs have been considered for GAN, all of them considering the best way to deal with high-resolution data, which is a complex matter in deep learning architectures due to vanishing gradients and high memory usage (Bond-Taylor et al., 2022). Stacked GAN (Huang et al., 2017) use two GAN sequentially, in order to first generate low-resolution data, and then improve the resolution. Another type of architecture is BigGAN (Donahue et al., 2016) and BigBiGan (Donahue and Simonyan, 2019). BigGAN uses several strategies to achieve high resolutions, including using very large mini-batches for variation reduction, normalisation to discourage collapse, and using large datasets to prevent overfitting (Bond-Taylor et al., 2022). The BigBiGAN builds on top of this architecture by adding an encoder and modifying the discriminator, achieving State-of-the-Art (SOTA) performance in unsupervised representation learning. 2.3

Variational Autoencoders

Variational Autoencoders (VAE) (Kingma and Welling, 2013) are DGM with an encoder-decoder network design. The aim of the encoder network is to convert the data from the real space into a latent one, while the decoder network’s goal is to reconvert the data in the continuous latent space back to the real space. A characteristic of VAE is that, during the encoding, the distributions are regularized in order to avoid overfitting, making the latent space more robust and generalisable. In brief, regularization encodes each data sample is as a distribution in the latent space rather than a single point. Regularization occurs both locally and globally, to guarantee completeness (meaningful post-decoding) and continuity (proximity means similarity) in the latent space. In terms of advantages, VAE have in-built prevention of overfitting due to regularization, making it more robust to noise in the real data. The continuous nature of the latent space also allows for easier interpolation between different samples and generate new data, as well as make inferences to predict missing or corrupted data. However, the continuous nature of the latent space can constitute a limitation when dealing with discrete data. Another disadvantage is the difficulty to find a good loss function, as the marginal likelihood is intractable, and as such it is learned by optimizing a tractable “evidence lower bound”, obtained using a tunable inference distribution (Zhao et al., 2017). This can lead to the generation of blurry pictures (Dosovitskiy and Brox, 2016), which has been attributed to insufficient approximation to the true posterior, as the Gaussian posterior oversimplifies the model, and may map different data points in the real space to the same encoding in the latent space, leading to averaging (Bond-Taylor et al., 2022).

Deep Generative Models in Supply Chains

2.4

449

Autoregressive Models

Autoregressive generative models (Behrmann et al., 2019) have the basic idea of generating each element of a data sequence one at a time, conditioned on the previously generated elements. The theory behind them is the chain rule of probability, where the probability of a variable can be decomposed in several components, allowing to directly maximize the likelihood of the data by training a recurrent neural network (Bond-Taylor et al., 2022). A major development in these types of model has been the introduction of the concept of attention, which enables the method to select the previous steps that are more relevant for the processing of the current step. In particular, self-attention has been very impactful, as it allows parallelization, stable-learning, and can learn long-distance dependencies (Bond-Taylor et al., 2022). When self-attention mechanisms are used with an encoder-decoder neural network structure, they are called transformers (Vaswani et al., 2017). The aim of transformers is to convert a sequence into another sequence, by predicting the most likely next element in a sequence. Self-attention mechanisms are used, as they allow parallelization, and can also determine relations between elements of the same sequence. This way, context is perceived more efficiently than using recurrent networks. The above-mentioned parallelization is one of the main advantages of transformers, as it enables an improvement in the processing efficiency. They can be used in many applications, but are particularly proficient with discrete data and text, achieving SOTA performance in several Natural Language Processing (NLP) tasks (Acheampong et al., 2021). However, the attention mechanism can only deal with fixed-length sequences, which requires the separation of long sequences into shorter ones, which can cause context fragmentation. As all DGM, transformers require large amounts of data to be trained. This can be alleviated through the use of transfer learning, as it is possible to use previously trained model coefficients, requiring only some fine-tuning to the specific task’s dataset. Given its potentialities and limitations, several transformer architectures have emerged. One of the most popular, with many variants is the Bidirectional Encoder Representations from Transformers (BERT) (Devlin et al., 2018). BERT makes use of transfer learning to improve its training speed. BERT’s capacity to understand language derives from its use of Masked Language Modeling (MLM) and the Next Sentence Prediction (NSP) mechanisms. BERT by taking as input random sentences, masking some of the words in the sentences, and reconstructing the masked words from the surrounding texts at the output. This allows it to determine the order of the sentences, achieving NSP, which helps it understanding long-distance relationships between texts. The model can be trained to several NLP tasks by performing supervised training on a dataset and replacing the BERT’s fully connected output with a new set of output layers. Several variants of BERT exist, namely Robustly Optimized BERT pre-training Approach (RoBERTa) (Liu et al., 2019), which has better performance with the cost of being more resource intensive, and DistilBERT (Sanh et al., 2019), which

450

E. Oliveira and T. Pereira

presents a lighter version of BERT, making it much faster than the original version. The most well-known version of transformers is the Generative Pre-trained Transformers (GPT) (Radford and Narasimhan, 2018), as its third and fourth versions are the basis of ChatGPT (OpenAI, 2022). GPT is composed by 12 transformer layers, and 12 attention heads, and uses the high volumes of unlabeled data for pre-training and fine-tuning. Its main advantage is an improved lexical robustness, flexibility, and performance, which come with the cost of long training times due to the high volume of data it has to process.

3

Applications in Supply Chain

In this section we review papers that apply DGM to SC. The papers reviewed are the result of several queries in on Clarivate’s Web of Science indexing service, with the key words [“supply chain” AND (“deep generative models” OR “generative” OR “GAN” OR “generative adversarial network” OR “VAE” OR “variational autoencoder” OR “transformer” OR “diffusion model”)]. A total of 12 papers were found, all of them using either GAN or transformers. Starting with GAN, and in chronological order, Zhao and You (2020) propose a framework based on GAN for robust chance constrained programming, and applies it to supply chain optimization under demand uncertainty. GAN are used to learn an estimated density model of the training data. The authors consider GAN an efficient way to model uncertainty in the data-driven optimization frameworks, as GAN do not require a priori approximations and assumptions, and are an unsupervised method that can learn complex relations within the historical data. The applicability of the proposed framework is illustrated through a case-study in biofuel SC in Illinois. Fu and Lee (2021) propose a framework for route planning in road networks based on GAN, called Progressive GAN for Route Planning (ProgRPGAN). The authors take advantage of the transfer learning capabilities of GAN to develop a chain of GAN that gradually increase the map resolution, creating ever more refined routes, generating realistic paths. Between each subsequent network, the parameters are passed to improve the efficiency and stability of the model training. ProgRPGAN also pre-trains the model using information on the road network and topology in order to facilitate even further the model training. The framework was tested using route information from taxi drivers in the cities of Porto and Chengdu, and outperformed previous SOTA models. The experiments also indicated that road network information is useful for route planning. More related to manufacturing, Deeluea et al (2022) use GAN to generate abnormal data to improve fault prediction models. An issue of fault prediction is the imbalance in the dataset, as normal data vastly outnumbers abnormal data, which leads to issues when training the model. The authors propose the use of GAN to artificially augment the datasets, balancing them, and improving early fault prediction in real-time on a low-frequency sensor device. The proposed system was tested in the operations of an injection molding machine. Results

Deep Generative Models in Supply Chains

451

indicate that GAN can capture the characteristics of the original signal, and that the additional data improves the performance of the model. Lin et al (2022) propose a dynamic supply chain member selection algorithm based on conditional generative adversarial networks (CGANs). The CGAN method proposed analyzes the index data of partners’ relevant factors. To determine this relevant factors, the method mines the alternative schemes that the decision makers of the core enterprises in the supply chain use to build the cooperative relationship. This is done to improve the accuracy and comprehensiveness of the decision-making. A core feature that led to the choice of GAN was its ability to balance the data, improving generalizability. Results indicate that CGAN exhibits a better classification and processing effect for small sample unbalanced data. Although it is more related with ensuring resilient orchestration than with SC operations, Rawat et al (2022) focus on the dangers in terms of cybersecurity that can exist when using GAN and other DGM. The authors examine backdoor attacks, identifying attack patterns that may be used, and ways to protect the models, recommending extensive sampling when using DGM acquired from unverified third parties. Cai et al (2023) also focus on cybersecurity issues, namely in privacy issues. To improve privacy in Internet of Things (IoT) SC, specifically in healthcare, the authors propose a dual GAN to address to problem of insufficient data, followed by using federated learning to train the models without requiring users to upload their sensible data, by only needing to upload the model’s parameters. The proposed approach was validated in a skin cancer image recognition casestudy. Simulation results showed that the quality of the images generated by the DualGAN was excellent, elevated the detection rate of skin cancer to 91%, while ensuring the patients’ privacy. There is one paper who mixes GAN with transformer models. Kalifa et al (2022) aim to improve prediction under anomalies by leveraging external knowledge in the form of text. The proposed methodology is based on transformers to code information about a day, based on the relations of the day’s events. That information is then used to forecast consumer behavior. The methodology uses a novel adversarial encoder framework based on transformers. This is called GAN-Event, and has two tasks: 1) reconstruct the events of the day (performed by the generator); 2) learn the strength of association between the day’s events (performed by the discriminator). Using transformers instead of regular neural networks allows GAN-Event to process input of varying size of events, which is useful because the number of events per day can vary among days. After GANEvent is used to add the context provided by the events, an LSTM model is used for sales forecasting. The methodology is empirically tested in large e-commerce datasets, and the results surpass the previous SOTA. The authors claim that learning the events’ association as compared to merely leveraging all the day’s events semantics has a significant impact on performance. Moving now for the papers applying transformers to SC, Bresson and Laurent (2021) adapt transformers from NLP to tackling Traveling Salesman Problem

452

E. Oliveira and T. Pereira

(TSP). Comparing its use to solve NLP, when transformers are used to tackle TSP, words become locations (2D points), and phrases and/or text become a tour (a sequence of locations). Another difference is that the order of the input sequence is irrelevant for TSP but relevant in NLP, while the order of the output sequence relevant for both. In the experiments with smaller problem sizes (50 and 100 location), the proposed method improves the results in relation to recent learned heuristics with an optimal gap of 0.004% for TSP50 and 0.39% for TSP100. However, traditional solvers still outperform the proposed method in terms of performance and generalization, although neural network solvers are faster. In Aejas et al (2022), BERT is used for entity recognition and clause extraction in SC contracts and the automation of smart contract creation using the extracted clauses. This type of transformer is used in conjunction with blockchain technology to develop an application that automatizes the creation of smart contracts. Focused on manufacturing, Fafard et al (2022) use a transformer designed for computer vision, called DINO (Caron et al., 2021). The objective is to cluster similar printed circuits according to their functionality. It helps detecting counterfeit printed circuit boards, and detect unreported functions. The authors report that without the pre-training step, the network does not converge as strongly into a well defined clusters. Vall´es-P´erez et al (2022) use transformers for sales (or time-series) forecasting. By using deep learning algorithms with better generalization capacity, they can use a single model for all the different locations and products, instead of needing separate, more specific models for each of those combinations, as with traditional forecasting methods. They also introduce a training trick that allows to train the model to adapt to different time steps, not requiring to retrain the model every time a prediction is needed. It is tested on the Corporaci´ on Favorita dataset from fashion retail, and is competitive in terms of performance with other more specific models. Finally, in Gammelli et al (2022), they use a variational Poisson recurrent neural network model (VP-RNN) to forecast future pickup and return rates for inventory management of a bike-sharing platform. Although it is not a transformer, it is an autoregressive generative model. From their case-study experiments, the proposed method outperformed benchmarks in terms of system efficiency and demand satisfaction. The authors also show that more accurate predictions do not necessarily translate into better inventory decisions, and recommend that predictive and prescriptive performance goals should be carefully aligned when designing new predictive model.

4

Discussion

In this section, we will discuss the main trends that emerge from the literature review performed in the previous section. The most glaring issue is the fact that only papers applying GAN or autoregressive models to SC have been published.

Deep Generative Models in Supply Chains

453

Applications of DGM to SC seem to favor the lack need for previous assumptions that GAN provides, and also the sequential nature of autoregressive models, especially transformers. We can also hypothesize if there is a predominance of the use of discrete or integer data in SC literature (e.g., sequences of locations, forecasting demand) instead of continuous data, to which EBM and VAE would be a better fit, given their capacity to interpolate in continuous latent spaces. In terms of applications, GAN are used for: – Generating data (to deal with uncertainty, to augment and balance data) [Zhao and You (2020); Deeluea et al (2022); Cai et al (2023)] – Route planning (using transfer learning) [Fu and Lee (2021)] – Supplier selection (balancing data) [Lin et al (2022)] – Processing non-structured data (image and text) [Kalifa et al (2022)] Autoregressive models are used for: – Route/tour planning (sequence generation) [Bresson and Laurent (2021)] – Processing non-structured data (image and text) [Kalifa et al (2022); Aejas et al (2022); Fafard et al (2022)] – Forecasting time-series (e.g., demand) [Vall´es-P´erez et al (2022); Gammelli et al (2022)] We can see that DGM are mostly used to: i) improve methods for existing tasks (supplier selection, forecasting); ii) process non-structured data, such as text and images; iii) generate new data, to deal with uncertainty or balance existing data. From these application, it is possible to see that DGM do more than just improving the performance in terms of solution quality and speed, expanding the tasks that can be performed, either by considering types of data that were not considered before, or by improving the quality of the existing data, making the methods more robust to exogenous issues (i.e., issues that are not directly controllable or even within the SC). Each type of DGM has its own characteristics and inner workings, which confers them distinct advantages and disadvantages. GAN are used for their lack of need to parameterize and transfer learning capabilities, while autoregressive models are used when we need to process sequences. VAE could be used in situation where interpolation is required, and diffusion models when intermediate versions of the generation process need to be verified. The last part of this section is reserved for a discussion on barriers and drivers on the application of DGM to SC, and what may constitute future research directions on this topic. Currently, the main barrier to a greater use of DGM and all their potentialities is the lack of datasets that are large enough to train DGM, which limits their application. An example of this can be seen in Wu et al (2022), where the authors comment that, when trying to apply the transformerbased solution developed in Bresson and Laurent (2021) to the 2021 Amazon Last Mile Routing Research Challenge dataset (Merch´ an et al., 2022), the results do not perform close to the best methods. The authors (Wu et al., 2022) suspect the reason for this is that training deep learning-based methods often requires

454

E. Oliveira and T. Pereira

a large number of samples. In this case, Bresson and Laurent (2021) used tens of millions of simulated tours in order to reach a very small optimality gap. However, in the Amazon last mile dataset there are only 6,112 training examples, which does not allow for proper training of deep-learning methods. Despite the exponential increase in the capacity to capture data, most real-world datasets openly available are several magnitudes smaller than necessary for the training of deep-learning models (several thousands vs. tens of millions). For this reason, most deep-learning methods are trained on simulated data. One of the difficulties lies in the validation and labeling of such high volume of data. In this aspect of supervision, the self-supervised nature of DGM may actually be beneficial, as it can enable the completion of certain task without requiring human-intensive labeling. A current and future trend in SC that might be connected with the further use of DGM in the future is SC visibility, which can be defined as useful information about demand and supply that actors from a SC have access to in a timely and accurate manner (Kalaiarasan et al., 2022). With the pressure for increased visibility in order to improve the SC’s performance and resilience, this may lead to an upsurge in the amount of available data, which will mostly be unlabeled. This represents a great opportunity for the use of DGM given the self-supervised characteristics discussed above. Therefore, improvements in SC visibility can lead to improved feasibility of the use of DGM, which in turn may improve processing the vast amount of information generated, creating a virtuous cycle.

5

Conclusion

In this study, we discuss the applications of Deep Generative Models (DGM) to Supply Chains (SC). The contributions of this study are: i) an introduction to DGM models and their different types; ii) a literature review on the application of DGM to SC, identifying the main areas of application; iii) a discussion on the main barriers and drivers for future application of DGM to SC. Given the novelty of DGM techniques, only 12 papers applying DGM to SC were identified. All of these papers used either GAN or Autoregressive models. Application focused on: i) improved performance; ii) processing non-structured data (images or text); iii) generating new data to improve models’ robustness and fix balancing issues in the data. The main barrier identified to the application of DGM to SC is the lack of available data, which the DGM requires in high volume. However, the self-supervised nature of DGM is quite relevant when this data is made available, as it can use data without human labeling. We speculate that the increase of importance of SC visibility may lead to a greater need for the application of DGM methods to SC. Acknowledgements. The authors acknowledge Funda¸ca ˜o para a Ciˆencia e a Tecnologia (FCT) for its financial support via the grant CEECINST/00096/2021 and the project UIDB/50022/2020 (LAETA Base Funding).

Deep Generative Models in Supply Chains

455

References Acheampong, F.A., Nunoo-Mensah, H., Chen, W.: Transformer models for text-based emotion detection: a review of bert-based approaches. Artif. Intell. Rev. 54(8), 5789– 5829 (2021). https://doi.org/10.1007/s10462-021-09958-2 Aejas, B., Bouras, A., Belhi, A., Gasmi, H.: Smart contracts implementation based on bidirectional encoder representations from transformers. In: Canciglieri Junior, O., No¨el, F., Rivest, L., Bouras, A. (eds.) Product Lifecycle Management, pp. 293–304. Springer International Publishing, Cham, Green and Blue Technologies to Support Smart and Sustainable Organizations (2022) Alain, G., Bengio, Y., Yao, L., Yosinski, J., Thibodeau-Laufer, Z.S., Vincent, P.: Gsns: generative stochastic networks. Inform. Inference: a J. IMA 5(2), 210–249 (2016). https://doi.org/10.1093/imaiai/iaw003 Behrmann, J., Grathwohl, W., Chen, R.T., Duvenaud, D., Jacobsen, J.H.: Invertible residual networks. In: Chaudhuri, K., Salakhutdinov, R, (eds) Proceedings of the 36th International Conference on Machine Learning, PMLR, Proceedings of Machine Learning Research, vol 97, pp. 573–582 (2019). https://proceedings.mlr.press/v97/ behrmann19a.html Bengio, Y., Yao, L., Alain, G., Vincent, P.: Generalized denoising auto-encoders as generative models. Curran Associates Inc., Red Hook, NY, USA, NIPS’13, pp. 899– 907 (2013) Bond-Taylor, S., Leach, A., Long, Y., Willcocks, C.G.: Deep generative modelling: a comparative review of vaes, gans, normalizing flows, energy-based and autoregressive models. IEEE Trans. Pattern Anal. Mach. Intell. 44(11), 7327–7347 (2022). https:// doi.org/10.1109/TPAMI.2021.3116668 Bresson, X., Laurent, T.: The transformer network for the traveling salesman problem (2021). arXiv:2103.03012 Brownlee, J.: A gentle introduction to generative adversarial networks (gans) (2019). https://machinelearningmastery.com/what-are-generative-adversarial-networksgans/. Accessed 12 Dec 2022 Cai, X., Lan, Y., Zhang, Z., Wen, J., Cui, Z., Zhang, W.: A many-objective optimization based federal deep generation model for enhancing data processing capability in Iot. IEEE Trans. Industr. Inf. 19(1), 561–569 (2023). https://doi.org/10.1109/TII.2021. 3093715 Caron, M., et al.: Emerging properties in self-supervised vision transformers. In: Proceedings of the IEEE/CVF International Conference on Computer Vision, pp. 9650– 9660 (2021) Deeluea, S., Jeenanunta, C., Tunpun, A.: Fault prediction model for motor and generative adversarial networks for acceleration signal generation. In: 2022 17th International Joint Symposium on Artificial Intelligence and Natural Language Processing (iSAI-NLP), pp 1–5 (2022). https://doi.org/10.1109/iSAI-NLP56921.2022.9960281 Devlin, J., Chang, M.W., Lee, K., Toutanova, K.: Bert: Pre-training of deep bidirectional transformers for language understanding. https://doi.org/10.48550/ARXIV. 1810.04805 Donahue, J., Simonyan, K.: Large Scale Adversarial Representation Learning. Curran Associates Inc., Red Hook, NY, USA (2019) Donahue, J., Kr¨ ahenb¨ uhl, P., Darrell, T.: Adversarial feature learning (2016). https:// doi.org/10.48550/ARXIV.1605.09782 Dosovitskiy, A., Brox, T.: Generating images with perceptual similarity metrics based on deep networks (2016). https://doi.org/10.48550/ARXIV.1602.02644

456

E. Oliveira and T. Pereira

Du, Y., Mordatch, I.: Implicit generation and modeling with energy based models. Adv. Neural Inform. Process. Syst. 32 (2019) Duvenaud, D., Wang, J., Jacobsen, J., Swersky, K., Norouzi, M., Grathwohl, W.: Your classifier is secretly an energy based model and you should treat it like one (2020) Fafard, A., Mohieldin, S., Spielberg, J.: Single image printed circuit board functional similarity clustering using vision transformers. In: 2022 IEEE Physical Assurance and Inspection of Electronics (PAINE), pp 1–7 (2022). https://doi.org/10.1109/ PAINE56030.2022.10014831 Fu, T.Y., Lee, W.C.: Progrpgan: Progressive gan for route planning. In: Proceedings of the 27th ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD 2021) (2022). https://doi.org/10.1145/3447548.3467406 Gammelli, D., Wang, Y., Prak, D., Rodrigues, F., Minner, S., Pereira, F.C.: Predictive and prescriptive performance of bike-sharing demand forecasts for inventory management. Transportation Research Part C: Emerging Technologies 138, 103,571 (2022). https://doi.org/10.1016/j.trc.2022.103571, https://www.sciencedirect.com/ science/article/pii/S0968090X22000183 Goodfellow, I.: Generative adversarial nets. In: Advances in Neural Information Processing Systems, pp. 2672–2680, http://papers.nips.cc/paper/5423-generativeadversarial-nets.pdf Hinton, G.E.: Training products of experts by minimizing contrastive divergence. Neural Comput. 14(8), 1771–1800 (2002) Hinton, G.E., Sejnowski, T.J.: Optimal perceptual inference. Proc. IEEE Conf. Comput. Vision Patt. Recogn. Citeseer 448, 448–453 (1983) Ho, J., Jain, A., Abbeel, P.: Denoising diffusion probabilistic models. Adv. Neural. Inf. Process. Syst. 33, 6840–6851 (2020) Huang, X., Li, Y., Poursaeed, O., Hopcroft, J., Belongie, S.: Stacked Generative Adversarial Networks. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 5077–5086 (2017) Jing, L., Tian, Y.: Self-supervised visual feature learning with deep neural networks: a survey. IEEE Trans. Pattern Anal. Mach. Intell. 43(11), 4037–4058 (2021). https:// doi.org/10.1109/TPAMI.2020.2992393 Kalaiarasan, R., Olhager, J., Agrawal, T.K., Wiktorsson, M.: The abcde of supply chain visibility: A systematic literature review and framework. Int. J. Prod. Econ. 248, 108,464 (2022).https://doi.org/10.1016/j.ijpe.2022.108464, https:// www.sciencedirect.com/science/article/pii/S0925527322000573 Kalifa, D., Singer, U., Guy, I., Rosin, G.D., Radinsky, K.: Leveraging world events to predict e-commerce consumer demand under anomaly. In: Proceedings of the Fifteenth ACM International Conference on Web Search and Data Mining (WSDM 22), pp. 1–9 (2022). https://doi.org/10.1145/3488560.3498452 Kingma, D.P., Welling, M.: Auto-encoding variational bayes (2013). https://doi.org/ 10.48550/ARXIV.1312.6114, LeCun, Y., Chopra, S., Hadsell, R., Ranzato, M., Huang, F.: A tutorial on energy-based learning. Predicting structured data 1(0) (2006) Lin, H., Lin, J., Wang, F.: An innovative machine learning model for supply chain management. J. Innov. Knowl. 7(4), 100,276 (2022)https://doi.org/10.1016/j.jik.2022. 100276, https://www.sciencedirect.com/science/article/pii/S2444569X22001111 Liu, Y., et al.: Roberta: A robustly optimized bert pretraining approach (2019). https:// doi.org/10.48550/ARXIV.1907.11692, https://arxiv.org/abs/1907.11692 Merch´ an, D.: 2021 amazon last mile routing research challenge: Data set. Transp. Sci. 0(0):1–4 (2022). https://doi.org/10.1287/trsc.2022.1173

Deep Generative Models in Supply Chains

457

Murphy, K.P.: Machine Learning: A Probabilistic Perspective. The MIT Press (2012) Open AI (2022) Dall-e 2. https://openai.com/dall-e-2/. Accessed 15 Dec 2022 OpenAI (2022) Chatgpt: Optimizing language models for dialogue. https://openai. com/blog/chatgpt/ Accessed 14 Dec 2022 Radford, A., Narasimhan, K.: Improving language understanding by generative pretraining (2018) Rawat, A., Levacher, K., Sinn, M.: The devil is in the gan: Backdoor attacks and defenses in deep generative models. In: Atluri, V., Di Pietro, R., Jensen, C.D., Meng, W. (eds.) Computer Security - ESORICS 2022, pp. 776–783. Springer Nature Switzerland, Cham (2022) Ren, Y., et al.: Deep clustering: A comprehensive survey (2022). https://doi.org/10. 48550/ARXIV.2210.04142, https://arxiv.org/abs/2210.04142 Ruthotto, L., Haber, E.: An introduction to deep generative modeling (2021). arXiv:2103.05180 Sanh, V., Debut, L., Chaumond, J., Wolf, T.: Distilbert, a distilled version of bert: smaller, faster, cheaper and lighter (2019). https://doi.org/10.48550/ARXIV.1910. 01108 Sohl-Dickstein, J., Weiss, E., Maheswaranathan, N., Ganguli, S.: Deep unsupervised learning using nonequilibrium thermodynamics. In: Bach F, Blei D (eds) Proceedings of the 32nd International Conference on Machine Learning, PMLR, Lille, France, Proceedings of Machine Learning Research, vol 37, pp 2256–2265 (2015). https:// proceedings.mlr.press/v37/sohl-dickstein15.html Thanh-Tung, H., Tran, T.: Catastrophic forgetting and mode collapse in gans. In: 2020 International Joint Conference on Neural Networks (ijcnn), IEEE, pp 1–10 (2020) Vall´es-P´erez, I., Soria-Olivas, E., Mart´ınez-Sober, M., Serrano-L´ opez, A.J., G´ omezSanch´es, J., Mateo, F.: Approaching sales forecasting using recurrent neural networks and transformers. Expert Syst. Appl. 201, 116,993 (2022). https://doi. org/10.1016/j.eswa.2022.116993, https://www.sciencedirect.com/science/article/ pii/S0957417422004146 Vaswani, A., et al.: Attention is all you need (2017). https://doi.org/10.48550/ARXIV. 1706.03762, https://arxiv.org/abs/1706.03762 Wu, C., Song, Y., March, V., Duthie, E.: Learning from drivers to tackle the amazon last mile routing research challenge (2022). arXiv:2205.04001 Yang, L., et al.: Diffusion models: A comprehensive survey of methods and applications (2023). arXiv:2209.00796 Zhao, S., You, F.: Distributionally robust chance constrained programming with generative adversarial networks (GANs). AIChE J. 66(6), (2020). https://doi.org/10. 1002/aic.16963 Zhao, S., Song, J., Ermon, S.: Towards deeper understanding of variational autoencoding models (2017). https://doi.org/10.48550/ARXIV.1702.08658

Business Context-Based Approach for Managing the Digitalization of Biopharmaceutical Supply Chain Operational Requirements Elena Jelisic1,2(B) , Milos Drobnjakovic1 , Boonserm Kulvatunyou1 , Nenad Ivezic1 , and Hakju Oh1 1 Engineering Laboratory, National Institute of Standards and Technology, Gaithersburg, MD,

USA {elena.jelisic,milos.drobnjakovic,boonserm.kulvatunyou, hakju.oh}@nist.gov 2 Faculty of Organizational Sciences, University of Belgrade, Belgrade, Serbia

Abstract. The COVID-19 pandemic has brought to attention several supply chain challenges in the globalized biopharmaceutical industry. Demand fluctuations and the ability to effectively repurpose manufacturing facilities, and to dynamically adapt scheduling and capacity to respond to those fluctuations are among the important challenges. The biopharmaceutical industry is highly segmented by geography and competency. Companies often must cooperate with other companies dispersed across the globe. While the manufacturing of final products is concentrated in a few western locations, raw materials may come from other parts of the world and the final products have to be distributed globally. Such a diverse environment is further complicated by stringent regulatory requirements and the lack of standardization for many operational raw materials and customization of product recipes. Therefore, it is crucial to have an effective way to document and use the requirements of different supply chain actors. Still, the requirements are usually written in paper form where one actor specifies a set of rules that other actors must follow to communicate specific information effectively. Digitalization, while promising to address this issue, is a stumbling block because of the complexity of the biopharmaceutical industry. Motivated by this challenge, this paper critically evaluates the novel Business Context-based approach for effective (digitalization) management of supply chain (operational) requirements. Biopharmaceutical supply chain networking is used as a specific case study. The investigation results show that, with identified improvements, the approach has the potential to manage supply chain requirements effectively. Keyword: Supply chain requirements · Supply Chain · Biopharmaceutical · Business Context

© IFIP International Federation for Information Processing 2023 Published by Springer Nature Switzerland AG 2023 E. Alfnes et al. (Eds.): APMS 2023, IFIP AICT 689, pp. 458–470, 2023. https://doi.org/10.1007/978-3-031-43662-8_33

Business Context-Based Approach

459

1 Introduction The biopharmaceutical market is one of the fastest-growing bioeconomy sectors, projected to reach an 11.3% Compound Annual Growth Rate (CAGR) by 2030 [1]. The market’s rapid growth due to ever-expanding treatment options has also led to the globalization of target markets and manufacturing. Namely, companies often outsource parts of their operations and must cooperate with other companies dispersed globally. Consequently, the complexity of the global biopharmaceutical network has led to the need for increased supply chain agility and resilience. Supply chain agility and resilience refer to the ability of a supply chain to quickly respond to changes in supply, demand, and other external disruptions. Yaroson et al. investigated supply chain agility and identified four dimensions of agility - alertness, accessibility, visibility, and willingness [2]. Tukamuhabwa et al. defined supply chain resilience as “the adaptive capability to prepare for and/or respond to disruptions, to make a timely and cost-effective recovery, and therefore progress to a post disruption state of operations – ideally, a better state than prior to the disruption” [3]. Many disruptions in logistics and transportation occurred during the COVID-19 pandemic [4]. These disruptions identified several potential vulnerabilities and optimization opportunities that must be addressed. One of the primary areas of improvement is the digitalization of the biopharma supply chain. Namely, digitalization would permit easier identification and establishment of backup supply routes and help anticipate disruptions in the existing ones. Digitalization of the supply chain is also imperative for one of biopharma’s primary development objectives – Biopharma 4.0 [5, 6]. However, the biopharmaceutical industry still needs to achieve this. To understand why digitalization is a stumbling block for the biopharmaceutical industry, we reviewed previous papers and identified four categories of digitalization challenges. This was important to understand the complexity of the biopharmaceutical industry. One of the challenges is the segmentation of the biopharmaceutical market, so it is crucial to have a mechanism to handle the heterogenous supply chain operational requirements (or capabilities if viewed from inside the organization) of various supply chain actors (e.g., obligatory regulations). The literature identifies two approaches – market mapping [6] and configuration of a supply chain [7–9] – that help analyze such requirements. Still, in practice, these requirements are usually documented in a paper form where one supply chain actor specifies a set of rules, as a free-form text, that other actors must follow to effectively communicate certain information (e.g., Certificate of Analysis). The contribution of the paper is a preliminary evaluation of the novel Business Context-based approach for effective management of supply chain operational requirements. We use OAGIS Express Pack for Real-Time Release of Raw Materials (RTRRM) [7], recently released by Open Application Group Inc. (OAGi) for evaluation. RTRRM has predefined supply chain messages specialized for application integration in the biopharmaceutical industry. Specifically, we use the Certificate of Analysis message from the package for our evaluation. RTRRM was created by using the Score platform [8, 9], which was developed in cooperation between the National Institute of Standards and Technology (NIST) [10] and OAGi [11].

460

E. Jelisic et al.

The remaining parts of the paper are structured as follows. Section 2 introduces the needed background. Section 3 describes the envisioned approach. Section 4 offers a preliminary evaluation of the approach using a biopharmaceutical domain use case. Section 5 discusses the results and future work. Section 6 closes the paper with concluding thoughts.

2 Background 2.1 Supply Chain Requirements According to [12], “the biopharmaceuticals market is estimated to be USD 407.72 billion in the current year”. The COVID-19 pandemic significantly impacted the biopharmaceutical industry, and the biopharmaceutical market is expected to grow considerably by the end of 2028 [12]. Acumen Research and Consulting recently published a report that estimates the growth to be at an 11.3% Compound Annual Growth Rate (CAGR) by 2030 [1]. This growing biopharmaceutical market is highly segmented by product type, therapeutic application, and geography [12]. Documenting supply chain operational requirements (e.g., obligatory regulations) for each specific actor is essential in such a segmented operating environment. According to Boehlje [13], six dimensions can be used to describe the supply chain requirements: (1) processes; (2) product flow; (3) financial flow; (4) information flow; (5) incentive systems; and (6) governance. Analyzing and documenting these dimensions should assure supply chain sustainability and competitiveness. Zamora defined a market mapping mechanism to document supply chain operational requirements that identify and associate all supply chain actors to their environment [14]. The environment defines “the operating conditions within which the actor operates, such as infrastructure, policies, and regulations, as well as institutions and processes that shape the market ecosystem” [14]. Recently, a supply chain configuration has been established as an emerging research topic [15–17]. The configuration of a supply chain identifies a set of supply chain requirements that should support operation and information sharing between supply chain actors, across heterogeneous languages, cultures, and regulations. Hernández and Pedersen investigated this topic and concluded that the configuration of a supply chain is complex and requires constant modifications to address the changes in specific countries, industries, or companies themselves [15]. The supply chain configuration in the biopharmaceutical industry is further complicated by custom designs, and single sources due to the strict regulatory requirements as well as the inherent “natural variability” typically present in biopharmaceutical production processes. Currently, supply chain requirements are usually documented in a paper or unstructured electronic form where one supply chain actor specifies a set of rules that other actors must follow to effectively communicate certain information (e.g., Certificate of Analysis - CoA). Examples of supply chain requirements are information elements that must be included in the CoA covering acceptance test procedures for specific raw material types (e.g., media, supplements, and buffers), acceptable test results, etc. Paper-based requirements are difficult to streamline in any subsequent analytical processing, which makes efficient utilization of the contained information challenging. Digitalizing supply chain requirements can streamline existing procedures and facilitate

Business Context-Based Approach

461

data-driven decision-making. For example, the reported values for a particular attribute of a raw material that impacts a product Critical Quality Attribute (CQA), or process Key Performance Indicators (KPIs) can be utilized to automatically make production adjustments before raw material utilization within the process. Nevertheless, there are several challenges to digitalization specific to the biopharmaceutical industry. These are discussed in the next section. 2.2 Digitalization Challenges During the COVID-19 pandemic, many disruptions in logistics and transportation have been experienced [4]. This has hastened the digitalization of the biopharmaceutical supply chains. However, there are obstacles to achieving this. We reviewed the recent literature to identify the most significant digitalization issues and divided them into four categories – complexity, resistance, regulatory compliance, and data integrity. Complexity. Biopharmaceutical chains are complex [18], segmented [12], and highly globalized [16]. Due to their globalized nature, biopharmaceutical companies must cooperate with small and large partners dispersed across the globe. According to [16], biopharmaceutical companies can be divided into three categories: 1) Biotechnology firms, 2) Traditional pharmaceutical companies, and 3) Generic drug companies. Often, the biopharmaceutical industry requires “just-in-time delivery” because of the limited shelflife and to ensure timely patient treatment [19] and timely delivery [20]. Moreover, to prevent possible interruptions (e.g., natural disasters, raw material shortage, regulatory changes), many companies practice a “dual-supply” strategy [21]. However, a dual supply chain strategy is difficult because there is a lack of regulatory standardization and, in some cases, highly customized product and production designs. Therefore, they must meet the diverse requirements of scattered manufacturing sites that usually differ in products, technology usage, development process, and obligatory regulations that drive this process [22, 23]. Resistance. Yaroson et al. investigated supply chain agility and identified four dimensions of agility - alertness, accessibility, visibility, and willingness [2]. The supply chain agility depends on its digitalization readiness, and digitalization assumes infrastructural, technological, managerial, and cultural changes. However, the biopharmaceutical industry is still resistant [24] and conservative [2] in accepting these fundamental changes. The main reasons for that are risk reduction [17], finance [25], and information sensitivity [20]. Regulatory Compliance. Grossman and Bates, in their research, stressed that regulatory compliance is a big obstacle for the pharmaceutical industry [26]. To register a new product, a pharmaceutical company must follow product registration procedures and comply with the current compendial requirements defined by local regulatory agencies. Pharmacopeias are used to describe these compendial requirements. Some of the widely-used pharmacopeias are - United States Pharmacopeia–National Formulary (USP-NF) [27], European Pharmacopoeia [28], Japanese Pharmacopoeia [29], British Pharmacopoeia [30], and Chinese Pharmacopoeia [31]. While there are some overlap (for example, British Pharmacopoeia mainly relies on European Pharmacopoeia but has some specific, additional requirements), there are many differences in international standards

462

E. Jelisic et al.

[32]. Since many pharmaceutical companies outsource their production processes, compliance becomes an even more significant challenge. Pharmacopeia compliance affects the product throughout its life cycle, from development and testing to commercialization. The acceptance of materials used in a production process and the validity of tests these materials undergo are just some of the variations that pharmaceutical companies need to address [33]. Wiggins and Albanese stressed the need to harmonize and develop global pharmacopeia standards [34]. The World Health Organization’s (WHO) initiative to establish Good Pharmacopeial Practices (GPhP) is a step towards this. The last International Meeting of World Pharmacopoeias was held in September 2022 [35]. Data Integrity. The Food and Drug Administration (FDA) agency proposes ALCOA (Attributable, Legible, Contemporaneous, Original, and Accurate) and ALCOA + (Complete, Consistent, Enduring, and Available, in addition) principles that should help govern the data integrity [36]. Data integrity is a big issue that prevents the digitalization of the biopharmaceutical industry [18, 37]. A recent study in 2020 indicated that documents were still stored as hard copies (e.g., contracts), making tracking and validating the procurement process highly demanding and time-consuming [25]. To resolve this, the first step is the digitalization of paper-based systems with electronic ones [38] to assure data transparency, accuracy, and improved information exchange among the supply chain actors [18]. However, several challenges hinder faster digitalization. First, drug serialization, even though mandated by authorities in Europe and the United States, is still an open issue [18]. The second challenge stems from the lack of standards for a specific portion of the process (e.g., eCoAs currently only have one standard, which is not necessarily fully applicable to complex materials). The need for interoperable standards [39] makes integration of systems protracted and, in many cases, business case specific, hindering the adoption of existing standards and incorporation of new standards. This specific issue is essential, considering that the amount of data pharmaceutical companies need to handle is considerably growing [40].

2.3 Business Context The Core Component Technical Specification (CCTS) is a meta-model standard that enables the development and use of Data Exchange Standards (from here on, standards) [41]. The CCTS uses three critical concepts – Core Components, Business Information Entities, and Business Context. While Core Components represent standard building blocks, Business Information Entities represent profiles of these standard components (i.e., Core Components) restricted for a specific integration use case (e.g., restriction of a value domain, cardinalities, data format, etc.). According to Novakovic, Business Context is “any information that can be used to characterize the situation of an entity within a scope where the business operates. An entity is a person, place, or object that is considered relevant to the execution of a business process in a business environment, including the business process and business environments themselves” [42]. The Business Context is defined by categories describing specific aspects of an integration use case. CCTS proposes eight business context categories, but additional ones can be defined if they are deemed as needed by an integration architect. Recent studies have investigated Business

Business Context-Based Approach

463

Context [43], proving it can improve standardization and integration efforts [44]. Moreover, these studies demonstrated that Business Context could be employed to develop the previously mentioned standard component profiles (i.e., Business Information Entities). Three Business Context models are identified in the literature – the UN/CEFACT Context Model (UCM), the Enhanced UN/CEFACT Context Model (E-UCM), and Business Context Ontology (BCOnt) [44]. Each of these models defines a set of operators (e.g., union, intersection) and predicates (e.g., ‘ < ‘(less than), ‘ >’ (more than)) that can be employed to describe more complex integration use cases.

3 Business Context-based Approach for Managing Supply Chain Operational Requirements Novakovic made a distinction between external and internal contexts [42]. The external context is detected using various physical sensors (e.g., Radio Frequency Identification (RFID), Internet of Things (IoT), Quick Response (QR) code, etc.) that enable traceability of the supply chain. In contrast, the internal context (i.e., Business Context) is detected using logical sensors. This paper focuses on Business Context, as it could lend itself to an approach for managing supply chain requirements. Figure 1 illustrates the proposed approach as a five-step process. Step 1 identifies all factors that regulate the structure and the content of exchanged messages (e.g., COA) in a biopharmaceutical supply chain. Examples of these factors are the geographical location of the biopharmaceutical factory and the target market, pharmacopeias that those markets comply with, raw material type, etc. These factors are to be represented as Business Context categories. Once the Business Context categories are identified, step 2 creates the Business Context knowledge base. According to [44], each category is associated with one or multiple schemas (i.e., sources of values), and the Business Context knowledge base comprises a collection of all possible values for each schema. These steps are crucial as other steps depend on them and must be conducted in cooperation with a domain expert.

Fig. 1. An approach for supply chain requirements management.

After creating the Business Context knowledge base, step 3 creates standard component profiles. This results in an association of each standard component to a specific Business Context expression and a restriction of that component. Such restriction determines the component’s structure and content that are valid for the associated biopharmaceutical integration use cases (e.g., acceptance test procedures, test results, exchange, etc.). Step 3 is repeated for each standard component1 . 1 The proposed approach assumes a collection of standard components exists.

464

E. Jelisic et al.

Step 4 expresses supply chain requirements for the required biopharmaceutical integration use case. The supply chain requirements are defined as a Business Context expression where each category is associated with a specific value selected from the Business Context knowledge base. Finally, step 5, filters the standard component profiles, by using the supply chain requirements, to be embedded in a message profile (e.g., CoA profile). This filtering is based on the Effective Business Context calculation proposed in [44]. The Effective Business Context intersects business context expressions assigned in step 3.1 and the supply chain requirements expressed in step 4. Those standard components for which Effective Business Context is calculated as an empty set are treated as nonrelevant and removed for the specific supply chain requirements.

4 Use Case An OAGIS Express Pack for the Real-Time Release of Raw Materials (RTRRM) developed by Open Application Group Inc. (OAGi) supports digitalization in the biopharmaceutical industry [7]. The Express Pack contains predefined message profiles for the biopharmaceutical industry. The Batch Certificate of Analysis is one of the Express Pack messages; it will be used for the evaluation. Certificates of Analysis (CoA) are fundamental for good manufacturing practice (GMP) and are defined as documents that “provide a summary of testing results on samples of products or materials together with the evaluation for compliance to a stated specification” [45]. Step 1: Identify Business Context Categories. We interviewed a domain expert to identify all critical aspects (i.e., business context categories) that drive the CoA’s structure and content. While all message types should be analyzed, we focused on a few CoA aspects, making sure this simplification does not affect the integrity of the evaluation. Four business context categories were chosen after the interviews and analysis of existing documents: Factory Location, Target Market, (product) Purpose of Use, and Pharmacopeia Compliance. Step 2: Create a Business Context Knowledge Base. Each business context category is associated with an appropriate business context schema that provides a source of business context values for the category. After consulting with domain experts, the classification schema for the Purpose of Use category was defined according to [46, 47], while [48] was used for the Pharmacopeia Compliance category. Figure 2 presents an excerpt of the Business Context knowledge base. The Business Context knowledge base can be supplemented with logical axioms that bring additional knowledge. Examples of such logical axioms are illustrated in Table 1. A typical axiom pattern is as follows: A raw material (rm) that complies with European pharmacopeia does not have to comply with British pharmacopeia. Step 3: Create Standard Components Profiles. Using a previously created Business Context knowledge base, Business Context can be defined by a combination of values from classification schemas. Each component from the CoA OAGIS Express Pack schema was contextualized using business contexts as shown in Fig. 3. The contextualization assumes the association of a component to a specific Business Context(s). For illustration purposes in this paper, components are prefixed with associated Business

Business Context-Based Approach

465

Fig. 2. Business Context knowledge base.

Table 1. Logical axioms. No

Logical axiom

1

not (compliesWith (rm, European) → compliesWith (rm, British))

2

not (compliesWith (rm, European) → compliesWith (rm, Italian))

3

not (compliesWith (rm, WHO) → compliesWith (rm, European))

Context(s) (e.g., BC_1_Manufacture Date Time, meaning the component is valid only for BC_1).

Fig. 3. Business Information Entity – data model.

466

E. Jelisic et al.

The definition of each Business Context is provided in Table 2. Table 2. Business Context. BC

Factory Location

Target market

Pharmacopeia Compliance

Purpose of Use

BC_1

China

USA

USP

GMP Manufacturing

BC_2

China

UK

British

GMP Manufacturing

BC_3

China

China

Chinese

GMP Manufacturing

Step 4: Express Required Supply Chain Requirements. In the pharmaceutical industry, a CoA is used as a source of information for different raw material information utilization purposes, from sourcing and testing to regulatory filing, where the manufacturer must report compliance with a specific regulatory agency. Because each utilization purpose has particular requirements, the manufacturer needs to adjust the content of the CoA accordingly. Table 3 illustrates three supply chain requirements defined by previously chosen Business Context categories. Table 3. Supply chain requirements. Utilization Purpose

Factory Location

Target market

Pharmacopeia Compliance

Purpose of Use

Sourcing & Testing

China

USA & UK

USP & British

GMP Manufacturing

Regulatory filing

China

UK

British

GMP Manufacturing

Regulatory filing

China

USA

USP

GMP Manufacturing

Let us consider a situation illustrated in Table 3 where a manufacturing company located in China wants to put a particular product on the US and UK markets. The product must comply with both the US and UK regulatory requirements for sourcing and testing purposes. The company needs information about all required tests and acceptable values for both markets to ensure compliance. However, for the regulatory filing purpose, where the company needs to report compliance to each specific market (i.e., US or UK), it requires only a subset of the CoA information. Step 5: Create a Message Profile. Finally, based on the previous steps, creating a specific CoA message profile for each utilization purpose is possible based on the provided supply chain requirements (as defined in Table 3). For that purpose, Effective Business Context is calculated as described in Sect. 3. As a result, three CoA message profiles

Business Context-Based Approach

467

are created, as presented in Fig. 4. Each profile contains only information relevant to the corresponding specific supply chain requirements (e.g., requirements for the Regulatory filing purpose). This is important because such profiling permits companies to accelerate and further standardize the exchange of information with suppliers and regulators. Accelerated, standardized, and digitalized communication can ease achieving the raw material Real-Time Release Testing objective. Namely, from the supplier to the manufacturer side, standardized and digitalized communication permits easier access to information on testing results that can lead to decisions such as manufacturing modifications or subsequent in-house testing planning. Also, our approach to standardization and digitalization of communication allows the manufacturer to clearly and easily specify any requirements that the supplier must comply with. Conversely, from the manufacturer to regulatory perspective, standardized and digitalized communication enhances end-to-end data traceability and permits a more streamlined regulatory filing.

Fig. 4. CoA message profiles.

5

Discussion and Future Work

The use case shown in Sect. 4 has demonstrated that the business context-based approach can be used for managing supply chain operational requirements particularly related the supply chain digitalization. Four categories have been identified as important context categories based primarily on the eCOA requirement driven by the regulatory filing and compliance in raw material. With further analyses of use cases related to the raw material impact on drug products, more context categories such as manufacturing method and product type, will likely be needed. As an example, trace metal concentrations (e.g., manganese) may have different effects depending on the mAb type and could thus have different optimal concentration ranges or testing requirements. Gaining access to this information in a timely manner is crucial for early detection of any out-of-specification conditions as well as, for providing timely insights and data as inputs for adaptive control strategies in the future.

468

E. Jelisic et al.

Second, the paper only addresses the supply chain part of the entire value chain. However, there are more types of messages that are fundamental to efficient Biopharma 4.0. Tech transfer where production process information needs to be transferred from the process development to full-scale production is one important area. Future work will focus on inter-company (e.g., to a contract manufacturer) and intra-company tech transfer use cases. Third, for evaluation purposes, an assumption was made that a collection of standard components is available. The paper used a Certificate of Analysis message from the OAGIS Express Pack for Real-Time Release of Raw Materials. Our previous research efforts have investigated the ability of standards to cover intended integration use cases [49]. The investigation results showed that it is impossible to expect that the standard can cover all variabilities in integration use cases. The main reason is the way standards are developed today. To address this issue, we have been working on a distributed standards development approach that would continuously monitor and collect the requirements of all supply chain actors [44]. Such an approach would ensure that a library of standard components is constantly updated and capable of supporting data exchanges for required variations in a biopharmaceutical domain. Fourth, in our previous research [49] we have also developed and demonstrated that Business Context can be employed to measure the quality of existing Data Exchange Standard components profiles (see Sect. 2.3). Future work will focus on investigating Business Context as a mechanism to develop similar measures to evaluate the quality of.

6 Concluding Remarks The paper identified several challenges that delay digitalization in the biopharmaceutical industry and hence the realization of Biopharma 4.0. One of the challenges is the segmentation of the biopharmaceutical market, so it is crucial to have a mechanism to handle the operational requirements of various supply chain actors. In the literature, there are different approaches to analyzing supply chain requirements. However, these requirements are usually documented in papers or free-form text files. As such, the requirements are challenging to streamline any subsequent analytical processing. An envisioned, Business Context-based approach for managing supply chain operational requirements is outlined to address this issue. The approach is evaluated on a supply chain digitalization use case using the electronic Certificate of Analysis from the OAGIS Express Pack for the RealTime Release of Raw Materials. The evaluation results showed that the approach has the potential to support managing complex supply chain digitalization requirements. To that end, several future work topics have been identified. Disclaimer and Acknowledgement. Specific commercial systems and applications identified in this paper are not intended to imply recommendation or endorsement by the National Institute of Standards and Technology, nor is it intended to indicate that they are necessarily the best available for the purpose.

References 1. 2022, Biopharmaceutical Market, Analysis Report and Region Forecast 2022 - 2030

Business Context-Based Approach

469

2. Yaroson, E.V., Breen, L., Hou, J., Sowter, J.: Resilience Strategies and the Pharmaceutical Supply Chain: The Role of Agility in Mitigating Drug Shortages. In: Barbosa-Povoa, A.P., Jenzer, H., de Miranda, J.L. (eds.) Pharmaceutical Supply Chains-Medicines Shortages, pp. 249–256. Springer (2019) 3. Tukamuhabwa, B.R., Stevenson, M., Busby, J., Zorzini, M.: Supply chain resilience: definition, review and theoretical foundations for further study. Int. J. Prod. Res. 53(18), 5592–5623 (2015) 4. Pujawan, I.N., Bah, A.U.: Supply chains under COVID-19 disruptions: literature review and research agenda. Supply Chain Forum 23(1), 81–95 (2022) 5. Erickson, J., et al.: End-to-end collaboration to transform biopharmaceutical development and manufacturing. Biotechnol. Bioeng. 118(9), 3302–3312 (2021) 6. Reinhardt, I.C., Oliveira, D.J.C., Ring, D.D.T.: Current perspectives on the development of industry 4.0 in the pharmaceutical sector. J. Ind. Inf. Integr. 18, 100131 (2019) 7. Figura, M.: Express Pack: Real-Time Release of Raw Materials for the Biopharmaceutical Industry. https://oagiscore.atlassian.net/wiki/spaces/EXPKS/pages/4564287489/Exp ress+Pack+Real-Time+Release+of+Raw+Materials+for+the+Biopharmaceutical+Industry. Accessed: 07 Mar 2023 8. Score Open-Source Project https://oagiscore.org/. Accessed 29 Mar 2023 9. Kulvatunyou, B. (Serm), Oh, H., Ivezic, N., Nieman, S.T.: Standards-based semantic integration of manufacturing information: past, present, and future. J. Manuf. Syst. 52, 184–197 (2018) 10. National Institute of Standards and Technology https://www.nist.gov/. Accessed 23 Jan 2023 11. “Open Application Group Inc.” https://oagi.org/. Accessed: 23-Jan-2023 12. “Biopharmaceuticals Market - Growth, Trends, COVID-19 Impact, and Forecasts (2023 2028),” Mordor Intell. https://www.mordorintelligence.com/industry-reports/global-biopha rmaceuticals-market-industry#:~:text=The biopharmaceuticals market is estimated,impact on the biopharmaceutical industry. Accessed: 29 Mar 2023 13. Boehlje, M.: Structural changes in the agricultural industries: how do we measure, analyze and understand them? Am. J. Agric. Econ. 81(5), 1028–1041 (1999) 14. Zamora, E.A.: Value chain analysis: a brief review. Asian J. Innov. Policy 5(2), 116–128 (2016) 15. Hernández, V., Pedersen, T.: Global value chain configuration: a review and research agenda. BRQ Bus. Res. Q. 20(2), 137–150 (2017) 16. Louis, B., Ruslan, R.: Global Value Chains and Smart Specialisation Strategy . Thematic Work on the Understanding of Global Value Chains and Their Analysis within the Context of Smart Specialisation . Global Value Chains and Smart Specialisation Strategy Thematic Work on The (2015) 17. Kim, Y.C., Atukeren, E., Lee, Y.: A new digital value chain model with PLC in biopharmaceutical industry: the implication for open innovation. J. Open Innov. Technol. Mark. Complex. 8(2), 63 (2022) 18. Sim, C., Zhang, H., Chang, M.L.: “Improving end-to-end traceability and pharma supply chain resilience using blockchain”, blockchain healthc. Today 5, 1–10 (2022) 19. 2022, “Tools of Tomorrow,” CELL GENE Ther. INSIGHTS, 8(11) 20. Crawford, M.: Blockchain: Will It Transform the Pharmaceutical Supply Chain?,” GMP Publ (2018) 21. Elder, D.P., Kuentz, M., Holm, R.: Pharmaceutical excipients - quality, regulatory and biopharmaceutical considerations. Eur. J. Pharm. Sci. 87, 88–99 (2016) 22. Pan, L.B.: Implementation Roadmap and Real Options Analysis for Biopharmaceutical Technology Introduction in Partial Fulfillment of the Requirements for the Degrees of Master of Business Administration and Master of Science in Chemical Engineering in Conjunction Wit (2013)

470

E. Jelisic et al.

23. Lamberti, M.J., et al.: An examination of eclinical technology usage and CDISC standards adoption. Ther. Innov. Regul. Sci. 49(6), 869–876 (2015) 24. Upton, J.: Setting Sights on the Smart Supply Chain. Pharm. Exec 25. Faridi, M.R., Malik, A.: Digital transformation in supply chain, challenges and opportunities in smes: a case study of al-rumman pharma. Emerald Emerg. Mark. Case Stud. 10(1), 1–16 (2020) 26. Grossman, M., Bates, S.: Knowledge capture within the biopharmaceutical clinical trials environment. Vine 38(1), 118–132 (2008) 27. USP-NF. https://www.uspnf.com/. Accessed: 29 Mar 2023 28. European Pharmacopoeia. https://pheur.edqm.eu/home. Accessed: 29 Mar 2023 29. Japanese Pharmacopoeia 18th Edition. https://www.pmda.go.jp/english/rs-sb-std/standardsdevelopment/jp/0029.html. Accessed: 29-Mar-2023 30. British Pharmacopoeia. https://www.pharmacopoeia.com/. Accessed 29 Mar 2023 31. 2020, Chinese Pharmacopoeia 2020 Edition: Key Points, Access Extra https://www.accestra. com/chinese-pharmacopoeia-2020-edition-key-points/. Accessed 10 Apr 2023 32. Bucerius, M., 2020, “Riding China’s Regulatory Rapids,” Med. Mak. [Online]. Available: https://themedicinemaker.com/business-regulation/riding-chinas-regulatory-rapids. Accessed 29-Mar-2023 33. Albanese, J.A., Wiggins, M.J.: A Case Study in Pharmacopoeia Compliance: Excipients and Raw Materials. BioPharm Int (2020) 34. Wiggins, J.M., Albanese, J.A.: Pharmacopoeia Compliance: Putting It All Together; What Is on the Horizon. 35. 2022, Thirteenth International Meeting of World Pharmacopoeias https://www.who.int/ news-room/events/detail/2022/09/27/default-calendar/thirteenth-international-meeting-ofworld-pharmacopoeias. Accessed 30 Mar 2023 36. Schniepp, S. J.: ALCOA+ and Data Integrity, Pharm. Technol. p. 77 37. Alosert, H., et al.: Data integrity within the biopharmaceutical sector in the era of industry 4.0. Biotechnol. J. 17(6), 1–8 (2022) 38. Markarian, J.: Tracking Raw Materials in Biopharma Manufacturing. Pharm. Technol., p. 54 (2019) 39. Pal, A., Kumar Rath, H., Shailendra, S., Bhattacharyya, A.: IoT Standardization: The Road Ahead, Internet of Things - Technology, Applications and Standardization, J. Sen, ed., BoD – Books on Demand, pp. 53–74 (2018) 40. Jain, T., Pandey, B.: Knowledge Management Implementation. BioPharm Int (2012) 41. 2009, Core Components Technical Specification CCTS, Version 3.0., p. 162 https://unece. org/DAM/cefact/codesfortrade/CCTS/CCTS-Version3.pdf. Accessed 10 Mar 2021 42. Novakovic, D.: Business Context Aware Core Components Modeling, Vienna University of Technology (2014) 43. Ivezic, N., Jelisic, E., Jankovic, M., Kulvatunyou, B., Kehagias, D., and Marjanovic, Z.: Advancing Data Exchange Standards for Interoperable Enterprise Networks (2022) 44. Jelisic, E., Ivezic, N., Kulvatunyou, B., Milosevic, P., Babarogic, S., Marjanovic, Z.: A novel business context-based approach for improved standards-based systems integration—a feasibility study. J. Ind. Inf. Integr., 30(June) (2022) 45. 2011, The Rules Governing Medicinal Products in the European Union 46. 2021, What Changes When a Drug Candidate Moves from the Research Stage to GMP Production – and Why?. Biovian https://biovian.com/news/the-whys-and-the-whats-of-gmpproduction-of-biopharmaceuticals/. Accessed: 07 Apr 2023 47. Quinn, K.S.: A Holistic Approach to Raw Material Risk Assessments through Industry Collaboration (2019) 48. 2021, Index of World Pharmacopoeias and Pharmacopoeial Authorities 49. Jelisic, E., Ivezic, N., Kulvatunyou, B., Nieman, S., Marjanovic, Z.: Business Context-Based Quality Measures for Data Exchange Standards Usage Specification, Springer Cham, Valencia

Volunteering Service Engineering in Non-profit Organizations Michael Freitag1(B)

and Oliver Hämmerle2

1 Fraunhofer Institute for Industrial Engineering IAO, Nobelstr. 12, Stuttgart 70569, Germany

[email protected]

2 Institut für Arbeitswirtschaft und Technologie Management IAT, University of Stuttgart,

Nobelstr. 12, Stuttgart 70569, Germany [email protected]

Abstract. Non-profit organizations are also affected by the digital transformation and should pursue it systematically. To this end, this paper presents and de-scribes a reference model for transformation. The “Volunteering Service Engineering” model focuses specifically on the use of digital tools with the strong involvement of volunteers. The reference model is based on the approaches of service engineering and design thinking and supports the collaborative, interactive work of volunteers. The process model was validated in a use case of volunteers at a German engineering company. This use case highlights the developed VDI app. Keywords: Service Engineering · Digital Transformation · Social Entrepreneurship · Public Innovation · Service Innovation

1 Introduction The impact of digital transformation on the work and organizational processes of the future is significant, affecting not only traditional production and service companies but also non-profit organizations such as associations, foundations, and unions. A number of studies [1–7] have been conducted on the status and development of digitalization in non-profit organizations in Germany, providing empirical data on the use of digital technologies and the need for change in terms of digital transformation. In the study “Digital Report 2020 - Non-Profits & IT” [2] more than 5,000 Non-ProfitOrganizations from Germany participated via online questionnaires between August 2019 and April 2020. The aim of the study is to make empirically reliable developments and trends visible about the degree of digitalization in civil society in Germany. The results make it clear: • The potential of digitization is not being fully exploited. • There are a lot of missed opportunities in NPOs. • The role of dialog with stakeholders is very important.

© IFIP International Federation for Information Processing 2023 Published by Springer Nature Switzerland AG 2023 E. Alfnes et al. (Eds.): APMS 2023, IFIP AICT 689, pp. 471–483, 2023. https://doi.org/10.1007/978-3-031-43662-8_34

472

M. Freitag and O. Hämmerle

The study “Digital Report 2020 - Non-Profits & IT” provides a general overview of the current state of digitization at NPOs in Germany as well as challenges and potentials based on examples. The three success factors are: • Evidence-based strategy - Collecting and analyzing data to test goal achievement, assess the quality of decisions, and enhance the evolution of offerings. • Innovative strength - Testing and openness to new ideas, seeing them not less as obstacles but as part of positive organizational development. • Stakeholder orientation - Alignment to communicate with external stakeholders (members, employees, donations, press, suppliers, universities, communities, influencers, etc.) and involve them in decision-making processes. This paper presents a reference model in order to use digital tools by volunteers in non-profit organizations [8] to address three success factors above. It is structured in the following chapters: • Literature analysis for social business models and social entrepreneurship • Developing of a reference model for “Service Engineering for Non-ProfitOrganizations” and • The use case “VDI Württembergischer Ingenieurverein” with an app for volunteers. • The paper concludes with a brief summary in Sect. 5.

2 Literature Analysis In this chapter a literature analysis for social business models and social entrepreneurship will be described. • Section. 2.1 describes a literature review on social business models. • Section 2.2 describes a literature review on social entrepreneurship. A search of the relevant literature via Google Scholar using the search terms “social business model” and “social entrepreneurship” led to an initial pre-selection of 60 papers. These were scanned based on the outline and abstracts. The 10 most relevant papers were selected based on the decision criteria “relation to NPO” and “practicality”. Only papers from the last 4 years were considered in the selection (2018–2022) to ensure that the approaches were up to date. The approaches are briefly described below as announced. 2.1 Selected Approaches for Social Business Model Table 1 lists the eight selected approaches, which are briefly described in order to derive the “volunteering service engineering” approach in Sect. 4. Kucher and Raible [9] define the term “social entrepreneurship” as “entrepreneurial activity that works for positive change in society, for the common good and the solution of social problems or for the environment. Companies are often organized on a non-profit basis, so that there is no profit motive in social entrepreneurship. The authors cite “social innovation”, "social enterprise” and “social business model” as synonyms. Kucher and Raible [9] take a look at different definitions of “social entrepreneurship” as a process:

Volunteering Service Engineering

473

Fig. 1. Procedure model of the literature analysis Table 1. Relevant approaches Year

Source

Description

2018

Dufft & Kreutter [1]

Digitization in Non-Profit Organizations: Strategy, Culture and Competencies in the Digital Transformation

2022

Kucher & Raible [9]

Social Entrepreneurship

2018

Olfe et al. [10]

Study about digitalization in NPOs

2019

Deman, R. [11]

Digital Transformation and Digital Maturity of Nonprofit Organizations Using the Example of Amnesty International Austria

2018

Beier, M. [12]

Digital strategies for Nonprofit Organizations in the early 21st century

2020

Schmid Foundation [13]

Canvas Business Model for Non-Profit Organizations

2018

Erpf, P., & Maring, N. C. [14]

Digitization as an opportunity for Nonprofit Organizations

2022

Bork & Tahmaz [15]

Actively shaping the digital transformation in civil society organizations - a guide

474

M. Freitag and O. Hämmerle

central to this is the "identification of a stable but inherently unjust equilibrium that causes the exclusion, marginalization, or suffering of segments of humanity that lack the financial resources or political influence to achieve transformative benefits on their own." The goal is to identify an opportunity in this unjust equilibrium and developing a social value proposition. The authors offer a comprehensive guide that explores how core elements contribute to the success or failure of a nonprofit organization. It uses a practical approach to analyze the key competencies needed to combine effective business practices with effective “social innovation.“ Olfe et al. [10] investigate the awareness of the need for change as well as the ability to change in the nonprofit sector with regard to digitization. To this end, more than 160 employees of nonprofit organizations were surveyed online and by telephone in summer 2017. The study shows what support nonprofits need in order to better assess the opportunities of digitization and use them positively for themselves. The results make it clear that digitization is both an opportunity and a challenge for nonprofits. Deman [11] structures the digital transformation of nonprofit organizations based on the 5 phases: Preparation/Basics, Analysis/Diagnosis, Planning/Conception, Implementation, and Control. First, he describes the creation of organizational and cultural prerequisites for the change process. The basis for this is an understanding of the initial situation. Basically, the author emphasizes three areas here: analysis of the strategy, analysis of the cultural system, analysis of the technical-organizational system. The conception phase is about the precise planning and control of the change process. Depending on the chosen method, precise goals are defined, an overall design is created and communication is planned accordingly. The next step is to implement the measures developed at the various levels. The author names the following as the most important prerequisites for successful digital transformation: training and participation, information and communication, commitment and targets/timelines, dealing with opponents and conflicts. Finally, the success of the digital transformation of a non-profit organization can only be determined by the sustainability of the results. The progress of the project must therefore be continuously tracked and the results achieved must be continuously improved. Beier [12] focuses on the 4 core questions in his remarks on digital strategies for non-profit organizations: • • • •

What do we want to achieve by using digital technologies? What do we need from whom? How do we reach the individual target groups in a targeted manner? How do we nurture our existing relationships?

In addition, an important part of the digital strategies of non-profit organizations is the “social business model”. The Schmid Foundation [13] has addressed this issue and adapted the Business Model Canvas specifically for non-profit organizations.

Volunteering Service Engineering

475

Erpf and Maring [14] highlight another important aspect of the digitization of nonprofit organizations by emphasizing the membership structure and, in particular, the importance of volunteers for NPOs. Here, the authors structure their approach on three levels: • Changes internal to the organization (micro level) • Changes in exchange with the organizational environment (meso level) • Organizational changes (macro level) Bork and Tahmaz [15] offer, among other things, a toolbox to support civil society organizations in their digital transformation with their guide “Actively shaping the digital transformation in civil society organizations” as part of the project “Those responsible #digital”. This offers design fields of digital engagement, digital strategy and shows the essential steps of the digitization process and gives a small insight into the mentioned toolbox. The work is based on the approach of Dufft et al. [1], which describes five fields of change in digitization: Strategy, Culture and mode of operation, Communication, Technology and Data, Organization and Processes. All the input, the ideas, the process steps and methods of the 8 relevant approaches are part of phases “Define strategic goals” and “Identify and examine digitalization potential” in the Fig. 1. 2.2 Selected Approaches for Social Entrepreneurship This chapter describes the two approaches of Meyer [16] and Yamakami [17, 18]. Meyer defines the goal of developing solutions and approaches to the design of the service system and the work of people within or with these systems. The approach is based on an example from the manufacturing industry, the need to set impulses for an interlinking of service engineering and work design as well as to point out resulting questions and possible starting points is to be clarified. With service orientation and digitalization, the tasks and activities of employees and users in service systems are changing. In order to design interaction relationships and work today and in the future in a way that is both economical and human-centered, taking into account technological and social developments, requirements must be identified and methodically integrated into the development of these systems. Digitalization has a significant impact on the design of interaction work. While the personnel effort can be reduced by technical support and optimization effects are possible, the individual degree of interaction increases at the same time. Yamakami [17] developed the approach “Servicenics theory”. That is an extension of Gamenics theory to facilitate service engineering [18]. The goal is to extend game design to the general principles of service engineering. Mobile social game design is based on the generally applicable design paradigms of service engineering. The author proposes a six-level view of the principles of service engineering complemented by the four-level golden rules of mobile social game design. All the input, the ideas, the process steps and methods of the two relevant approaches are part of phases “Develop smart service ideas for volunteers”, “Select digital tools and test prototypically” and “Implementation” in Fig. 2.

476

M. Freitag and O. Hämmerle

3 Volunteering Service Engineering The reference model “Service Engineering for Non-Profit-Organization” in Fig. 2 is a synthesis of the approaches described above. The focus of the synthesis is on the generation of new digital service offers and interface management between core staff and volunteers. The aim is to ensure the participatory involvement of volunteers in the digital transformation of NPOs. The reference model in Fig. 2 shows the key phases in introducing smart services for volunteers in non-profit organizations [19]. The reference model is a combination of the service engineering approach [20–24] and design thinking [25]. It focuses on the structured involvement of volunteers in the service development process based on the selected approaches in chapter 2.1 and 2.2. The knowledge and experience gained in service engineering can be transferred in simplified form, and adapted if required [26, 27].

Fig. 2. Reference model “Volunteering Service Engineering for Non-Profit-Organizations”

The reference model “Service Engineering for Non-Profit-Organization” in Fig. 2 consists of the five following phases: • • • • •

Define strategic goals, Identify and examine digitalization potential, Develop smart service ideas for volunteers, Select digital tools and test prototypically and Implementation.

The five phases are described in the following paragraphs. First three phases were discussed more in detail in a previous paper by Freitag and Hämmerle [28], here in this paper the first three phase are described in a briefly form. Define Strategic Goals. The first phase “Define strategic goals” is to define the strategic objective of the non-profit organization, for example, an increase in the target group of volunteers, engaging younger members or increasing digital support for volunteer work [1, 3, 5]. The methods “SWOT analysis”, “Stakeholder analysis” and “Pains & Gains analysis” are briefly described as examples:

Volunteering Service Engineering

477

• SWOT-Analysis. The analysis and assessment of the strengths, weaknesses, opportunities and potentials of non-profit organizations is a widely used method to present the current state. In particular, the aspects of strategy, the competencies of full-time and voluntary staff, and the IT infrastructure should be included. • Stakeholder Map. Visualization of all internal and external stakeholders of a non-profit organization and their wishes and relationships among each other. • Pains & Gains Analysis. The purpose of Pains & Gains is to collect and visually depict the difficulties and expectations of customers in connection with customer tasks. The representation serves the better understanding of the requirements of the target group for a non-profit organization. Identify and Examine Digitalization Potential. In the second step “Identify and examine digitalization potential” the potential for digitalization—for example, through a customer journey or a detailed process analysis—is identified and reviewed within the NPO. The “customer journey” describes the path for a potential volunteer takes through various “touchpoints” with a nonprofit organization until he or she takes a desired target action. This “journey” is visualized with a map. The touchpoints are important for this and must be defined in advance [25]. Develop Smart Service Ideas for Volunteers. Next, the ideas for Smart Services for Volunteers are brought together and reviewed to their suitability for the non-profit organization [2–4, 10, 14, 15]. This phase is based on the approach of design thinking [25]. The design thinking consists of the follow steps: • • • • • •

Empathize, Define, Ideate, Prototype, Test and Implement smart service ideas.

This phase is focusing on ideas not solutions. This is very important to understand this difference. Select Digital Tools and Test Prototypically. Based on the smart service ideas of the previous phase NPOs must look for the existing tools that fits to the generated ideas. NPOs don´t have time or money to develop an own tool. NPO should only use existing tools or a customizing solution of this tools. The best-known method in this phase is “Prototyping”. It refers to testing the selected tools before final release. Prototyping is especially important to see to what extent, for example, a software meets the needs of users. Prototyping thus makes ideas tangible. It is an experimental process to concepts with the needs of the customers. Another tool is the “Interaction training”. In a broader understanding, interaction refers to the mutual interaction and involvement of individual actors or systems, such as in interpersonal or human-technology interaction. Interaction thus goes beyond communication - it consists of situation-related actions and can only come about if the actors are

478

M. Freitag and O. Hämmerle

willing and able to do so. Interaction also refers to the interpersonal contact and exchange that is at the heart of person-centered service. It is the central moment in which the value of person-centered service is created. Implementation. In the final phase “Implementation” the selected ideas are implemented and tested as prototypes by using digital tools. This is done through subscriptions to online cloud services or the use of open-source tools, for example. Here, the selected tool is to scale, and ensure that all volunteers are qualified to use it. If the prototypical test run was successful, then the digital tool can be implemented in the organization. An important step would be the creation of a “guideline”. Such a guideline is a concise written description of how to carry out certain procedures or how to deal with certain situations, which helps people (e.g., service providers and service recipients) to carry out procedures properly or to deal with certain situations competently. Action guidelines refer to a limited subject area for which they provide orientation and show concrete options for action (e.g., dealing with complaints, behavior in hazardous situations). Guidelines for action can be provided by external bodies responsible for specific topics or they can be created internally within the company. Content marketing is also important in the launch. The term “content marketing” refers to a marketing or business process as part of a communications strategy in which relevant and valuable content is created that is aimed at a specific target group. The aim is to attract potential customers through this content and also to generate profitable actions. Throughout all 5 phases of the reference model in Fig. 2, the focus is on involving volunteers [2, 3, 10, 14, 15].

4 Use Case–Validation of the Reference Model The reference model “Service Engineering for Non-Profit-Organizations” in Fig. 2. Was validated for the use case “VDI Württembergischer Ingenieurverein”. The use case “VDI Württembergischer Ingenieurverein” is an association of engineers in the southwest of Germany. The association has more than 13,000 members. It includes all personal and supporting members of the VDI who have their place of residence or business in the district of southwest of Germany. “VDI Württembergischer Ingenieurverein” has 8 staff persons and over 2.000 members as volunteers in different technical working groups. Figure 3 shows the validation of the use case, just as the targets and methods in the different phases. Define Strategic Goals. During this phase strategic targets were defined as part of a workshop using the SWOT-Analysis. Two key targets emerged here: • Increasing the number of members, especially increasing appeal among young students and • Improvement and expansion of member offerings fostering, especially online collaboration and networking during the Covid-19 pandemic.

Volunteering Service Engineering

479

Fig. 3. Validation of the reference model by the use case VDI

Identify and Examine Digitalization Potential. Based on these targets an empirical survey (questionnaire) was conducted. The empirical survey consisted of a quantitative data collection and analysis of 30 questions, answered by 845 members of the use case VDI [29]. The data set of the members were: • • • •

7% of the respondents were students, 4% were employed for max. 4 years, 66% were employed for more than 4 years and 20% were no longer employed.

The needs and requirements of volunteer members could be identified by analyzing and assessing the data collected. For example, the following beneficial digital offers were identified [29]: • Design of hybrid committee meetings that combine face-to-face and live online meetings, • Self-services for volunteers in event planning and event promotion, • Possibility to connect interesting speakers for meetings and • Scheduling with digital calendar of events. For example, one question in the questionnaire was to develop an app in order to support hybrid meetings, planning and event promotion. This would give VDI members the opportunity to read contributions from other members and to publish their own contributions. Figure 4 shows how highly the respondents rated the benefits of such an app. As well as providing information on events that can be selected at any time, the app should also allow users to network with one another. In addition, push messages should be customizable. 46% of all VDI respondents rated the benefits of a VDI community app as high to very high. This proportion is particularly pronounced among students and young professionals at 64%, 44% of whom rated the benefit as very high.

480

M. Freitag and O. Hämmerle

Fig. 4. Survey results “Community app” [19]

Develop Smart Service Ideas for Volunteers. Based on the results of the questionnaire three different online workshops were taking part. Ten volunteers from the use case VDI were participated in the three workshops. Based on method Design Thinking the volunteers were looking for new features of the app in order to implement smart services in the use case VDI with the help of two moderators. Select Digital Tools and Test Prototypically. Several quotes were obtained from different providers for the development of the app. The app “VDI Connect” – a Screenshot is shown in Fig. 5 – was implemented on the basis of the requirements analysis. The app “VDI Connect” can be downloaded in the Apple AppStore or in the Android PlayStore. Up to now it is free of charge. Every person can download the app and test it even if the person is not a member of VDI. Implementation. As already described, the app is on the market and can be downloaded. “VDI Connect” currently offers VDI members, among other things, the option of a messenger service, an event calendar and current news. However, the app is still in the test phase and is to be supplemented with further features.

Volunteering Service Engineering

481

Fig. 5. VDI App

5 Summary In order to successfully navigate the digital transformation of non-profit organizations, it is crucial to have a strategy and reference model for transformation that can inform, mobilize and engage members effectively. This requires a targeted and accessible approach, with a strong emphasis on involving volunteers and utilizing digital tools. The presented reference model is based on service engineering and design thinking approaches, which aim to foster collaborative and interactive cooperation among volunteers. The reference model was tested and validated through the use case of an engineering association in the southwest of Germany. It was found that digitalization in the area of virtual volunteering has been extensively studied, and a survey conducted at “VDI Württembergischer Ingenieurverein” identified important needs for action. Among the wishes expressed by volunteers was an organization-specific app with events calendar, newsletter and news.

482

M. Freitag and O. Hämmerle

This was implemented so that a stronger involvement of young volunteers should also succeed.

References 1. Dufft, N., Kreutter, P.: Digitalisierung in Non-Profit-Organisationen: Strategie, Kultur und Kompetenzen im digitalen Wandel. In: Zukunftsorientiertes Stiftungsmanagement, pp. 105– 115. Springer Gabler, Wiesbaden (2018) 2. des Stiftens, H.: Digital-Report 2020, Non-Profits & IT, abgerufen am: 30.03.2022 unter (2020). https://www.hausdesstiftens.org/wp-content/uploads/Digital-Report-2020.pdf 3. Reiser, B.: Organisatoren des Gemeinwohls Über die Beteiligung Ehrenamtlicher am Organisationsgeschehen. BdW Blätter der Wohlfahrtspflege 164(2), 67–69 (2017) 4. Simsa, R., Meyer, M., Badelt, C. (eds.) Handbuch der Nonprofit-Organisation: Strukturen und Management, Schäffer-Poeschel (2013) 5. Eimhjellen, I., Steen-Johnsen, K., Folkestad, B., Ødegård, G.: Changing patterns of volunteering and participation. In: Scandinavian Civil Society and Social Transformations, pp. 25–65. Springer, Cham (2018). https://doi.org/10.1007/978-3-319-77264-6_2 6. IW Consult: Digitalisierung in NGOs. Eine Vermessung des Digitalisierungsstands von NGOs in Deutschland. Abgerufen am: 30.02.2022 unter (2018). https://www.iwconsult.de/filead min/user_upload/projekte/2018/Digital_Atlas/Digitalisierung_in_NGOs.pdf 7. Jabło´nski, A., Jabło´nski, M.: Social Business Models in the Digital Economy. Palgrave Macmillan (2020) 8. Freitag, M.: Digitale Transformation von Non-Profit-Organisationen – Status quo und Handlungsbedarfe. Fraunhofer IAO, Stuttgart (2021) 9. Kucher, J.H., Raible, S.E.: Social Entrepreneurship: A Practice-based Approach to Social Innovation. Edward Elgar Publishing (2022) 10. Olfe, F.: Digitalisierung von Non-Profit-Organisationen. Newsletter für Engagement und Partizipation in Deutschland (2018). https://www.b-b-e.de/fileadmin/Redaktion/05_Newsletter/ 01_BBE_Newsletter/2018/newsletter-13-olfe.pdf 11. Deman, R.: Digitale Transformation und digitaler Reifegrad von Nonprofit-Organisationen am Beispiel Amnesty International Österreich (2019) 12. Beier, M.: Digitale Strategien für Nonprofit-Organisationen Anfang des 21. Jahrhunderts (Digital Strategies for Nonprofit Organizations at the Beginning of the 21st Century), Beier, Michael (2018). Digitale Strategien für Nonprofit-Organisationen Anfang des, vol. 21, pp. 101–118 (2018) 13. Stiftung, S.: Canvas Business Modell für Non-Profit-Organisationen (2021). https://bib liothek.isb-w.eu/alfresco/d/d/workspace/SpacesStore/7567bde6-5984-4a3d-b1bf-df0be714c bef/Business%20Canvas%20Model_NPOs_Text.pdf 14. Erpf, P., Maring, N.C.: Digitalisierung als Chance für Nonprofit-Organisationen. VerbandsManagement 44, 6–13 (2018) 15. Bork, M; Tahmaz, B.: Den digitalen Wandel in zivilgesellschaftlichen Organisationen aktiv mitgestalten – ein Leitfaden (2021). https://www.ziviz.de/sites/ziv/files/leitfaden_digitaler_ wandel_in_zivilgesellschaftlichen_organisationen.pdf 16. Meyer, K.: Vom Service Engineering zum Social Service Engineering–Anforderungen an die Schnittstelle zwischen Dienstleistungsentwicklung und Arbeitswissenschaft. Zeitschrift für Arbeitswissenschaft 74(1), 52–58 (2020) 17. Yamakami, T.: Servicenics approach: a social service engineering framework. In: Eighth International Conference on Digital Information Management (ICDIM 2013), pp. 358–362. IEEE (2013)

Volunteering Service Engineering

483

18. Yamakami, T.: A Four-Stage Gate-Keeper Model of Social Service Engineering: Lessons from Golden Rules of Mobile Social Game Design 9th, International Conference on Autonomic and Trusted Computing, S. 160 (2012) 19. Falkner, J., Kett, H., Castor J.: Ein Leitfaden – Smarte Produkte und Dienstleistungen. Wie Sie als Entscheider Schritt für Schritt ins Thema einsteigen. Mittelstand 4.0 Kompetenzzentrum, Fraunhofer Verlag, Stuttgart (2020) 20. Freitag, M., Schiller, C.: Approach to test a product-service system during service engineering. Procedia CIRP 64, 336–339 (2017) 21. Freitag, M., Westner, P., Schiller, C., Nunez, M.J., Gigante, F., Berbegal, S.: Agile productservice design with VR-technology: a use case in the furniture industry. Procedia CIRP 73, 114–119 (2018) 22. Freitag, M.; Hämmerle, O.: Agile guideline for development of smart services in manufacturing enterprises with support of artificial intelligence. In: IFIP International Conference on Advances in Production Management Systems. Springer, Cham, 645–652 (2020) 23. Freitag, M., Schiller, C., Hämmerle, O.: Guideline to develop smart service business models for small and medium sized enterprises. In: IFIP International Conference on Advances in Production Management Systems, pp. 369–375. Springer, Cham, (2021). https://doi.org/10. 1007/978-3-030-57993-7_73 24. Pezzotta G., Pirola F., Pinto R., Akasaka F., Shimomura Y.: A service engineering framework to design and assess an integrated product- service. Mechatronics, 169–179 (2015) 25. Plattner, H., Meinel, C., Weinberg, U.: Design-thinking. Mi-Fachverlag, Landsberg am Lech (2009) 26. Qu M., Yu S., Chen D., Chu J., Tian B.: State-of-the-art of design, evaluation, and operation methodologies in product service systems. Comput. Industry (2016) 27. Ojasalo, J., Ojasalo, K.: Using service logic business model canvas in lean service development. In: Gummesson, E, Mele, C, Polese, F. (eds.) Service Dominant Logic, Network and Systems Theory and Service Science: Integrating three Perspectives for a New Service Agenda, Naples: Conference Proceedings 5th Naples Forum on Service (2016) 28. Freitag, M., Haemmerle, O.: Digitalization of services for volunteers in non-profit organization. In: Advances in Production Management Systems. Smart Manufacturing and Logistics Systems: Turning Ideas into Action: IFIP WG 5.7 International Conference, APMS 2022, Gyeongju, South Korea, 25–29 September 2022, Proceedings, Part I, pp. 329–334. Springer Nature Switzerland, Cham (2022). https://doi.org/10.1007/978-3-031-16407-1_39 29. Gutmann, O., Schäfer, P.-M.: Auf dem Weg in eine digitale Zukunft – Erfahrungen des VDI Württembergischer Ingenieurverein e.V. In: Studie Digitale Transformation von Non-ProfitOrganisationen – Status quo und Handlungsbedarfe, Stuttgart, Fraunhofer IAO, pp. 17–26 (2021)

Workforce Evolutionary Pathways in Smart Manufacturing Systems

The Role of Organizational Culture in the Transformation to Industry 4.0 Rogerio Queiroz de Camargo , Márcia Terra da Silva(B) , Ana Lucia Figueiredo Facin , and Rodrigo Franco Gonçalves Universidade Paulista, São Paulo, SP 04036-002, Brazil [email protected]

Abstract. Industry 4.0 is based on implementing new technologies to make companies globally competitive, but, as is known, technological change generates organizational change challenges. Given this scenario, this study aims to identify the organizational culture characteristics aligned with Industry 4.0. The method used is a systematic literature review in a search in the Web of Science and SCOPUS databases, which resulted in 26 articles after exclusions, read in full. As a result, the relevance of organizational culture is confirmed, and the organizations’ characteristics most cited are continuous learning to adapt to innovations, combined with sharing and transferring knowledge. Other cultural characteristics are related to entrepreneurship to accept and face risks and uncertainties, based on flexibility, creativity, and innovation. Also considered crucial for Industry 4.0 were social responsibility and the skills of employees necessary for human beings to play a central role in receiving technologies and serving as a guide for technological progress in organizations. However, although its relevance to academia and organizations, the theme is not exhausted due to its complexity. So, new searches on change management and the investigation of cases study are suggested to deepen knowledge of the organizations’ challenge to get Industry 4.0 in commission. Keywords: Organizational Culture · Adhocratic Culture · Learning Organization · Industry 4.0

1 Introduction We face changes comprising culture, society, and the economy, while disruptive technological transformations also impact manufacturing. In these, the technologies that support Industry 4.0 make it possible to increase levels of flexibility and efficiency in production but require adaptation on the human and social side in companies, not only to learn to deal with new processes and equipment but to create solutions for new problems, resulting from these innovations. It is not news that, from the organizational and management point of view, implementing new technologies generates a series of uncertainties, as it impacts production, structure, and organizational culture (OC). Among the challenges that this transition © IFIP International Federation for Information Processing 2023 Published by Springer Nature Switzerland AG 2023 E. Alfnes et al. (Eds.): APMS 2023, IFIP AICT 689, pp. 487–500, 2023. https://doi.org/10.1007/978-3-031-43662-8_35

488

R. Q. de Camargo et al.

brings, one can mention the commitment to frequent innovation and the development of learning capacity and knowledge transfer [1], in addition to the fear of losing jobs and resistance to change [2]. Furthermore, it should be noted that these challenges are interrelated and can be mutually reinforcing, with OC being a central aspect for facing the barriers or exacerbating the problems, depending on the configuration it assumes [3]. Santos and Martinho [3] consider that OC can be one of the facilitating elements for the transformation process for Industry 4.0 if it presents traits related to adaptability and flexibility. For these authors, a culture focused on adaptability would guide employee behaviors that facilitate the acceptance and even the proposition of innovations. However, this organizational structure of companies focused on innovation and capable of dealing with constant changes is not new. It was already widely defended as an Adhocratic or Innovative Organization in the second half of the 20th century [4]. It gained momentum with the idea of an organization that learns or Learning Organization. In 2008, Garvin [5] consolidated practices that guide the company towards a learning organization, encompassing the role of the leader, middle management, processes and technique es, and the Learning Culture, that is, the environment, behavior, and values that encourage learning in the company. Moreover, later articles, building on the notion of learning organization and learning culture, refer to the importance of the concept to understanding social issues related to emerging technology adoption [6]. The organization’s learning capacity is also one of the items addressed in the maturity models and roadmaps for the transition to Industry 4.0 created by professional associations, such as ACATECH (German National Academy of Science and Engineering). For this institution, accepting change is seen as being open to innovations, adapting behavior, and using open communication, so that the company can react to unexpected events; for this, “the employees should ideally have instant access to the necessary explicit and implicit knowledge” [7]. The mention of implicit and explicit knowledge refers to the area of knowledge management and to articles that point out the importance of knowledge management to leverage innovation and, simultaneously, how digital technologies can facilitate the transfer of knowledge [8]. However, although there is already important literature, the appropriate OC is not described in detail, leaving space for divergence and doubts. Therefore, it is necessary to advance and consolidate the understanding of how Industry 4.0 can impact the OC of companies and vice versa, identifying the cultural characteristics that support this movement. Thus, this investigation applies the Systematic Literature Review method (SLR) to identify the cultural characteristics most aligned with the organizational change process necessary for implementing Industry 4.0. The article is structured as follows: after this introduction, there is a theoretical foundation section presenting classic concepts of Industry 4.0 and Organizational Culture that help to analyze the results, the third section describes the application of the Systematic Literature Review methodology (SLR), and in Sect. 4 the results are presented and discussed. Finally, the conclusion presents the main contributions and suggestions for future work.

The Role of Organizational Culture in the Transformation to Industry 4.0

489

2 Theoretical Background 2.1 Organizational Culture OC is universally present in organizations, influencing all its aspects [9]. Schein [9] defines OC as “a set of basic assumptions developed, created or discovered by a group, through a learning process, which aims to promote adaptation to the internal and external demands of a specific organization.“ Once stabilized, the culture will be taught and shared with new participants/entrants as the correct way to perceive, think, and feel in an organization [9]. In organizations, OC influences strategy, goals, future vision, and technological choices. For Schein [9], OC comprises three interdependent levels: on the surface, artifacts that are easily observable but difficult to interpret. A second level comprehends shared and conscious values, declared but which do not always underlie people’s actions. Finally, at a deeper level, Schein [9] points out that basic assumptions guide the group’s behavior in order to make communication easier and integrate actions and decisions. To effectively unveil a group’s OC, it is necessary to understand these assumptions - introjected and not made explicit, indicating the difficulty of unveiling the characteristics of the culture and managing a cultural change. According to Cameron and Quinn [10], “an organizational culture is composed of values, leadership styles, language, symbols, routines and practices that differentiate an organization.“ For them, OC can be created by the founder, formed over time, and also consciously developed, being a key factor in long-term effectiveness. In this way, the possibility of carrying out a conscious movement of cultural change opens up, identifying the company’s current culture with its respective characteristics and proposing how the future should be. For several authors, this OC management is based on the conception of types of culture, an analysis model according to which organizational cultures can be understood from assumptions such as leadership styles, the definition of success, and business relationships, among others [10, 11, and 12]. Wallach [11], for example, proposed a supportive, bureaucratic, and innovative culture. Schneider [12], one of the first to propose a typology, presented four types of culture: control, collaboration, competence, and cultivation. A recent typology that typifies the culture of innovation more clearly is that of Cameron and Quinn [10]. In this typology, the authors distinguish four types of culture: clan, in which success is perceived as the result of human development and group cohesion; adhocracy, where people consider effectiveness is reached through innovation; market, which aggressively competes for market share; and hierarchy, in which one believes that only control and efficiency are capable of producing effectiveness. For a given company, when assessing how it fits into each assumption, one can visualize a type of culture that best describes it.

490

R. Q. de Camargo et al.

Figure 1 presents, in a structured and simplified way, the characteristics of Cameron and Quinn’s four types of culture [10]. Notably, the axes of the figure have specific foci, which for the authors, are the basis for differentiating the types of culture. So, the vertical axis points up and down to the main capabilities the organization strives to gain and maintain: upward, flexible production and the capacity to choose between production alternatives; downward, a stable position in the market and control over the environment and resources. The horizontal axis points left to the focus on internal resources integration and right to the external market, seeking differentiation from competitors. This way, each culture type is the arrangement of constructs that relate to and reinforce the others: • Orientation reveals the strategic competence stakeholders perceive as being inherent to the company. • Leadership type refers to the beliefs, values, and behavior of the central leader; • Value drivers point out the vision shared among employees and leaders.

Fig. 1. The four types of Organizational Culture. Adapted from Cameron and Quinn 2006

The Role of Organizational Culture in the Transformation to Industry 4.0

491

• The aggregator factor is the glue that holds the group together. For instance, feeling committed to innovation is essential for a person to sense belonging to an Adhocratic organization. • Theory of effectiveness is how they explain to themselves success and failure. 2.2 Industry 4.0 and Organizational Culture For Zhong, Xu, Klotz, and Newman [13], the availability of digital technology can increase the flexibility of industries, mass customization, improve quality and productivity, and reduce time to market, giving rise to what is called Industry 4.0. The analysis of the Industry 4.0 implementation process has been the subject of several academic articles, many of which propose maturity models to assess the readiness of companies or their positioning in the transformation to Industry 4.0. It is noteworthy that maturity models emphasize OC as an issue to be faced by companies, as can be seen in the work of Simetinger and Zhang [14], who analyzed and compared 17 maturity models, pointing out that the Culture theme was identified in 10 of the analyzed models. Likewise, Santos and Martinho [3] analyzed in depth three maturity models, showing that the dimensions used by the different models are related not only to technological aspects but also to organizational ones so that the roadmap for the implementation of Industry 4.0 include issues related to leadership, culture, and human resources. As for OC, ACATECH’s Industrie 4.0 Maturity Index [7] presents two dimensions that must be followed to improve suitability for Industry 4.0: Willingness to change and Social collaboration. Each one comprises a series of capabilities as follows. • Willingness to change • • • • •

Recognize the value of mistakes Openness to innovation Data-based learning and decision making Continuous professional development Shaping change

• Social collaboration • Democratic leadership style • Open communication • Confidence in processes and information systems Comparing the capabilities of this model with the Cameron and Quinn culture types [10] in Fig. 1, one can identify some similarities with adhocracy, as in Table 1 ahead.

492

R. Q. de Camargo et al. Table 1. Comparison between Adhocracy Culture Type [10] and acatech Study [7]

Adhocracy (Cameron & Quinn) [10]

acatech STUDY (Schuh et al.) [7]

The work environment is dynamic, entrepreneurial, and creative The employees are encouraged to take risks

Recognize the value of mistakes

Leadership is entrepreneurial and innovative

Democratic leadership style

The organization’s glue is its commitment to innovation and development

Openness to innovation

The organization values new challenges, prospecting opportunities, and trying new things

Shaping change

Success means having the most unique or newest products and being ahead of their time Data-based learning and decision making Continuous professional development Open communication Confidence in processes and information systems

3 Methods This research seeks to answer the question: What characteristics of Organizational Culture are most aligned with the organizational change process for Industry 4.0? For that, we applied a systematic literature review (SLR) based on the work of Okoli [15] and Petticrew and Roberts [16]. Thus, the protocol for this review, which can be seen in Table 2, is based on Okoli’s Systematic Guide for the Development of Literature Review [15]. Applying the search criteria in March 2023 resulted in 93 articles, 41 from Scopus Elsevier and 52 from the Web of Science. Of these, 29 were excluded for being duplicated, 10 for lack of access, and 28 for analyzing non-industrial sectors or for not focusing on Industry 4.0 or organizational culture. The remaining 28 articles were entirely read with the research question and the Organizational Culture and Industry 4.0 framework in mind. Of these, 2 were discarded for only marginally mentioning the Organizational Culture theme. Then, the 26 articles were organized, the perspectives on the characteristics of the Organizational Culture and its relations with Industry 4.0 were highlighted, and the different views were compared. As a result, perspectives on OC and its relationship to Industry 4.0 emerge more clearly.

The Role of Organizational Culture in the Transformation to Industry 4.0

493

Table 2. Search, inclusion, and exclusion criteria for the SLR. Search Criteria 1

Databases

Choose databases: Scopus Elsevier and Web of Science. These databases were defined because they contain scientific articles considered “state of the art” due to their global scope, ease of access, and volume of articles published in English

2

Search of strings

Choose keywords: the boolean string used is “Industry 4.0” OR “Fourth industrial revolution” OR “Smart Factory” AND “Organizational Culture”

3

Search fields

Title, Abstract, and Keywords

Inclusion Criteria 1

File Type

Include only complete/finished articles in journals (excluding abstracts, book chapters, events proceedings, editorials, patents, etc.)

2

Language

Include only articles in English

3

Period

Include only studies published from January 2012 to February 2023

Exclusion Criteria 1

Duplication

Exclude duplicate articles in both databases (Scopus Elsevier and Web of Science)

2

Access

Exclude inaccessible articles

3

Focus on non-industrial sectors

Exclude articles from other areas such as services or agriculture, in the medical field, Human Resources, Accounting or other sectors

4

Different focus of the theme Organizational Culture in Industry 4.0

Exclude articles centered on other subjects that mention Organizational Culture peripherally, such as Cybersecurity or other focuses

4 Results and Discussion The 26 selected articles are listed in Table 3. The complete reference of each one can be found in the item References. Reading the articles revealed the importance of the interrelation between organizational culture, industry 4.0, knowledge, and innovation, as OC is a precondition for the effectiveness of knowledge transfer, which is fundamental for the rate of innovation in Industry 4.0. The learning culture was the most mentioned, referring to continuous learning, practical learning, knowledge sharing, and knowledge transfer, among other

494

R. Q. de Camargo et al. Table 3. List of articles selected in the Systematic Literature Review

N

Theme/Focus

Autor(es)

Ano

1

The barriers to data-driven Omar, Minoufekr & Plapper [17] decision-making include the lack of technological competencies, inadequate culture, and the data monetization model

2019

2

Barriers to implementing Industry 4.0 practices and the need for an organizational culture focused on innovation and continuous development

2021

3

Analysis of the factors that influence the Sansabas-Villalpando et al. [19] implementation of the innovation culture

2019

4

OCTAPACE profiling of the organizational culture of organizations in Serbia and California

Draškovi´c et al. [20]

2020

5

Impact factors on organizational innovation in the industrial areas

Fan et al. [21]

2021

6

Need for employee training, support from top management, and organizational culture

Gupta, Singh & Gupta [22]

2022

7

Implementing innovative technologies requires an organizational culture focused on continuous development

Bizubac & Hoermann [23]

2021

8

The culture of innovation must be Batz, Kunath & Winkler [24] oriented towards acquiring, assimilating, transforming, and exploiting knowledge

218

9

A culture of learning is the key to facing Ivaldi et al. [25] the changes of Industry 4.0 and putting human beings at the center

2022

10

Organizational culture is one of the items that support the transformation of Industry 4.0

Kohnová, Papula & Salajová [26]

2019

11

Industry 4.0 requires a corporate culture that values entrepreneurship spirit

Veile et al. [27]

2020

12

Industry 4.0 requires continuous Mohelska & Sokolova [28] innovation and education that depend on people’s skills and organizational culture

2018

13

The relationship between innovation, organizational culture and knowledge sharing

2020

Terra, Berssaneti & Quintanilha [18]

Michna & Kmieciak [29]

(continued)

The Role of Organizational Culture in the Transformation to Industry 4.0

495

Table 3. (continued) N

Theme/Focus

Autor(es)

14

Industry 4.0 suggests an OC model with Zakharova et al. [30] components from the four types of culture: hierarchical, market-oriented, adhocracy, and clan

2020

15

The Organizacional Learning Culture and work engagement in Industry 4.0

2021

16

Innovative culture facilitates the sharing Ziaei Nafchi & Mohelská [32] of knowledge necessary for Industry 4.0

2020

17

The appropriate organizational culture enables trust in knowledge sharing necessary for Industry 4.0

Pietruszka-Ortyl et al. [33]

2021

18

Industry 4.0 relies on leadership, organizational culture, and employee feedback on technologies

Liu et al. [34]

2022

19

Quality Culture as a part of Organizational Culture

Durana et al. [35]

2019

20

Organizational culture is a strong influencer in developing the strategy for Industry 4.0

Sivarethinamohan et al. [36]

2021

21

Social culture influences the exchange of Aguilar Rodríguez et al. [37] knowledge, an important factor for the success of Industry 4.0

2021

22

Barriers to implementing Industry 4.0: lack of leadership skills, fragile organizational cultures, lack of digital awareness, and lack of awareness of the benefits

Mukhuty, Upadhyay & Rothwell [38]

2022

23

An adaptive culture, learning, and competence provide means for implementing Industry 4.0

Carlsson et al. [39]

2022

24

Smart factories imply disruptive technologies that will change business processes, the value chain, organizational culture, and human resource policies

Roblek et al. [40]

2021

25

Industry 4.0 demands an organizational culture that encourages creativity

Stacho et al. [41]

2021

26

Organizational cultures oriented toward results and analysis of data and facts facilitate the operation of Industry 4.0

Cadden et al. [42]

2022

Urrutia Pereira et al. [31]

Ano

496

R. Q. de Camargo et al.

topics (Articles 2, 4, 5, 8, 13, 16, 17, and 24 of Table 3). For example, article 9 points out that the continually transforming environment has some organizational implications, such as the necessity of conceiving learning paths for the whole company. This way, a learning culture should be nurtured so that knowledge transfer can breed. The authors conclude that some aspects of the organization’s culture are adaptability, meaning the capability to modify a project during its life cycle; visibility of the process to multiple stakeholders; value generation, considering not only economic but also sustainable and ethical dimensions; and facing risks and uncertainty [25]. Cameron and Quinn’s Adhocracy [10] was also cited, a type of culture aimed at innovative organizations and entrepreneurial leaders that fits the characteristics of Industry 4.0. We identified that other articles that point to the importance of learning culture for Industry 4.0 also emphasize that social responsibility and employee skills are crucial to Industry 4.0 (Articles 9, 11, 20, 23, and 24 of Table 3). Regarding social responsibility, there is a need to place the human being at the center, as the one who will receive the technology and act as a guide for technological progress in the organization. The culture is described as open-minded and flexible, adequate to interdisciplinary thinking and entrepreneurship spirit (Articles 11, 18, and 13). In addition, some articles point out that it is necessary to recognize employees’ resistance and seek to gain their trust in the transformation process, helping them overcome insecurity with a socially responsible orientation (Articles 7 and 23 of Table 3). A theme that is very present among these articles seeks to identify which management strategies are the most effective for the development of an OC that supports an environment conducive to continuous education, to learn and adapt to the innovation resulting from Industry 4.0 (Articles 10, 12, 15, 19, 22 and 25 of Table 3). All the articles above emphasize the innovative aspect necessary for the company to improve the process or create new products and point to a type of culture that highlights some elements present in the Adhocracy Culture [10], in the ACATECH study [7] or Learning Culture [5 and 6]. On the one hand, the entrepreneurial spirit for the management and leadership of the Adhocracy Culture, and on the other hand, the appreciation and encouragement of constant learning from the Learning Culture complement each other in the Organizational Culture panorama. However, several complementary aspects are highlighted. Article 1 [17], for example, describes OC’s properties, emphasizing that it requires employee skills and computational infrastructure. A significant challenge to this kind of organization is a departmentalized structure, which can hinder the information sharing essential to evidence-based problem recognition and solution. According to Omar et al. [17], the Organizational Culture stresses data-based decision-making, continuous professional development, open communication, and confidence in processes and information systems. The analysis is similar to Schuh [7] and refers to the necessary competence for data processing. Finally, some articles focus on regional contexts, comparing cultures across countries. Article 14, for example, compares the contexts of three Asian countries – Russia, China, and Iran – and analyzes cultures according to Cameron and Quinn’s typology. Article 4 compares cultural traits more present in two cultures – Serbia and California – and argues that for Industry 4.0, it is essential to have creative human capital. Finally,

The Role of Organizational Culture in the Transformation to Industry 4.0

497

Article 10 compares five European countries, focusing the analysis on Human Resources training. Some articles used a different (but not contradictory) OC framework than those described above. Article 3, for example, does not fit Adhocracy or Learning Culture since, based on the Literature Review, it proposes a specific type of OC for Industry 4.0, emphasizing that innovation is crucial. The article explains the culture for Industry 4.0, detailing six areas of management: Financial, Organizational, Technology, Intellectual property, Process, and Knowledge, and concludes that organizational management is the most relevant area to raise innovation. Also, Article 19 presents the concept of quality culture and does not fit into any group. Consolidating the topics addressed for the OC and comparing them with Table 1, one can see the complementarity of views so that each SLR author sheds light on an aspect addressed by the classics and intersects with other aspects cited by the other. However, the articles in the systematic literature review favor the Learning Organization view, which, although it articulates well with Cameron and Quinn [10] and Schuh et al. [7] (see Table 1), emphasizes aspects more directly related to the learning process. For example, for Cameron and Quinn, adhocratic leadership is entrepreneurial; for Schuh, it is democratic; and for Garvin [5] reinforces learning. Moreover, for Learning Organization authors, the work process is a critical locus of learning [43], so it should be central when discussing organizational culture. Egan [43], for instance, shows the relevance of Job Satisfaction to motivation to transfer learning. Likewise, Garvin [5] points out the centrality of learning from the own experience and remarks that it demands time. For the author, everyone should be allowed time to think, reflect and analyze past actions and consequences. Unfortunately, this possibility does not seem to fit the accelerated pace of Industry 4.0. However, it is hard to confirm since there is little data on that since only one article mentioned the work organization in its analysis.

5 Conclusion Using the Systematic Literature Review, it was possible to answer the initial research question and identify the characteristics of the Organizational Culture more aligned with the principles of Industry 4.0. Moreover, the adopted reference based on two classic authors [5 and 10] and, compared with the vision of a Maturity Model aimed at practical use by industries [7], proved to be quite comprehensive and assertive to encompass the analyzed authors with rare exceptions. The ability to deal with innovation, whether in terms of products or improvements in the production process, stands almost unanimously in the literature and can be identified as a decisive aspect of the required cultural characteristics. This aspect leads to the Adhocratic culture type [10] and fits the ACATECH study [7]. On the other hand, most authors also state that it is necessary to renew workers’ skills, whether at the operational level or in the technical staff. This need seems not to be concentrated in the moment of transition, in which the technologies not yet mastered require the retraining of operators. Indeed, it extends over a long period, predicting that this transition and technological change will continue. Thus, one innovation will follow another, and in addition to the adhocratic culture that privileges creativity and the ability

498

R. Q. de Camargo et al.

to face risks, this organizational culture also emphasizes the capacity for continuous learning, which justifies the need for learning paths, knowledge transfer mechanisms, and a spirit for information and knowledge sharing. As a main contribution to the academy, one can mention the merging of the concepts of organizational culture, consolidating and relating the characteristics presented by several authors. Furthermore, identifying cultural characteristics contributes to industry managers, as it provides a way to assess the company’s current culture and suggests ways to manage organizational change as a basis for the technological, structural, and organizational changes inherent in the Industry 4.0 transformation process. However, culture is a very abstract theme that involves different definitions; thus, a limitation of the work was the analysis based on authors who privilege the concept to the detriment of measurement, which reduced the relative presence of valuable articles that used a form of measurement of Organizational Culture. Moreover, Industry 4.0 is a recent movement that proposes a revolutionary and advanced process with a broad application area, dealing with product development, production process, logistics, and marketing; yet, the Theoretical Background in this study relies on classical concepts, as it also seems to be the case of several articles read. Thus, the future challenge remains to find the conditions for people and organizations to act in an integrated and efficient way along the transformation process inherent to Industry 4.0 and investigate if a new set of cultural types can contribute to the Cameron and Quinn one [10]. Finally, this study showed the need to consider long-term organizational change management in implementing Industry 4.0, as culture evolves, is subject to internal and external efforts, and employees need to adapt to each change. Therefore, investigating barriers and facilitators to the organizational change management process could be a topic for future research. Moreover, few SLR authors mentioned work organization beyond employee competencies, as said in the previous item, and it can be the subject of a case study on work organization, to investigate how workers deal with learning from practical experiences. Acknowledgment. This study was financed in part by the Coordenação de Aperfeiçoamento de Pessoal de Nível Superior - Brasil (CAPES) - Finance Code 001.

References 1. Nosalska, K., Pi˛atek, Z.M., Mazurek, G., Rz˛adca, R.: Industry 4.0: coherent definition framework with technological and organizational interdependencies. J. Manuf. Technol. Manag. 31(5), 837–862 (2019) 2. Bentley, T., Green, N., Tappin, D., Haslam, R.: State of science: the future of work–ergonomics and human factors contributions to the field. Ergonomics 64(4), 427–439 (2021) 3. Santos, R.C., Martinho, J.L.: An Industry 4.0 maturity model proposal. J. Manuf. Technol. Manag. 31(5), 1023–1043 (2020) 4. Mintzberg, H.: Structure in 5’s: a synthesis of the research on organization design. Manage. Sci. 26(3), 322–341 (1980) 5. Garvin, D.A.: Learning in action: a guide to putting the learning organization to work. Harvard Business Review Press (2003), Garvin, D.A., Edmondson, A.C., Gino, F.: Is yours a learning organization?. Harvard Bus. Rev. 86(3), 109 (2008) 6. Song, J.H., Chermack, T.J.: A theoretical approach to the organizational knowledge formation process: Integrating the concepts of individual learning and learning organization culture. Hum. Resour. Dev. Rev. 7(4), 424–442 (2008)

The Role of Organizational Culture in the Transformation to Industry 4.0

499

7. Schuh, G., Anderl, R., Gausemeier, J., Hompel, M., Wahlster, W.: Industrie 4.0 maturity index. managing the digital transformation of companies (acatech STUDY). Herbert Utz Verlag, Munich (2017) 8. De Bem Machado, A., Secinaro, S., Calandra, D., Lanzalonga, F.: Knowledge management and digital transformation for Industry 4.0: a structured literature review. Knowl. Manag. Res. Practice 20(2), 320–338 (2022) 9. Schein, E.H.: Organizational Culture and Leadership, 5th edn. Wiley & Sons, Atlas (2017) 10. Cameron, K.S., Quinn, R.E.: Diagnosing and changing organizational culture: based on the competing values framework. Jossey-Bass, San Francisco, CA (2006) 11. Wallach, E.J.: Individuals and organizations: The cultural match. Train. Dev. J. 37(2), 29–36 (1983) 12. Schneider, W.E.: The reengineering alternative: A plan for making your current culture work. Professional Publishing, Inc., Burr Ridge, IL (1994) 13. Zhong, R.Y., Xu, X., Klotz, E., Newman, S.T.: Intelligent manufacturing in the context of industry 4.0: a review. Engineering 3(5), 616–630 (2017) 14. Simetinger, F., & Zhang, Z.: Deriving secondary traits of industry 4.0: a comparative analysis of significant maturity models. Systems Res. Behav. Sci. 37(4), 663–678 (2020) 15. Okoli, C.: A guide to conducting a standalone systematic literature review. Commun. Assoc. Inf. Syst. 37(1), 43 (2015) 16. Petticrew, M., Roberts, H.: Systematic reviews in the social sciences: a practical guide. John Wiley & Sons (2008) 17. Omar, Y.M., Minoufekr, M., Plapper, P.: Business analytics in manufacturing: Current trends, challenges and pathway to market leadership. Operations Research Perspectives 6, 100127 (2019) 18. Terra, J. D., Berssaneti, F.T., Quintanilha, J.A.: Challenges and barriers to connecting world class manufacturing and continuous improvement processes to Industry 4.0 paradigms. Eng. Manag. Production Serv. 13(4), 115–130 (2021) 19. Sansabas-Villalpando, V., Pérez-Olguín, I.J.C., Pérez-Domínguez, L.A., Rodríguez-Picón, L.A., Mendez-González, L.C.: CODAS HFLTS method to appraise organizational culture of innovation and complex technological changes environments. Sustainability 11(24), 7045 (2019) ´ c, Ð, Petrov, V., Uzelac, Z.: Comparison of organizational cultures from 20. Draškovi´c, Z., Celi´ a transitional economy and a knowledge economy: empirical study from Serbia and Southern California. Strategic Manag. Inter. J. Strategic Manag. Decision Support Syst. Strategic Manag. 25(4), 17–23 (2020) 21. Fan, Y.J., Liu, S.F., Luh, D.B., Teng, P.S.: Corporate sustainability: Impact factors on organizational innovation in the industrial area. Sustainability 13(4), 1979 (2021) 22. Gupta, A., Singh, R.K., Gupta, S.: Developing human resource for the digitization of logistics operations: readiness index framework. Int. J. Manpow. 43(2), 355–379 (2022) 23. Bizubac, D., Hoermann, B.O.: Digital disruptive innovation effects in the manufacturing industry. Revue Roumaine Des Sciences Techniques—Série Électrotechnique Et Énergétique 66(1), 41–46 (2021) 24. Batz, A., Kunath, M., Winkler, H.: Discrepancies between cluster services and SMEs’ needs constraining the creation of a culture of innovation amidst Industry 4.0. LogForum 14(3), 387–405 (2018) 25. Ivaldi, S., Scaratti, G., Fregnan, E.: Dwelling within the fourth industrial revolution: organizational learning for new competences, processes and work cultures. J. Work. Learn. 34(1), 1–26 (2022) 26. Kohnová, L., Papula, J., Salajová, N.: Internal factors supporting business and technological transformation in the context of Industry 4.0. Verslas: Teorija ir praktika. Bus. Theory Practice 20, 137–145 (2019)

500

R. Q. de Camargo et al.

27. Veile, J.W., Kiel, D., Müller, J.M., Voigt, K.I.: Lessons learned from Industry 4.0 implementation in the German manufacturing industry. J. Manuf. Technol. Manag. 31(5), 977–997 (2020) 28. Mohelska, H., & Sokolova, M.: Management approaches for Industry 4.0–the organizational culture perspective. Technol. Econ. Develop. Econ. 24(6), 2225–2240 (2018) 29. Michna, A., Kmieciak, R.: Open-mindedness culture, knowledge-sharing, financial performance, and industry 4.0 in SMEs. Sustainability 12(21), 9041 (2020) 30. Zakharova, L., Shilova, L., Gadbedji, Z., Liuchuan, Z.: Organizational cultures of vocational schools and firms in Russia, China and Iran as perceived by students and teachers. Educ. Stud. Moscow 3, 234–254 (2020) 31. Urrutia Pereira, G., de Lara Machado, W., Ziebell de Oliveira, M.: Organizational learning culture in industry 4.0: relationships with work engagement and turnover intention. Hum. Res. Develop. Inter. 25(5), 557–577 (2022) 32. Ziaei Nafchi, M., Mohelská, H.: Organizational culture as an indication of readiness to implement industry 4.0. Information 11(3), 174 (2020) ´ 33. Pietruszka-Ortyl, A., Cwiek, M., Zi˛ebicki, B., Wójcik-Karpacz, A.: Organizational culture as a prerequisite for knowledge transfer among IT professionals: the case of energy companies. Energies 14(23), 8139 (2021) 34. Liu, S.F., Fan, Y.J., Luh, D. B., Teng, P.S.: Organizational culture: the key to improving service management in industry 4.0. Appli. Sci. 12(1), 437 (2022) 35. Durana, P., Kral, P., Stehel, V., Lazaroiu, G., Sroka, W.: Quality Culture of Manufacturing Enterprises: A possible way to adaptation to Industry 4.0. Soc. Sci. 8(4), 124 (2019) 36. Sivarethinamohan, R., Kavitha, D., Koshy, E.R., Toms, B.: Reimagining future of future by redesigning talent strategy in the age of distraction and disruption. Inter. J. Systematic Innovation 6(4), 33–45 (2021) 37. Aguilar Rodríguez, I.E., Bernal Torres, C.A., Aldana Bernal, J.C., Acosta Aguinaga, A., Artieda Cajilema, C.H., Chalá, P.: Relationship between social culture, industry 4.0, and organizational performance in the context of emerging economies. J. Indus. Eng. Manag. 14(4), 750–770 (2021) 38. Mukhuty, S., Upadhyay, A., Rothwell, H.: Strategic sustainable development of Industry 4.0 through the lens of social responsibility: the role of human resource practices. Bus. Strategy Environ. 31, 2068–2081 (2022) 39. Carlsson, L., Olsson, A.K., Eriksson, K.: Taking responsibility for industrial digitalization: navigating organizational challenges. Sustainability 14(2), 866 (2022) 40. Roblek, V., Meško, M., Pušavec, F., Likar, B.: The role and meaning of the digital transformation as a disruptive innovation on small and medium manufacturing enterprises. Front. Psychol. 12, 592528 (2021) 41. Stacho, Z., Stachová, K., Cagáˇnová, D., Blštáková, J.: The key managerial competencies tendencies application in the business environment in Slovakia within the context of industry 4.0. EAI Endorsed Trans. Energy Web 8(32), e11 (2020) 42. Cadden, T., Dennehy, D., Mantymaki, M., Treacy, R.: Understanding the influential and mediating role of cultural enablers of AI integration to supply chain. Int. J. Prod. Res. 60(14), 4592–4620 (2022) 43. Egan, T.M., Yang, B., Bartlett, K.: The effects of organizational learning culture and job satisfaction on motivation to transfer learning and turnover intention. Hum. Resour. Dev. Q. 15(3), 279–301 (2004)

A Reflective Framework for Understanding Workforce Evolutionary Pathways in Industry 5.0 Alexandra Lagorio1(B)

, Chiara Cimini1

, and David Romero2

1 Department of Management, Information and Production Engineering,

University of Bergamo, Viale Marconi 5, Dalmine, BG, Italy {alexandra.lagorio,chiara.cimini}@unibg.it 2 Department of Industrial Engineering and Mechatronics, School of Engineering and Sciences, Tecnológico de Monterrey, 14380 Mexico City, Mexico

Abstract. The transition towards Industry 5.0 standards characterized by an increasing and responsible use of digital and smart technologies as well as by the development of more sustainable, human-centric, and resilient industrial systems, requires a careful reflection on the competencies and skills that will be needed in its workforce. Indeed, changing production processes and systems following an organizational renovation and/or the introduction of new technology can have positive impacts in terms of productivity and efficiency in a company’s operating model, but also can lead to negative effects in terms of resistance to change, increased workload, and mismatches between the workforce skills and the skills needed for the newly created types of jobs and duties. To prevent these possible issues, proper workforce management and operators’ training strategies should be implemented to not only contain costs and improve the efficiency of manufacturing and logistics processes but also to preserve operators’ satisfaction. This research work aims to support the transition towards Industry 5.0 standards by analysing the main trends of workforce evolution and by designing suitable pathways that companies could pursue for crafting proper talent development strategies and for facing current skill-related challenges such as skill shortages, skill gaps and shortfalls, and skill mismatches. To do so, this paper presents a Workforce Evolution Framework, with six Workforce Evolutionary Pathways, which are described to guide academics and practitioners in their understanding of the evolution of specific job profiles in the field of manufacturing to set targeted training strategies. Keywords: Industry 5.0 · Operator 5.0 · Workforce Development · Skills Development · Skill Shortages · Skill Gaps · Skill Mismatches · Training Strategies

1 Introduction In the last ten years, the manufacturing field has been characterized by many sociotechnical and environmental challenges, which have promoted significant advances in the adoption of digital and smart technologies in emerging cyber-physical production © IFIP International Federation for Information Processing 2023 Published by Springer Nature Switzerland AG 2023 E. Alfnes et al. (Eds.): APMS 2023, IFIP AICT 689, pp. 501–512, 2023. https://doi.org/10.1007/978-3-031-43662-8_36

502

A. Lagorio et al.

systems, raising great expectations about a future transformation towards more sustainable, human-centric, and resilient industrial systems. Nevertheless, more recently, it has been widely recognized that to realize the “Industry 4.0” vision, it is of utmost importance the development of proper talent development strategies [1]. Indeed, human-centricity has been appointed as one of the main pillars of the next step in industrial development referred to as: “Industry 5.0”, or “Fifth Industrial Revolution” [2]. Developing human skills and capabilities, by empowering learning and education, is expected to have a positive impact on social well-being and create a competitive advantage for sustainable economic success as well [3, 4]. Also, Work Design strategies such as “Job Enlargement” and “Job Enrichment” [5] will be increasingly crucial to maintain high levels of workers’ satisfaction and motivation even in changing work characteristics in terms of tasks’ number and complexity [6]. Nevertheless, the research in “Human-centric Manufacturing” is still underdeveloped and further investigation should be carried out including aspects that vary from the more technical (e.g., human-machine interface design, human-robot collaboration) to the more strategical and organizational ones (e.g., workforce management, skills development) [7]. Moreover, it is necessary to consider that, apart from the main technological disruptions that have been introduced in many industrial systems, several other social trends are characterising an evolution of the workforce that is challenging companies to consider simultaneously issues related to inclusiveness, social sustainability, ageing population, and so on [8]. To this purpose, this research work aims to support the transition towards Industry 5.0 standards by analysing the main trends of workforce evolution and by designing the most interesting pathways that companies could pursue for crafting proper talent development strategies and for facing current skill-related challenges such as skills mismatches and skills shortages. To do so, a Workforce Evolution Framework has been developed with six Workforce Evolutionary Pathways that have been identified. The framework has been conceptualized with the support of expert researchers and can be used as a guide to design the most suitable evolutionary pathways for specific job profiles in the field of manufacturing. This paper is organized as follows: Sect. 2 presents the research methodology used to build and validate the proposed framework; Sect. 3 discusses the main manufacturing workforce evolution trends that are pushing for new skill requirements in new or renewed job profiles; Sect. 4 presents the framework developed, and Sect. 5 finally discusses the main results of this research together with its limitations and further development.

2 Research Methodology To build and validate the proposed Workforce Evolution Framework, a three-step research methodology was developed (see Fig. 1), and is described in the following paragraphs. Step 1: Main Manufacturing Workforce Trends Identification – First of all, the main workforce trends impacting nowadays the manufacturing sector were analysed and were considered as possible causes of the skill- and job profile-related challenges. In particular, external trends such as the advancement of digital and smart technologies and workforce demographic tendencies, and internal trends like talent management and training development strategies, have been identified and described starting from a literature review

A Reflective Framework for Understanding Workforce

503

analysis. The results of this first step are described in detail in Subsects. 3.1 and 3.2 of this paper. Step 2: Skill-related Challenges Identification – Based on the manufacturing workforce trends analysed in Step 1, different skill-related challenges were identified. Through further analysis of the literature, three main skill-related challenges were identified [8]: (i) skill shortages, which denote a deficit in the number of skilled workers available versus those that are needed for specific job profiles; (ii) skill gaps or shortfalls, which are represented by workers that do not possess the skill and competence levels needed for specific job profiles, and (iii) skill mismatches, which refer to a workforce that does not possess the overall skill levels and skill sets needed to address the specific supply and demand of skilled workers for an explicit job profile. Subsection 3.3 will explain in depth this result. Step 3: Workforce Evolution Framework Description – To address the skill-related challenges discussed in Step 2, and to not be unprepared concerning the manufacturing workforce trends described in Step 1, companies have two different choices to make: (i) hiring a new workforce, even if it does not meet the job profile requirements of the company (in terms of skills and skill levels) to address the “skill shortages”, or (ii) provide the necessary skills to the already hired workforce through training strategies such as cross-skilling, up-skilling, re-skilling, and expert-skilling [8] to address “skill gaps” and “skill mismatches”. The proposed Workforce Evolution Framework aims to help companies to understand which “workforce dynamics” are necessary (along the different framework axes following different possible workforce evolutionary pathways) to support the implementation of different workforce management strategies. Section 4 will describe in detail each component of the framework.

Fig. 1. Research Methodology

504

A. Lagorio et al.

3 Main Manufacturing Workforce Trends Identification The first step of this research work is represented by the analysis of the current (main) trends that are characterising the manufacturing workforce. A literature review has been conducted by combining the keywords “workforce” AND “trend*” AND “manufacturing” in the Scopus meta-database. Starting from 256 results, filters have been applied to limit these to the English language, publication date in the last ten years (2013–2023), and belonging to the disciplines of Engineering, Business, Social Science, Computer Science, Economics, and Decision Sciences – reaching at the end a corpus of 122 papers. Finally, following a selection based on reading the titles and abstracts of these 122 papers, 24 papers have been fully reviewed. By developing a preliminary “text analysis” of the abstracts and keywords of these full-read papers, a word cloud was obtained and is presented in Fig. 2, which summarizes the 30 most frequent words in these 24 papers. This was useful to point out the most important trends to consider, which are further discussed in the following and have been classified into external trends (i.e., not depending on the single company workforce management strategy), and internal trends (i.e., depending on the company approach to human resource management).

Fig. 2. Workforce Manufacturing Trends

3.1 External Manufacturing Workforce Trends Reviewing the scientific literature, it emerged that the manufacturing workforce is evolving under the push of both technological and social trends. Addressing the technological perspective, several scholars discussed the need of developing a “Workforce 4.0” [6, 9] by investigating how Industry 4.0 technologies, which are becoming increasingly widespread, are changing the nature of many human activities, with the introduction of assistive, collaboration, and augmentation tools and devices that manufacturing workforce have to integrate within a smart working approach [10]. Even if digitalization solutions will be used to support work and not to control it totally [11], several job profiles will be required to update the set of related knowledge and skills, in particular in the field of data science, automation, and robotics [12]. Moreover, Galaske et al. [13] point out how production workers will be affected mostly by digitalization, considering that operators will be required to cover the role of conductor and coordinator of the factory and should act mainly as decision-makers.

A Reflective Framework for Understanding Workforce

505

Considering the social perspective, the most relevant trend affecting manufacturing is the ageing of the population which is already reflected in the ageing of the industrial workforce [14]. Since demographic change is significantly affecting the workforce age profile, the scientific community as well as the policymakers and the business leaders must put great effort into facing the challenges related to the impact of age on the operators’ performance [15]. Indeed, further research is needed to understand the capabilities (i.e., physical and cognitive) of older workers and highlight the relevance of their experience in terms of strategies and transferred knowledge to exploit and support the ageing workforce [16]. 3.2 Internal Manufacturing Workforce Trends From the internal perspective, managing evolving trends in the workforce generates several issues for the human resources development field. Indeed, managing human resources properly in industrial environments is crucial to maintaining high levels of productivity in the workforce, and for reducing the risks of employees quitting their job. Issa et al. [17] state that enhancing the employee’s welfare through the adoption of policies that include performance-based remuneration to personnel or career development strategies ensures a reduction of employee turnover. To this purpose, other methods that can be employed by companies are: to encompass workers’ qualifications-based job rotation [18], training programs designed to address specific job profiles and skills [19], and increased organizational flexibility to allow industrial smart working practice characterized by flexible working time and space [20]. In the literature review, other minor yet important trends emerge. For instance, LópezManuel et al. [21] clarify the importance of considering that temporary labour is widely spread and should be considered as a factor affecting workforce productivity. Other criticalities emerge, in addition, when dealing with the management of workers with disabilities. In this case, inclusion should be promoted through the design of friendly workspaces and workplaces and skill-fit assessment for optimal operations design and allocation [22]. 3.3 Skill-Related Challenges Identification The main internal and external trends analysed in the previous subsections are leading to different skill-related challenges. To find out which of these different skill-related challenges are occurring in the manufacturing sector, a literature review was conducted using the Scopus meta-database and searching for publications including different keywords such as “skill* issue*”, “skill* problem*”, “skill* challenge*” and “manufacturing”. The three main skill-related challenges emerging from this research are: “skill shortages”, “skill gaps or shortfalls”, and “skill mismatches”. However, first of all, it is necessary to define the term “skill” that has a consolidated definition in the scientific literature. Skills are defined as “cognitive and non-cognitive abilities that are specific to a particular job, occupation, or sector” [23]. Skill shortages arise when “employers are unable to recruit staff with the required skills in the accessible labour market and at the ongoing rate of pay” [24]. Linked to skill shortages it is also possible to find the two concepts of skill surplus (when the supply

506

A. Lagorio et al.

of skills is higher than the demand) and skill gaps (when skill levels of the existing workforce are insufficient to meet the requirements of the firms). The term skill gaps should also be sometimes found as a synonym for “skill shortages”. Skill mismatch refers to “the gap between supply and demand for skills and to the fact that observed matches between available workers and available jobs offered by firms in terms of skills and qualifications are sub-optimal” [25]. These issues are particularly critical because skills constraints affect both the ability to innovate and adopt technological development for companies and the employability prospects and access to quality jobs for workers [26]. Moreover, it is necessary to consider that identifying skills required for a job or possessed by a worker is not easy. Often a different terminology is used to identify the same skill by the educational providers and companies. Also for these reasons, they are often undervalued by job seekers [27]. Sometimes, skill mismatches are related to older workers that own outdated skills and are less familiar with technology [28]. Skill mismatches and Skill shortages usually force companies to hire workers that are not fitted completely for the job position offered. This means that a worker could be “under-skilled” to perform a certain task or “over-skilled” for a different one. Both circumstances negatively affected wages and job satisfaction, however, it is more critical for an over-skilled workforce. If under-skilled workers can be upskilled and slowly adapt their skills to a higher job requirement, the negative impacts could be contained. On the contrary, if over-skilled workers are assigned under-challenged tasks, this produces larger negative impacts on job satisfaction increasing absenteeism and resignations. For this reason, maintaining a certain control of the worker’s workload and the tasks’ challenging level should be a way to mitigate the negative effects produced by “under-skilling” and “over-skilling” issues [27]. The main solution to face skill-related challenges refers to skills development through investment in training and education. In past years, companies were not willing to resort to this solution because they had feared to faces the hold-up problem (trained workers asking for higher wages) or the poaching problem (other companies could hire the trained workers with a loss of money invested) [25]. However, nowadays it is clear that investing in higher education and vocational education and training (VET) is necessary as much as forecasting future skill needs and supply by using occupational and skill models and frameworks. To follow this research direction, a Workforce Evolution Framework has been developed to help companies in understanding which “workforce dynamics” occur in their cases and which workforce management and operators’ training strategies (for skills development) are necessary to improve workforce productivity and satisfaction.

4 Workforce Evolution Framework Description Figure 3 presents the proposed Workforce Evolution Framework that has been developed to represent the Workforce Evolutionary Pathways that companies need to consider to overcome the skill-related challenges explained previously and therefore face the discussed recent external and internal trends in the manufacturing workforce field. The Workforce Evolution Framework axes and the Workforce Evolutionary Pathways have been conceptualized based on three consensus-based sessions, involving the authors of

A Reflective Framework for Understanding Workforce

507

this paper and two other professors who are experts in the domains of (i) manufacturing and logistics systems, and (ii) production workers management strategies. The group discussion was conducted based on the questions: (i) What are the main evolution pathways that manufacturing job profiles can undertake? and (ii) What are the skill development strategies that support each pathway?

Fig. 3. Workforce Evolution Framework: Axes & Workforce Evolutionary Pathways

4.1 Workforce Evolution Framework Axes The proposed Workforce Evolution Framework stems from the intersection of two axes relating to the job competencies and skills needed by the workforce. On the vertical axis, ideal competencies and skills range from the most “operational” to the most “strategic”. Operational skills are related to working tasks that are generally conducted on the shop floor and enable workers to undertake short-term actions to operate manufacturing systems. Strategical skills, on the other hand, refer to capabilities related to management, collaboration, and decision-making, that are, relevant to pursuing a strategic goal – in the meaning discussed by [29]. On the horizontal axis, on the other hand, the competencies and skills range from the most “basic” to the most “advanced”. This axis refers to increasing levels of knowledge and experience that can be required to conduct the work activities. The intersection of these two axes forms the four quadrants of the framework. On the lower left-hand side, there is the quadrant in which the job profiles that have been defined as “Operative Job Profiles” can be placed. These job profiles require “basic” and “technical” skills such as the case of machine and equipment operators, assembly workers, and warehouse storekeepers. Generally, Operative Job Profiles are in charge

508

A. Lagorio et al.

of physical activities that are performed on the shop floor for which “basic-technical” (operational) skills are very often required to operate specific machines and equipment or conduct particular manual operations like assembly, stocking, picking, and transport activities. Moving counter-clockwise, lower-right-hand side, it is possible to move to the quadrant comprising jobs requiring advanced “operational” skills, which have been defined as “Specialized Operational Job Profiles”. To this category of job profiles belong, for instance, logistics technicians, maintenance technicians, quality analysts, and automation technicians. These job profiles own deeper knowledge about technological systems, but their work content is still belonging to the “practical” domain. Continuing counter-clockwise along the framework quadrants, it is possible to find in the quadrant located on the upper-right-hand side of the framework the job profiles requiring “advanced” and “strategic” skills. These job profiles have been defined as “Professional Job Profiles” and belong, for instance, to production engineers and quality and logistics managers. They require broader “technical” skills but are in charge of “technical” decision-making over manufacturing and logistics processes, therefore configuring as “managerial” positions. Finally, concluding the path around the framework in a counter-clockwise direction, the quadrant located in the upper-left-hand relates to “Managerial Job Profiles”. These job profiles require “basic” and “strategic” skills associated with the organization of manufacturing and logistics processes such as production planning and control, and resources like raw materials, workers, and work centers. To this quadrant belongs, for instance, the figures of operations managers, production planners, and team leaders. 4.2 Workforce Evolutionary Pathways Many reasons may lead to shifts along the different axes of the Workforce Evolution Framework. As was illustrated in Sect. 3.3, there are cases in which, following the introduction of new technologies or strategic-organizational changes (see Sect. 3.1), a mismatch is generated between the skills required by the company and those possessed by the existing workforce. As seen in Sect. 3.2, moreover, the mismatch may also arise as a result of an ageing workforce or changes in the work environment and industry (e.g., an increase in workload, flexible working, the possibility of agile working) that lead to a decrease in employee satisfaction and sometimes even to resignations. This mismatch is not always easily solved by hiring new workers due to economic or social reasons and the scarcity of workers with those specific skills needed. In this case, through the application of different training strategies, workers could follow different “Evolutionary Pathways” to move in the different quadrants of the framework matrix. In particular, six different Workforce Evolutionary Pathways can take place between the different quadrants of the proposed framework. The First Workforce Evolutionary Pathway goes from Operational Job Profiles to Managerial Job Profiles. This type of movement may occur when, for example, there is a need for a figure who could manage a team of Operative Job Profiles or a figure who has the skills needed not only to perform “operational” tasks but also control and “managerial” activities over an entire process that may include several operational tasks. The Second Workforce Evolutionary Pathway is from Specialized Operative Job Profiles

A Reflective Framework for Understanding Workforce

509

to Professional Job Profiles. This movement occurs when it is decided that someone who already has “specific” skills in technological-related tasks can also start to be involved in more “strategic” choices of technology design and adoption. The most typical case concerns automation technicians that become chief technology officers or innovation managers, in charge of a company’s digital transformation. For both of these Workforce Evolutionary Pathways, it is necessary to apply a Job Enlargement strategy. Job Enlargement refers to the possibility to expand the operators’ tasks and duties, to obtain a polyvalent workforce [30]. This strategy must be deployed together with an Up-skilling and a Re-skilling set of training programs to let operators acquire more “strategic” competencies and skills. This means that employees undertaking these pathways will have to acquire new skills related to managing, planning, and organisation (up-skilling) or need to align their current ones to manage more strategic goals (re-skilling). If the first two Workforce Evolutionary Pathways are on the y-axis, the third and fourth pathways are on the x-axis and are respectively the pathways from Managerial Job Profiles to Professional Job Profiles (third pathway), and from Operational Job Profiles to Specialized Job Profiles (fourth pathway). The Third Workforce Evolutionary Pathway occurs when there is a need to “specialize a managerial profile”, such as when an operations manager is promoted and becomes responsible for the entire digital transformation of a company, therefore, he/she requires to acquire specialized competencies and skills related to (digital) processes and (digital and smart) technologies. Similarly, in the Fourth Workforce Evolutionary Pathway, there is an “increased specialization of knowledge” and this may occur in the case where, for example, an operator is asked not only to perform a certain task on a piece of machinery but also to carry out problemsolving and maintenance actions on board the same machinery. Both Job Enlargement and Job Enrichment strategies can be applied to carry out these two pathways. Indeed, in addition to Job Enlargement, Job Enrichment consists in increasing specialistic and deep knowledge about processes to perform tasks more autonomously and with increased responsibility [30]. In particular, Cross-skilling and Expert-skilling training programs should be particularly effective for these typologies of job profile transitions. This means that workers should acquire new knowledge in other fields concerning the current tasks (cross-skilling) or gain a deeper knowledge about the currently performed tasks’ domain (expert-skilling) [6]. The Fifth Workforce Evolutionary Pathway is less frequent than the other four pathways analysed so far, but it was still decided to be included in the framework since it is possible. This Workforce Evolutionary Pathway goes from Specialised Operative Job Profiles to Managerial Job Profiles. It occurs when “strategic” competencies and skills are required but without the need for “specialised” knowledge. This may be the case when a maintenance technician becomes a maintenance manager. In this case, the knowledge must be more “basic” than that required for performing maintenance interventions but more “strategic” because it requires the organization of maintenance activities and teams. For this pathway, it is necessary to apply a Job Enrichment strategy and to implement Up-skilling and Expert-skilling training programs for the workforce. Finally, it was decided to include a Sixth Workforce Evolutionary Pathway within the same (four) quadrants as well, an internal pathway. This is because there may be cases

510

A. Lagorio et al.

in which it is necessary to perform different tasks inherent to the same job profiles that do not require a shift to a different type of skill. This may happen, for example, in the case of warehouse workers who are asked to do both the acceptance of incoming orders, material handling, picking and packing, and preparation of outgoing orders. These are different tasks that always require “basic” and “operational” skills, even though they are activities with different characteristics. In this case, Job Enlargement strategies are adopted and a Multi-skilling training program must take place so that the workforce becomes more flexible and able to perform various activities as required.

5 Conclusions and Further Research Due to the innumerable transitions/evolutions that companies are currently facing in terms of digital transition, ecological transition, and the evolution towards the Industry 5.0 paradigm that also takes into account new standards for resilience and humancentricity for industrial systems development, it is becoming more and more crucial to ensure that firms have the appropriate competencies to not fall behind their competitors and stakeholders. Because of exogenous factors such as the increasing “digitalization” and “smartification” of production processes and systems and the ageing of the population, or endogenous factors like the satisfaction of employees and their willingness to improve their job position, skill mismatches are created between the skills required by companies and the skills they have available among their employees. In addition, it is not always possible to recruit new employees due to their cost, or to continue to invest in employees who have already been trained over the years. It is therefore essential to find strategies to train the workforce according to their new skill requirements and to ensure that those workers who occupy certain job profiles are able, after appropriate training, to take on new job profiles or job tasks created by new skill requirements. The Workforce Evolution Framework proposed in this paper is designed as a guiding tool to understand when, how, and by applying which strategies it is most appropriate to make an employee with a certain job profile undertake an evolutionary pathway towards another type of job profile depending on the requirements of the company. This framework, with specific operationalization company characteristics and human resource management approaches, can become an important decision-support tool for companies operating in the manufacturing sector. Furthermore, along with a managerial application, the framework could be used to set proper training and lifelong learning programs to align the industrial workforce skills to the companies’ requirements, both at technical schools and university-level. For instance, it can be used to identify proper skill development strategies starting from different needs, concerning both the company and the workers’ perspectives. An example could be the need to change job profiles about the introduction of an automation technology that substitutes the operator in some tasks but requires enriched skills to manage supervision operations. Other examples can concern external trends, such as the aging of workers, that could drive an evolution of their working activities towards more managerial tasks to reduce the content of heavy manual activities. The framework in its current version has limitations, in particular, it is generic and has not yet been extensively validated with the involvement of industrial stakeholders.

A Reflective Framework for Understanding Workforce

511

Therefore, future research developments will aim at both the validation of the framework through its adoption on real business cases and its deployment to specific business areas (e.g., manufacturing, logistics, maintenance), to ensure its generalisation and reliability. Moreover, in this way it will be possible to specify more precisely which “basic”, “advanced”, “operational”, and “strategical” competencies and skills are needed and to identify in detail which actions to undertake to successfully deploy Job Enlargement and Job Enrichment strategies and their related training programs for skill developments such as cross-skilling, up-skilling, re-skilling, expert-skilling, and even multi-skilling, transforming the Workforce Evolution Framework into an effective decision support tool.

References 1. Li, L.: Reskilling and Upskilling the Future-ready Workforce for Industry 4.0 and Beyond. Inform. Syst. Front. (2022). https://doi.org/10.1007/s10796-022-10308-y 2. Breque, M., De Nul, L., Petridis, A.: European Commission, Directorate-General for Research and Innovation, Industry 5.0 – Towards a Sustainable, Human-centric and Resilient European Industry (2021). https://data.europa.eu/doi/, https://doi.org/10.2777/308407 3. Taisch, M., et al.: World Manufacturing Foundation, The 2019 World Manufacturing Forum Report – “Skills for the Future of Manufacturing” (2019). https://worldmanufacturing.org/ wp-content/uploads/WorldManufacturingFoundation2019-Report.pdf/ 4. World Economic Forum: The Future of Jobs Report 2020 (2020). https://www3.weforum. org/docs/WEF_Future_of_Jobs_2020.pdf 5. Cimini, C., Lagorio, A., Gaiardelli, P.: The evolution of operators’ role in production: how lean manufacturing and industry 4.0 affect job enlargement and job enrichment. Inter. J. Production Res. (2022). https://doi.org/10.1080/00207543.2022.2152894 6. Cimini, C., Romero, D., Pinto, R., Cavalieri, S.: Task classification framework and job-task analysis for assessing the impact of smart and digital technologies on the operators 4.0 job profiles. Sustainability, 15(5), 3899 (2023) 7. Cimini, C., Lagorio, A., Cavalieri, S., et al.: Human-technology integration in smart manufacturing and logistics: current trends and future research directions. Comput. Ind. Eng. 169, 108261 (2022) 8. Romero, D., Stahre, J.: Social sustainability of future manufacturing – challenges & strategies. In: The 2019 World Manufacturing Forum (2019), ISBN: 978–88–94386–12–7 9. Leesakul, N., et al.: Workplace 4.0: Exploring the implications of technology adoption in digital manufacturing on a sustainable workforce. Sustainability 14, 3311 (2022) 10. Dornelles, J.d.A., Ayala, N.F., Frank, A.G.: Smart working in industry 4.0: how digital technologies enhance manufacturing workers’ activities. Comput. Indus. Eng. 163, 107804 (2022) 11. Baethge-Kinsky, V.: Digitized industrial work: requirements, opportunities, and problems of competence development. Front. Soc. 5 (2020) 12. Li, G., et al.: Data science skills and domain knowledge requirements in the manufacturing industry: a gap analysis. J. Manuf. Syst. 60, 692–706 (2021) 13. Galaske, N., et al.: Workforce Management 4.0 – Assessment of human factors readiness towards digital manufacturing. In: Advances in Ergonomics of Manufacturing: Managing the Enterprise of the Future, pp. 106–115. Springer (2018). https://doi.org/10.1007/978-3-31960474-9_10 14. ; de Vere, I., et al.: Designing to enable an ageing workforce. In: 24th International Conference on Engineering and Product Design Education (2022). https://doi.org/10.35199/EPDE.202 2.23

512

A. Lagorio et al.

15. Di Pasquale, V., et al.: Ageing and human-system errors in manufacturing: a scoping review. Int. J. Prod. Res. 58, 4716–4740 (2020) 16. Panagou, S., Fruggiero, F., Lambiase, A.: The sustainable role of human factor in I4.0 scenarios. Proc. Comput. Sci. 180, 1013–1023 (2021) 17. Issa, L., et al.: Employee retention in agile project management. In: 10th International Conference on Information and Communication Systems, pp. 160–165. IEEE (2019) 18. Hochdörffer, J., Hedler, M., Lanza, G.: Staff scheduling in job rotation environments considering ergonomic aspects and preservation of qualifications. J. Manuf. Syst. 46, 103–114 (2018) 19. Pontes, J., et al.: Relationship between trends, job profiles, skills and training programs in the factory of the future. In: 2nd IEEE International Conference on Industrial Technology, pp. 1240–1245. IEEE (2021) 20. Cimini, C., Cavalieri, S.: Industrial smart working: a socio-technical model for enabling successful implementation. IFAC-PapersOnLine 55, 505–510 (2022) 21. López-Manuel, L., et al.: Making a virtue of necessity once again: assessing the effect of temporary labour on lean practices in highly routinized environments. Inter. J. Lean Six Sigma (2022). https://doi.org/10.1108/IJLSS-04-2022-0091 22. Narayanan, S., Terris, E., Cole, D.: disability inclusion in operations. In: Swaminathan, J.M., Deshpande, V. (eds.) Responsible Business Operations. SSSCM, vol. 10, pp. 115–135. Springer, Cham (2021). https://doi.org/10.1007/978-3-030-51957-5_6 23. OECD: Getting Skills Right. https://www.oecd-ilibrary.org/employment/getting-skills-right_ 25206125 24. Quintini, G.: Right for the Job: Over-Qualified or Under-Skilled? OECD, Paris (2011). https:// doi.org/10.1787/5kg59fcz3tkd-en 25. Brunello, G., Wruuck, P.: Skill shortages and skill mismatch: a review of the literature. J. Econ. Surv. 35, 1145–1167 (2021) 26. Pater, R., Cherniaiev, H., Kozak, M.: A dream job? skill demand and skill mismatch in ICT. J. Educ. Work. 35, 641–665 (2022) 27. Santiago-Vela, A., Hall, A.: Distinguishing challenging and over-challenging jobs: cognitive and affective skills mismatches and their impact on job satisfaction. Res. Comparative Inter. Educ. (2022). https://doi.org/10.1177/17454999221116486 28. Kim, H., Song Lee, B.: Aging workforce, wages, and productivity: do older workers drag productivity down in Korea? Journal of the Economics of Ageing 24, 100444 (2023) 29. Khalifa, A.S.: Strategy and what it means to be strategic: redefining strategic, operational, and tactical decisions. J. Strateg. Manag. 14, 381–396 (2021) 30. Lagorio, A., Cimini, C., Gaiardelli, P.: Reshaping the concepts of job enrichment and job enlargement: the impacts of lean and industry 4.0. In: Dolgui, A., Bernard, A., Lemoine, D., von Cieminski, G., Romero, D. (eds.) Advances in Production Management Systems. Artificial Intelligence for Sustainable and Resilient Production Systems: IFIP WG 5.7 International Conference, APMS 2021, Nantes, France, September 5–9, 2021, Proceedings, Part I, pp. 721–729. Springer International Publishing, Cham (2021). https://doi.org/10.1007/9783-030-85874-2_79

Managing Change Towards the Future of Work - Clustering Key Perspectives Katrin Singer-Coudoux1(B) , Greta Braun2 , and Johan Stahre2 1 Fraunhofer IPK, Pascalstr. 8-9, 10587 Berlin, Germany

[email protected] 2 Chalmers University of Technology, Hörsalsvägen 7a, 412 96 Göteborg, Sweden

Abstract. Manufacturing companies face rapid changes impacting the way their employees work and the way work is organized. Future of Work serves as a collective term to describe future working environments that today are characterized by increasing digitalization, requirements on sustainability, and resilience. This literature review synthesis nine topics for organizations to address regarding Future of Work. The contribution is a state-of-the-art analysis of the research area Future of Work in the manufacturing industry, clustering the nine identified topics into the established organizational framework Technology, Organization, People. These nine topics are: (1) Knowledge, Skills, and Competence Development; (2) Leadership; (3) Collaboration and Communication; (4) Corporate Culture; (5) Work Forms; (6) Workplace and Work Environment; (7) Technology Infrastructure and Strategy; (8) Occupational Health and Wellbeing; and (9) Digital Organisation and Network. This study gives theoretical implications about the research topic Future of Work, broadly used but yet under development, by synthesizing relevant work about the Future of Work in the manufacturing industry and providing a framework where the identified topics are clustered. From a managerial perspective, this study supports companies in understanding the topics to address and start change processes. The identified topics can be used by managers to organize their efforts towards the Future of Work to prepare their employees and their organization. Keywords: Future of Work · New Work · Workforce · Industry · Organizational Change

1 Introduction In the last years, research about the future of work has rapidly developed, leading to new trends in literature. Reasons for this include fast technological development, the climate crisis, and the COVID-19 pandemic. In parallel, demographic change is impacting Europe and other parts of the world and leading to a decrease in the working population [1]. For manufacturing companies, this results in an increased focus on sustainability and more frequent discussions about resilience. People working in industry are facing a change in purpose, content, activities, and tasks within their work. “Future of Work” serves as a collective term to describe what companies entail to know about the change © IFIP International Federation for Information Processing 2023 Published by Springer Nature Switzerland AG 2023 E. Alfnes et al. (Eds.): APMS 2023, IFIP AICT 689, pp. 513–527, 2023. https://doi.org/10.1007/978-3-031-43662-8_37

514

K. Singer-Coudoux et al.

in work, considering trends such as digitization, and in addition to that, how the workforce and the place where people work can be prepared for this transformation [2]. This concept, although still in development, includes elements such as a technological perspective, organizational changes, shifting managerial structures, and the adaptation required at corporate culture level. For manufacturing companies, exploring the Future of Work can offer opportunities to secure competitiveness by improving work conditions, increasing productivity, and augmenting the workers. Simultaneously, challenges such as occupational health, the qualification of employees for new tasks, and the acceptance of change must be addressed. However, changing definitions of what work means, is a complex transformation which poses major challenges for individuals, companies, and society as a whole. These players must find their place in a new system where they can create environmental, economic and social sustainability. The aim of this paper is to identify the current state-of-the-art of research concerning the Future of Work and synthesise it in a way that provides guidance to manufacturing companies who wish to transform their current work definition. The research questions posed in this article are: 1) What topics are addressed in literature regarding Future of Work in manufacturing industry? and 2) How can they be clustered to be practically usable in a change process?

2 Methodology In this literature review, we examine research on New Work and the Future of Work, focusing on research that is relevant for manufacturing companies from a management perspective. We used a multi-step methodology, including a database search, snowball technique, and backward search, which led to the identification of nine key topics. The Scopus database was queried using relevant search terms, and we limited the search to English articles and conference papers. We utilized the search terms: “New Work”, “Future of Work”, OR “work AND Digital*”, also in combination with the terms “manufactur*”, “industr*”, OR “production”. Snowball and backward search techniques were used for comprehensive coverage of the literature. We utilized the platform Rayyan to screen titles and abstracts for relevance and selected articles based on specific inclusion criteria related to our research objectives. For the full-text review, articles were included if they 1) discuss or define one of the terms “New Work” or “Future of work” or 2) discuss one specific or various research areas within the research topic New Work or Future of Work from a management or organizational perspective which are relevant for manufacturing companies. We excluded articles that did not address our research themes, that were not empirical or conceptual in nature, if there was no access to the full paper and papers that were published before 2016. Initially, we obtained 2,530 potentially relevant publications, which were reduced to 1,944 after removing duplicates. Screening titles for relevance further narrowed this number to 952. By screening the abstracts focusing on literature relevant for the Future of Work in manufacturing from a management perspective, we identified 280 relevant publications. After screening full texts and applying the snowball technique and backward search, we included 68 articles in the literature review. In a following step, we applied Braun and Clarke’s (2006) six-stage process for thematic analysis [3]. First, we familiarized ourselves with the data by noting preliminary

Managing Change Towards the Future of Work

515

ideas. We then indicated initial keywords from the extracted data in the Rayyan platform. Keywords that shared similar patterns were grouped into potential themes. These themes were then reviewed and refined, and this process continued until no new themes emerged. During this process, we engaged in continual reflection and discussion until consensus was reached to ensure the themes represented the data. To ensure the reliability of the thematic analysis, we conducted a process of validation. This involved cross-verification of themes and findings by a third researcher. Discrepancies were discussed and resolved to ensure the final themes represented the data. This approach’s qualitative nature ensured an exploration of the literature, allowing us to uncover themes that might not be evident through automated or quantitative methods. Each stage of the process was performed with transparency in the Rayyan platform to ensure the validity and reliability of the findings. The data extraction and analysis process revealed nine key topics in the literature. These themes were central to research on New Work, the Future of Work, and the digitalization of work and its impacts from a business perspective.

3 Results of Literature: Future of Work The literature review results reveal nine topics framing the Future of Work in manufacturing companies from a holistic organizational perspective. These topics are: (1) Knowledge, Skills, and Competence Development, (2) Leadership, (3) Collaboration and Communication, (4) Corporate Culture, (5) Work Forms, (6) Workplace and Work Environment, (7) Technology Infrastructure and Strategy, (8) Occupational Health and Wellbeing, and (9) Digital Organization and Network (Fig. 1).

Fig. 1. Future of Work - Key topics

516

K. Singer-Coudoux et al.

3.1 Knowledge, Skill and Competence Development The digitalization trend which has drastically impacted manufacturing contexts, along with the increased requirements for sustainability and resilience are transforming “work” as we know it. These trends have created new challenges regarding employee competencies, and prompt researchers to explore the skills which will be key to support companies in their digital transformation journeys. New key skills will need to support increasingly automated working contexts, and to do so, there is a need for new digital learning approaches that support the development of relevant competencies in increasingly digitalized work environments. In their work, Bühler, Jelinek, Nübel [4] foresee the future as a world where engineering students must be prepared for a volatile, uncertain, complex and ambiguous future. The complex relationship between technology, jobs, and skills has the potential to drive business growth and job creation, but also to displace entire roles when certain tasks become obsolete or automated [5]. Routineness has emerged as a key factor in understanding the relationship between digitalization and employee competencies [6]. In light of these challenges, numerous scientists are exploring the identification of necessary competencies that play a central role under the changed framework conditions of digitalization [4, 6–9], often divided into transferable skills like lifelong learning, creativity, and technology-related skills. Other studies show the lack of skills as a risk of implementing new business models, such as servitization. Lack of skills could pose a risk to the successful implementation of new and more sustainable business models [10]. In response to the challenges posed by digitalization, researchers are investigating effective methods for competence development in an increasingly digitalized world of work, e.g. examining new learning approaches and technologies, such as mixed reality in training unskilled workers in Industry 4.0 and smart factories for manual assembly tasks [11]. E-Learning [12], Blended Learning [13], and Gamification [14] are identified as important strategies for enhancing competence development in a digital era. Researchers are also exploring the application of Artificial intelligence, machine learning and natural language processing technologies in software solutions for Human Resource Management and HR-Development Focused on Skills Management and Skills Development [15]. They conclude that it is yet to be determined how effectively these solutions will fulfil their potential in assisting organizations to more accurately identify, align, and assess employee competencies [16]. Bothmer, Schlippe [15] suggest a solution with AI supporting recruiters and people searching for jobs, by using natural language processing to understand which skills an applicant to a job advertisement is missing and match them to relevant courses to upskill. Caused by continuous transformations in industry, lifelong learning is becoming more important. Employees have to find ways to motivate themselves to learn and upskill throughout their lifetime, which makes it important for learning providers and leaders to discuss the motivational challenges of adult learners [17].

Managing Change Towards the Future of Work

517

3.2 Leadership Digital transformation is a significant research field in the realm of leadership and new work, with numerous studies examining its effects on the world of work and the competencies required for leaders to successfully guide their organizations through this change. Aspects such as changing work processes and methods, leading virtual teams, and effective leadership styles in a digital work environment are addressed [17–20]. The leadership of virtual teams constitutes another research field. Due to digitalization and globalization, more teams work remotely and primarily communicate through digital media. This presents unique challenges for leaders, as they must build trust and relationships without personal contact and direct communication. Instead, they must develop virtual leadership competencies to maintain team performance and motivation and resolve conflicts. This area examines the specific challenges arising from the spatial separation of team members, as well as communication, coordination, and motivation [21, 22]. Furthermore, research addresses cultural change and leadership by investigating the role of leaders in promoting cultural change within organizations. Studies have shown that that transformational leadership and organizational culture were positively and significantly related to change management among virtual team employees’ [23]. Diversity and inclusion in leadership is another research field, with organizations increasingly relying on diverse workforces to remain innovative and competitive. Leaders must be capable of recognizing and leveraging the differences and similarities of employees to create an inclusive work environment where all can reach their full potential. Research in this area focuses on fostering diversity and creating an inclusive work environment that accommodates employees from different backgrounds and abilities [24, 25]. Lastly, leadership competence development plays a crucial role in the context of new work. With rapid changes and increased complexity in the world of work, leaders must possess a wide range of skills and competencies to face future challenges. Research in this area aims to prepare leaders for the demands of a rapidly changing work environment and equip them with the necessary skills to lead their employees successfully level [26, 27]. 3.3 Collaboration and Communication When addressing the changes in the world of work, researchers also investigate the impact of these changes on communication and collaboration. One focus is on the use of digital technologies such as virtual meetings, email management, instant messaging, social media platforms, as well as collaboration and project management tools to improve collaboration in teams and organisations [28–30]. Besides, research examines changes in communication from the digital transformation of teams. In this context, Vuchkovski, Zalaznik, Mitr˛ega, Pfajfar [31] identified communication differences in three areas (informal vs. formal, spontaneous vs. structured, and synchronous vs. sequential) when comparing communication in conventional teams to virtual teams. Besides, several challenges related to the digital transformation of teams (e.g., onboarding challenges, barriers in the virtual environment, structural challenges, new team roles such as change manager, chief officer of happiness, etc.)

518

K. Singer-Coudoux et al.

were identified. The research underscores the need for digital communication training and virtual team resilience to enhance communication and knowledge transfer in the digital transformation of teams. 3.4 Corporate Culture The future of work is characterized by an increasing emphasis on purpose, corporate social responsibility, and self-organization. Research in these areas aims to better understand how organizations can create meaningful and engaging work environments that promote employee satisfaction, productivity, and loyalty. A clear sense of purpose has been shown to have a positive effect on employee engagement. Research found that autonomously motivated employees may find inspiration and further enhancement in an appealing, broader purpose although purpose per se may not prove an antecedent to motivation [32], Thus, companies that effectively communicate their purpose and align it with employees’ values can foster engagement among their workforce. Corporate social responsibility research investigates how companies can act in socially and environmentally responsible ways. By taking responsibility for the environment, employees, and the common good, organizations can positively impact their workforce and society. According to Santana, Cobo [33], future of work research in this field explores strategies and measures that companies can adopt to fulfil their social responsibility and the implications for corporate governance and sustainability and sustainable HRM. The concept of self-organization focuses on creating work environments where employees have greater responsibility and autonomy over their work. Researchers are examining organizational structures and work methods that best support selforganization, as well as the associated challenges, such as balancing autonomy and control or coordinating between individual teams [34, 35]. 3.5 Work Forms The future of work is characterized by the emergence of new work forms, business models, and flexible work arrangements. These transformations, which include the sharing economy, gig economy, remote work, agile work methods, self-organization, and democratization, have significant implications for employee rights, social protection, and working conditions. New work forms and business models, such as the sharing economy and gig economy, are changing how employees work and earn money. Research in this area examines the impact of these developments on workers’ rights, social protection, and working conditions, as well as the broader implications for labor markets and the economy [36– 38]. Remote work, also referred to as telework or home office, involves working from a remote location, often with the help of digital technology. Research in this domain focuses on the effects of remote work on employee health, work effectiveness, and the challenges of collaboration and communication among remote teams [28, 39, 40]. Also, flexible work models, such as flextime, part-time hours or job-sharing, are becoming increasingly important in the context of New Work. Research in this area

Managing Change Towards the Future of Work

519

investigates their impact on employee performance, specifically through job satisfaction and organizational commitment. Further, research analyses if these impacts differ based on whether the arrangement was formalized or informally negotiated between the employee and their direct manager [41]. In the context of the future of work research, agile work methods emphasise flexibility and adaptability. Agile management, Kanban, Scrum, Radical Collaboration, etc., for example, focus on exploiting the productivity potential of digital processes. The resulting increase in productivity creates economic freedom that can also be used to serve the expectations of employees. Research investigates employees’ experience of meaning and commitment in agile new work environments by looking at the effects of self-organisation, flat hierarchies, and elaborate agile system architectures in the light of personal relationships, group dynamic processes and aspects of psychological empowerment [42]. Participation and democratization in the context of the future of work involve employees in decision-making processes within the company. Research in this domain investigates how employee and team characteristics, collaboration modes, information resources, and advanced technologies can be combined to create collective intelligence for effective decision-making [43]. 3.6 Workplace and Work Environment The future of work is increasingly influenced by the design of workspaces, with a focus on creating productive and pleasant environments for employees. Key aspects of modern office design, workplace design, co-working spaces, and ergonomics are critical in shaping the employee experience and fostering productivity, engagement, and well-being. Modern office design refers to the arrangement of workspaces and office spaces to create a productive and pleasant working environment. Central to modern office design is open space planning, the use of natural lighting, improved room acoustics, and the integration of technology in the workplace. Research in this area investigates the impact of modern office design on work productivity, workplace well-being as well as how job motivation is affected by technologies in the changing workplace [44, 45]. Co-working spaces are shared work areas used by independent workers or small businesses to share a workspace. These spaces offer a flexible working environment and promote collaboration and idea exchange among workers. Research in this area examines the effects of co-working spaces on work productivity, creativity, and employee engagement [46]. Ergonomics relates to the design of workplaces to create a safe and comfortable working environment that reduces injuries and physical discomfort among workers. Research in this field focuses on developing methods to optimize ergonomics in the workplace to maintain the physical health and work productivity of employees [47, 48]. 3.7 Technology Infrastructure and Strategy The increasing importance of human-machine interaction and the impacts of artificial intelligence and automation on the job market have profound implications for the workforce. Research in the area of IT strategy focuses on creating and implementing IT

520

K. Singer-Coudoux et al.

strategies that align with an organization’s overall business objectives. Ensuring that IT strategies are aligned with broader organizational goals is crucial for companies to remain competitive and agile in the digital age. Research explores how the functionality of information systems aligns with strategic objectives, the processes for institutionally integrating these strategically coordinated systems, and methods for effectively carrying out strategic transformations linked to system utilization [49]. Besides, cybersecurity research focuses on developing strategies and technologies to protect against cyber threats, including data breaches, phishing, and ransomware attacks. This research is particularly important given the growing reliance on digital technologies and the increasing frequency and sophistication of cyber-attacks. Ensuring the security and integrity of data and systems is vital for organizations to maintain trust and prevent costly incidents [50, 51]. Human-machine interaction has gained importance with the increasing use of artificial intelligence and other technologies. Research in the field of new work investigates how collaboration between humans and machines can be successfully designed and how this affects the world of work [52]. Romero, Stahre, Taisch [53] highlight the prevalence of new technologies and the importance of understanding how to effectively collaborate with machines to maximize productivity and to ensure the successful integration of technology into the workforce. Lastly, research on artificial intelligence and automation focuses on the impacts on the future of work, including changes in job design, the displacement of workers, and the need for re-skilling and up-skilling [54, 55]. 3.8 Occupational Health and Wellbeing The future of work is increasingly focused on the health and well-being of employees in the digital age. This includes enhancing workplace health and safety through digital technologies, addressing psychological health, improving workplace safety, optimizing workplace design and ergonomics, and fostering a better work-life balance through digital tools. Digital health and safety technologies are being analysed on how to improve workplace health and safety. Examples include wearables that monitor employees’ health and issue alerts when specific risks arise, as well as tools for optimizing the ergonomics of workplaces [56]. Research on psychological health in the workplace focuses on how digital technologies can help reduce but also increase stress and burnout [57–59]. Workplace safety research examines the prevention of accidents and injuries on the job. Researchers are exploring the use of digital technologies, such as sensors and augmented reality, to reduce the risk of accidents and increase safety [60, 61]. Lastly, research on work-life balance investigates how digital technologies can contribute to achieving a harmonious balance between work and leisure. Flexible working models, where employees can determine their work hours to better plan their leisure time, are one example of how digital technologies can facilitate improved work-life balance [62, 63].

Managing Change Towards the Future of Work

521

3.9 Digital Organisation and Network The rise of platform economies, the development of digital organizational structures, and cross-organizational cooperation are all relevant factors shaping the way work is conducted in the digital age. The field of digital process organization focuses on applying digital technologies to optimize work processes and procedures within an organization. Tools such as automation and artificial intelligence show potential to improve industry productivity and competitiveness, as well as process efficiency and products’ quality. As digital technologies continue to evolve, a larger and three-fold focus on technology, process, and organization is necessary to enable the achievement of the full benefits of digitalization [64]. Research on agile organizational structures examines the introduction of agile practices that enable companies to respond more flexibly to market changes. According to Prikladnicki, Lassenius, Carver [65], business agility is concerned with businesses and individuals to be more adaptive, creative, and resilient to pivot and remain necessary to customers and changing market needs. The development of platform economies is another area of research. Platform economies allow companies to collaborate with other businesses and organizations to create shared innovations. The growth of platform economies fosters interconnectivity and collaboration, which are vital for innovation in the digital age [66]. Digital organizational design research focuses on shaping organizational structures and hierarchies in a digitized world. Flat hierarchies and decentralized decision-making structures can be employed to enable faster and more effective decision-making. As digital technologies transform the way organizations operate, it is necessary to re-evaluate traditional organizational structures to maintain efficiency and adaptability [67, 68]. Lastly, research reflects on boundaries between organizations which becoming more fluid, leading to different cross-organizational and multi-stakeholder collaborations [69].

4 Discussion This study aimed to provide an overview of the current state of research on the Future of Work by conducting a literature review. The review identified nine key topics essential for the change process in companies: (1) Knowledge, Skills, and Competence Development; (2) Leadership; (3) Collaboration and Communication; (4) Corporate Culture; (5) Work Forms; (6) Workplace and Work Environment; (7) Technology Infrastructure and Strategy; (8) Occupational Health and Wellbeing; and (9) Digital Organization and Network. These topics serve as guidelines for emerging trends in future research and are relevant for companies’ change management. The nine themes identified in this study, while distinct, interact with each other in a complex, interconnected system. The topic Knowledge, Skills, and Competence Development is influenced by all the other topics. If the technological structure changes, employees need to learn new skills. Leadership emerges as a central influence, might shape corporate culture, driving technology strategy, and fostering occupational health and wellbeing. Conversely, leadership styles might be influenced by modifications in work forms and the workplace environment. Effective Collaboration and Communication can be influenced by technological infrastructure and in turn, enhance knowledge

522

K. Singer-Coudoux et al.

exchange within an organization. The Corporate Culture, influenced by leadership styles, work forms, and technology, might influence employee wellbeing and foster efficient communication. Variations in Work Forms, facilitated by technology and adaptive leadership, can potentially influence occupational health and wellbeing. In turn, changes in work forms might transform the workplace environment and necessitate new skills. The Workplace and Work Environment, influenced by work forms and technological changes, can influence the corporate culture. Technology Infrastructure and Strategy permeates through all other themes, shaping work forms, necessitating new skills, and influencing leadership strategies. Occupational Health and Wellbeing, influenced by leadership, corporate culture, work forms, and workplace environment, can reciprocally impact productivity, influencing areas such as collaboration and communication. Digital Organization and Network, affecting the manner of work execution and shaping work forms, relies heavily on the technology infrastructure. These themes’ interconnectedness creates a dynamic system shaping the Future of Work, where themes do not exist in isolation but continually impact each other. To fully understand and respond to the changing landscape of work, it is necessary to consider these themes not just individually, but as a complex web of interrelated elements. Nevertheless, the study is not without limitations. Due to the high number of search results in the database, the coverage might not have encompassed all potentially relevant studies due to the non-comprehensive nature of the literature review. This means certain pertinent insights could have been overlooked. Additionally, the thematic analysis was influenced by the researchers’ interpretation, which could carry inherent biases related to our backgrounds in business, management, and engineering. While we strived to reach a consensus, these individual perspectives could shape the clustering and interpretation of the themes. The Future of Work is an evolving concept, and this study offers insights into the current understanding of the topic, but further research is needed as it becomes more established. The study focuses on papers that can be relevant for the manufacturing sector but is expected to be transferable to other industrial sectors due to similar macro-trends and geopolitical situations. The transferability of the framework to other industries, however, was not investigated in-depth. Some articles in the literature review did not explicitly mention manufacturing but were still considered relevant. The literature review has both theoretical and managerial implications, contributing to the understanding of the Future of Work in the manufacturing industry and providing a framework for addressing the identified topics. Planned future research is about finding the challenges for companies in their change management towards the future of work. We plan to do studies with companies regarding their maturity regarding these topics and the challenges of the development. In general, the area would benefit from more studies about the skills development because it seems like this topic is affecting the other areas as well. This study could be beneficial in the manufacturing sector and broader industrial contexts, enabling comparisons of findings and the creation of a prescriptive framework for managing industrial companies towards the Future of Work.

Managing Change Towards the Future of Work

523

5 Conclusion It is tremendously important for companies today to find ways to create awareness and understanding of ongoing changes impacting the Future of Work and the work their employees do. Based on a literature review, this paper has identified a number of critical topics and clustered them into useful groups ready to be adopted by companies for their change process. The primary topics are Knowledge, Skills, and Competence Development; Leadership; Collaboration and Communication; Corporate Culture; Work Forms; Workplace and Work Environment; Technology Infrastructure and Strategy; Occupational Health and Wellbeing; Digital Organisation and Network. This clustering greatly simplifies the change processes needed by categorising the type of change needed, i.e. organisational change, process management, and technological development. By employing a rigorous and systematic methodology, we were able to provide an up-to-date overview of the current state of research on Future of Work, with a specific focus on relevance for manufacturing companies from an organizational perspective. The identification and analysis of the nine key topics offer valuable insights for both academics and practitioners working in this area, helping to advance the understanding of the complex and dynamic nature of work in the digital age.

References 1. European Commission. The Impact of Demographic Change (2022) 2. McKinsey, Company. What is the future of work? https://www.mckinsey.com/featured-ins ights/mckinsey-explainers/what-is-the-future-of-work. (2023) 3. Mihas, P.: Qualitative research methods: approaches to qualitative data analysis. In: International Encyclopedia of Education (Fourth Edition). Elsevier, pp 302–313 (2023) 4. Bühler, M.M., Jelinek, T., Nübel, K.: Training and preparing tomorrow’s work-force for the fourth industrial revolution. Educ. Sci. 12, 782 (2022) 5. World Economic Forum The future of jobs report 2018. Insight report. World Economic Forum, Cologny/Geneva 6. Kolade, O., Owoseni, A.: Employment 5.0: The work of the future and the future of work. Technol. Soc. 71, 102086 (2022). https://doi.org/10.1016/j.techsoc.2022.102086 7. Chalutz, H.-G., Cohen, Y.: The “new Fit”: skills and competencies for the future of work. IFAC-PapersOnLine 55, 511–515 (2022). https://doi.org/10.1016/j.ifacol.2022.04.245 8. Hecklau, F., Orth, R., Kidschun, F., et al.: Veränderte Kompetenzanforderungen im Rahmen von Digitalisierung und Industrie 4.0. WM 2019 - Wissensmanagement in digitalen Arbeitswelten: Aktuelle Ansätze und Perspektiven - Knowledge Management in Digital Workplace Environments: State of the Art and Outlook: 9–29 (2020) 9. Qiu, S., Natarajarathinam, M., Johnson, M.D., et al.: The Future of Work: Identifying Futureready Capabilities for the Industrial Distribution Workforce. In: ASEE Annual Conference and Exposition, Conference Proceedings (2021) 10. González Chávez, C.A., Unamuno, G., Despeisse, M., et al.: Analyzing the risks of digital servitization in the machine tool industry. Robot. Comput. Integr. Manufact. 82, 102520 (2023). https://doi.org/10.1016/j.rcim.2022.102520 11. Pimminger, S., Kurschl, W., Schonbock, J., et al.: Mixed reality workplace training systems for smart factories: challenges and future directions. In: Proceedings of the 2022 IEEE International Conference on Human-Machine Systems, ICHMS 2022 (2022). https://doi.org/10. 1109/ICHMS56717.2022.9980642

524

K. Singer-Coudoux et al.

12. Shahriar, S.H.B., Akter, S., Sultana, N., et al.: MOOC-based learning for human resource development in organizations during the post-pandemic and war crisis: a study from a developing country perspective. JRIT 16, 37–52 (2023). https://doi.org/10.1108/JRIT-09-20220054 13. Kelly, J.A., McNair, V.: Re-imagining Blended Learning. An Experience-Led Approach to Accelerate Student Future Skills Development. In: Auer, M.E., Hortsch, H., Michler, O., et al. (eds.) Mobility for Smart Cities and Regional Development - Challenges for Higher Education, vol. 389, pp. 558–564. Springer International Publishing, Cham (2022) 14. Keepers, M., Nesbit, I., Romero, D., et al.: Current state of research & outlook of gamification for manufacturing. J. Manuf. Syst. 64, 303–315 (2022). https://doi.org/10.1016/j.jmsy.2022. 07.001 15. Bothmer, K., Schlippe, T.: Skill Scanner: Connecting and Supporting Employers, Job Seekers and Educational Institutions with an AI-Based Recommendation System. In: Guralnick, D., Auer, M.E., Poce, A. (eds.) Innovative Approaches to Tech-nology-Enhanced Learning for the Workplace and Higher Education, vol. 581, pp. 69–80. Springer International Publishing, Cham (2023) 16. Ifenthaler, D., Seufert, S. (eds.): Artificial Intelligence Education in the Context of Work. Springer International Publishing, Cham (2022) 17. Braun, G., Järvinen, M., Stahre, J., et al.: Motivational challenges of engineers participating in an online upskilling program. In: European Conference on e-Learning; Kidmore End (2022) 18. Greimel, N.S., Kanbach, D.K., Chelaru, M.: Virtual teams and transformational leadership: an integrative literature review and avenues for further research. J. Innovation Knowl. 8, 100351 (2023). https://doi.org/10.1016/j.jik.2023.100351 19. Klein, M.: Leadership characteristics in the era of digital transformation. Bus. Manag. Stud. Int. J. 8, 883–902 (2020). https://doi.org/10.15295/bmij.v8i1.1441 20. Larson, L., DeChurch, L.: Leading teams in the digital age: four perspectives on technology and what they mean for leading teams. Leadersh Q 31, 101377 (2020). https://doi.org/10. 1016/j.leaqua.2019.101377 21. Efimov, I., Harth, V., Mache, S.: Health-oriented self- and employee leader-ship in virtual teams: a qualitative study with virtual leaders. Int. J. Environ. Res. Public Health 17, 6519 (2020). https://doi.org/10.3390/ijerph17186519 22. Groves, K.S., Feyerherm, A.E.: Leader cultural intelligence in context. Group Org. Manag. 36, 535–566 (2011). https://doi.org/10.1177/1059601111415664 23. Kaur Bagga, S., Gera, S., Haque, S.N.: The mediating role of organizational culture: transformational leadership and change management in virtual teams. Asia Pac. Manag. Rev. 28, 120–131 (2023). https://doi.org/10.1016/j.apmrv.2022.07.003 24. Gotsis, G., Grimani, K.: Diversity as an aspect of effective leadership: integrating and moving forward. Leadersh. Org. Dev. J. 37, 241–264 (2016). https://doi.org/10.1108/lodj-06-20140107 25. Gotsis, G., Grimani, K.: The role of servant leadership in fostering inclusive organizations. J. Manag. Dev. 35, 985–1010 (2016). https://doi.org/10.1108/jmd-07-2015-0095 26. Schiuma, G., Schettini, E., Santarsiero, F., et al.: The transformative leadership compass: six competencies for digital transformation entrepreneurship. Int. J. Entrepreneurial Behav. Res. 28, 1273–1291 (2022). https://doi.org/10.1108/ijebr-01-2021-0087 27. Ngayo Fotso, G.M.: Leadership competencies for the 21st century: a review from the Western world literature. Eur. J. Training Dev. 45, 566–587 (2021) 28. Späth, K., Storck, S., Nass Bauer, C., et al.: Fulfilling remote collaboration needs for New Work. Procedia Comput. Sci. 191, 168–175 (2021) 29. Azarova, M., Hazoglou, M., Aronoff-Spencer, E.: Just slack it: A study of multidisciplinary teamwork based on ethnography and data from online collaborative software. New Media Soc. 24, 1435–1458 (2022)

Managing Change Towards the Future of Work

525

30. Anders, A.: Team communication platforms and emergent social collaboration practices. Int. J. Bus. Commun. 53, 224–261 (2016) 31. Vuchkovski, D., Zalaznik, M., Mitr˛ega, M., et al.: A look at the future of work: the digital transformation of teams from conventional to virtual. J. Bus. Res. 163, 113912 (2023). https:// doi.org/10.1016/j.jbusres.2023.113912 32. van Tuin, L., Schaufeli, W.B., van den Broeck, A., et al.: A corporate purpose as an antecedent to employee motivation and work engagement. Front Psychol 11, 572343 (2020). https://doi. org/10.3389/fpsyg.2020.572343 33. Santana, M., Cobo, M.J.: What is the future of work? A science mapping analysis. Eur. Manag. J. 38, 846–862 (2020). https://doi.org/10.1016/j.emj.2020.04.010 34. Farkhondeh, M., Müller, B.: Holacracy: a new way of organizing? Manag. Rev. 32, 302–317 (2021). https://doi.org/10.5771/0935-9915-2021-4-302 35. Baten, R.A., Bagley, D., Tenesaca, A., et al.: Creativity in temporal social networks: how divergent thinking is impacted by one’s choice of peers. J. R. Soc. Interface. 17, 20200667 (2020). https://doi.org/10.1098/rsif.2020.0667 36. Bates, O., Lord, C., Alter, H., et al.: Lessons from one future of work: opportunities to flip the gig economy. IEEE Pervasive Comput. 20, 26–34 (2021). https://doi.org/10.1109/MPRV. 2021.3113825 37. Huang, N., Burtch, G., Hong, Y., et al.: Unemployment and worker participation in the gig economy: evidence from an online labor market. Inf. Syst. Res. 31, 431–448 (2020). https:// doi.org/10.1287/isre.2019.0896 38. Hudek, I., Širec, K.: The impact of digital transformation on freelancer well-being. In: Khosrow-Pour, D., Anshari, M., Razzaq, A., et al. (eds.) Digital Psychology’s Impact on Business and Society. IGI Global, pp 56–91 (2023) 39. Barabaschi, B., Barbieri, L., Cantoni, F., et al.: Remote working in Italian SMEs during COVID-19. learning challenges of a new work organization. J. Workplace Learn. 34, 497–512 (2022). https://doi.org/10.1108/JWL-10-2021-0132 40. Sengupta, D., Al-Khalifa, D.: Pandemic imposed remote work arrangements and resultant work-life integration, future of work and role of leaders—a qualitative study of Indian millennial workers. Adm. Sci. 12, 162 (2022). https://doi.org/10.3390/admsci12040162 41. de Menezes, L.M., Kelliher, C.: Flexible working, individual performance, and employee attitudes: comparing formal and informal arrangements. Hum. Resour. Manage. 56, 1051– 1070 (2017). https://doi.org/10.1002/hrm.21822 42. Schneider, T.: Agil, hierarchiefrei und selbstorganisiert im New Work oder überwältigt von Systemstrukturen und unterdrückten gruppendynamischen Prozessen im New Office. Gr. Interakt. Org. 51, 469–479 (2020) 43. Kudyba, S., Fjermestad, J., Davenport, T.: A research model for identifying factors that drive effective decision-making and the future of work. J. Intellect. Capital 21, 835–851 (2020). https://doi.org/10.1108/JIC-05-2019-0130 44. Schmid, Y., Dowling, M.: New work: new motivation? A comprehensive literature review on the impact of workplace technologies. Manag. Rev. Q. 72, 59–86 (2022) 45. Nelson, E.C., Wray, H.E., White, N.C.: The future of work, workplaces and smart buildings. In: Proceedings of the 9th ACM International Conference on Systems for EnergyEfficient Buildings, Cities, and Transportation, pp. 326–329. https://doi.org/10.1145/356 3357.3566136 46. Krause, I.: Coworking spaces: windows to the future of work? Changes in the organizational model of work and the attitudes of the younger generation. Foresight STI Gov. 13, 52–60 (2019) 47. Miguez, S.A., Hallbeck, M.S., Vink, P.: Participatory ergonomics and new work: reducing neck complaints in assembling. Work 41(Suppl 1), 5108–5113 (2012). https://doi.org/10. 3233/WOR-2012-0802-5108

526

K. Singer-Coudoux et al.

48. Zaalouk, A., Han, S.: Parameterized design optimization framework for worker-friendly workplaces in modular construction. J. Constr. Eng. Manage. 147(5), 04021030 (2021) 49. Arvidsson, V., Holmström, J., Lyytinen, K.: Information systems use as strategy practice: a multi-dimensional view of strategic information system implementation and use. J. Strat. Inf. Syst. 23, 45–61 (2014) 50. Quiroz, J.T., Oscategui, M.A.A., Armas-Aguirre, J.: Cybersecurity taxonomy: research and knowledge areas. In: 2021 IEEE 1st International Conference on Advanced Learning Technologies on Education & Research (ICALTER), pp.1–4 (2021). https://doi.org/10.1109/ICA LTER54105.2021.9675075 51. Mouloua, S.A., Ferraro, J., Mouloua, M., et al.: Trend analysis of cyber security research published in HFES proceedings from 1980 to 2018. Proc. Hum. Factors Ergonomics Soc. Ann. Meet. 63, 1600–1604 (2019) 52. Cheon, E., Zaga, C., Lee, H.R., et al.: Human-machine partnerships in the future of work: exploring the role of emerging technologies in future workplaces. In: Companion Publication of the 2021 Conference on Computer Supported Cooperative Work and Social Computing, pp. 323–326 (2021) 53. Romero, D., Stahre, J., Taisch, M.: The Operator 4.0: towards socially sustainable factories of the future. Comput. Ind. Eng. 139,106128 (2020) 54. Hammer, A., Karmakar, S.: Automation, AI and the Future of Work in India. ER 43, pp.1327– 1341 (2021). https://doi.org/10.1108/ER-12-2019-0452 55. Agrawal, A., Gans, J.S., Goldfarb, A.: Artificial intelligence: the ambiguous labor market impact of automating prediction. J. Econ. Perspect. 33, 31–50 (2019). https://doi.org/10. 1257/jep.33.2.31 56. Mettler, T., Wulf, J.: Physiolytics at the workplace: Affordances and constraints of wearables use from an employee’s perspective. Info. Syst. J. 29, 245–273 (2019). https://doi.org/10. 1111/isj.12205 57. Boers, E., Afzali, M.H., Conrod, P.: Temporal associations of screen time and anxiety symptoms among adolescents. Can J. Psychiatry 65, 206–208 (2020) 58. Carolan, S., de Visser, R.O.: Employees’ perspectives on the facilitators and barriers to engaging with digital mental health interventions in the workplace: qualitative study. JMIR Ment. Health 5, e8 (2018). https://doi.org/10.2196/mental.9146 59. Rauschenberg, C., Schick, A., Goetzl, C., et al.: Social isolation, mental health, and use of digital interventions in youth during the COVID-19 pandemic: a nationally representative survey. Eur. Psychiatry 64, e20 (2021) 60. Svertoka, E., Saafi, S., Rusu-Casandra, A., et al.: Wearables for industrial work safety: a survey. Sensors (Basel) 21, 3844 (2021). https://doi.org/10.3390/s21113844 61. Tamers, S.L., Streit, J., Pana-Cryan, R., et al.: Envisioning the future of work to safeguard the safety, health, and well-being of the workforce: a perspective from the CDC’s National Institute for Occupational Safety and Health. Am. J. Ind. Med. 63, 1065–1084 (2020). https:// doi.org/10.1002/ajim.23183 62. Shagvaliyeva, S., Yazdanifard, R.: Impact of flexible working hours on work-life balance. Am. J. Ind. Bus. Manag. 04, 20–23 (2014). https://doi.org/10.4236/ajibm.2014.41004 63. Khallash, S., Kruse, M.: The future of work and work-life balance 2025. Futures 44, 678–686 (2012). https://doi.org/10.1016/j.futures.2012.04.007 64. Talamo, C., Bonanomi, M.M.: The Impact of Digitalization on Processes and Organizational Structures of Architecture and Engineering Firms. In: Daniotti, B., Gianinetto, M., Della Torre, S. (eds.) Digital Transformation of the Design, Construc-tion and Management Processes of the Built Environment, pp. 175–185. Springer International Publishing, Cham (2020) 65. Prikladnicki, R., Lassenius, C., Carver, J.C.: Trends in agile: business agility. IEEE Softw. 37, 78–80 (2020). https://doi.org/10.1109/MS.2019.2945652

Managing Change Towards the Future of Work

527

66. Wei, R.: Research on the operating mechanism and supervision of platform economy. In: Proceedings of the 2021 3rd International Conference on Economic Management and Cultural Industry (ICEMCI 2021). Atlantis PressParis, France (2021) 67. Ackermann, M., Schell, S., Kopp, S.: How mercedes-benz addresses digital transformation using holacracy. J. Organ. Chang. Manag. 34, 1285–1299 (2021). https://doi.org/10.1108/ JOCM-12-2020-0395 68. Bonanomi, M.M., Hall, D.M., Staub-French, S., et al.: The impact of digital transformation on formal and informal organizational structures of large architecture and engineering firms. Eng. Constr. Architectural Manag. 27, 872–892 (2020). https://doi.org/10.1108/ECAM-032019-0119 69. Barrane, F.Z., Ndubisi, N.O., Kamble, S., et al.: Building trust in multi-stakeholder collaborations for new product development in the digital transformation era. Benchmarking 28, 205–228 (2021). https://doi.org/10.1108/BIJ-04-2020-0164

Development of a Task Model for Artificial Intelligence-Based Applications for Small and Medium-Sized Enterprises Florian Clemens1(B) , Fabian Willemsen2 , Susanne Mütze-Niewöhner2 , and Günther Schuh1 1 FIR an der RWTH Aachen, 52074 Aachen, Germany

[email protected] 2 Institute of Industrial Engineering and Ergonomics, RWTH Aachen University,

52062 Aachen, Germany

Abstract. The adoption of artificial intelligence (AI) technologies in manufacturing companies is challenging, particularly for SMEs that lack the necessary skills to develop and integrate AI-based applications (AI applications) into their existing IT system landscape. To address this challenge, the research project VoBAKI (IGF-Project No.: 22009 N) aims to enable SMEs to identify and close skill gaps related to AI application development and implementation using proper sourcing strategies. This paper presents the interim results from the second phase of the project, which involves identifying the tasks in the lifecycle of AI applications and determining the specific skills required for executing these tasks. The presented results provide a detailed lifecycle including the phases for the development and usage of AI applications, as well as the specific tasks that SMEs must consider when implementing an AI application. These results serve as the foundation for future research regarding the required skills to execute the presented tasks and provide a roadmap for SMEs to close skill gaps and successfully implement AI applications. Keywords: Artificial Intelligence · Machine Learning · Manufacturing Company · SME · Decision-Maker · Artificial Intelligence Lifecycle

1 Introduction 1.1 Problem Definition Existing development processes of AI applications include various tasks from data collection and model development to the evaluation of the trained models [1]. The use of these models in manufacturing companies presents a special challenge, as there are significantly more tasks to be performed compared to tasks included in existing development processes (e. g. integration of AI applications in existing IT-systems). To address these tasks, a variety of skills are required [2]. These skills are not present in many companies, especially not in SMEs [3]. Based on research conducted for the German federal © IFIP International Federation for Information Processing 2023 Published by Springer Nature Switzerland AG 2023 E. Alfnes et al. (Eds.): APMS 2023, IFIP AICT 689, pp. 528–542, 2023. https://doi.org/10.1007/978-3-031-43662-8_38

Development of a Task Model for Artificial Intelligence

529

ministry for economic affairs and climate action, most of the SMEs surveyed reported that their adoption of AI technologies is impeded by a lack of skills [3]. Furthermore, SMEs lack an overview of these tasks and the skills required to close the described gap. 1.2 Objective and Project Description The objective of the research project VoBAKI (IGF-Project No.: 22009 N) is to enable SME to independently identify skill gaps regarding the development and implementation of AI applications and close them using a proper sourcing strategy. To achieve this, the project is structured in 5 steps. The results are elaborated in cooperation with a user committee that is composed of 18 companies of different industries and sizes (see Table 1). Table 1. Companies of the user committee

Branch Manufacturing IT & Service IT & Service Manufacturing Consulting IT & Service Consulting IT & Service Manufacturing Manufacturing Consulting IT & Service IT & Service IT & Service Manufacturing Consulting Manufacturing Manufacturing

Number of Employees 11-50 201-500 1-10 51-200 1-10 11-50 1001-5000 201-500 11-50 11-50 11-50 >100.000 10-50 1001-5000 10-50 1-10 10-50 201-500

SME x x x x x

x x x x x x x

First, the project examines the objectives of SMEs regarding the use of AI applications. Second, the project identifies tasks in the lifecycle of AI applications and determines the specific skills required for the execution of these tasks. Third, potential sourcing strategies will be described and evaluated with respect to their practical relevance for SMEs. Eventually, the results of the project will be combined in an approach for the specific identification of skill gaps and the selection of the proper sourcing strategy to close these gaps. This paper presents the interim results from the second step of the project. A lifecycle model is introduced that includes phases for the development and usage of AI

530

F. Clemens et al.

applications. Furthermore, these phases are detailed by AI application-specific tasks that SMEs must consider when implementing an AI application. The presented results are the foundation for future research regarding the required skills to execute the presented tasks. 1.3 Structure of the Paper Section 2 presents the existing development processes of AI applications as they are the foundation of the designed lifecycle. Furthermore, it outlines the research gap. Section 3 displays the research approach to develop a lifecycle for AI applications as well as the identification and description of lifecycle-related tasks. Section 4 describes the results achieved regarding the design of a lifecycle as well as the identification of tasks in the lifecycle. Eventually, Sect. 5 outlines the future research steps in the project VoBAKI.

2 Related Work 2.1 Development Processes for AI Applications Over the years, many processes for the development of AI applications were designed [4]. The origins of most processes are the Cross-Industry-Stand (Crisp-DM) and the Knowledge Discovery in Databases (KDD) [4]. Therefore, Sect. 2.1 presents Crisp-DM and the KDD in detail as well as SEMMA, a process that is equally acknowledged like Crisp-DM and the KDD. Eventually, Sect. 2.2 presents the resulting research gap. Cross Industry Standard Process in Data Mining (CRISP-DM) The most used top-down process for data analysis is the CRISP-DM process, which emerged during research in the year 2000 [5]. Top-down refers to the application-oriented approach to data analysis. Based on the intended application, it is determined which data must be available in which quantity, format and quality. CRISP-DM is a six-stage process for the use of data mining with the goal of pattern recognition and process optimization. The first process phase deals with the analysis and identification of the business field relevant to the company. In particular, this involves the definition of goals and success-related criteria or requirements [6]. Building on this, the second phase deals with the investigation of the data and the understanding regarding the information obtained from it. In addition to data collection and investigation, data quality testing is an essential part of the subsequent third phase. After ensuring the data quality, the selection, cleansing, integration, and formatting of the data required for modeling takes place. The central phase of the CRISP-DM process is the modeling, in which data analysis is performed using an appropriate modeling technique or algorithm. The applied model uses a distinction between training and evaluation data. This is used to generate an optimal performance of the model [7]. The evaluation deals with the assessment of the achieved results (e. g. assessment of the prediction accuracy). The last step is the transfer of the evaluated model as an application to practice [5].

Development of a Task Model for Artificial Intelligence

531

The CRISP-DM process is characterized by its application-oriented reference. Based on the selected requirements, the selection of the optimal modeling procedure or algorithm is made after the analysis of the necessary data. This ensures an efficient way for data analysis. Through this top-down procedure only the data necessary for the selected use case and thus also the appropriate algorithm is used and identified. Knowledge Discovery in Databases (KDD) Another procedure for data analysis is the KDD process according to Fayyad. This process for detecting knowledge in databases is composed of the phases of goal definition, data acquisition data preprocessing, data transformation, data mining and evaluation [8]. In the first step of the process, the relevant goals are identified and defined. While the data ingestion phase deals with data collection and selection, the essential task of data preprocessing is the analysis of the selected data. This data is transformed into the optimal data format for a selected process after the phase is completed. The transformation represents the basis for the central phase of the KDD process, the data mining [9]. Data mining is particularly concerned with pattern detection and extraction and the selection and application of methods, such as cluster analysis, that are necessary for this purpose. In the last step of the process, the patterns generated from this are displayed graphically, interpreted in a discussion, and finally evaluated [6]. The goal of KDD is the specific analysis of data to identify subject-specific correlations and generate knowledge. Particularly with large and complex data sets, pattern recognition or knowledge extraction is a major challenge. By means of data mining a significant reduction of complexity can be realized [10]. SEMMA In addition to CRISP-DM and KDD, SEMMA is another top-down method for processing data. SEMMA, developed by the SAS Institute, is an acronym made up of the terms Sample, Explore, Modify, Model and Access [11]. The first step of the procedure deals with the identification of factors influencing the process. Thus, a sample relevant for the further procedure is to be determined. In the next step, the selected data sample is examined about the interrelationships of the individual data and possible influencing factors. The considered data are documented as well as visually represented. Subsequently, the selected data are cleaned, selected, and transformed into the format required for the selected model [12]. The fourth step of SEMMA represents the core of the top-down procedure. Here, techniques and algorithms are investigated, based on which the targeted model is developed. The goal here is to identify insights and patterns from the available data. The data analysis techniques used for this purpose are, for example, decision trees and neural networks [13]. The final step addresses model accessibility [13]. Unlike CRISP-DM and KDD, SEMMA does not deal meticulously with the business processes of a company. 2.2 Research Gap Currently used development processes like CRISP-DM primarily focus on the development of AI models. However, there is a gap in research regarding the implementation of these models as applications in a productive environment of SMEs. Specifically, there is

532

F. Clemens et al.

a lack of research on how unexperienced decision makers in SME, particularly managers, can effectively use those processes to get an overview of the entire development process, including important tasks like setting up the technical architecture and organizational development for an AI application. One potential reason for this gap is that development processes are often designed by and for developers, with less consideration for the needs of decision-makers without any experience in the field of AI. The implications of this research gap are significant. Without a comprehensive understanding of the entire development process, decision-makers in SMEs may struggle to effectively allocate resources, manage timelines, and make informed decisions regarding AI projects. Additionally, the failure to consider the required technical architecture and organizational development may result in AI models that are not scalable, do not integrate with existing IT systems, and are not user-friendly. Further research is needed to address this gap and provide decision-makers of SME with the tools and information needed to effectively manage AI development projects.

3 Approach This chapter provides a detailed description of the approach and methods used to elaborate the results presented in this paper (see Fig. 1). The methodology is separated in two main steps, the design of an AI application lifecycle and the identification and description of tasks in the lifecycle. Section 2.1 shows that the currently available development processes lack certain critical phases of an AI application lifecycle. Thus, they are not usable for decision makers in SME (see Sect. 2). As a result, a comprehensive lifecycle was designed. The approach included three steps. First, existing development processes were researched through a literature analysis (see Sect. 2.1). Second, the phases usable for a lifecycle were extracted from the existing development processes. Eventually, missing phases were added and validated with the user committee. After the design of the AI application lifecycle, AI-specific tasks in the phases of the lifecycle had to be identified and described. At first, common roles of AI projects were identified. For this purpose, literature research and searches on common job portals were carried out. Based on the role or job description, the roles were compared and categorized in terms of content (according to Mayring’s qualitative content analysis [14]). As an interim result, a selection of 11 relevant AI roles was derived (data scientist, data engineer, project manager, software developer, decision maker, HR developer, change manager, domain expert, subject matter expert, AI consultant and AI potential researcher). The user committee used these roles to identify proper interview partners in their companies. 13 interview partners from 9 companies of the user committee were identified. Because some interview partners were able to cover more than one role, interviews for all roles were conducted. 97 tasks were identified and described with the knowledge of the experts as well as further information from the literature. Afterward, the tasks were grouped and assigned to the lifecycle. The user committee validated the results.

Development of a Task Model for Artificial Intelligence

533

1.1 Identify data science frameworks 1. Design an AI application lifecycle

1.2 Extract phases from the frameworks 1.3 Create a lifecycle based on the existing phases 2.1 Identify roles in an AI project 2.2 Identify role owners in the user comittee

2. Identify tasks in the lifecycle

2.3 Identify and describe tasks in AI projects 2.4 Assign tasks to the lifecycle 2.5 Validation of the lifecycle and tasks with the user committee

Legend Literature research

Expert interview

Content analysis (Mayring)

Fig. 1. Approach

4 Phases and Tasks in the Lifecycle of an AI Application Through the creation of the AI application lifecycle, our aim is to provide a more comprehensive and adaptable guide for the successful development and usage of AI applications in SMEs. This lifecycle serves decision-makers of SMEs as an overview to avoid false estimations while planning the development and use of AI applications. The lifecycle uses a three-level phases, tasks, and sub-tasks framework that provides a hierarchical structure for organizing and categorizing project activities. At the highest level, the phases represent major stages in the application lifecycle. Tasks, which fall under each phase, delineate specific objectives or goals to be accomplished. Lastly, sub-tasks provide a detailed breakdown of the tasks, outlining the specific actions or sub-objectives required to complete each task. The AI application lifecycle comprises two main decision points as well as six phases which are presented below. However, the development of AI applications is an iterative process, where each phase informs and influences the others (see Fig. 2). For example, if the environment of the AI application changes over time in the phase “Monitoring & Operation” the applied model might have to be re-trained in the phase “Model Training”. That means it must be adapted to the changed database.

534

F. Clemens et al.

Operation decision

Planning & System Design

Data Collection & Processing

Model Training

Monitoring & Operation

Verification & Integration in Validation production

Disposal decision

Fig. 2. AI application lifecycle

AI applications not only involve a significant amount of data processing and analysis. Their development requires a combination of expertise from different domains [15]. Therefore, the lifecycle of an AI application is more complex and requires a more detailed view of the different tasks involved (see Fig. 3). By identifying the specific tasks in the lifecycle of an AI application and assigning them to the different phases, it is ensured that all the necessary steps are considered while planning the development and implementation of an AI application. Planning & Operation Decision System Design

Data Collection & Processing

Model Training

Verification & Integration in Validation Production

Initialize Record as-is IT project architecture Requirements engineering

Monitoring & Operation

Disposal Decision

Ensure operation of the AI system Digitize manual Establish data entries data pipeline

Design AI system

Establish AI system

Integrate models in AI system

Develop machine learning models Connect the domains data science and business Adjust the organization

Assess the organizational impact

Support change Coordinate trainings Close project Communicate (interim) results Control project Plan and control resources

Fig. 3. Tasks in the lifecycle of an AI application

4.1 Operation Decision The “Operation Decision” in an AI application lifecycle refers to the management decision made on whether an AI application should be developed and used in the company. The operation decision includes the following tasks. Before the decision is made, the economic impact and the required resources to conduct the project successfully must be estimated. Furthermore, SMEs must be aware of the objectives they want to achieve by implementing an AI application. These subtasks

Development of a Task Model for Artificial Intelligence

535

are included in the tasks “Plan and control resources” as well as “Control project”. Right after the decision is made, the project must be initialized, and SMEs must communicate the decision. Furthermore, the effects on the organization should be assessed to be able to derive necessary changes that need to be made in the affected processes. The controlling, communication and the planning and controlling of resources are ongoing tasks (see Fig. 3). In conclusion, the operation decision is a critical step in the AI application lifecycle, as it involves deciding whether to develop and use an AI application within the organization. To make an informed decision, SMEs must carefully consider the economic impact, goals, and resources required for the project. 4.2 Planning and System Design The phase “Planning & System Design” is the first phase of the AI application lifecycle. The requirements engineering includes the continual management of requirements regarding the back-end functionalities, front-end, data and IT-architecture. To achieve the greatest possible acceptance of the AI application, SMEs should involve their affected employees in this task. Moreover, the consideration of the IT-architecture is highly relevant as projects tent to fail because the machine learning models often remain prototypes [16]. That means they are not integrated in the productive environment such as processes and IT-systems. Furthermore, the requirements engineering includes the consideration of ethical, data security and cyber security issues. It includes analyzing the privacy and personal data protection implications of the AI application and verifying compliance with laws and regulations. It is important to assess these aspects in advance to ensure that the AI application is implemented in an ethical, privacy, and legal manner. Ethical aspects may include avoiding decisions based on characteristics such as age, gender, race, or ethnicity that violate applicable laws or ethical standards [17]. Another important element from a compliance perspective is data protection. Here, laws and regulations must be complied with to ensure the protection of personal data, including the processing, storage, and use of data. A further aspect is compliance with laws and regulations that apply to the AI application, including compliance with industry standards. The recording of the as-is IT architecture is an additional task in this phase. Depending on the AI application to be developed, it includes the check of the current network infrastructure, available machine, or product interfaces as well as available sensory. Furthermore, the application related IT system landscape must be captured. That includes the recording of the IT systems, the interfaces of these systems, and the provided data. Another important task during the phase of planning and system design is to connect the different views of the departments in SMEs, normally the data science view and business view. SMEs must gain an understanding of the field of application of the AI application to be able to conceptualize an architecture and get an idea which data should be collected. Based on the requirements and as-is IT architecture the AI application must be designed. First, the data model must be developed. It must be clarified which data will be used for the AI application and where it will be stored. That includes the decision whether to use a cloud, hybrid, or on-premises architecture.

536

F. Clemens et al.

Afterwards, the required software components (e. g. libraries, interfaces) and hardware components (e. g. server) will be selected. Cybersecurity-relevant requirements must always be considered in this step. It must be clarified to what extent data must be encrypted and access rights must be clarified. In addition, further cybersecurity measures should be taken depending on the sensitivity of the processed data. The “Planning & System Design” phase sets the foundation for the successful development of an AI application by identifying the requirements, ethical considerations, and compliance regulations that must be adhered to. By carefully analyzing these aspects in advance, organizations can ensure that their AI applications are implemented in a responsible and legally compliant manner, while also meeting their needs. 4.3 Data Collection and Processing The phase “Data Collection & Processing” is a crucial step in the AI application lifecycle. In SMEs in particular, processes are often not sufficiently digitalized, and data is collected manually [18]. The primary objective for the use of an AI application is the automation of the data collection to ensure good data quality. Whether there are possibilities for automation, e.g., through sensor technology, is case-dependent. Another important foundation for good data quality is to obtain time-synchronous data. For example, the manual management of data on handouts, which are transferred to an IT system at the end of the shift, leads to great asynchrony with sensor data in a production process. Such process steps should be adapted or automated. Useful measures are the training of employees, demonstrating the purpose of data collection, and in some cases, gamification approaches, helpful to ensure time-synchronous and high-quality data. Digitalization is the basis for the use of AI applications. More complex digitalization projects should be conducted before the implementation of an AI application is considered. Once the base for the implementation of the AI application is set, the designed architecture must be established. This task includes the implementation of interfaces to existing IT systems and the connection of machines or products through sensory. Furthermore, the SMEs must set up the required hardware, such as servers, edge computers and cloud components. At the same time, the SMEs must establish the data pipeline. That includes the identification of data sources and the filtering and pre-processing of data. Often, raw data is not usable. Accordingly, filtering of relevant data must be performed in the data pipeline (e. g. already on a server next to the machine). Especially for time series data (e. g. vibration data) the Fourier Transformation is suitable to store only a spectrum of the collected values [19] or simple statistical values (e. g. average, max, min) of a certain timeframe. It is recommended to just store raw data when it is not yet clear what AI application shall be developed. The collected and stored data must be evaluated from the business perspective. This includes enriching data with context such as start-up situations for machines, bridge days or batch changes. This information helps to identify the valuable data for the correct training of the model in subsequent phases. In conclusion, the phase of data collection and processing plays a crucial role in the successful implementation of AI applications in SMEs. Digitalization is the foundation

Development of a Task Model for Artificial Intelligence

537

for automating the data collection process, which ensures good data quality. The establishment of the data pipeline and the identification of relevant data sources are also vital for the development of effective AI models. By evaluating the collected data from a business perspective and enriching it with context, SMEs can identify the valuable data necessary for training the model. 4.4 Model Training The phase “Model Training” is a crucial phase in the development of an AI application. In this phase, it is important to continuously improve and evaluate the model’s performance and make necessary adjustments to improve its accuracy and efficiency. Successful model training is a critical step towards the successful implementation of AI applications. The task “Develop machine learning models” includes various subtasks that overlap with the phases before and after the model training. The data preparation and cleansing are important steps in the process of data processing. They involve a variety of activities aimed at bringing the data into a format suitable for further analysis and use. The data consolidation involves merging data from different sources to ensure consistency and completeness. The data cleansing involves removing duplicates, invalid or erroneous data, as well as cleaning up data that is not necessary for analysis. Enrichment involves adding additional information to the data, e.g., by adding geographic coordinates or demographic data. Validation involves checking the data for plausibility and integrity to ensure it is correct and complete. The tools and technologies that are used to perform these tasks can vary depending on the scope and scale of the data, but they may use popular tools such as Python libraries such as Pandas and Numpy. [20]. Exploratory data analysis helps to better understand the data and underlying problem and create a better foundation for developing suitable machine learning models. During exploratory data analysis, various techniques such as statistical analysis, visualizations, and machine learning methods are used to examine and understand the data. Visualizations can be used to discover and interpret patterns and trends in the data during exploratory data analysis, monitor the performance of the model during development. [20]. Feature engineering refers to the process of creating, selecting, and modifying features for use in a machine learning model. Features are the variables or input attributes that the model uses to make decisions or create forecasts. Feature engineering is an important step in the development of machine learning models, as the quality and number of features used can directly affect the performance of the model. Feature engineering often requires a deep understanding of the problem and data and can require many iterations and experiments to find the best features for a model. [20]. Once a model is trained, its performance is verified which will be part of the following chapter.

538

F. Clemens et al.

4.5 Verification and Validation Especially during the development of a machine learning model, SMEs will iterate between the phases “Model Training” and “Verification & Validation” because a crucial part of the development of a machine learning model is the evaluation of the model. The evaluation of trained models refers to the process of assessing the performance of a machine learning model using metrics such as accuracy, error rate, and other metrics. This step is important to determine how well the model solves the given problem and to decide if further optimization is needed. [20] During the development many iterations between the two phases are necessary as many parameters effect the performance of the model while each adjustment requires once again the training as well as the evaluation. While the first models are trained and evaluated, organizational tasks can be started. Once the AI application is ready for the integration, the employees must be educated to use it. To ensure that, SMEs must coordinate trainings regarding the use of the AI application as well as skills such as data literacy. If the employees understand how the AI application works, the acceptance and trust will likely increase [21]. Furthermore, the change of the adjustment of the organization should be supported as processes and workflow will change due to the AI application. Practical experience has shown that especially enterprises that invested a high budget in the change management in AI application developments were successful with it [22]. In conclusion, the development of a machine learning model involves a continuous iteration between the phases of model training and verification & validation. The evaluation of trained models is a crucial step to determine their performance and identify opportunities for optimization. With proper planning and investment in change management, organizations can successfully leverage AI to improve their operations and gain a competitive edge in their respective industries. 4.6 Integration in Production In the phase “Integration in Production” the main objective is to integrate the machine learning model into the established architecture as well as test it. During integration, the machine learning model is first connected to the data sources it needs for processing and decision-making. The data from the sources is extracted for this purpose and passed to the model. Subsequently, the model must be called, and the obtained results returned to the application. To ensure smooth operation, mechanisms are developed to detect and handle errors in the model or in the integration logic. In addition, it may be necessary to perform optimizations to ensure that the model performs as required in the production environment. Finally, measures are developed and implemented to ensure that the model and the data it works with are protected from external influences. Subsequently, the AI application must be tested. The testing of a developed machine learning model in a production environment is crucial to ensure its accuracy, reliability, and robustness [20]. The testing process involves preparing necessary test data, functional, performance, and robustness testing, application testing, error handling, and documentation. Especially in this phase the described organizational tasks are crucial as it decides if the AI application will be accepted by the employees or not.

Development of a Task Model for Artificial Intelligence

539

In conclusion, the “Integration in Production” phase is a critical step in the deployment of an AI application. It involves connecting the machine learning model to the necessary data sources, testing it, and implementing measures to ensure its protection. Ultimately, the success of this phase depends on the effective collaboration of all stakeholders involved in the process, as it determines whether the AI application will be embraced and accepted by employees. 4.7 Monitoring and Operation The phase “Monitoring & Operation” includes a variety of tasks. This phase is often underestimated by SMEs as the development processes normally focus the development and deployment but not on the usage of AI applications (see Sect. 2.1). SMEs must monitor the established data pipeline to ensure that the data flowing through the pipeline is processed correctly and that the pipeline itself is running stably and reliably. This can be achieved by monitoring metrics such as turnaround times, error rates, and data quality metrics (e.g., completeness), as well as monitoring resource utilization and the status of each component in the pipeline [23]. In the event of an error or failure, a quick response can then be made, and errors corrected. Moreover, SMEs must maintain and optimize the data pipeline. Maintaining a data pipeline means regular care and maintenance of the pipeline to ensure that it runs stably and reliably and that the data is processed correctly [23]. This includes updating dependencies and libraries to ensure that the pipeline has the latest security updates and bug fixes. Optimizing a data pipeline involves various measures to improve the performance and reliability of the pipeline and minimize resource utilization. For example, profiling code can identify which parts of the pipeline take the most time and optimize them [23]. Furthermore, if certain steps of the pipeline take time, it may be useful to parallelize them or scale the resources for execution. In addition, adjusting the configuration of the pipeline can help optimize performance and minimize resource utilization. Not just the data pipeline must be monitored. Also, the applied model requires continual adjustments. The subtask refers to improving and evolving the model used in the AI application to enhance the performance and extend or adapt its capabilities. Existing machine learning models can be optimized in operation by using new algorithms and techniques or by having a larger data set available. Moreover, existing machine learning applications are adapted to changing requirements or environments to ensure that the AI application remains effective and reliable. Not just metrics but especially the feedback of the benefit of the AI application from employees can help to identify potential for improvement. Furthermore, the IT-architecture should be monitored and maintained. Included are the monitoring of hardware performance, the execution of maintenance work, the troubleshooting of e.g., servers, edge computers, computing capacities, or storage solutions. The task refers to ensuring the smooth operation of the hardware and software components required to implement an AI application. It may also include monitoring the scalability and performance of the application, as well as monitoring the integrity and security of the data. This involves building and monitoring the IT architecture of the AI application so that it runs securely and stably and is continuously available and scalable.

540

F. Clemens et al.

In conclusion, SMEs must recognize the importance of the “Monitoring and Operation” phase in AI development. Proper monitoring and maintenance of the data pipeline, the applied model, and the IT architecture are crucial to ensure the stability, reliability, and scalability of the AI application. This requires continual adjustments, optimization, and adaptation to changing requirements and environments. By prioritizing the “Monitoring and Operation” phase, SMEs can maximize the benefits of their AI application and avoid potential problems or failures in the future. 4.8 Disposal Decision The “Disposal Decision” in the lifecycle of an AI model refers to the process of deciding what to do with the model once it has reached the end of its useful life or is no longer needed. There are two options for disposing of an AI model, including archiving, or deleting the model. The disposal decision should be made based on the specific needs and requirements of the organization, also considering potential future needs.

5 Conclusion In conclusion, the success of an AI application relies not only on AI-specific tasks but also on a multitude of non-AI-specific tasks that are essential for its seamless integration into the broader ecosystem of SME. These tasks encompass a wide spectrum of activities, from designing and implementing robust IT infrastructure to effectively communicate project results to stakeholders (see Sect. 4). Further research may also point out the relevance of AI-specific skills these non-AI-specific tasks. In the context of SME, it is important to acknowledge that they may not be able to undertake all tasks of AI projects themselves as the diverse nature of AI-specific and non-AI-specific tasks requires a wide range of skills and knowledge that may not be readily available within SMEs. As a result, SMEs will face the challenge of creating their own company-specific roles based on the identified tasks in Sect. 4 by assigning them to their employees. In combination with the aimed assignment of AI-specific skills to the tasks this evaluation can help to identify areas where they may lack the necessary skills to successfully execute AI projects. Thus, SMEs can gain a clearer understanding of the skills they need to acquire externally or gain through trainings.

6 Outlook The next step in the VoBAKI project is to identify and describe the necessary skills for carrying out the identified tasks in the lifecycle of an AI application and their associated tasks. This will provide SMEs with a clear understanding of the required skill sets for successfully implementing and maintaining an AI application. By identifying these skills, SMEs will be able to evaluate their own capabilities and identify any gaps that need to be filled. This will enable SMEs to make informed decisions about whether to invest in building these skills in-house or to seek external support. Ultimately, this will help SMEs to successfully implement AI applications that are efficient, effective, and reliable, and that deliver real value to their organizations.

Development of a Task Model for Artificial Intelligence

541

Acknowledgements. We would especially like to thank the user committee for attending several interviews and providing us with insights and feedback that helped to shape our research. Their contributions were essential in validating the results of our study and ensuring their relevance to the practical needs of SMEs. Moreover, we would like to express our gratitude to the Federal Ministry for Economic Affairs and Climate Action for their support of VoBAKI (IGF-Project No.: 22009 N). Their funding and support were essential in enabling us to carry out this research and develop practical recommendations for SMEs on the lifecycle management of AI applications.

References 1. Azevedo, A.: Manuel Filipe Santos (2008) KDD, SEMMA and CRISP-DM: a parallel overview. In: IADIS European Conference on Data Mining 2. Keim, D., Sattler, K.-U.: Von Daten zu KI – Intelligentes Datenmanagement als Basis für Data Science und den Einsatz Lernender Systeme Von Daten zu KI – Intelligentes Datenmanagement als Basis für Data Science und den Einsatz Lernender Systeme (2020) 3. Seifert, I., Bürger, M., Wangler, L., et al.: Potenziale der künstlichen Intelligenz im produzierenden Gewerbe in Deutschland. Hg. v. Begleit-forschung PAiCE (2018) 4. Martinez-Plumed, F., Contreras-Ochando, L., Ferri, C., et al.: CRISP-DM twenty years later: from data mining processes to data science trajectories. IEEE Trans. Knowl. Data Eng. 33, 3048–3061 (2021). https://doi.org/10.1109/TKDE.2019.2962680 5. Chapman, P., Clinton, J., Kerber, R., et al.: CRISP-DM 1.0: Step-by-step data mining guide (2000) 6. Freitag, M., Kück, M., Ait Alla, A., et al.: Potenziale von Data Science in Produktion und Logistik: Teil 2 - Vorgehensweise zur Datenanalyse und Anwendungsbeispiele. Ind. 4.0 Manag. 35, 39–46 (2015) 7. Schröer, C., Kruse, F., Gómez, J.M.: A systematic literature review on applying CRISP-DM process model. Procedia Comput. Sci. 181, 526–534 (2021). https://doi.org/10.1016/j.procs. 2021.01.199 8. Fayyad, U., Piatetsky-Shapiro, G., Smyth, P.: From Data Mining to Knowledge Discovery in Databases (1996) 9. Göhring, A.: Anwendung KDD-basierter Methoden zur Interpretation multi-dimensionaler Isotopen-Fingerabdrücke, Ludwig-Maximilians-Universität München (2019) 10. Mezger, M., Schindler, W.: Zur unscharfen Klassifikation von Datenbanken mit fCQL (2003) 11. Wunderlich, J.: Web Mining im Dienste der Personalisierung — der Schlüssel zur digitalen Kundennähe. In: Salmen, S.M., Gröschel, M., Schel, M.G. (eds.) Handbuch Electronic Customer Care: Der Weg zur digitalen Kundennähe, pp. 89–103. Springer-Verlag, Berlin (2004) 12. Alawi, J.S.: Predicting Student Performance Using Data Mining Approach: A Case Study in Oman (2022) 13. Tariq, H.I., Sohail, A., Aslam, U., et al.: Loan default prediction model using sample, explore, modify, model, and assess (SEMMA). J. Comput. Theor. Nanosci. 16, 3489–3503 (2019). https://doi.org/10.1166/jctn.2019.8313 14. Mayring, P.: Qualitative Inhaltsanalyse: Grundlagen und Techniken, 12, überarb Beltz, Weinheim (2015) 15. Sambasivan, N., Veeraraghavan, R.: The deskilling of domain expertise in AI development. In: Barbosa S (ed.) CHI Conference on Human Factors in Computing Systems. Association for Computing Machinery, New York, NY, United States, pp 1–14 (2022) 16. Homeyer, A., Lotz, J., Schwen, L.O., et al.: Artificial intelligence in pathology: from prototype to product. J. Pathol. Inform. 12, 13 (2021). https://doi.org/10.4103/jpi.jpi_84_20

542

F. Clemens et al.

17. Brendel, A.B., Mirbabaie, M., Lembcke, T.-B., et al.: Ethical management of artificial intelligence. Sustainability 13, 1–18 (2021). https://doi.org/10.3390/su13041974 18. Park, J.: Evaluating a mobile data-collection system for production information in SMEs. Comput. Ind. 68, 53–64 (2015). https://doi.org/10.1016/j.compind.2014.12.006 19. Zhang, Z., Verma, A., Kusiak, A.: Fault analysis and condition monitoring of the wind turbine gearbox. IEEE Trans. Energy Convers 27, 526–535 (2012). https://doi.org/10.1109/TEC. 2012.2189887 20. Vanderplas, J.T.: Python data science handbook: Essential tools for working with data, 1st edn. O’Reilly Media Inc, Sebastopol CA (2017) 21. Saßmannshausen, T.M., Heupel, T.: Vertrauen in KI – Eine empirische Analyse innerhalb des Produktionsmanagements. In: Buchkremer, R., Heupel, T., Koch, O. (eds.) Künstliche Intelligenz in Wirtschaft & Gesellschaft: Auswirkungen, Herausforderungen & Handlungsempfehlungen. Springer Gabler, Wiesbaden, Heidelberg, pp. 169–192 (2020) 22. Fleming, O., Fountaine, T., Henke, N., et al.: Ten red flags signaling your analytics program will fail. McKinsey & Company (2018) 23. Munappy, A.R., Bosch, J., Olsson, H.H.: Data Pipeline Management in Practice: Challenges and Opportunities. In: Morisio, M., Torchiano, M., Jedlitschka, A. (eds.) PROFES 2020. LNCS, vol. 12562, pp. 168–184. Springer, Cham (2020). https://doi.org/10.1007/978-3-03064148-1_11

Indoor Positioning-based Occupational Exposures Mapping and Operator Well-being Assessment in Manufacturing Environment Gergely Hal´ asz1 , Tibor Medvegy2 , J´ anos Abonyi1 , and Tam´ as Ruppert1(B) 1

2

ELKH-PE Complex Systems Monitoring Research Group, Department of Process Engineering, University of Pannonia, Veszprem, Hungary [email protected] Mechatronics and Measurement Techniques Research Group, Research Centre for Engineering Sciences, University of Pannonia, Veszprem, Hungary Abstract. This research was motivated by the need for detailed information about the spatial and contextualized distribution of occupational exposures, which can be used to improve the layout of the workspace. To achieve this goal, the study emphasizes the need for position-related information and contextualized data. To address these concerns, the study proposes the use of Indoor Positioning System (IPS) sensors that can be further developed to establish a set of metrics for measuring and evaluating occupational exposures. The proposed IPS-based sensor fusion framework, which combines various environmental parameters with position data, can provide valuable insights into the operator’s working environment. For this, we propose an indoor position-based comfort level indicator. By identifying areas of improvement, interventions can be implemented to enhance operator performance and overall health. The measurement unit installed on a manual material handling device in a real production environment and collected data using temperature, noise, and humidity sensors. The results demonstrated the applicability of the proposed comfort level indicator in a wire harness manufacturing setting, providing location-based information to enhance operator wellbeing. Overall, the proposed framework can be used as a tool to monitor the industrial environment, especially the well-being of shop floor operators. Keywords: Operator 4.0 · operator well-being sensors · mobile sensor · IPS

1

· environmental

Introduction

The main objective of this paper is to propose an approach for assessing the well-being of operators in indoor environments, with the operator comfort level serving as a measurable indicator. To achieve this, we propose a framework c IFIP International Federation for Information Processing 2023  Published by Springer Nature Switzerland AG 2023 E. Alfnes et al. (Eds.): APMS 2023, IFIP AICT 689, pp. 543–555, 2023. https://doi.org/10.1007/978-3-031-43662-8_39

544

G. Hal´ asz et al.

based on the utility function and information fusion to enhance the assessment of operator well-being. Additionally, we have developed a mobile measurement device that utilizes moving units to obtain a detailed layout of the shop floor. Understanding the specific areas and conditions where workers are exposed to various environmental factors is essential. Distributed information is essential to capture the spatial and contextualized distribution of occupational exposures. Unlike concentrated parameters, this exposure varies between operators and workplaces. To formulate a clear objective, it is necessary to articulate the need for detailed information that encompasses both spatial and temporal dimensions. To achieve this, it is necessary to have position-based information that is contextualized and linked to relevant data. We incorporate indoor positioning data to create location-based sensor information. The application and further development of the Indoor Positioning System (IPS) serve as a solution to address these motivations. By employing IPS sensors and associated analysis techniques, it becomes possible to gather precise position-based information. The integration of IPS sensors and related technologies offers a promising avenue for advancing workplace monitoring and design, ultimately promoting a safer and more efficient working environment. This transition leads us to the core concept of the study, which aims to develop a comprehensive framework for collecting and analyzing precise data that captures the detailed nature of occupational exposures. We employed utility functions to effectively represent the measurement data obtained from the environment. These utility functions serve as mathematical models that assign values or scores to the measured parameters, such as temperature, humidity, and noise levels. By utilizing utility functions, we are able to translate the raw data into a standardized representation that captures the perceived comfort level of the operator. The utility functions play a crucial role in quantifying the operator’s comfort by assigning higher scores to favorable conditions and lower scores to unfavorable conditions. This allows us to assess and compare different environmental scenarios in terms of their impact on the operator’s well-being and comfort. For this characterization, we have determined a so-called comfort level. The value represents the extent to which the operator can handle environmental load in each zone of the shop floor and how comfortable they feel in these areas. To measure this new indicator, we have also developed a mobile sensor unit as an extension of the standard IPS tag to make the measurement more cost-effective. With this idea, we can utilize all the moving units on the shop floor, such as AGV (automated guided vehicles) or AMR (autonomous mobile robots), or even the material handler operators equipped with the developed sensor. As a result, we can identify specific areas that require attention or modifications to create a more comfortable and conducive workspace for the operators. The utilization of utility functions in representing the measurement data greatly contributes to our understanding of the operator’s comfort and aids in the development of strategies to prioritize their well-being. To demonstrate the

Operator Well-being Assessment

545

applicability and effectiveness of the proposed framework, a detailed case study is conducted. The state-of-the-art is described in Sect. 2. In Sect. 3, the developed sensors and the proposed comfort level indicator are detailed. Section 4 presents a wire harness industrial case study where we demonstrate the applicability of the developed mobile sensor unit and the relevance of the proposed comfort level indicator.

2

Assessments of Well-being and Comfort Level

The aim of Operator 4.0 is to give operators all the information they need to discover, use, and benefit from support systems that can help them increase their critical skills for their work, thus increasing their productivity [10]. With these in mind, Operator 4.0 is intrinsically linked to the comfort of operators, minimizing negative influences and negative stimuli in their working environment. A sub-type of this is Healthy Operator 4.0, which aims to assess the stress levels and potential physical risks of the worker in a cyber-physical environment. To achieve this, it uses wearable devices to monitor the health and well-being of operators. The measured information will be used to model the operator’s habits and behavior, which will help to optimize the operator’s working environment. The idea behind Resilient Operator 5.0 is to combine human creativity, ingenuity, and manufacturing-informed innovation with new, sustainable, energy-saving ways to keep operators thriving in the face of exhausting and unpredictable conditions. Resilient Operator 5.0 can be broken down into two main groups, Self-resilience and System-resilience. Self-resilience broadly addresses the health, safety, and productivity of operators on the shop floor. Self-resilience is a powerful toolkit for protecting the health of operators. These include a variety of intelligent health wearables, personal protective equipment, exoskeletons technology, augmented reality technology, and virtual reality technology [8]. Well-being itself is a multi-component concept, and can refer to an individual’s sense of happiness, as well as their comfort in a particular environment or situation, their health, or their financial security. A comprehensive summary of worker well-being is provided constructs, categorizes them, and offers a brief characterization based on the relevant academic literature [16]. Worker wellbeing encompasses the overall well-being of working individuals, distinguishing it from concepts like employee well-being and well-being at work. It also differs from work-specific well-being, which focuses on constructs originating and applicable within the work context [16]. Well-being can be assessed through both objective measures related to the“standard of living” and subjective measures based on an individual’s cognitive and affective judgments about their life, with objective aspects influencing subjective well-being levels [14]. A comprehensive overview of worker well-being, including its multi-faceted nature and various constructs is done in Ref. [16]. It addresses questions about the definition of worker well-being, measurement approaches, and considerations for selecting appropriate measures,

546

G. Hal´ asz et al.

aiming to enhance researchers’ understanding and facilitate effective strategies for promoting worker well-being. Among these factors, we have examined and quantified the sense of comfort. Acoustic comfort goes beyond simply providing an acceptable indoor environment for occupants; it involves the examination of various factors that influence acoustic comfort and mitigate discomfort [3]. Comfort is the combination of many factors so that even people who are in the same geographical area at the same time and the same age can also show conflicts in their perception of thermal comfort. The lack of inadequate indoor comfort can cause dissatisfaction in the occupants and negatively affect their productivity and performance, as well as various problems such as dryness, health and morality. These factors collectively contribute to overall comfort, which in turn affects the productivity and energy consumption of the factory, ultimately influencing its profitability. Several investigations have shown that the overall comfort of occupants depends on the ambient conditions, the characteristics of the building, and the individualities linked to the occupant. Human productivity is strongly believed to be highly dependent on acoustic, visual, thermal, and air quality conditions of the built environment [3] It follows that it would be optimal for the industry if the operator felt as comfortable as possible at work, thus preventing health problems and improving performance. The temperature, humidity and noise levels are usually measured considered the comfort-level [12]. Occupational hazards [18] and exposures [11] play a significant role in determining the well-being and comfort level of operators. Understanding and managing these occupational hazards is crucial for creating a safer and more conducive working environment that promotes operator wellbeing and ensures optimal comfort levels during their daily tasks [2]. However, the thermal sensation is not significantly altered by humidity. Too low a humidity level will result in dry air, which will cause various respiratory problems on a closed and hot shop floor. Too high, on the other hand, increases perspiration, can cause breathing difficulties, and promotes mold growth on the shop floor, thus increasing the risk of allergies and reducing comfort levels. Noise, as a significant factor, can greatly impact the ability of individuals to work effectively, with high noise levels hindering concentration. To ensure optimal working conditions, it is crucial to measure and control noise levels, as prolonged exposure to noise can lead to severe issues. For instance, even a noise level of 80 dB can result in long-term hearing damage, while 140 dB can cause immediate harm. To mitigate these risks, different protective measures are utilized, emphasizing the importance of identifying areas with high noise levels. Local exposure refers to the level of exposure to a particular environmental harmful substance [17], such as noise, in a specific location or area. An example of local exposure to noise in an industrial setting could be the noise level experienced by workers on a factory floor near a particularly loud machine. In this case, the noise exposure is specific to the workers in that area and may be higher than in other parts of the factory. Daily exposure [9], on the other hand, refers to the total level of exposure to a particular environmental harmful substance, such

Operator Well-being Assessment

547

as noise, over a period of time, typically a day. An example of daily exposure to noise in an industrial setting could be the noise level experienced by workers throughout their workday, including breaks and rest periods. In this case, the noise exposure is not specific to any particular location but rather represents the total exposure throughout the day. The perception of comfort is broadly defined for the indoor environment of residential buildings, offices, hospitals, schools, workshops, shopping centers, etc. Maintaining comfort in the occupied zone is the daily activity of man. The average person in an urban area spends about 85–90% of the time indoors. [3] To accurately measure and assess environment information, which serve as crucial indicators of cognitive and physical well-being in individuals [6], the utilization of an IPS becomes essential. By employing a location-driven sensing approach, the monitoring and measurement of these activities can be effectively carried out, enabling a comprehensive understanding of individuals’ well-being and supporting personalized care and intervention strategies. Sensors and wireless sensor networks have emerged as a vast research and development domain, garnering significant attention and prompting numerous reviews and studies across various disciplines, due to their potential to revolutionize data collection, monitoring, and control processes by enabling real-time and remote sensing capabilities, facilitating the measurement and analysis of diverse environmental parameters, ranging from temperature and humidity to motion and pollution levels, and offering extensive applications in fields such as environmental monitoring, healthcare, agriculture, infrastructure management, and industrial automation [1,4,5,13]. In our study, we propose a novel method to measure the operator comfort by utilizing environmental data and mobile measurement units, eliminating the need for individual sensor units for each person. By incorporating indoor positioning data, we can assess the real-time conditions in the manufacturing space, capturing both the spatial and temporal aspects of operator well-being. This approach not only enhances our understanding of operator comfort but also enables us to efficiently evaluate the layout, identifying critical areas that require attention from a well-being perspective. The integration of environmental sensors with the IPS ensures a cost-effective and continuous monitoring process, underscoring the significance of ongoing evaluations in promoting operator well-being. The developed framework and the proposed mobile measurement unit is introduced in the next section.

3

Developed Framework to Discover the Location-based Comfort Level

In this section, we employ temperature, humidity, and noise sensors, accompanied by time and position stamps, to capture comprehensive data on the working environment. To ensure a holistic understanding of the operator’s experience, it is crucial to fuse this information effectively. Therefore, we propose a comprehensive methodology for data fusion based on the utility functions, integrating the various sensor inputs to provide a unified and contextualized analysis (Sect. 3.1).

548

G. Hal´ asz et al.

In addition to the sensor development, we have also designed and implemented a location-based evaluation technique to enhance the accuracy and relevance of the collected data (Sect. 3.2). This involves capturing the precise position in relation to the environmental parameters, allowing for a more nuanced assessment of the operator’s exposure and comfort level. The integration of these sensor technologies and the development of a location-based evaluation technique significantly contribute to the overall effectiveness of the proposed framework (Sect. 3.3), enabling a more comprehensive analysis of the working environment and its impact on operator well-being and performance. 3.1

Methodology

In this section, we present the methodology for fusing the sensors we use. The sensor fusion methodology is represented in Fig. 1. This framework describes the sensor information process, where we have a measurement point that includes humidity (ht ), temperature (wt ), the noise (st ) and the actual position data (pt ). 2D position from the IPS is denoted by pt = [xt , yt ]. The noise level is denoted by st , the temperature is wt and the humidity is ht . Here, t is the actual time point. We defined the values of the three environmental sensors as utility functions to define the mathematical relationship that combines the different variables to calculate the overall utility to get the comfort level. Figure 2 shows these three utility functions. First, the utility function of the temperature is calculated on the basis of the following equation: ⎧ ⎪ ⎪ ⎨0,   if wt < wmin Utw =

min ⎪ ⎪ ⎩0,

wt 1 ||wideal −wt || wideal , 1

, if wmin ≤ wt ≤ wmax

(1)

if wt > wmax

where Utw is the utility function for temperature, wt is the temperature in the t-th timepoint and wideal is the ideal temperature assign to the maximum utility

Fig. 1. The steps of the developed framework

Operator Well-being Assessment

549

Fig. 2. Utility functions of the temperature (left), humidity (middle) and noise (right)

score (1). In Fig. 2 (left), the ideal temperature is 22C ◦ , wmin is 15C ◦ , and wmax is 29C ◦ . The utility of the humidity (Unh ) is calculated on the same formula as we did with the temperature (Unw ). In Fig. 2 the ideal humidity (middle) is 50%, the minimum (hmin ) and maximum (hmax ) humidity are 20% and 80%, respectively. The utility function of noise is different, as there is some standard limit of the noise acceptance. In Fig. 2 (right), scomfort is 60 dB and smax is 110 dB. ⎧ ⎪ ⎪ ⎨1,   if st < scomfort s sn 1 Ut = min ||scomfort −sn || scomfort , 1 , if scomfort ≤ st ≤ smax (2) ⎪ ⎪ ⎩0, if s α

(b) ry offset > β

(c) rz offset > γ

(d) rx offset

Fig. 5. The impact of geometrical parameters on the finished product quality. α, β, γ, δ represent the optimal values for considered parameters.

By adhering to these considerations, we can ensure efficient and reliable parameter identification and usage in the solution’s present and future applications. Geometrical parameters are used to find the perfect alignment between the tweezers and the grinding belt. Figure 5 reports examples from sub-optimal robotic belt polishing processes. To see if the parameters are properly tuned, the operator can paint tweezers with a marker and set in offset to 0 to have the minimum contact force possible. Then, the polishing process can evaluate if the marker has been properly removed. The goal behind the BRILLIANT App is to support the operator in finding the optimal value of the parameters through a simple and user-friendly GUI, together with its guided procedure. The BRILLIANT App provides two different setup procedures. The first deals with mobile station repositioning. The second supports the introduction of new tweezers types. Both procedures support the operator in setting up all the cobotic tasks represented in Fig. 4, providing specific instructions and performing the operations in test mode, mitigating the risks due to improper setup. A video of the work cell and the BRILLIANT App is available at YouTube. With its modular structure and architecture, the BRILLIANT App makes it easy to apply the same approach and adapt the solution to other robotics

712

E. Montini et al.

Fig. 6. BRILLIANT App user interfaces. Behind, the station setup wizard. In front, the operational dashboard.

applications and processes, reducing the need for extensive reprogramming or customisation. Part programs, procedures and the GUI can be easily adapted to different applications and contexts, allowing developers to quickly reconfigure tasks and parameters through configuration files, saving development cycles. These features will be exploited when porting the BRILLIANT App to new applications. Indeed, thanks to its flexibility, the BRILLIANT App is also suitable for other types of tasks and parts, as demonstrated by the different processes (e.g., part handling and positioning, grinding) involved in the presented use case. In general, the solution and the methodology behind the application are well-suited for applications that can be managed by varying key points positions and/or key parameters (e.g., number of cycles, forces, speed) before starting the motion.

5

Validation and Results

The smart work cell has been extensively tested at the tweezers manufacturing company. The tests were assessed against the key goals reported in Sect. 2. The main results are increased productivity, increased well-being for workers and increased production system flexibility. Increased Productivity. Despite the cycle time having been slightly increased (+10%), the productivity of the work cell, due to an increment of parts quality, increased by 30%. Throughput time and WIP have been relevantly reduced. The test performed on four types of tweezers and various batches showed a 97% success rate. The process takes between 1.85 to 2.2 min, depending on parameters and band conditions. Continuous monitoring and reduced workload on the operator allowed for more focus on routine maintenance tasks.

A Smart Work Cell to Reduce Adoption Barriers of Collaborative Robotics

713

Increased Well-being. The process was composed of eight tasks; six of them are now executable by the cobot. The cobot has been assigned to welding, tweezers aligning, tail grinding, and polishing welding spots, which are repetitive tasks. The operator is now in charge of the initial single plate cleaning, the supervision of the work cell and quality control. This leads to a reduction of 75% of the repetitive and non-ergonomic activities. Increased Flexibility. The solution has been tested with four different types of tweezers, with completely different shapes and materials. Thanks to the BRILLIANT App, the operator can now set up new types of tweezers with a structured and quick procedure, requiring less than 30 min. Moreover, by re-positioning the collaborative robot in the workstation, he/she can easily calibrate all the processes. An operator has tested and completed ten setups, requiring 10 min (e.g., to calibrate the parameters) to run the process with the expected quality and reliability. The validation scenario in the manufacturing sector, characterised by skilled operators who excel in their job but lack expertise in robotics, has provided compelling evidence that the proposed solution effectively addresses the research questions outlined in 1 and 1. Regarding question 1, the validation results have solidified that a well-designed support system and effective program parameterization enable operators to work effortlessly with a user-friendly dashboard. Importantly, this approach proves to be significantly simpler than traditional robot programming methods. For research question 1, an exploration of force control strategies for the belt polishing process has been undertaken. However, a feasibility analysis has identified several challenges hindering its implementation. One such challenge is the lower accuracy of the F/T sensor installed on the UR10, which falls short of the application requirements and complicates force control implementation. High-quality F/T sensors with the desired precision typically cost around 4.000€-5000€, which accounts for over 10% of the cobot’s price. Despite this setback, the combination of position control, a geometric parameterization of the process, and an indirect force compensation derived from a flexible beam approximation of the tweezers have successfully led to the attainment of the expected results.

6

Conclusions

The proposed solution is crucial for obtaining high-quality tweezers with small batch sizes and collaborative robot assignments to multiple workstations. Without the proposed step-by-step procedure for fine-tuning the program parameters and smart solution, finding the optimal parameters can require up to 1 h of trials, leading to delays and errors that impact the quality of the final product. The proposed work also demonstrates collaborative robots’ capability to achieve humans’ flexibility and dexterity with the repeatability of cobots towards artisanal manufacturing 4.0.

714

E. Montini et al.

The system exhibits certain limitations in its current state of development and operation. The operators’ knowledge and experience can limit the ability of the cell to achieve optimal performance. Insufficient understanding of the process or lack of expertise among operators can hinder proper parameterization and attaining optimal production levels. Additionally, human variability in parameter settings may introduce inconsistencies in the production process, potentially affecting product quality and system performance consistency. Manual parameter setup by operators can also result in prolonged setup times compared to fully automated systems, reducing the system’s flexibility to adapt to rapid production changes or product variety. The system’s scalability may be constrained by the need for highly skilled and specifically trained operators to set parameters, making it challenging to replicate the system across multiple cells or production lines without adequate expert operators. Furthermore, the current implementation of the collaborative robotic cell lacks integration with existing enterprise information systems, such as machinery or production management software. This absence of connectivity can give rise to challenges and potential issues regarding data exchange, coordination, and synchronisation. To overcome these limitations, the next steps aim to enhance operator support during the setup phase of the work cell by developing a solution that combines AI-based optimisation, active learning and integration with enterprise information systems. This comprehensive approach will facilitate knowledge transfer from the operator to the robotic system, leveraging machine vision technology to analyse images of sub-optimal machining. The optimisation tool, integrated with the BRILLIANT App, will provide the operator with optimal parameters for robotic belt grinding, considering real-time data from the interconnected enterprise information systems. Moreover, the BRILLIANT App will be integrated with the broader organisational information infrastructure. This integration will enable the smooth flow of information, enhance decision-making processes, and foster a more interconnected and intelligent manufacturing environment. Acknowledgements. This work was supported by EU’s Horizon 2020 research and innovation program (Grant numbers 825196).

References 1. Aaltonen, I., Salmi, T.: Experiences and expectations of collaborative robots in industry and academia: barriers and development needs. Procedia Manufact. 38, 1151–1158 (2019). https://doi.org/10.1016/j.promfg.2020.01.204 2. Abdous, M.A., Delorme, X., Battini, D., Sgarbossa, F., Berger-Douce, S.: Assembly line balancing problem with ergonomics: a new fatigue and recovery model. Int. J. Prod. Res. 61, 1–14 (2022). https://doi.org/10.1080/00207543.2021.2015081 3. Bettoni, A., et al.: Mutualistic and adaptive human-machine collaboration based on machine learning in an injection moulding manufacturing line. Procedia CIRP 93, 395–400 (2020). https://doi.org/10.1016/j.procir.2020.04.119

A Smart Work Cell to Reduce Adoption Barriers of Collaborative Robotics

715

4. Daniele, F., et al.: Collaborative robotics adoption: KITT4SME report 2022 in collaboration with trinity robotics. Tech. Rep., University of Applied Science and Arts of Southern Switzerland (2022). https://doi.org/10.13140/RG.2.2.24439.50081 5. Fast-Berglund, ˚ A., Romero, D.: Strategies for implementing collaborative robot applications for the operator 4.0. IFIP Adv. Inf. Commun. Technol. 566, 682–689 (2019). https://doi.org/10.1007/978-3-030-30000-5 83 6. Fritzsche, L., Wegge, J., Schmauder, M., Kliegel, M., Schmidt, K.H.: Good ergonomics and team diversity reduce absenteeism and errors in car manufacturing. Ergonomics 57(2), 148–161 (2014). https://doi.org/10.1080/00140139.2013.875597 7. Liu, D., Cao, J.: Determinants of collaborative robots innovation adoption in small and medium-sized enterprises: an empirical study in China. Appl. Sci. 12(19), 10085 (2022). https://doi.org/10.3390/app121910085 8. Montini, E., et al.: A framework for human-aware collaborative robotics systems development. Procedia CIRP (2023) 9. Niu, Y., Schulte, F.: Human aspects in collaborative order picking – what if robots learned how to give humans a break? In: Dolgui, A., Bernard, A., Lemoine, D., von Cieminski, G., Romero, D. (eds.) APMS 2021. IAICT, vol. 632, pp. 541–550. Springer, Cham (2021). https://doi.org/10.1007/978-3-030-85906-0 59 10. Park, J., Sørensen, L.C., Mathiesen, S.F., Schlette, C.: A digital twin-based workspace monitoring system for safe human-robot collaboration. In: 2022 10th International Conference on Control, Mechatronics and Automation (ICCMA), pp. 24–30. IEEE (2022). https://doi.org/10.1109/ICCMA56665.2022.10011622 11. Raza, M., Malik, A.A., Bilberg, A.: PDCA integrated simulations enable effective deployment of collaborative robots: case of a manufacturing SME. Procedia CIRP 104, 1518–1522 (2021). https://doi.org/10.1016/j.procir.2021.11.256. 54th CIRP CMS 2021 - Towards Digitalized Manufacturing 4.0 12. Salunkhe, O., Fager, P., Fast-Berglund, ˚ A.: Framework for identifying gripper requirements for collaborative robot applications in manufacturing. In: Lalic, B., Majstorovic, V., Marjanovic, U., von Cieminski, G., Romero, D. (eds.) APMS 2020. IAICT, vol. 591, pp. 655–662. Springer, Cham (2020). https://doi.org/10. 1007/978-3-030-57993-7 74 13. Schnell, M., Holm, M.: Challenges for manufacturing SMEs in the introduction of collaborative robots. In: SPS2022: Proceedings of the 10th Swedish Production Symposium, vol. 21, pp. 173–183. IOS Press (2022). https://doi.org/10.3233/ ATDE220137 14. Steinmetz, F., Weitschat, R.: Skill parametrization approaches and skill architecture for human-robot interaction. In: 2016 IEEE International Conference on Automation Science and Engineering (CASE), pp. 280–285. IEEE (2016). https:// doi.org/10.1109/COASE.2016.7743419 15. Xia, G., Li, C., Zhu, Q., Xie, X.: Hybrid force/position control of industrial robotic manipulator based on kalman filter. In: 2016 IEEE International Conference on Mechatronics and Automation, pp. 2070–2075 (2016). https://doi.org/10.1109/ ICMA.2016.7558885 16. Xu, X., Chen, W., Zhu, D., Yan, S., Ding, H.: Hybrid active/passive force control strategy for grinding marks suppression and profile accuracy enhancement in robotic belt grinding of turbine blade. Roboti. Comput.-Integr. Manufact. 67, 102047 (2021). https://doi.org/10.1016/j.rcim.2020.102047

Optimizing Performance-Allocation Trade-Off: The Role of Human-Machine Interface Technology in Empowering Multi-skilled Workers in Industry 4.0 Factories Federica Costa(B)

, Alireza Ahmadi , and Alberto Portioli-Staudacher

Department of Management, Economics, and Industrial Engineering, Politecnico di Milano, Via Lambruschini 4/B, 20156 Milan, Italy [email protected]

Abstract. Companies today face growing demands for customization, necessitating the maintenance of high quality, low cost, and short lead times. Workload management plays a crucial role in achieving these objectives, and it primarily involves two forms: input control and output control. Existing literature has significantly focused on input control while often overlooking output control. This research focuses on the critical lever of workers’ flexibility as an essential output control element and investigates its impact on lead time performance. This research focuses on recognizing workers’ flexibility as a key competitive opportunity for companies, allowing them to cope effectively with temporary job demand imbalances. Companies can optimize direct labor utilization and production lead times by reallocating workers from underloaded stations to overloaded ones, thereby driving higher profit margins and customer satisfaction. To evaluate the benefits of increased flexibility, we employ discrete-event simulation techniques demonstrating a significant reduction in lead time. Additionally, we examine the impact of limiting workers’ movement across workstations. Surprisingly, These results demonstrate that, despite reducing the number of workers’ relocations, the advantages outweigh the negative effect on lead time. Companies can achieve optimal output control strategies that enhance their overall performance by improving the trade-off between lead time and the number of workers’ relocations. Overall, this research underscores the importance of output control in workload management and highlights workers’ flexibility as a critical lever to improve lead time performance. By adopting effective output control strategies, companies can efficiently align capacity and orders, enhancing operational performance and customer satisfaction. This study provides valuable insights into the dynamics of workload management and offers practical recommendations for manufacturing organizations seeking to optimize their processes. Keywords: Flow-shop · workload control · multi-skilled workers · labor flexibility · Human-machine interface

© IFIP International Federation for Information Processing 2023 Published by Springer Nature Switzerland AG 2023 E. Alfnes et al. (Eds.): APMS 2023, IFIP AICT 689, pp. 716–729, 2023. https://doi.org/10.1007/978-3-031-43662-8_51

Optimizing Performance-Allocation Trade-Off

717

1 Introduction In companies, the seamless collaboration between humans and machines has revolutionized various functions such as assembly, quality assurance, logistics, and maintenance, amplifying the level of automation to unprecedented heights as Industry 4.0 continues to proliferate. By harnessing the power of advanced technologies and artificial intelligence, this symbiotic relationship between humans and machines has ushered in a new era of productivity, efficiency, and innovation. All functions focus on human-machine-robot collaboration and efficient interfaces to better connect human workers with machines [1]. The use of robots as machines is increasing yearly in almost all industries. Especially for tasks that require heavy lifting or repetitive work, robots can offer long-term savings by reducing labor costs and increasing consistent speed and accuracy, thereby boosting productivity [2]. Humans contribute their cognitive abilities, creativity, and adaptability, while machines offer accuracy and tireless execution. This trend exemplifies transitioning towards a human-centric Industry 5.0, focusing on synergy [3]. With the increasing need for customized products and batch sizes, production is becoming more complex and flexible. As a result of the increasing demand for the variability of products, manufacturing companies must maintain high quality, low costs, and a short production lead time while responding to a rising need for customization [4]. Flexibility is critical in this environment, characterized by short product life cycles and wide variety, and companies are switching from “make-to-stock” to “make-to-order” policies. A key element of successfully implementing a “make-to-order” policy is to have shorter and more reliable lead times [5]. This policy shortens customer waiting time and provides trustworthy feedback on when their order will be completed and shipped on time. Given the increase of automation in manufacturing and the need for an extreme level of flexibility, workers must supervise and operate multiple machines, making multiskilling essential. Rule-based machines with fixed programming schemes often need human interaction to handle production complexity. Thus, operators must be multiskilled and able to manage multiple machines. Production planning and control synchronizes capacity and demand, improving lead time performance by regulating job inflow and outflow through capacity adjustments. However, leveraging workers’ flexibility can be a competitive opportunity for manufacturing companies [6]. It allows coping with temporary job demand imbalances by reallocating workers from underloaded stations to overloaded ones [7]. In addition, this flexibility improves direct labor utilization and production lead times, which drive higher profit margins and customer satisfaction. Furthermore, strict employment legislation and no-layoff policies lead to long-term work commitments [8]. Managers in the automotive industry commonly refer to labor flexibility as the means for the plant to adjust its capacity in the short term. It is often viewed as the only viable option for maintaining operations [4]. Industry 4.0’s factory automation requires multi-skilled workers who can supervise and operate multiple machines. Thus, managing multi-skilled workers and their movements between machines is worth studying. The research investigates how to manage multi-skilled workers on a production line and how performances are affected, precisely

718

F. Costa et al.

the trade-off between worker movements amongst different machines and lead time reduction. We know that workers’ flexibility has been abundantly studied in dual resourceconstrained (DRC) literature [9]. However, in DRC, researchers have conducted limited research on the “where rules” that limit the number of workers who can work simultaneously on the same machine or restrict workers’ movements based on the task capacity of the machine. However, it is important to address these two parameters as workers supervising on the same machine and frequently shifting between machines may cause interferences that can degrade efficiency and worsen the overall system performance [10]. Researchers have extensively studied worker movements in assembly lines. Assembly line balancing involves planning and implementing a line “ex-ante” based on a forecast structural demand. In an assembly line problem with the traveling worker, a worker is statically assigned to more than one workstation if he can provide the task capacity necessary to operate the two workstations. Instead, in the presented “when/where/who” DRC rules, workers can dynamically travel to different workstations to cope with temporary imbalances in the job demand of the workstations. Furthermore, researchers seldom included skilled workers and workers’ movement across workstations in their studies when considering multi-manned assembly lines [11]. This study investigates how managing multi-skilled workers in a production line affects system performance, specifically focusing on the trade-off between worker shifts amongst machines and lead time reduction. This study addresses the research gap by examining how the restriction of workstation relocation affects system performance. By developing a realistic simulation model with multiskilled workers, the study aims to provide insights into optimizing the management of multi-skilled workers in a production line. Specifically, it seeks to understand the trade-off between worker movement and lead time reduction, which are crucial factors in balancing system efficiency and productivity. The findings of this research will contribute to a deeper understanding of the effects of managing multi-skilled workers and provide practical implications for enhancing performance in manufacturing environments, particularly in the context of Industry 4.0 and factory automation.

2 Literature Review The Literature Review section discusses three subtopics: Workload Control, Worker’s Flexibility, and Human Machine Interface Technologies. The main focus was on thoroughly examining the topic of workers’ flexibility, which led us to explore the literature on workload control and dual-resource-constrained concepts. Through these resources, we gained in-depth insight into the interactions between workers and their workstations. Additionally, we analyzed human-machine interface technologies to identify any potential limitations. To address the coherence and integration of ideas, we have provided a more comprehensive analysis of the existing literature. We emphasized common themes, controversies, and unresolved issues that led to the identified research gap.

Optimizing Performance-Allocation Trade-Off

719

2.1 Workload Control Workload control involves aligning capacity with demand and improving lead time performance through input and output control. Input control focuses on planning job release and shop floor flow, considering job demand and process control. Output control adjusts system output using production data and capacity levers, such as adding machines or subcontracting. Leveraging workforce flexibility, including workers’ overtime and machine movement, can effectively adjust production volume, lead time, and efficiency. 1. Input control uses manufacturing job information to plan release and shop floor flow [12]. It ensures jobs meet deadlines. Input control considers job demand and process control. This data calculates job impact on station load. We used input control as a “workload limiting algorithm” to regulate job release [13]. This algorithm prioritizes jobs by dispatching rules, like first-in-first-out. The highest priority job is released if its contribution to system task capacity is below a workload norm. 2. Output control adjusts system output using production data and production capacity levers [14]. Adding machines and/or flexibility can boost production. Subcontracting to a supplier increases system output. Finally, this research shows that leveraging the workforce lever, workers’ flexibility as overtime, and machine movement can adjust production volume, lead time, and efficiency. 2.2 Worker’s Flexibility According to Gupta and Jain [15], cross-training workers, allowing them to perform different kinds of tasks, enables superior performance for companies. For example, the flexibility permitted by workers cross-training reduces the impact of turnover, absenteeism, and illness, as well as the variability of changes in product mix and demand in seasonal markets and budgeting cycles [16]. DRC literature includes many studies about workers’ flexibility, so we decided to include it in this research. Studies have demonstrated that cross-training workers to perform different tasks enhances company performance by reducing the impact of turnover, absenteeism, and changes in product mix and demand. The DRC literature extensively investigates workers’ flexibility, which encompasses decisions regarding the transfer of workers between workstations in terms of “where” and “when,” as well as determining which workers are eligible (“who”). Increasing workers’ flexibility enhances their productivity by reducing idle time and improving task processing time. This research aims to bridge the gap in knowledge regarding the effectiveness of multi-skilled workers’ movements and their influence on system performance. The DRC literature addresses three crucial decisions that need to be made regarding workers’ movement across workstations. • “Where” includes moving the worker to the upstream station with a non-empty queue and more complex rules, such as transferring the worker to the first station with a non-empty queue with the “oldest” job currently in the system. • “When” include centralized control, according to which workers are transferred to other workstations after the completion of the current job [17], and decentralized control, by which workers can be transferred to other workstations when they are idle [18].

720

F. Costa et al.

• “Who” concerns the definition of eligible workers, which can be transferred to other departments. Research focuses on a) workers’ flexibility, which concerns multiskilled workers, and b) worker efficiency, which represents the proficiency of workers in performing tasks. An increase in workers’ flexibility reduces their idleness since it increases the probability that their work is needed on at least one machine [19], while workers’ efficiency affects the time taken to process a job. 2.3 Human-Machine Interface Technologies Integrating human-machine interface (HMI) technologies with automation systems are increasingly common in industrial settings. HMI technologies allow operators to manage and monitor multiple machines through a single interface, providing real-time feedback on parameters. Studies suggest integrating HMI technologies with automation systems improves operator performance, productivity, workload control, and worker flexibility. By enabling operators to manage multiple machines and tasks simultaneously, HMI technologies address the challenges posed by complex industrial processes. Integrating human-machine interface (HMI) technologies with automation systems has become increasingly common in today’s industrial settings. HMI technologies allow operators to manage and monitor multiple machines through a single interface, providing real-time feedback on parameters such as temperature, pressure, and speed. Recent studies have shown that HMI technologies can significantly improve operator performance in managing multiple machines and tasks. These findings suggest that integrating HMI technologies with automation systems can help address the challenges posed by the increasing complexity of industrial processes and the need for operators to manage multiple machines and tasks simultaneously [20]. Additionally, HMI technologies can enable workload control and worker flexibility in industrial settings, allowing operators to adjust their task capacity and prioritize tasks by providing real-time feedback on machine parameters and task progress. Finally, integrating HMI technologies with automation systems can improve operator performance, increase productivity, and enhance workload control and worker flexibility in industrial settings [21]. The Literature Review provides a more coherent and integrated analysis of the subtopics. It highlights the research gap related to managing multi-skilled workers and their movements in the context of workload control, worker flexibility, and HMI technologies. By addressing this research gap, the study aims to better understand system performance in manufacturing environments, particularly in the era of Industry 4.0 and factory automation.

3 Methodology To answer the research question, we used discrete-event simulation in Python since it allows to model human-machine interactions in real time and the shifts of workers from one station to another. The basis from which we started is the model presented by Costa et al. [5]. Moreover, in the DRC area of research, discrete-event simulation has been already vastly applied [4, 5, 7, 9].

Optimizing Performance-Allocation Trade-Off

721

Simulation Model The simulation model represents a pure flow shop made up of five stations. Each station has a pool/buffer in front of it to store the work in process. The whole “System” is completed by adding in front of the shop floor a “Pre-Shop” pool and by adding another pool after the shop floor (“Order Delivered” pool) (Fig. 1).

Fig. 1. Shop configuration

This shop configuration is the basis and general representation of all the simulation models. Received customer orders are always accepted, placed in the pre-shop pool, and released in the system according to the selected release rule. All the machines must process the jobs in the same sequence, following the first-in-first-out rule. The system assigns a default workstation to each of the five workers operating in the system. Depending on the selected level of flexibility, they may also work on other workstations for a maximum of two workers operating in a single workstation. The processing time needed to process a job on a workstation depends on the workers’ efficiency and the number of workers assigned to the workstation. The system operates for 480 min daily, equivalent to an “eight net worked hours” shift. The model does not consider production defects, system failures, or raw materials stockout, which might cause stops or delays. The time needed to complete a job is delivered to the “Order Delivered” pool, referred to as “gross throughput time,” It is the best proxy to measure customer lead time, even if it does not consider logistics. Jobs Arrival and Release The order arrival to the pre-shop pool follows an exponential distribution, which is considered a good approximation of a confirmed order arrival process [22]. There is no limit to the order arrival; once they arrive, jobs remain in the pre-shop pool until release. The system uses a workload-limiting algorithm for job release to the shop floor. The algorithm prioritizes jobs based on the first-come-firstserved rule. Only when a job’s contribution is accounted for in the task capacity, already introduced, and does not surpass the “workload norm” the highest priority job is released into the system. Workers Two types of algorithms are used to model workers’ allocation: a) “Static,” when workers cannot relocate to other workstations, and b) “Reactive,” in which workers can be relocated when idle. The Reactive algorithm includes three possible flexibility levels: “chain,” when workers can process jobs at their default station and the station right downstream their default station, “triangular,” when workers can process jobs at their

722

F. Costa et al.

default station at the one upstream and the one downstream, and “flat” when workers can process jobs on any of the workstations. From now on, we will only refer to the four flexibility levels: static, chain, triangular, and flat. Workers’ efficiency on a given workstation regulates the time the worker takes to process a job on that workstation: efficiency is fixed at 100% for the entire simulation time. Machines Five machines exist in the system. The time taken to process one job (“actual processing time”) on a machine depends both on the number of workers (one or two) and on their efficiency. The relevant calculation is shown below. Actualprocessingtime(i, j) = 

Processingtime(i, j) Workerposition(k, j) ∗ Workerefficiency(k, j) k

(1)

where processing time (i, j) is the processing time of job i at station j defined through a log-normal distribution with a mean of 30 min and a variance of 900 min^2 truncated between 0 and 360 [5], worker efficiency (k, j) is the efficiency of worker k at station j and worker position (k, j) is a binary variable equal to 1 whether worker k is allocated at station j and, otherwise, to 0. Workers Association with Machines A worker remains at their default station for as long as he/she has jobs to process. Otherwise (if flexibility is not Static), the worker can be allocated to other machines. A machine is included in the set of possible “next machines” for the worker if all the following apply: • The worker can work at the workstation according to the rules of the selected level of flexibility (chain/triangular/flat). • The workstation has a job demand to process higher than zero. • The number of workers currently working on the workstation is zero or one (two is the maximum number of workers allowed on a single station simultaneously). If the set of possible “next machines"/workstations is empty, the worker is reallocated to its default station. Otherwise, the next station is chosen randomly among the possible stations. The transfer time is always assumed to be zero. Model 1 The initial model, Model 1, forms the basis for subsequent models. In this model, workers are assigned to a default machine, but they can relocate to other machines when idle, depending on the level of flexibility chosen. The set of possible machines where the worker can move is determined based on certain rules, such as the machine having a job demand to process higher than zero and the number of workers currently working on the machine being zero or one. The next machine for relocation is randomly selected from this set of possible machines, with all machines having the same efficiency of 100%. Model 2 Model 2 differs from Model 1 in the “Workers Association to the Machines” rule. In Model 2, a new rule is added to limit the set of possible machines where workers can move. This rule requires the workstation to have a job demand to process higher than the queue length limit. This change is made to study the trade-off between the number of worker relocations and the average gross throughput time (GTT) for the simulation time, as workers’ movement can lead to interferences and time wastage.

Optimizing Performance-Allocation Trade-Off

723

Model 3 Model 3 is similar to Model 2 but with a new rule introduced to reduce the number of worker relocations. When choosing the next machine at the end of a job, if the worker is already outside its default workstation and no job is at its default workstation, the current machine is preferred if it is included in the set of possible next machines.

3.1 Measured Performance The performances measured in all the models are: • Gross throughput time (GTT), also known as manufacturing lead time: GTT is the time elapsed from when a customer order is received and placed in the pre-shop-pool to when the production of the related job is completed • Shop floor throughput time (SFT): SFT is the gross throughput time of the job i minus the time that the job i spends in the pre-shop pool, so it is the time in which job i is processed on the shop floor. When speaking about GTT and SFT performance, we refer to the average time, measured in minutes, for all the jobs processed in the system during the simulation time. • Number of Relocations: the sum of all the worker transfers during the simulation time. The objective is to achieve a good trade-off between minimizing GTT and the number of relocations. However, we consider GTT as the main performance to consider since it is a good proxy of the time a customer waits to receive its order, even though GTT does not include some logistics lead times, such as stock picking time, transport time, and so on. 3.2 Simulation Parameters Overall, the research question was addressed using Python discrete-event simulation. The simulation model was based on the model proposed by Costa et al. [5] to create a realistic working environment. The model represents a flow shop with five stations and a pool/buffer for work in progress. The system configuration includes a pre-shop pool and an order-delivered pool. In this simulation model, customer orders are accepted and placed in the pre-shop pool. A release rule determines the job release, and the machines process jobs in a strict first-in-first-out order. The system operates for 480 min daily, equivalent to an “eight net worked hours” shift. However, the simulation model does not consider production defects, system failures, or raw material stockouts that could cause delays or disruptions. Hence, the model captures the workflow in an idealized setting without these real-world constraints. The arrival of confirmed orders follows an exponential distribution, which provides a good approximation. Jobs remain in the pre-shop pool until they are released to the shop floor, following a first-come-first-served rule and workload norms. The simulation model assigns workers using two algorithms: “Static” and “Reactive”. In the Static algorithm, workers remain on their default machines, while in the Reactive algorithm, workers can be relocated when idle. The model offers three levels of worker allocation flexibility: “chain,” “triangular,” and “flat”. Workers can switch workstations

724

F. Costa et al.

within each flexibility category. The simulation model consists of five machines, and the processing time of each machine depends on the number and efficiency of its workers, which is assumed to be 100% throughout the simulation. The study introduces three models: Model 1 allows idle workers to switch machines, Model 2 adds an additional rule limiting the machines where workers can be relocated, and Model 3 includes a rule to reduce worker relocations. The performance metrics measured in all models include gross throughput time (GTT), shop floor processing time (SFT), and the number of relocations, as in other studies of the DRC field [4, 7, 9]. GTT indicates customer lead time, which is the time between customer order and job completion. SFT represents the time taken for processing on the shop floor. Although the goal is to strike a balance between minimizing GTT and reducing relocations, the study primarily focuses on GTT as the main performance measure since it serves as a good proxy for customer waiting time, even though it does not consider all logistics lead times such as stock picking and transport time. The methodology adopts a discrete-event simulation approach to model a flow shop system with different worker allocation strategies (Static and Reactive) and machine associations. The performance metrics of interest are GTT, SFT, and the number of relocations. Welch’s moving average method was employed to determine simulation time, and it was observed that GTT performance converges after 100,000 time units of simulation. To ensure stability and eliminate any variability resulting from the initial transient phase, a warm-up period of 200,000 time units was established. The total simulation length was set at 500,000 time units (equivalent to 1041.6 days), with the simulation time referred to as the functioning period of our system, which includes 300,000 time units of simulation after the warm-up period.

4 Results and Discussions As shown in the literature review, this research shows that workers’ flexibility, one of the levers of output control, can significantly benefit performance. This is because it allows the managing of temporary job demand imbalances by reallocating workers from underloaded stations to over-loaded ones. The present work shows that gross throughput time can be improved to a great extent even when realistic constraints (i.e., a maximum limit of two workers at the same time on the same workstation and variable efficiency due to learning curves) are included present in the system. Although we have simulated all the models, we can only provide an example of Model 2 due to limited space. Figure 2 shows four curves representing, for Model 2, workload norm 3000, the performance in terms of GTT versus the total number of relocations for the four flexibility levels: each of the dots is a tested queue length limit that increases from left to right. The static “curve” is only a single dot since it obviously cannot vary with queue length limits. The maximum performance in terms of GTT is achieved with flat flexibility and queue limit zero: this means that, by training workers on four additional machines on top of their default one and putting no constraints on their relocations when idle, GTT can be reduced to less than half. One of the novel aspects of this work is the introduction of a “where rule” that regulates workers’ relocations to other machines when idle, i.e., when there is no task

Optimizing Performance-Allocation Trade-Off

725 3000

2000 1500 1000

Av. GTT

2500

500 0 60000

50000

40000

30000

20000

10000

0

Relocations Chain

Triangular

Flat

Static

Fig. 2. GTT versus the number of relocations for model 2

capacity to process on the default machine they are assigned to. A machine can be considered for the provision of help by idle workers only if the job demand it must process is higher than a “queue length limit”. This rule was tested to assess the trade-off between the average gross throughput time of jobs in the system and the total number of workers relocations across machines during the simulation time. We retained that it was necessary to balance the trade-off because having many workers relocated in a real working environment would certainly lead to interferences and waste of time, which might affect the GTT in turn. Table 1. GTT and Relocations variations by queue length limit, model 2, chain. Queue length limit

Chain GTT

GTT % variation

Relocations

Relocations variation

0

1916.26935

-

24683

-

3

1968.57252

3%

15366

-38%

5

1994.13761

1%

12262

-20%

7

2023.40101

1%

9958

-19%

10

2067.75555

2%

7453

-25%

15

2142.3783

4%

4656

-38%

20

2210.83954

3%

2983

-36%

25

2267.51482

3%

1963

-34%

30

2317.06033

2%

1315

-33%

40

2392.12393

3%

604

-54% (continued)

726

F. Costa et al. Table 1. (continued)

Queue length limit

Chain GTT

GTT % variation

Relocations

Relocations variation

50

2436.34625

2%

279

-54%

70

2477.63228

2%

63

-77%

100

2495.52

1%

5

-92%

Table 1 depicts Model 2, showcasing chain flexibility and examining the trade-offs between two performance metrics. The table presents information for various queue length limits, including a) the optimal gross throughput time (GTT) achieved by adjusting the workload norm (not included in the table), b) the percentage variation in GTT compared to the previous queue length limit, c) the number of relocations, and d) the percentage variation in relocations compared to the previous queue length limit. Notably, increasing the queue length limits leads to a decrease in relocations and a subsequent increase in GTT. This occurs because workers remain idle when relocations are prevented due to the queue length limit, resulting in an unused working capacity. It is important to highlight that even with a small queue length limit, such as 3 time units, there is a slight increase in GTT while relocations decrease significantly. For example, with flat flexibility, GTT increases by 5% while relocations decrease by 35%. This implies that in a simulation time of 300,000 time units, at a queue limit of zero, each worker relocates approximately 4.9 times every 60 time units, amounting to nearly 40 relocations in an 8-h shift. In contrast, relocations reduce to around 25 in an 8-h shift at a queue limit of three. Similarly, for triangular flexibility, GTT increases by 14%, and relocations decrease by 43%. As we move across different queue length limits, the percentage reduction in relocations becomes smaller, and GTT increases for any queue limit and flexibility. These observations hold true for each of the proposed models. This section effectively presents and discusses the obtained results to provide a clear understanding of the findings. The research highlights the significance of workers’ flexibility as a critical factor in improving performance through output control. The research demonstrates effective management of temporary job demand imbalances by reallocating workers from under-loaded stations to over-loaded ones. Importantly, significant improvements in GTT can be achieved even with realistic constraints, such as a maximum limit of two workers per workstation and variable efficiency due to learning curves. To illustrate these findings, the research introduces Model 2 and Fig. 2, which illustrate the relationship between GTT and the number of relocations for different levels of flexibility. The curves in the figure correspond to a workload norm of 3000, and each dot on the curves represents a tested queue length limit. Notably, the highest GTT performance is observed with flat flexibility and a queue limit of zero. This indicates that by training workers on additional machines and allowing unrestricted relocations during idle time, GTT can be reduced by more than half.

Optimizing Performance-Allocation Trade-Off

727

An innovative aspect of this research is introducing the “where rule,” which governs workers’ relocations to other machines based on the job demand of each machine when idle. The research emphasizes the trade-off between the average GTT of jobs and the total number of worker relocations across machines during the simulation. Striking a balance in this trade-off is crucial, as excessive relocations in a real working environment can lead to interferences and wasted time, potentially affecting GTT. Table 1 provides an overview of the variations in GTT and relocations for different queue length limits in Model 2. The table illustrates that increasing the queue length limit leads to a decrease in relocations but an increase in GTT. For instance, even a small queue length limit of 3 time units results in a 35% reduction in relocations but a 5% increase in GTT. This indicates that workers relocate approximately 40 times in an 8-h shift with a queue limit of zero, whereas with a queue limit of three, relocations decrease too.

5 Conclusion Our study utilized simulation to demonstrate the significant reduction in lead time associated with greater flexibility. Additionally, we examined the effects of restricting worker movement between workstations and found that reducing worker moves had a more pronounced impact on lead time than any negative effects. Furthermore, we successfully improved the trade-off between lead time and worker movement. We also investigated the impact of input control on system performance. Our findings revealed that the best performance was achieved when a lower workload was released into the system with high flexibility. This allowed workers to operate on two or three additional machines with low queue length limits. However, in cases where flexibility was low due to a static or chain structure and queue length limits were high, the best performance was observed with a higher workload in the system. This was attributed to the ability of workers to move to other workstations and process jobs there, even with a low workload, thereby avoiding idleness and congestion. Conversely, a higher workload was necessary to enhance their utilization when worker relocation was more restricted. It is important to note that our study focused specifically on a flow shop configuration, and future research should explore the applicability of these findings to other shop configurations, such as job shops. Additionally, while simulation allowed us to study a range of scenarios, conducting real-world experiments in future research would provide valuable validation of our main findings. In future research, we propose examining the trade-off between workers’ learning speed and its impact on system performance. This can be done by considering the influence of new digital technologies, such as IoT wearable devices. Furthermore, investigating the positive effects of digital technologies on system performance, particularly by testing a higher slope of the learning curves enabled by such technologies, would contribute to a deeper understanding. It would also be worthwhile to assess if investing in digital technologies is optimal for certain systems. This research concludes the Results and Discussions section by highlighting the key findings. It emphasizes the significant reduction in lead time achieved through greater flexibility, the relatively minor impact of restricting worker movement on the lead

728

F. Costa et al.

time, and the importance of input control in system performance. The study acknowledges the limitations of focusing on a flow shop configuration and suggests avenues for future research, including exploring different shop configurations, conducting real-world experiments, and investigating the impact of digital technologies on system performance.

References 1. Enrique, D.V., Druczkoski, J.C.M., Lima, T.M., Charrua-Santos, F.: Advantages and difficulties of implementing Industry 4.0 technologies for labor flexibility. Procedia Comput. Sci. 181, 347–352 (2021). https://doi.org/10.1016/j.procs.2021.03.130 2. International Federation of Robotics. (2022). Positive Impact of Industrial Robots on Employment 3. Pasparakis, A., De Vries, J., De Koster, R.: Assessing the impact of human-robot collaborative order picking systems on warehouse workers. J. Oper. Manag. (2023), Advance online publication.https://doi.org/10.1080/00207543.2023.2183343 4. Francas, D., Löhndorf, N., Minner, S.: Machine and labor flexibility in manufacturing networks.Int. J. Prod. Econ. (2011).https://doi.org/10.1016/j.ijpe.2010.03.014 5. Costa, F., Kundu, K., Rossini, M., Portioli-Staudacher, A.: Comparative study of bottleneckbased release models and load-based ones in a hybrid MTO-MTS flow shop: an assessment by simulation. Oper. Manag. Res. 16(1), 33–48 (2023) 6. Portioli-Staudacher, A., Costa, F., Thürer, M.: The use of labour flexibility for output control in workload controlled flow shops: a simulation analysis. In: Proceedings of the 16th International Conference on Factory Communication Systems, vol. 12543, pp. 429–442 (2020). https://doi.org/10.5267/j.ijiec.2019.11.004 7. Costa, F., Portioli-Staudacher, A.: Labor flexibility integration in workload control in Industry 4.0 era. Oper. Manag. Res. 14(3–4), 420–433 (2021). https://doi.org/10.1007/s12063-021-001 84-6 8. Askar, G., Sillekens, T., Suhl, L., Zimmermann, J.: Flexibility planning in automotive plants. In: Günther, H.-O., Mattfeld, D.C., Suhl, L. (eds.) Management Logistischer Netzwerke, pp. 235–255. Physica, Heidelberg (2007). https://doi.org/10.1007/978-3-7908-1916-9_14 9. Thürer, M., Zhang, H., Stevenson, M., Costa, F., Ma, L.: Worker assignment in dual resource constrained assembly job shops with worker heterogeneity: an assessment by simulation. In: Dornberger, R., Jung, R., Steinmetz, R. (Eds.) Proceedings of the 16th International Conference on Factory Communication Systems, pp. 6336–6349. Springer Lecture Notes in Computer Science, vol. 12543 (2020). https://doi.org/10.1007/978-3-030-55184-2_492 10. Kher, H., Malhotra, M.: Acquiring and operationalizing worker flexibility in dual resource constrained job shops with worker transfer delays and learning losses. Omega 22, 521–533 (1994). https://doi.org/10.1016/0305-0483(94)90032-9 11. Malhotra, M.K., Narasimhan, R., Giri, B.C.: The impact of learning and labor attrition on worker flexibility in dual resource constrained job shops. Decis. Sci. 24, 641–664 (1993). https://doi.org/10.1111/j.1540-5915.1993.tb01296.x 12. Kingsman, B., Hendry, L.: The relative contributions of input and output controls on the performance of a workload control system in make-to-order companies. Prod. Plann. Control. 13, 579–590 (2002). https://doi.org/10.1080/0953728021000026285 13. Thürer, M.J., Land, M., Stevenson, M., Fredendall, L.D.: On the integration of due date setting and order release control. Prod. Plann. Control. 28, 420–430 (2017). https://doi.org/10.1080/ 09537287.2017.1302102 14. Huang, Y.: Information architecture for effective workload control: an insight from a successful implementation. Prod. Plann. Control. 28, 351–366 (2017). https://doi.org/10.1080/ 09537287.2017.1288278

Optimizing Performance-Allocation Trade-Off

729

15. Gupta, S., Jain, S.K.: A literature review of lean manufacturing. Int. J. Manag. Sci. Eng. Manag. 8, 241–249 (2013). https://doi.org/10.1080/17509653.2013.825074 16. Sammarco, M., Fruggiero, F., Neumann, W.P., Lambiase, A.: Agent-based modeling of movement rules in DRC systems for volume flexibility: human factors and technical performance. Int. J. Prod. Res. 52(3), 633–650 (2014). https://doi.org/10.1080/00207543.2013.807952 17. Araz, Ö.U., Salum, L.: A multi-criteria adaptive control scheme based on neural networks and fuzzy inference for DRC manufacturing systems. Int. J. Prod. Res. 48(1), 251–270 (2010). https://doi.org/10.1080/00207540802471256 18. Małachowski, B., Korytkowski, P.: Competence-based performance model of multi-skilled workers. Comput. Ind. Eng. 91, 165–177 (2016). https://doi.org/10.1016/j.cie.2015.11.018 19. Boysen, N., Fliedner, M., Scholl, A., Bock, S.: A classification of assembly line balancing problems. Jenaer Schriften zur Wirtschaftswissenschaft (2006). Available at: https://pdfs.sem anticscholar.org/c9c5/1dd66ddf14b0a0af0fa62e86ff49fb2dac5d.pdf 20. Wang, X.V., Wang, L.: Augmented reality enabled human–robot collaboration. Adv. HumanRobot Collaboration Manuf. 395–411(2021). https://doi.org/10.1007/978-3-030-69178-3_16 21. Esengün, M., Üstünda˘g, A., ˙Ince, G.: Development of an augmented reality-based process management system: the case of a natural gas power plant. IISE Trans. 55, 201–216 (2023). https://doi.org/10.1080/24725854.2022.2034195 22. Moreira, M.R.A., Alves, R.A.F.S.: Input-output control order release mechanism in a jobshop: how workload control improves manufacturing operations. Int. J. Comput. Sci. Eng. 7(3), 214–223 (2012)

Towards Industry 5.0: Empowering SMEs with Blockchain-Based Supplier Collaboration Network Prince Waqas Khan1 , Imene Bareche2 , and Thorsten Wuest1(B) 1

Department of Industrial and Management Systems Engineering, West Virginia University, Morgantown, WV, USA {princewaqas.khan,thwuest}@mail.wvu.edu 2 School of Computer Science and Technology, Chongqing University of Posts and Telecommunications, Chongqing 400065, China [email protected]

Abstract. Industry 5.0 integrates human-centered production systems with cyber-physical systems, the Internet of Things (IoT), and machine learning technologies to create flexible, efficient, and environmentally friendly manufacturing processes. Small and medium-sized enterprises (SMEs), despite their importance in the manufacturing sector, typically need more resources and skills to implement Industry 5.0 technology. This article presents a blockchain-based supplier collaboration network to support small and medium-sized enterprises migrating to Industry 5.0. Small suppliers, manufacturers, and other supply chain stakeholders can engage securely and transparently within the proposed framework. The underlying blockchain technology allows for secure, immutable storage and real-time data sharing and collaboration. The proposed approach uses smart contracts to automate cooperation procedures and integrates blockchain with emerging technologies like IoT and machine intelligence. Moreover, the underlying method can boost SMEs businesses’ competitiveness and market access. The proposed approach will therefore be of benefit to companies that have limited resources to build their own supply chain management systems. Finally, we discuss the challenges and limitations of the proposed system for implementing a blockchainenabled supplier collaboration network for SMEs in Industry 5.0. Keywords: Industry 5.0 · Internet of Things (IoT) · Blockchain · Smart contracts · Transparency · Small and medium-sized enterprises

1

Introduction

Industry 5.0 is a concept that builds on the foundation of Industry 4.0 and emphasizes the role of human-centered design, social responsibility, and sustainability in the manufacturing process [21]. The exploration of Industry 5.0 is still in its infancy, and research findings are relatively scarce. c IFIP International Federation for Information Processing 2023  Published by Springer Nature Switzerland AG 2023 E. Alfnes et al. (Eds.): APMS 2023, IFIP AICT 689, pp. 730–744, 2023. https://doi.org/10.1007/978-3-031-43662-8_52

Towards Industry 5.0

731

Industry 5.0 seeks to create a manufacturing ecosystem that benefits humans and the environment and places the worker’s well-being at the center of the production process. Small and medium-sized enterprises (SMEs) play a vital role in the manufacturing sector, accounting for a significant portion of global employment and output [23]. However, SMEs often face several challenges when it comes to implementing Industry 5.0 technology. Unlike large firms, SMEs typically have limited financial and technical resources to invest in new technologies, making it difficult to compete with larger companies [12]. As Industry 5.0 promotes the integration of advanced technologies in manufacturing processes, SMEs are expected to face increasing amounts of highly dynamic spatial data [4,5]. Moreover, implementing Industry 5.0 technologies, such as blockchain-based supplier collaboration networks, requires a high level of technical expertise and specialized skills that may be limited in SMEs. As a result, SMEs may struggle to leverage advanced technologies to improve their manufacturing processes and fall behind their competitors in terms of efficiency, productivity, and competitiveness. Therefore, developing solutions that enable SMEs to adopt Industry 5.0 technologies is crucial for the long-term growth and sustainability of the SME sector and the manufacturing industry. Blockchain technology offers SMEs a new way to access secure, transparent, and decentralized systems for conducting transactions and managing supply chains [18]. By eliminating the need for intermediaries such as banks, blockchain technology can reduce transaction costs and increase the efficiency of SMEs’ operations. Moreover, blockchain technology can give SMEs greater control over their data and enhance the transparency and accountability of their business practices. This can help SMEs build trust with their customers, suppliers, and investors, which can be crucial for their growth and long-term success. Finally, blockchain technology enables SMEs to access new markets and expand their business opportunities beyond traditional boundaries. 1.1

Contributions

This article presents a blockchain-based supplier collaboration network to help SMEs migrate to Industry 5.0. The proposed system enables small suppliers, manufacturers, and other stakeholders to collaborate securely and transparently. The use of blockchain technology provides a secure, immutable, and real-time data-sharing environment that can boost the competitiveness of SMEs businesses and provide them with greater market access. Smart contracts are used to automate cooperation procedures and integrate blockchain with emerging technologies. This article explores how smart contracts and a blockchain-based framework can be used to derive feasible benefits from supply chain process design for SMEs. By presenting a practical example of how blockchain technology can be used to facilitate supplier collaboration in Industry 5.0, this article aims to contribute to the growing stream of research studies on the use of blockchain in the manufacturing sector. This article presents the following main contributions: – Introducing a blockchain-based supplier collaboration network that aims to assist and support SMEs transition to Industry 5.0.

732

P. W. Khan et al.

– Discussion of the key parts of the solution, including smart contract automation and blockchain. – The proposed private blockchain built on Hyperledger Fabric intends to promote human-centered design, social responsibility, and sustainability in the manufacturing process for SMEs. The rest of the article is structured as follows: Sect. 2 provides a comprehensive literature review on the relevant topics. In Sect. 3, we propose a blockchainbased collaboration network, and a detailed description of the structure and the private blockchain platform utilized in its implementation is provided. The performance analysis of the proposed system is presented and discussed in Sect. 4. Finally, Sect. 5 summarises the main findings and conclusions of the study and draws out the perspectives for future work.

2

Literature Review

SMEs have long faced financing challenges due to the contradiction between strong demand and weak trust [14]. The resultant financing constraints are closely related to the production efficiency of manufacturing enterprises. Jiang et al. [13] have proposed a solution to the financing problem of SMEs through the use of supply chain finance (SCF) with the help of blockchain technology. The authors have identified the limitations of the traditional SCF system, which only allows Tier 1 suppliers to access financing through the high credit scores of core enterprises, leaving Tier 2 and Tier N suppliers in the long tail market with no access to funding. To overcome this limitation, the authors have proposed establishing a trust transitivity model based on electronic payment vouchers (EPVs) issued by the blockchain. They have used intuitionistic fuzzy set theory to consider direct and indirect trust and introduced a dynamic weight allocation strategy to incentivize and punish SMEs. The proposed model aims to provide a new idea for trust evaluation and relieve the financing difficulties of SMEs. The authors have also verified the rationality of the model through numerical examples. The manufacturing industry is constantly evolving, with new challenges requiring innovative solutions. Leng et al. [19] discuss how resilient manufacturing is a vision in the Industry 5.0 blueprint for satisfying sustainable development goals during pandemics or the rise of individualized product needs. The Industrial Internet of Things (IIoT) contains confidential data and private information that needs to be protected from cyber-attacks and unauthorized access. This article proposes a secure blockchain middleware for decentralized IoT geared towards Industry 5.0. Blockchain technology has emerged as a promising solution for various issues the manufacturing industry faces, including financing, security, and collaboration [16]. In [6] Bellavista et al. discuss about how blockchain technology is a key part of Industry 4.0. It is seen as a huge boost to the digitalization of companies, so a lot of money is being put into blockchain. But there isn’t a single blockchain technology; there are different options, but they can’t work together. The ecosystem envisioned by the authors of this article is an interoperable blockchain network that can support highly integrated supply chains in

Towards Industry 5.0

733

collaborative manufacturing. While in [22], authors talk about how the spread of COVID-19 has had a big effect on the growth of businesses. In the time after the pandemic, blockchain technology has become one of the most important tools for businesses to become more competitive in the market quickly. Blockchain technology hasn’t been widely used or adopted because it requires a lot of money from people in the supply chain. To tackle this issue, this article proposes a collaborative adoption model to help stakeholders share costs and benefits while adopting blockchain technology. Industry 5.0 represents the next stage of manufacturing, where human intelligence and machine intelligence are integrated to create a more efficient and sustainable production system [21]. A recent research study by Leng et al. [20] discusses how Industry 5.0 is the next generation of manufacturing that integrates human and machine intelligence to create a more efficient and sustainable production system. The study also investigates how blockchain technology can help small and medium-sized enterprises (SMEs) collaborate to create a more efficient supply chain. The proposal in this work involves a blockchain smart contract pyramid-driven multi-agent autonomous process control system to help SMEs achieve resilient, individualized manufacturing in Industry 5.0.

3

Proposed Methodology

The proposed private blockchain built on Hyperledger Fabric intends to encourage human-centered design, social responsibility, and sustainability in the manufacturing process for SMEs. Small suppliers, manufacturers, and other supply chain players can interact securely and transparently using this platform, facilitating collaboration in the context of Industry 5.0. This system offers a safe and transparent platform for cooperation between small suppliers, manufacturers, and other supply chain players. 3.1

Empowering Human Actors in SMEs with Blockchain

Using private blockchain technology in SME supply chains can give people more power by making things more clear, traceable, and under their control. Blockchain creates digital copies of samples for approval, saving time and money by not having to do the same work twice and making it easier to keep track of many samples [25]. Private blockchains ensure that private data is safe and can only be accessed with permission. This lets SMEs build trust with stakeholders and strengthen their relationships. By using the benefits of a private blockchain, SMEs can speed up processes, cut down on paperwork, and get rid of middlemen. This lets people make better choices, improve efficiency, and help their supply chains grow. By leveraging blockchain’s potential, the supply chain can prioritize human well-being, sustainability, and collaborative ecosystem development, aligning with the objectives of Industry 5.0. The private blockchain-based collaboration network that is being proposed can help meet the goals of humancentered design in several ways, some of which are as follows:

734

P. W. Khan et al.

– User Empowerment: By providing real-time and accurate data, blockchain enables users to make informed decisions that align with their preferences and values. – Data Ownership: With blockchain, individuals have ownership and control over their data, allowing them to determine who can access and utilize their information. – Privacy Protection: Blockchain provides privacy-enhancing features such as encryption and pseudonymity, safeguarding individuals’ personal and transactional data. – User-centric data access: Blockchain allows users to selectively disclose specific data to trusted parties, ensuring that individuals have control over the disclosure and utilization of their information. – Shared decision-making: Blockchain facilitates consensus mechanisms, empowering participants to engage in shared decision-making processes, leading to more inclusive and democratic outcomes. 3.2

Structure of the Proposed Blockchain-Based Collaboration Network

The use of blockchain technology enables real-time data exchange and collaboration while maintaining data immutability and security. Smart contracts automate cooperation procedures, resulting in a more seamless experience for supply chain players [17]. This platform will help SMEs improve their competitiveness and market access by utilizing a strong, flexible, and environmentally friendly production method. The convergence of human-centered design, social responsibility, and sustainability in the context of Industry 5.0 will help SMEs fulfill changing customer and stakeholder demands while contributing to society’s longterm development. Figure 1 shows a simplified overview of how participants interact with the blockchain network using the REST API and Fabric Client. The participants include suppliers, manufacturers, distributors, logistics providers, customers, stakeholders, and administrators. In Hyperledger Fabric, only clients permitted to use the system in a private blockchain network could interact with it. Participants communicate with the Hyperledger Fabric network via the REST server [17]. Hyperledger Fabric is a permissioned blockchain platform comprising various components designed to facilitate the development of distributed ledger applications. The platform is comprised of user interface components, chaincode, ledger, consensus protocol, membership services, peer nodes, ordering service nodes, channels, software development kits (SDKs), application programming interfaces (APIs), membership service providers (MSP), and certificate authorities (CA). These components work together to provide a secure, efficient, and flexible blockchain solution that can be tailored to the specific needs of an enterprise [1]. The user interface component allows users to interact with the blockchain network, while the chaincode specifies the transaction rules. The ledger records all transactions, and the consensus protocol ensures

Towards Industry 5.0

735

Fig. 1. Interaction of different participants with blockchain

that all nodes agree on the ledger’s state. Membership services manage participants’ identities, while peer nodes execute transactions and maintain the ledger. Ordering service nodes create blocks and distribute them to the network, while channels provide a mechanism for creating sub-networks with restricted access. SDKs and APIs allow developers to build applications that interact with the blockchain network, while MSP offers a means for managing access control policies. Finally, the CA issues digital certificates and manages public keys to secure communication between network participants. Figure 2 shows the transaction flow for the interaction of participants with smart contracts on the blockchain. Manufacturers create smart contracts for their products and deploy them on the blockchain network, which distributors, logistics providers, and customers join. These participants interact through the blockchain network to send and confirm purchase orders, shipping requests, and product receipts. Customers place orders with distributors, who confirm and create shipping requests for logistics providers. The smart contract ensures that transactions are validated against predefined rules and regulations and, if approved by the consensus algorithm, are added to the blockchain ledger. 3.3

Private Blockchain Platform

Blockchain is a term for the technology of sharing encrypted data with decentralized peers. It is a way to keep track of information using cryptographic hash functions. Blockchain also has an integral part of a smart contract that makes it possible to automate a trade between two parties without a mediator [8]. Hyperledger Fabric is used in this work as a private blockchain platform. It was created by IBM as an open-source platform. The Hyperledger Composer is an open-source framework for building and deploying private blockchains [9]. The Hyperledger Composer is a tool for creating and deploying private blockchains. It includes a command-line interface (CLI) that allows users to create smart

736

P. W. Khan et al.

Fig. 2. Transaction flow for the interaction of participants with smart contracts on the blockchain

contracts, participants, and assets. The four main components of a blockchain, namely the blocks, the smart contracts, the transactions, and the channels, are described in the following sections. Block. The first block in the blockchain is called the “genesis block.” It is where the ledger starts, but it does not have any transaction records in it [7]. Alternatively, it puts the network’s initial state into the configuration transaction. A block has three components: the header, the transaction, and the metadata. When a block is made, metadata is written to it. Important information about the transaction is stored in the block’s header. It has the number of the current block, the hash of the current block, and the hash of the block before it. The information stored in the blockchain comprises different transactions [10]. Each transaction contains a unique ID, header, cryptographic signature, proposal, response, and endorsement. Transactions. Before a transaction begins, the request is verified by the membership service provider (MSP) [26]. Figure 3 demonstrates how transactions are processed in Hyperledger Fabric by application. The transaction process begins when the authenticated client app connects to the peer. It sends the transaction proposal to the peer after connecting. Next, the peer sends this proposal to chaincode, and chaincode sends the ledger query that was asked for. Chaincode can also suggest a change to the ledger. Then, the ledger’s response returns

Towards Industry 5.0

737

Fig. 3. Sequence diagram of application transaction processing

to the peer, sending it to the proposal. After the response application gets the request, it sends a claim that a transaction should be ordered to the person who made the request. The orderer gets all of the network’s transactions, puts them into blocks, and sends them to all peers. The peer gets the block, checks it, and then updates the ledger. Channels. In Hyperledger Fabric, channels play a crucial role in the secure and efficient collaboration of organizations [2], making them an essential component of Industry 5.0. Our Manufacturing Supplier Collaboration Network for SMEs also utilizes this powerful feature, with two channels specifically designed for the production and distribution phases of the supply chain process. Channel “Production” has one organization named “Manufacturer,” while Channel “Distribution” has two organizations named “Distributor” and “Logistics Provider.” Each organization within a channel has its copy of the ledger, and only authorized members can access the channel. In order to illustrate the connection between the two channels, Fig. 4 shows the Fabric runtime connecting the two channels through the anchor peers of each organization. An anchor peer is a node that is a member of multiple channels and is responsible for maintaining a reference to the other channels in which it participates. Each organization has selected an anchor peer to represent them on the production and distribution channels. When the organizations need to interact, they send messages to their anchor peers, who then relay the messages to the anchor peers on the other channel. This allows the two channels to communicate through a secure and reliable network while maintaining the confidentiality and privacy of each organization’s data.

738

P. W. Khan et al.

Fig. 4. Fabric runtime connecting the two channels through the anchor peers

Smart Contract. A smart contract is a computer program that handles transactions and gets to blocks and records on the blockchain. It adds something useful to the blockchain. A smart contract is kept in a database that is shared among many computers [11]. A smart contract lets the organization put limits or checks on transactions that they want. Smart contracts also have different rules and business logic written into them. It can be seen as an example of a deal between two people or groups. Both sides agree on the rules for how products or services will be traded. Smart contracts, on the other hand, use programming code to ensure that these rules are followed. Smart contracts are made with computer code that has parts for starting the agreement, doing things, and ending it.

Towards Industry 5.0

4

739

Performance Analysis

This section demonstrates an overview of the proposed method’s evaluation using a Hyperledger Fabric composer playground, along with the results. Table 1 details the simulation environment utilized to develop the system. The system was developed on a computer with an AMD Ryzen 9 5950X 16-Core Processor and 32 GB of RAM running Ubuntu 20.04.3 LTS. Table 1. Blockchain simulation environment Sr # Component

Description

1

Processor

AMD Ryzen 9 5950X 16-Core Processor

2

RAM

32 GB

3

OS

Ubuntu 20.04.3 LTS

4

Hyperledger Fabric v 2.4

5

CLI

Hyperledger Fabric CLI

6

IDE

Visual Studio Code with Hyperledger Fabric extension

The latest version of the open-source framework, Hyperledger Fabric version 2.4, was used to take advantage of the most recent updates. Hyperledger’s Composer is used as a development and testing tool to rapidly create and test a configurable blockchain interface [15]. The Playground’s Composer CLI tool is used to code business network definitions. Table 2. User-centric access controls and data management capabilities Actor

Access Level Data Visibility Data Modification

Admin

Read/Write

Full

Full

Suppliers

Read/Write

Limited

Limited

Manufacturers Read/Write

Limited

Limited

Distributors

Read/Write

Limited

Limited

Retailers

Read/Write

Limited

Limited

Customers

Read

Limited

No

Regulators

Read

Full

No

Table 2 presents access levels and data management capabilities for different actors in a proposed system. It provides evidence of a user-centric approach to data access and control. Admin has full access to data visibility and modification, while other actors, such as Suppliers, Manufacturers, Distributors, and Retailers, have limited access. Customers have limited visibility and no data modification rights, while Regulators have full visibility but no modification rights.

740

P. W. Khan et al.

Fig. 5. Average transaction latency evaluation

We have also carried out an investigation into the blockchain network to examine how well our proposed system would work. Many experiments were conducted using different performance methods to test how well the proposed blockchain system works through various metrics evaluation. To do this, we use Hyperledger Caliper, which is an open-source simulator from IBM [3]. Three groups of 250, 500, and 750 parts were created for testing. We wrote down at least a percentage of the maximum and maximum delay (in milliseconds) on the proposed blockchain platform. Figure 5 shows how long a transaction takes on average when transaction latency increases. Latency Lt is calculated using the Eq. 1, where Tc is the confirmation time, Nt is the network threshold, and Ts is the submission time. The underlying transaction refers to the case where the time taken by a user to make a request on the network has increased. After the transmission rate of the optimal tps, which is also called the best transmission rate, has reached its optimal tps, the delay has become much worse. Lt = Tc × Nt − Ts

(1)

The average transaction throughput of the proposed private blockchain network has been analyzed, as shown in Fig. 6. The x-axis displays the throughput (t) in transactions per second (TPS), while the y-axis represents the send rate. The Eq. 2 has been utilized to calculate the transaction throughput T Pt , where Vt denotes the total number of valid transactions, and T indicates the total time in seconds. A higher throughput indicates the blockchain network can handle more transactions per second, while a lower throughput signifies slower processing. The graph shows that the proposed blockchain network has a high throughput, allowing faster and more efficient transaction processing. This increased throughput would enhance the platform’s usability and user experience by reducing transaction times and increasing transactional efficiency.

Towards Industry 5.0

741

Fig. 6. Average transaction throughput evaluation

 (Vt ) T Pt =  (T )

(2)

Figure 7 is a radar chart that shows how the delay of the network changes in regard to a varying number of people using the network. The median time for 250, 500, and 750 users was 117, 115, and 369 ms, respectively. The highest was 213, 352, and 761 ms, sequentially. Get request transaction latency evaluation showed

Fig. 7. Get request transaction latency evaluation

742

P. W. Khan et al.

that as the number of users increased, so did the time taken for the system to work. Depending on the network’s number of nodes, Hyperledger Fabric could handle a certain number of transactions simultaneously. If there were six nodes in the network, it could handle up to 10,000 concurrent transactions [24]. Therefore, implementing a blockchain-enabled supplier cooperation network for SMEs in Industry 5.0 is feasible. However, the limitations and challenges require careful consideration. Hence, SMEs must assess their resources, expertise, and readiness for change before implementing such networks. When implemented correctly, blockchain-enabled supplier collaboration networks can provide SMEs with transparency, cost-effectiveness, and efficiency, improving their competitiveness and sustainability.

5

Conclusion

Blockchain technology can help fulfill the goals of human-centered design, social responsibility, and sustainability in the manufacturing process by enabling transparency and traceability of the whole supply chain. The article introduces a blockchain-based supplier collaboration network to assist small and mediumsized firms (SMEs) in transitioning to Industry 5.0. The suggested solution allows small suppliers, manufacturers, and other supply chain players to collaborate securely and transparently. Blockchain technology creates a secure, immutable, real-time data-sharing environment, increasing SMEs’ competitiveness and offering them wider market access. Smart contracts are used to streamline collaboration operations and link blockchain with new technologies such as the Internet of Things and machine learning. The proposed private blockchain built on Hyperledger Fabric intends to encourage human-centered design, social responsibility, and sustainability in the manufacturing process for SMEs. The use of the REST API and Fabric Client ensures that the various participants can interact with the blockchain network securely. Limitations and possible pitfalls of this study include the potential transaction fees associated with executing smart contracts and network actions, which can vary based on network congestion and resource consumption. The findings are based on a hypothetical scenario and would benefit from real-world implementation and evaluation. Implementing new technology may face resistance from stakeholders due to a preference for traditional methods. Future research can also examine the feasibility of implementing such networks in other industries to determine if the results are consistent across different contexts. Furthermore, the study’s findings can be utilized to provide guidelines and best practices for the deployment of blockchain-enabled supplier cooperation networks. Acknowledgements. This material is based upon work supported by: i) the U.S. Department of Energy’s Office of Energy Efficiency and Renewable Energy (EERE) under the Advanced Manufacturing Office Award Number DE-EE0007613. Disclaimer: This report was prepared as an account of work sponsored by an agency of the United States Government. Neither the United States government nor any agency thereof, nor any of its employees, makes any warranty, express or implied, or assumes any legal

Towards Industry 5.0

743

liability or responsibility for the accuracy, completeness, or usefulness; ii) the National Science Foundation under Grant No. 2119654. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation.

References 1. Abbas, K., Afaq, M., Ahmed Khan, T., Song, W.C.: A blockchain and machine learning-based drug supply chain management and recommendation system for smart pharmaceutical industry. Electronics 9(5), 852 (2020) 2. Abbas, K., Tawalbeh, L.A., Rafiq, A., Muthanna, A., Elgendy, I.A., Abd El-Latif, A.A.: Convergence of blockchain and IoT for secure transportation systems in smart cities. Secur. Commun. Networks 2021, 1–13 (2021) 3. Ampel, B., Patton, M., Chen, H.: Performance modeling of hyperledger sawtooth blockchain. In: 2019 IEEE International Conference on Intelligence and Security Informatics (ISI), pp. 59–61. IEEE (2019) 4. Bareche, I., Xia, Y.: Selective velocity distributed indexing for continuously moving objects model. In: Wen, S., Zomaya, A., Yang, L.T. (eds.) ICA3PP 2019. LNCS, vol. 11945, pp. 339–348. Springer, Cham (2020). https://doi.org/10.1007/978-3030-38961-1_30 5. Bareche, I., Xia, Y.: A distributed hybrid indexing for continuous KNN query processing over moving objects. ISPRS Int. J. Geo Inf. 11(4), 264 (2022) 6. Bellavista, P., Esposito, C., Foschini, L., Giannelli, C., Mazzocca, N., Montanari, R.: Interoperable blockchains for highly-integrated supply chains in collaborative manufacturing. Sensors 21(15), 4955 (2021) 7. Decker, C., Wattenhofer, R.: Information propagation in the bitcoin network. In: IEEE P2P 2013 Proceedings, pp. 1–10. IEEE (2013) 8. Feng, Q., He, D., Zeadally, S., Liang, K.: BPAS: blockchain-assisted privacypreserving authentication system for vehicular ad-hoc networks. IEEE Trans. Industr. Inf. 16, 4146–4155 (2019) 9. Gaur, N., Desrosiers, L., Ramakrishna, V., Novotny, P., Baset, S.A., O’Dowd, A.: Hands-On Blockchain with Hyperledger: Building Decentralized Applications with Hyperledger Fabric and Composer. Packt Publishing Ltd. (2018) 10. Gupta, S., Sadoghi, M.: Blockchain transaction processing (2019) 11. Hang, L., Kim, D.H.: Design and implementation of an integrated IoT blockchain platform for sensing data integrity. Sensors 19(10), 2228 (2019) 12. Jasra, J., Hunjra, A.I., Rehman, A.U., Azam, R.I., Khan, M.A.: Determinants of business success of small and medium enterprises. Int. J. Bus. Soc. Sci. 2(20) (2011) 13. Jiang, R., et al.: A trust transitivity model of small and medium-sized manufacturing enterprises under blockchain-based supply chain finance. Int. J. Prod. Econ. 247, 108469 (2022) 14. Karadag, H.: Financial management challenges in small and medium-sized enterprises: a strategic management approach. EMAJ Emerg. Markets J. 5(1), 26–40 (2015) 15. Khan, P.W., Byun, Y.C., Park, N.: A data verification system for CCTV surveillance cameras using blockchain technology in smart cities. Electronics 9(3), 484 (2020) 16. Khan, P.W., Byun, Y.C., Park, N.: IoT-blockchain enabled optimized provenance system for food industry 4.0 using advanced deep learning. Sensors 20(10), 2990 (2020)

744

P. W. Khan et al.

17. Khan, P.W., Byun, Y.: A blockchain-based secure image encryption scheme for the industrial internet of things. Entropy 22(2), 175 (2020) 18. Lanzini, F., Ubacht, J., De Greeff, J.: Blockchain adoption factors for SMES in supply chain management. J. Supply Chain Manag. Sci. 2(1–2), 47–68 (2021) 19. Leng, J., et al.: Secure blockchain middleware for decentralized IIoT towards industry 5.0: a review of architecture, enablers, challenges, and directions. Machines 10(10), 858 (2022) 20. Leng, J., Sha, W., Lin, Z., Jing, J., Liu, Q., Chen, X.: Blockchained smart contract pyramid-driven multi-agent autonomous process control for resilient individualised manufacturing towards industry 5.0. Int. J. Prod. Res. 1–20 (2022) 21. Leng, J., et al.: Industry 5.0: Prospect and retrospect. J. Manuf. Syst. 65, 279–295 (2022) 22. Liu, W., Liu, X., Shi, X., Hou, J., Shi, V., Dong, J.: Collaborative adoption of blockchain technology: a supply chain contract perspective. Front. Eng. Manag. 10(1), 121–142 (2023) 23. Mittal, S., Romero, D., Wuest, T.: Towards a smart manufacturing toolkit for SMEs. In: Chiabert, P., Bouras, A., Noël, F., Ríos, J. (eds.) PLM 2018. IAICT, vol. 540, pp. 476–487. Springer, Cham (2018). https://doi.org/10.1007/978-3-03001614-2_44 24. Nasir, Q., Qasse, I.A., Abu Talib, M., Nassif, A.B.: Performance analysis of hyperledger fabric platforms. In: Security and Communication Networks 2018 (2018) 25. Repsol: Repsol Technology Lab: Industry (2023). https://www.repsol.com/en/ technology-and-digitalization/technology-lab/industry/index.cshtml. Accessed 06 Jun 2023 26. Zhang, S., Zhou, E., Pi, B., Sun, J., Yamashita, K., Nomura, Y.: A solution for the risk of non-deterministic transactions in hyperledger fabric. In: 2019 IEEE International Conference on Blockchain and Cryptocurrency (ICBC), pp. 253–261. IEEE (2019)

A Stochastic-Based Model to Assess the Variability of Task Completion Times of Differently Aged and Experienced Workers Subject to Fatigue Andrea Lucchese1(B)

, Salvatore Digiesi1

, and Giovanni Mummolo2

1 Department of Mechanics, Mathematics and Management, Polytechnic University of Bari,

70126 Bari, Italy {andrea.lucchese,salvatore.digiesi}@poliba.it 2 Ionic Department in Legal and Economic System of Mediterranean: Society, Environment, Culture, Università degli Studi di Bari Aldo Moro, 70121 Bari, Italy [email protected]

Abstract. In current industrial scenarios, the new paradigm of Industry 5.0 (I5.0) is gaining interest: considering the Industry 4.0 paradigm as a base, I5.0 is aimed at reaching a more sustainable, human-centric, and resilient industry. Under the I5.0 human-centric perspective, behavioural issues assume high criticality, thus requiring a more reliable prediction of operators’ performances. In manual assembly lines, operators’ performances are characterized by a stochastic behaviour over time. System’s and operator’s features cause the variability of task completion time: the former is related to properties of the work environment (e.g., ergonomics, cycle time), the latter is related to the intrinsic stochastic behaviour of operators. Furthermore, workers’ features and their different attitudes to becoming fatigued, influence performance variability. In this context, the authors propose a new stochastic model that expresses the variability of execution times of operators involved in manual assembly lines by considering their differences in age, experience, and fatigue state. The novelty of the proposed model relies on considering the stochastic behaviour of workers influenced by age, experience as well as fatigue. The effectiveness of the proposed model is tested through numerical experiments of a job rotation scheduling problem to maximize productivity with proper worker-workstation assignments. Keywords: Manual Assembly · Human Performance Modelling · Task Completion Time · Stochastic Variability · Workforce Diversity · Job Rotation

1 Introduction Industry 4.0 has been characterized by implementing multiple technologies, contributing to build more productive and efficient systems. Nevertheless, while I4.0 is technologydriven, the new paradigm I5.0 is more focused on creating more sustainable, resilient © IFIP International Federation for Information Processing 2023 Published by Springer Nature Switzerland AG 2023 E. Alfnes et al. (Eds.): APMS 2023, IFIP AICT 689, pp. 745–759, 2023. https://doi.org/10.1007/978-3-031-43662-8_53

746

A. Lucchese et al.

and human-centric systems [1]. Although the use of technologies contributed to the development of operator 4.0 [2], in current manufacturing systems characterized by intensive manual activities, I4.0 technologies could be challenging to implement due to the high motor content and low cognitive content of tasks, movement and space limitations (complex implementation of exoskeletons), different level of familiarity with new technologies (e.g. experience with high-tech tools), and anthropometric features of workers (e.g., age) [3]. This scenario may occur in intensive manual assembly lines where operators’ performance is strongly influenced by their attributes, such as anthropometric features (e.g., age), skills and experience level. In these contexts, operators perceive high levels of fatigue, lowering their performance over time. Therefore, workers’ behaviour must be appropriately modelled by incorporating their specific qualities and features. In order to assign the correct operator to the specific activity to be performed, guaranteeing the best performance, it must be considered that the behaviour of operators is stochastic, and that the degree of variability changes over time. In estimating the time that operators require to carry on a manual activity (execution time), a specific time variability can be detected. Therefore, workers’ behaviour is variable, and their stochasticity can be expressed through a given probability distribution of execution times. The present paper aims to provide an innovative method to model the stochastic behaviour of workers by considering their differences quantitatively. The proposed model is finally tested through numerical experiments of a job rotation scheduling problem by defining the proper worker-workstation assignments able to maximize operators’ performance and productivity. The remainder of this paper is organized as follows: the second section is devoted to analysing features that characterize the workforce and impact on workers’ performance; the third section is devoted to describing the mathematical formulation of the job rotation problem and modelling the stochastic behaviour of operators. The third section is focused on the numerical experiments and discussion of results, while in the last section, conclusions are provided.

2 Workforce Diversity During manual activities, operators are subject to physical fatigue, leading to a lowering of the ability to execute motor activities [4–6]. In manual assembly lines, while operators perform manual handling activities, they must reach, grasp, move, position, and release multiple objects of a given weight, volume, geometry, and material. During the working day, these activities reduce the capacity to generate force and increase the reaction time, leading to lower performance over time; furthermore, physical fatigue increases when operators are subject to an intensive pace since they must perform manual tasks at higher speeds [7]. Physical fatigue can be detected by monitoring the energy expenditure through oxygen consumption and heart rate variability and by evaluating the maximum voluntary contraction and maximum endurance time through the EMG [8, 9]. Moreover, in manual assembly tasks, wearable sensors, such as inertial measurement units (IMUs) and heart rate monitoring, have been coupled to evaluate physical fatigue [10]. If appropriate strategies are not implemented, the fatigue accumulates over time, leading to the lowest

A Stochastic-Based Model to Assess the Variability of Task Completion Times

747

performance during the last part of the work shift. These effects bring to the increase of expected values and of variability of execution times [11]. To avoid the fatigue accumulation phenomenon, the recovery time, defined as the time needed to recover from the fatigue induced by work-related activities, is considered [12]. The recovery time location is defined strategically, giving operators short breaks, or longer breaks (lunch breaks), avoiding excessively long shifts. Moreover, the adverse effects induced by fatigue tend to increase faster in case of older workers, with a more significant influence in paced systems. Physical fatigue increases by tightening the pace of manual assembly tasks [13]. In particular, by increasing the intensity of assembly tasks, older workers find it more challenging to deal with shorter times [14]. Nowadays, the theme of ageing workforce is critical; for this reason, one of the main goals of I5.0 is to reach more human-centric inclusive manufacturing systems able to build a suitable working environment for an ageing workforce. It is estimated that in 2050 half of workers will be aged over 50, and the presence of aged workers in production roles will impact the performance of manufacturing systems more than ever [15]. This phenomenon is particularly observable when considering subjects differently aged with the same experience level [16]. Nevertheless, although it is commonly associated a decline of working performances (both motor and physical) with ageing, it must not be neglected that worker’s productivity also depends on the experience level. In several cases, it has been observed that older workers are also the most experienced ones [17]. In the context of manual assembly lines, there is evidence that experience increases with age, leading to increasing the overall productivity and decreasing the severity of errors: the experience acquired by older workers allows for the prevention of severe errors [18]. There are two opposite phenomena that characterize the ageing workforce: the first relates to the increase of fatigue effects in paced manufacturing systems, bringing to an increase in the execution times; the second links the enhancement of experience level with the increase of workers’ age: older workers tend to be the most experienced ones, bringing to the prevention of severe errors during the execution of manual tasks. In this work, these phenomena are jointly considered in modelling the variability of execution times of workers performing assembly activities. Their behaviour is affected by workforce differences and expressed in stochastic terms through a given probability density function. The proposed model is then applied to a job rotation scheduling problem, whose mathematical formulation is proposed in the following.

3 Mathematical Formulation In this section, the stochastic behaviour of operators is modelled, and the mathematical formulation of the job rotation scheduling problem is proposed. The following assumptions are defined for the job rotation scheduling problem: 1. A work shift consists of a certain number of time slots of equal length; job rotation is allowed only between time slots; when a worker is assigned to perform a task at a new workstation, both time to move to the new workstation and setup time are assumed to be negligible.

748

A. Lucchese et al.

2. The sets of working elements (basic movements) that define the manual assembly activity for each workstation is fixed; the sum of times associated to basic movements is equal to the nominal task completion time. 3. Assembly activities to be performed at workstations have the same complexity. 4. The number of workstations and workers is fixed: the number of workers can be equal to or greater than the number of workstations. 5. Each manual assembly activity can be executed only by one worker in each time slot. 6. For each time slot, the production target is fixed, as well as the cycle time. 7. Task execution times of workers are not constant over time; for each time slot and workstation, the expected execution time of workers can differ from the cycle time, depending on fatigue and experience level. 8. Task execution times of workers are not deterministic; for each time slot and workstation, execution time of workers is normally distributed, expressing the intrinsic stochasticity of the movement time; the standard deviation of execution time is affected by fatigue and age. 9. Only fatigue phenomenon is considered affecting operators’ performance over time; other phenomena like learning and forgetting are considered negligible. In the following all the indices, parameters, and variables, adopted are listed and described. Indices w index for workers j index for workstations k index for time slots in a work shift Input Parameters W J K Ow WSj Sk Tk Qk∗ CT tj CVmin σj,min fsw fq w fμwk fσwk B

number of workers (w = 1, . . . , W ) number of workstations (j = 1, . . . , J ) number of time slots in a work shift (k = 1, . . . , K) w th worker j th workstation k th time slot duration of Sk [min] theorical production target for each Sk [units/time slot] cycle time, maximum time allowed to perform the tasks in all workstations [s] nominal task completion time at WSj [s] minimum value of the coefficient of variation of the execution time minimum value of the standard deviation of the execution time at WSj [s] factor modelling the experience of Ow factor modelling the quality of production of Ow factor modelling the variation of the expected execution time (μwjk ) of Ow at WS j in Sk due to fatigue factor modelling the variation of the standard deviation of execution time (σwjk ) of Ow at WS j in Sk due to fatigue buffer size

A Stochastic-Based Model to Assess the Variability of Task Completion Times

749

Output Parameters twjk μwjk σwjk mintwj pwjk Qeff

execution time of Ow at WS j in Sk [s] expected value of twjk [s] standard deviation of twjk [s] incompressible execution time of Ow at WS j [s] probability that twjk ≤ CT actual production in a work shift [units/work shift]

Variables xwjk 1 if Ow is assigned at WS j in time slot Sk , 0 otherwise (xwjk ∈ {0; 1}) The objective function (O.F.) is: max P

W  K J  

xwjk · pwjk

(1)

w=1 j=1 k=1

The O.F. allows to evaluate the best worker-workstation assignments maximizing ∗ ). O.F. is subject to: the probability of performing the activity at WS j in Sk (pwjk J 

xwjk = 1∀w = 1, . . . , W ∀k = 1, . . . , K

(2)

j=1 W 

xwjk = 1∀j = 1, . . . , J ∀k = 1, . . . , K

(3)

w=1

Each Ow can be assigned only to WS j in Sk (Eq. 2). Each manual assembly activity at WS j can only be executed by an Ow in Sk (Eq. 3). ∗ into units produced during the work shift, information about To turn the solution pwjk the buffer size is required. In this context, two opposite scenarios are considered: in the former, buffers’ size is zero (B = 0), in the latter an infinite buffers’ size is considered (B = ∞). The following equations express the units produced in a work shift in the two cases: Qeff ,B=0 =

K  k=1

Qeff ,B=∞ =

K  k=1

∗ Qk∗ · min(pwjk · fq w )

(4)

∗ Qk∗ · max(pwjk · fq w )

(5)

j

j

∗ represents the maximum probability value that O performs activity at WS in S pwjk w j k with a completion time twjk ≤ CT . To consider the actual production of the work shift, one needs to estimate the scraps generated during each Sk ; therefore, factor fqw is adopted considering the quality of production, i.e., errors made by the worker w: the higher the value of fqw , the higher the quality (perfect quality in case of fqw = 1). fqw Depends only

750

A. Lucchese et al.

on worker’s experience and does not depend on WSj and Sk , since it is assumed that all the tasks have the same complexity. In case of an infinite buffer size (B = ∞), the one preceding the fastest operator on the line plays the role of an infinite source of workpieces. In this case the actual daily ∗ · f )). On the contrary, production depends on the fastest operator in each Sk (max(pwjk qw in case of no buffer, the bottleneck is given by the slowest operator. In this case the ∗ · f )). The actual daily production depends on the slowest operator in each Sk (min(pwjk qw ∗ theoretical production target of a time slot Qk is expressed as: Qk∗ =

Tk · 60 CT

(6)

3.1 Modelling the Stochastic Behaviour of Operators During assembly activities, execution times of workers are affected by variability, with corresponding values that are influenced by fatigue state, experience level and age. Therefore, the behaviour of workers is variable, and stochasticity of execution times can be expressed through a given probability distribution. In case of paced systems, workers’ execution times could be assumed to be normally distributed [19–22], i.e. the stochastic behaviour of workers can be represented by a normal probability density function with expected value μwjk , and standard deviation σwjk , with subscripts w, j, k denoting the influence of the specific worker, assembly activity, and fatigue state. μwjk is expressed as: μwjk = t j · fsw · fμwk

(7)

where t j is the nominal task completion time evaluated by the sum of times associated to the basic movements (i.e. reaching, grasping, moving, placing etc.…), given by MTM (Method-Time Measurement) [23]. In case of M basic  movements that define the manual activity to be performed at workstation WSj , t j = M m=1 MTMm,j . σwjk is expressed as: σwjk = σj,min · fσwk

(8)

To obtain σwjk , a reference standard deviation, expressed by σj,min , must be firstly identified. The procedure for estimating σj,min is described in Appendix A. The normal probability density function is expressed as:   f twjk =



σwjk

− 12 1 ·e √ · 2π

twjk −μwjk σwjk

2

(9)

Equation 9 expresses the probability distribution of execution times. The probability that Ow performs the activity at WSj in Sk within the cycle time CT can be expressed as:  CT pwjk =





mintwj f twjk dt ∞   mintwj f twjk dt

(10)

A Stochastic-Based Model to Assess the Variability of Task Completion Times

751

∞     where pwjk is normalized through mintwj f twjk dt. Moreover, f twjk refers to a distribution not limited inferiorly or superiorly, while in real cases of assembly activities, execution times of workers cannot obviously be lower than zero or, more often, lower than a certain value. Therefore, pwjk is evaluated by considering its lower limit (in terms of time). This value is often defined as the 75% of t j , expressing the minimum value of time below which Ow is not able to complete the assigned manual activity at workstation WSj in any case [24]. Since the nominal task completion time t j is associated to a fully experienced worker (fsw ,min = 1 in our case), the minimum incompressible time is influenced by the experience level of Ow , as defined in the following: mintwj = t j · fsw · 0.75

(11)

The higher the worker’s experience, the lower the value of fsw ; in case of a fully experienced operator (fsw ,min = 1), it is obtained mintwj = t j · 0.75, as in [24]. In the present section we detailed how the stochastic behaviour of workers in executing manual activities and affected by experience, age, and fatigue, can be expressed by the normal probability density function of execution times. The model defined to maximize the production in a work shift by considering the stochastic behaviour of workers has been tested by means of numerical experiments of a job rotation scheduling problem, detailed in the next Section.

4 Numerical Experiments Fatigue phenomenon causes changes in the expected values and variability of execution times [11]: in our model, both effects are modelled by means of a change in the values of fσwk and fμwk factors. For the numerical case, it is assumed a linear increase of fatigue effect over time, with values of fσwk and fμwk averaged for each Sk (Fig. 1). An 8-h work shift is divided into three time slots (K = 3) of 140 min each, with a short break of 15 min after S1 , and a lunch break of 45 min after S2 . Two different scenarios are considered: in

Fig. 1. Trends of fatigue accumulation in scenarios “B” and “G”, and corresponding evaluation method of fμwk and fσwk for the numerical experiments

752

A. Lucchese et al.

the first one, no recovery effect takes place during breaks (worst case, scenario “B”); in the second, an ideal recovery effect is observed during breaks independently from their durations (best case, scenario “G”) (Fig. 1). In both scenarios, the presence of breaks between consecutive time slots makes assumption 1 more realistic. Values of fμwk and fσwk are set based on scientific literature data. During manual assembly activities it is observed that due to fatigue effect, μwjk can increase up to + 20% (fμwk = 1.2), and σwjk up to +100% (fσwk = 2) [11]. Fatigue increases over time, and increases more in case of aged workers [16]: this scenario can refer to an aged worker at the last period of the work shift (time slot S3 ), where fμwk and fσwk reach the maximum values (1.2 and 2 respectively). Therefore, by knowing the duration of each slot Sk , and each break, values of fμwk and fσwk in case of an aged or ‘Senior’ worker (O1 ) can be estimated (Table 1a, rows corresponding to O1 ). To consider fμwk and fσwk for younger workers, scientific literature data of differently aged workers are considered. In case of repetitive assembly tasks, the expected execution time of workers increases with age, reaching the maximum increment in case of ‘Senior’ workers (O1 - 62 years old on average); for ‘Junior’ workers (33 years old on average), increments in observed execution times are around 12% of those observed for O1 , while for ‘Middle’ workers (47 years old on average) they are around 40% of those observed for O1 [16]. Therefore, in case of ‘Middle’ (O2 ) and ‘Junior’ (O3 ) workers categories, the corresponding values of fμwk and fσwk are estimated as the 12% and 40% of the values calculated for ‘Senior’ workers category (Table 1a). For the numerical experiments, three workers (W = 3) are considered, each one belonging to a different age category. Moreover, as discussed in the second section, older workers (O1 ) are assumed to be the most experienced ones [17], while the youngest workers (O3 ) the most unexperienced ones. Remembering that an unexperienced worker can bring to an increase of the expected execution time up to + 20% [25], values of the parameter fsw can be estimated for the three workers (Table 1b). Moreover, since the experience acquired enables the prevention of severe errors [18], the oldest worker is associated to the highest production quality (highest fqw ) (Table 1b). Table 1. 1a. (left): values of fμwk and fσwk for each operator, workstation and time slot; 1b. (right): values of fsw and fqw for each operator

“B”

“G”

1.034 1.014 1.004 1.034 1.014 1.004

1.110 1.044 1.013 1.034 1.014 1.004

1.200 1.080 1.024 1.034 1.014 1.004

1.171 1.068 1.020 1.171 1.068 1.020

1.549 1.220 1.066 1.171 1.068 1.020

2.000 1.400 1.120 1.171 1.068 1.020

1 1.1 1.2

0.99 0.96 0.93

A Stochastic-Based Model to Assess the Variability of Task Completion Times

753

For the numerical experiments, three workstations are considered (J = 3), each one with a specific nominal task completion time t j . Values adopted for the numerical case are in Table 2. Table 2. Input parameters and corresponding values adopted for numerical experiments Input parameters

Value(s)

W

3

J

3

K

3

Tk = T [min]

140

CT [s]

18, 20

Qk∗ [units/time slot]

466 (CT =18 [s]), 420 (CT =20 [s])

t j [s]

t 1 = 12, t 2 = 13, t 3 = 14

The value of CT is set a priori. For the numerical experiments, two different values of CT (18 and 20 s) are considered to estimate how properties of the system (i.e., production target Qk∗ (Eq. 6)) affect the stochastic behaviour of workers and impact their performance in terms of productivity. Moreover, these CT values are in line with literature data, since in case of assembly lines, it is observed that CT can be up to the 120% of the expected execution time [26, 27]. Therefore, by knowing μwjk and σwjk for each Ow at WSj in Sk , the stochastic behaviour of workers executing assembly activities can be expressed through Eq. 9. An example of the normal probability density function for the three workers at WS1 (t 1 = 12s) in the first (S1 ) and last (S3 ) time slot in scenario “B” with CT = 20s is depicted in Fig. 2 and Fig. 3 respectively. It can be observed that when workers are subject to the minimum fatigue state (first time slot S1 ), the effect of experience is clearly visible, with the minimum μwjk in case of the most experienced worker (O1 ) (Fig. 2). On the contrary, in the last time slot (S3 ), the positive effect due to experience is mitigated by the associated fatigue state (Fig. 3); the effect of fatigue on execution times is particularly evident for older workers: the increase of the expected value as well as the standard deviation of execution times is the greatest for the older workers, as observed in real cases. In Figs. 2 and 3 the incompressible time does not change since it is characterized only by the nominal task completion time of the specific WSj , and the experience level of OW (Eq. 11).

754

A. Lucchese et al.

Fig. 2. Probability distribution function of twjk for the three workers at workstation WS1 , in time slot S1 , in scenario “B”

Fig. 3. Probability distribution function of twjk for the three workers at workstation WS1 , in the last time slot S3 , in scenario “B”

4.1 Results The numerical experiments have been executed on a PC equipped with a 2.60 GHz Intel Core i7 CPU and 16.0 GB RAM using the MATLAB® software (version R2022b), and the total computational runtime was less than 1 min. By considering the stochastic behaviour of workers in the job rotation scheduling problem, solutions that define the worker-workstation assignments able to maximize the workers’ productivity (Eq. 1) are depicted in Table 3.

A Stochastic-Based Model to Assess the Variability of Task Completion Times

755

Table 3. Best worker-workstation assignments in each scenario considered CT = 18 [s]

CT =20 [s] “G”

“B”

WS1

WS2

WS3

WS1

WS2

WS3

S1

O3

O2

O1

O3

O2

O1

S2

O3

O2

O1

O3

O2

O1

S3

O3

O2

O1

O3

O2

O1

S1

O3

O2

O1

O3

O2

O1

S2

O1

O3

O2

O3

O2

O1

S3

O1

O2

O3

O1

O3

O2

Table 4. Actual production in a work shift and corresponding production efficiency for each scenario corresponding to the best worker-workstation assignment Qeff ,B=∞

Qeff ,B=0

Qeff ,B=∞ K ∗ · 100 k=1 Qk

Qeff ,B=0 K ∗ · 100 k=1 Qk

CT =20 [s] Qk∗ =420

“G”

1242

1170

98.57%

92.86%

“B”

1234

1142

97.94%

90.63%

CT =18 [s] Qk∗ =466

“G”

1335

1290

95.49%

92.27%

“B”

1285

1132

91.92%

80.97%

In scenario “G”, i.e., when a total recovery effect takes place during breaks, the best worker-workstation assignment is given by allocating the most experienced-aged operator at the workstation with the highest nominal task completion time (the closest to the CT ) (O1 −WS3 ), while the most unexperienced-youngest operator at the workstation with the lowest nominal task completion time (the furthest from the CT ) (O3 − WS1 ). This solution does not vary with the CT values considered (20 [s] and 18 [s]) and does not vary over time. In this case it is always preferrable to rely on the most experienced-aged operator for the most critical workstations, i.e., when the nominal task completion time is the closest to CT . In scenario “B” (when no recovery effect takes place during breaks) and CT = 20s, the best worker-workstation assignment is the same as in scenario “G” only for the first time slot (S1 ), while for the other two time slots (S2 , S3 ), the best solution allocates the most experienced-aged operator at the first workstation (O1 − WS1 ); from this result, it can be stated that when fatigue accumulates over time, and due to the different response of operators to fatigue, it is preferrable to assign younger operators at the workstations characterized by higher nominal task completion times (WS2 , WS3 ), rather than the most experienced-aged one. The effect of fatigue influences the performance of aged operators to such an extent that less experienced but younger operators are preferred in workstations with tightener times. The same phenomenon occurs in scenario “B” with CT = 18s, but with a minor extent, since the most experienced-aged operator O1 is

756

A. Lucchese et al.

assigned at WS1 and the most unexperienced-youngest operator O3 is assigned at WS3 only for the last time slot S3 . These results are mainly caused by the fact that the behaviour of operators is not deterministic, but stochastic: by changing the time constraints (e.g., CT ), pwjk does not change linearly, since it is evaluated from a normal probability distribution function characterized by μwjk and σwjk (Eq. 9–11). This observation further emphasises the importance of appropriately modelling the stochastic behaviour of an operator. Moreover, by evaluating the actual production during a work shift for each scenario, it can be observed the costs in terms of production efficiency (Table 4). As it can be observed in Table 4, the production efficiency, defined here as the ratio between the actual and the target production in the work shift, for scenario “B”, in case of CT = 18s and null buffer size, is at minimum 10% lower than the other scenarios. In Table 4 it can be observed that production efficiency values obtained for the job-rotation problem solutions are in general greater than the 90%, providing a good worker-workstation assignment. Moreover, the production efficiency is always lower in case of null buffer size (compared to infinite buffer size), while, for a given buffer size (0 or ∞), production efficiency is always greater in scenario “G” than in scenario “B”. Differences in terms of production efficiency, by comparing “G” and “B” scenarios are at maximum 2% in case of CT = 20s, while at maximum 12% in case of CT = 18s, further emphasizing that the different effect of fatigue on operators is even greater in case of tighter times. The same line of reasoning can be followed when comparing the production efficiency with different buffer sizes: the production efficiency can be up to 7% or 11% higher in case of infinite buffer size in the “B” scenario (CT = 20s for 7%, CT = 18s for 11%). Therefore, the different effect of fatigue on operators is even greater in case of tighter times by comparing different buffer sizes. These results highlight the influence of the buffer size, as confirmed by multiple authors such as [28]. In the numerical experiments presented, a simple line composed by only three workstations have been considered; for sure production efficiency differences between “G” and “B” scenarios would increase with the length of the line.

5 Conclusion and Further Research Nowadays, behavioural issues about workers assume high criticality, thus requiring a more reliable prediction of operators’ performances. In line with one of the main goals of I5.0 of reaching more inclusive human-centric manufacturing systems, the present paper presented an innovative method to model the stochastic behaviour of different operators involved in manual tasks through the probability density function of execution times, including the workforce diversity in terms of age, fatigue and experience. The proposed model has been tested on a job rotation scheduling problem, whose solutions are the ones that assign a specific worker to a given workstation by maximizing the probability of performing the activity. From preliminary results obtained it can be concluded that the stochastic behaviour of operators is affected by workforce diversity, and significantly impact their performance over time. This observation suggests that in real industrial contexts, a particular attention must be paid to the length and number of shifts and breaks: practitioners should analyse more carefully the management of shifts and breaks through working-time models more fitted to the operator.

A Stochastic-Based Model to Assess the Variability of Task Completion Times

757

To sharpen the model proposed in the present paper, the next step will be focused on considering new categories of workers (e.g., different experience levels with the same age, different age with same experience levels), as well as investigating the buffer size, since, from results obtained, it is one of the main system parameters responsible for the reduction of production efficiency. The proposed model can be employed as a base to pursue a more reliable prediction of operators’ performance, shaping their stochastic behaviour in current industrial contexts that require the execution of time-intensive manual activities. This model contributes to create more human-centric manufacturing systems. Nevertheless, to allow an age-independent social inclusion of operators in I5.0 environments, age-friendly workplaces and processes should be designed by considering the integration of ergonomics, neuro-ergonomics and the human factor principles, in order to empower the experience and high-level skills acquired by aged operators, and assist them through innovative supporting solutions (e.g., cobots) while meeting their specific needs. To increase the reliability of the model in expressing the stochastic behaviour of operators, other factors should be considered or better described; the latter case can be associated with fatigue. Fatigue is a complex phenomenon that affects operators variably and dynamically: by directly observing operators and monitoring oxygen consumption (energy expenditure), as well as heart rate variability, a better prediction of fatigue can be evaluated, and a more reliable stochastic model can be defined. Therefore, for future research, it is important to observe operators executing activities and monitor specific parameters through advanced sensors and I4.0 wearable technologies. Funding. This research is partially supported by the European Union Next-Generation EU (PNRR – MISSIONE 4 COMPONENTE 2, INVESTIMENTO 1.3) under the EP MICS (project no. PE00000004, CUP D93C22000920001) and partially supported by the Italian Ministry of Education, Universities and Research (MIUR), SO4SIMS project (Smart Operators 4.0 based on Simulation for Industry and Manufacturing Systems- Project PRIN-2017FW8BB4).

Appendix A σj,min Is the minimum value of the standard deviation at workstation WSj corresponding to the case of CVmin . In the scientific literature can be found that from the observation of operators executing manual assembly tasks, CV shows as minimum value 0.1 [11, 19, 29]. Bearing this in mind it must be evaluated what specific condition refers to the case of CVmin = 0.1. i.e., it must be evaluated the specific Ow and fatigue state (i.e., Sk ) to be associated to the minimum ratio between the standard deviation and the expected value. To achieve this goal, it must be remembered that during time, the fatigue of Ow tends to increase, reaching the maximum fσwk /fμwk ratio at the last time slot of the work shift; during manual assembly activities it is observed that due to fatigue effect, the expected execution time can increase up to + 20%, while the standard deviation of execution time can double (+100%) [11]. Therefore, being σwjk /μwjk proportional to fσwk /fμwk (Eq. 7–8), the condition of minimum fatigue that brings to the minimum value of fσwk /fμwk , is the one associated to the first time slot S1 . Focusing on the experience of a specific Ow , it is noted that the experience of worker affects solely its expected value of execution time, while keeping the corresponding standard deviation almost unchanged; in particular, an

758

A. Lucchese et al.

unexperienced worker can bring to an increase of the expected execution time up to + 20% [25]. Therefore, in case of unexperienced worker, fsw reaches its maximum value (fsw ,max ). Due to observations made, the condition associated to the CVmin is the one referring to an unexperienced worker (fsw ,max ) at the minimum fatigue level (minimum fσwk /fμwk ratio) (eq. A1):     σwjk σj,min fσwk CVmin = CVj,min = = · (A1) μwjk min fμwk min t j · fsw ,max By setting CVmin = 0.1, the value of σj,min can be evaluated, and the σwjk can be quantified (Eq. 8).

References 1. Neumann, W.P., Winkelhaus, S., Grosse, E.H., Glock, C.H.: Industry 4.0 and the human factor – a systems framework and analysis methodology for successful development. Int. J. Prod. Econ. 233, 107992 (2021) 2. Romero, D., Bernus, P., Noran, O., Stahre, J., Berglund, Å.F.: The operator 4.0: human cyberphysical systems & adaptive automation towards human-automation symbiosis work systems. IFIP Adv. Inf. Commun. Technol. 677–686 (2016) 3. Choi, T.M., Kumar, S., Yue, X., Chan, H.L.: Disruptive technologies and operations management in the industry 4.0 era and beyond. Prod. Oper. Manag. 31, 9–31 (2022) 4. Gawron, V.J., French, J., Funke, D.: An overview of fatigue. stress, workload, and fatigue. Lawrence Erlbaum Associates, 581–595 (2001) 5. Intranuovo, G., De Maria, L., Facchini, F., Giustiniano, A., Caputi, A., Birtolo, F., et al.: Risk assessment of upper limbs repetitive movements in a fish industry. BMC Res Notes. 12, 354 (2019) 6. Facchini, F., Mummolo, G., Vitti, M.: Scenario analysis for selecting sewage sludge-toenergy/matter recovery processes. Energies 14, 276 (2021) 7. Calzavara, M., Persona, A., Sgarbossa, F., Visentin, V.: A device to monitor fatigue level in order-picking. Ind. Manag. Data Syst. 118, 714–727 (2018) 8. Achten, J., Jeukendrup, A.E.: Heart rate monitoring: applications and limitations. Sport Med. 33, 517–538 (2003) 9. Zhang, Z., Li, K.W., Zhang, W., Ma, L., Chen, Z.: Muscular fatigue and maximum endurance time assessment for male and female industrial workers. Int. J. Ind. Ergon. 44, 292–297 (2014) 10. Sedighi Maman, Z., Alamdar Yazdi, M.A., Cavuoto, L.A., Megahed, F.M.: A data-driven approach to modeling physical fatigue in the workplace using wearable sensors. Appl. Ergon. 65, 515–529 (2017) 11. Murrell, K.F.H.: Operator variability and its industrial consequences. Int. J. Prod. Res. 1, 39–55 (1961) 12. Swaen, G.M.H., Van Amelsvoort, L.G.P.M., Bültmann, U., Kant, I.J.: Fatigue as a risk factor for being injured in an occupational accident: results from the Maastricht Cohort Study. Occup. Environ. Med. 60, i88–i92 (2003) 13. Kolus, A., Wells, R., Neumann, P.: Production quality and human factors engineering: a systematic review and theoretical framework. Appl. Ergon. 73, 55–89 (2018) 14. Xu, X., Qin, J., Zhang, T., Lin, J.H.: The effect of age on the hand movement time during machine paced assembly tasks for female workers. Int. J. Ind. Ergon. 44, 148–152 (2014)

A Stochastic-Based Model to Assess the Variability of Task Completion Times

759

15. Dupont, C., Benin, A., Belgi, M.U.: European Agency for Safety and Health at Work. Safer and healthier work at any age: Review of resources for workplaces. Anim. Genet. 39, 1–55 (2016) 16. Gilles, M.A., Guélin, J.C., Desbrosses, K., Wild, P.: Motor adaptation capacity as a function of age in carrying out a repetitive assembly task at imposed work paces. Appl. Ergon. 64, 47–55 (2017) 17. Gräßler, I., Roesmann, D., Cappello, C., Steffen, E.: Skill-based worker assignment in a manual assembly line. Procedia CIRP 100, 433–438 (2021) 18. Börsch-Supan, A., Weiss, M.: Productivity and age: evidence from work teams at the assembly line. J. Econ. Ageing 7, 30–42 (2016) 19. Öner-Közen, M., Minner, S., Steinthaler, F.: Efficiency of paced and unpaced assembly lines under consideration of worker variability – a simulation study. Comput. Ind. Eng. 111, 516– 526 (2017) 20. Dudley, N.A.: Work-time distributions. Int. J. Prod. Res. 2, 137–144 (1963) 21. Slack, N.: Work time distributions in production system modelling. Oxford Centre for Management Studies (1990) 22. Sury, R.J.: An industrial study of paced and unpaced operator performance in a single stage work task. Int. J. Prod. Res. 3, 91–102 (1964) 23. Dar-EI, E.M.: Human Learning: From Learning Curves to Learning Organizations. Hillier FS, editor. Kluver Academic Publisgers, Massachusetts (2000) 24. MTM-1 Analyst Manual. UK MTMA Ltd. (2000) 25. Franks, I.T., Sury, R.J.: The performance of operators in conveyor-paced work. Int. J. Prod. Res. 5, 97–112 (1966) 26. Folgado, R., Peças, P., Henriques, E.: Mapping workers’ performance to analyse workers heterogeneity under different workflow policies. J. Manuf. Syst. 36, 27–34 (2015) 27. Gökçen, H., Faruk, B.Ö.: A new line remedial policy for the paced lines with stochastic task times. Int. J. Prod. Econ. 58, 191–197 (1999) 28. Conway, R., Maxwell, W., McClain, J.O., Thomas, L.J.: The role of work-in-process inventories in serial production lines. Oper. Res. 36, 229–241 (1988) 29. Doerr, K.H., Arreola-Risa, A.: A worker-based approach for modeling variability in task completion times. IIE Trans. Inst. Ind. Eng. 32, 625–636 (2000)

A Proposal for Production Scheduling Optimization Method with Worker Assignment Considering Operation Time Uncertainty Daiki Nagata1(B) , Toshiya Kaihara1 , Daisuke Kokuryo1 , Toyohiro Umeda2 , and Houei Mizuhara2 1 Graduate School of System Informatics, Kobe University, 1-1 Rokkodai-Cho, Nada,

Kobe 657-8501, Hyogo, Japan [email protected], [email protected], [email protected] 2 Kobe Steel, Ltd, 1-5-5 Takatsukadai, Nishi, Kobe 651-2271, Hyogo, Japan {umeda.toyohiro,mizuhara.houei}@kobelco.com

Abstract. In a make-to-order factory, the top priority is to meet due date determined in advance through consultation with the customer. However, due to the low degree of repetitiveness of work in each process, work performance tends to fluctuate against the predetermined standard time. This uncertainty in operation time can prevent production from proceeding as planned and affect due dates. In addition, since the work in each process depends on the ability of the worker in charge of it, it is important to consider the worker’s ability appropriately. In this study, we propose a scheduling optimization method that includes worker assignment under operation time uncertainty. The method formulates a job shop scheduling problem that takes into account worker skill and operation time uncertainty, and minimizes the expected total tardiness and operation time variance. In experiments, we evaluate the effectiveness of the proposed method by solving problems with various levels of uncertainty and several combinations of workers. Keywords: Job shop scheduling · Worker assignment · Total tardiness · Uncertainty

1 Introduction Nowadays, many factories have begun involving advanced information technologies and knowledge discovering methods to facilitate information flow, with the main need to deal with the large volume of data which are generated and collected from manufacturing processes and its related operations [1]. One example is the tracking of products by RFID systems, which facilitates the acquisition and use of various data from processes in factories [1, 2]. On the other hand, the manufacturing industry has been increasing the variety of their products and shortening delivery times in order to meet diverse customer needs [3]. In a production environment that is becoming increasingly complex due to the © IFIP International Federation for Information Processing 2023 Published by Springer Nature Switzerland AG 2023 E. Alfnes et al. (Eds.): APMS 2023, IFIP AICT 689, pp. 760–774, 2023. https://doi.org/10.1007/978-3-031-43662-8_54

A Proposal for Production Scheduling Optimization Method

761

diversification of production types, production planning and scheduling are important in order to satisfy customer requirements [4]. In a make-to-order factory, the top priority is to meet due date determined in advance through consultation with the customer. In the case of individual production in the job shop process, not all tasks are performed automatically by machines, and some tasks require intervention by workers. However, in the case of individual production in the job store process, not all tasks are performed automatically by machines, and some tasks require intervention by workers. In addition, because the specifications vary greatly from product to product [5], the repetitiveness of work in each process is low. Therefore, the work performance tends to fluctuate against the predetermined standard time [6]. This variation in work performance could prevent production from proceeding as planned, and result in tardiness [6]. In addition, the work efficiency depends on the skills of each worker [4, 6]. And the difference in work efficiency can be visualized by the aforementioned recent sensing technology. Therefore, in general job shop scheduling problems, workers are ignored in the modeling, but it is important to consider them. Based on the above, in production scheduling, it is necessary to minimize the risk of tardiness by taking into account in advance the skill level of each worker and the variation in the processing time of each process. In this paper, we propose a production planning method that considers both worker allocation and production scheduling, and aims to minimize tardiness by taking into account the variation in work performance that occurs in manual operations with workers.

2 Target Model 2.1 Factory In this study, our target factory is a job shop-type process as shown in Fig. 1. The characteristics of job, machine and worker in the factory are as follows:

Fig. 1. Job shop model

Job: • • • •

All jobs are known at the beginning of the production. Each job is processed by multiple processes. The number and order of the processes required to complete a job varies for each job. Each process uses both human operators (workers) and automated machines.

762

D. Nagata et al.

Machine: • • • •

All machines can be used from the start of production. The machine that processes each process of a job is predetermined. No interruption occurs during job processing. No machine failure occurs.

Worker: • All workers can work from the start of production. • The worker can move between machines and can operate all machines. • Once a worker starts a manual operation on a machine, he cannot move from the machine until the manual operation is completed. • Travel time between machines is not considered. • Workers’ skill is classified into 3 levels: advanced, intermediate, and beginner. 2.2 Processing Time and Its Uncertainty in Manual Operations [7] The processing time for each process defines the following characteristics: • The processing time for each process of a job is the sum of the manual processing time by the worker and the automatic processing time by the machine. • In each process, the ratio of standard manual operation time is pre-determined.

Fig. 2. An example of manual operation time distribution in each skilled type

In individual production process, specifications may differ for each product, and the manual operation performed by workers in each process tends to be fluctuated. Workers’ abilities are not uniform, and work efficiency varies from worker to worker. Therefore, in this study, we model the uncertain manual operation time as a random variable. Since the actual work tends to be later than predetermined standard time, this study assumes that the manual operation time for each process follows an Erlang distribution [8]. The probability distribution differs depending on the task and the worker in charge. The higher the skill level of the worker, the smaller both the average and variance of the probability distribution are assumed to be Fig. 2 [9]. The probability distribution for each worker performing each process is assumed to be given.

A Proposal for Production Scheduling Optimization Method

763

3 Worker Allocation and Job Shop Scheduling Method Considering Uncertainty of Operation Time The target problem of this study is to simultaneously determine the allocation of workers in charge of each process in addition to the general job shop scheduling problem [10]. In this study, we propose a production scheduling method that includes worker assignment considering the variation of processing time for each job. Since the manual operation time of each process varies, the proposed method considers the variation as a probability distribution and proposes a schedule that minimizes the sum of expected total tardiness and total processing time variance of each job. The proposed method performs job shop scheduling including worker assignment after pre-processing of scheduling. Section 3.1 defines the symbols used in the formulation of the proposed method. Section 3.2 provides an overview of the proposed method and its specific formulation. 3.1 Notation The notations used in proposed method are shown as follows: Parameters: • • • • • • • • • • • • • • • • • • •

j, p : job number (j, p = 1, . . . , J )  k, q : process number of each job k, q = 1, . . . , Kj m, m : machine number (m, m = 1, . . . , M ) i, i : worker number (i, i = 1, . . . , I ) s : real number called scenario coefficient S (UB) : upper limit of scenario coefficients S (LB) : lower limit of scenario coefficients PM jk : machine number to process the k th process of job j Dj : due date of job j PT jk : standard total processing time of the k th process of job j APT jk : automatic processing time of the k th process of job j Modejkmi : mode of manual operation time for the k th process of job j at machine m by worker i Vjkmi : variance of manual operation time for the k th process of job j at machine m by worker i σj : assumed standard deviation of total processing time for job j γjk : standard ratio of manual operation time to total processing time PTjk BigM : sufficiently large positive number W1 : weight of the first term of the objective function W2 : weight of the second term of the objective function s : step size of scenarios (constant number)

Variables: • ST jk : processing start time of the k th process of job j • CT jk : processing completion time of the k th process of job j • Cjkmi : completion time of the k th process of job j by worker i at machine m

764

D. Nagata et al. (+sσ )

• Tj : tardiness of job j when scenario coefficient is s • ET j : expected value of tardiness of job j  1 : ifthekthprocessofjobjprecedestheqthprocessofjobp • Xjkpq : 0 : otherwise  1 : ifworkeriprocessesthekthprocessofjobjatmachinem • Yjkmi : 0 : otherwise In the proposed method, the decision variables are STjk , Xjkpq andYjkmi . 3.2 Algorithm of Proposed Method The proposed method creates a schedule in the following steps. STEP1. Initialization of standard deviation of processing time for each job. STEP2. Worker assignment and job shop scheduling. First, in STEP1, the standard deviation of processing time for each job is initialized in order to solve the scheduling problem in STEP2. Next, in STEP2, job shop scheduling including worker allocation is performed. In STEP2, the standard deviation of the processing time of each job calculated in STEP1 is taken into account, and the worker assignment and schedule that minimize the difference between the expected total tardiness and the processing time variance are obtained. Details of each are described below. Initialization of standard deviation of processing time for each job (STEP1) This method attempts to minimize the expected total tardiness by considering the probability distribution of the processing time of each job. However, the distribution of the completion time of each job is unknown unless the worker assignment is determined. In addition, if the standard deviation is formulated at the scheduling stage, it becomes a nonlinear problem and is difficult to find a solution. Thus, the scheduling problem is formulated as a linear programming problem by calculating the assumed value of the standard deviation in STEP1. In STEP 1, the worker processes each operation is unknown. Therefore, the standard deviation of the processing time for each job is calculated assuming that all processes are processed by the intermediate skill workers. The processing time for each process follows an independent Erlang distribution. Since the Erlang distribution is regenerative, the assumed standard deviation of the processing time for each job j (σj ) is given by the Eq. (1). Vjkmi uses the intermediate worker’s value.  σj =

 k

Vjkmi {m = PM jk }

(1)

Worker assignment and scheduling optimization phase (STEP2) The formulation of the job shop scheduling problem, including worker assignment, is described below. This method considers the uncertainty in the processing time of each process and aims to derive a schedule that is optimal with robustness. Therefore, the proposed method minimizes the sum of the expected total tardiness and the standard

A Proposal for Production Scheduling Optimization Method

765

deviation of the total processing time. The former is an index focusing on optimality, while the latter is an index focusing on improving robustness. In this study, the weighted sum of these two indices is used as the objective function.5 Objective functions min. W1 ∗ F1 + W2 ∗ F2 (2) where F1 = ETj =

S (UB) s=S (LB)

(+sσ )

Tj F2 =

J

ET j

(3)

∗ σj ∗ s{∀j}

(4)

j=1 (+sσ )

pjt ∗ Tj

= CTjKj + s ∗ σj − Dj {∀j}

J Kj I M j=1

(2)

k=1

i=1

m=1

(5) 

Vjkmi ∗ Yjkmi

(6)

Fig. 3. An example of approximate calculation of expected job tardiness

Equation (2) is the objective function and represents the weighted sum minimization of the two indices. Equation (3) shows the sum of the expected total tardiness, and Eq. (4) is the definition of the expected tardiness for each job. As shown in Fig. 3, these expected values are approximated using several possible scenarios of tardiness of jobs. The tardiness for each scenario used in Eq. (4) is defined by Eq. (5). The constant s (defined as the scenario coefficient) indicates how early or late the completion time of a job is relative to the standard completion time in a scenario. As an example, Fig. 4 shows the scenario with s = 1. CT jKj is the standard completion time. It is the value when all jobs are done in the standard processing time. And, pjt is the probability of tardiness of job j in the scenario coefficient s. The second term of the objective function is then the sum of the variance of the total processing times for each job defined by Eq. (6). Definition and Constraint:   (7) where APT jk = PT jk ∗ 1 − γjk {∀j, k}  CTjk = Cjkmi + APT jk ∀j, k, m = PMjk , i = PWjk

(8)

766

D. Nagata et al.

Fig. 4. An example of job completion time and tardiness for scenario +1σ (s = 1)

 Cjkmi = STjk + Modejkmi ∀j, k, m = PMjk , i = PWjk s.t.

J Kj I j=1

k=1

i=1

 Yjkmi = 1 m = PM jk (10)

STj(k+1) ≥ CTjk {∀j, k}(11) 

  STjk ≥ CTpq − BigM 1 − Xjkpq − BigM 2 − Yjkmi − Ypqmi

(9) (10) (11) (12)

{m = PM jk = PM pq , j = p, ∀k, q, i, i } STpq ≥ CTjk − BigM ∗ Xjkpq − BigM (2 − Yjkmi − Ypqmi )

 m = PM jk = PM pq , j = p, ∀k, q, i, i 

  STjk ≥ Cpqm i − BigM 1 − Xjkpq − BigM 2 − Yjkmi − Ypqmi

STpq

{m = PM jk , m = PM pq , j = p, ∀k, q, i, i } 

≥ Cjkmi − BigM ∗ Xjkpq − BigM 2 − Yjkmi − Ypqmi {m = PM jk , m = PM pq , j = p, ∀k, q, i, i }

(13)

(14)

(15)

Equation (7) defines the machine processing time, the length of which is a constant determined by the standard total processing time PTjk and standard ratio of manual operation time γjk . Equation (8) defines the completion time of the k th process of job j, and Eq. (9) defines the manual completion time of the k th process of job j. Equations (10) to (15) are constraints. Equation (10) is a constraint that guarantees that each process is always processed by one worker. Equation (11) is a constraint on the precedence relation between processes of the same job. Equations (12) and (13) are precedence constraints for jobs in the same machine, which guarantee that no more than two jobs are processed simultaneously by the same machine. Equations (14) and (15) are precedence constraints for jobs in the same worker, which guarantee that the same operator does not perform two or more operations at the same time.

A Proposal for Production Scheduling Optimization Method

767

4 Computational Experiments In order to evaluate the performance of the proposed method, two types of computational experiments are performed. First, we compare the performance of the proposed method with considering uncertainty of the processing time and the conventional deterministic method without considering uncertainty. Next, we apply the proposed method to combinations of workers with various skill levels and evaluate the characteristics of the proposed method. In each experiment, we solve the optimization problem expressed in the formulation described in Sect. 3 using the branch-and-bound method with IBM ILOG CPLEX 12.10 [11] to produce a schedule. Then, the planned schedules are evaluated by simulating the operation phase of actual factory with variable manual operation time for workers as shown in Sect. 4.1. 4.1 Operation Phase Simulation In order to calculate the operational results for the proposed schedule, an operational simulation is performed under the following conditions. The manual operation time for each process is randomly determined based on the Erlang distribution, which is determined by the planned worker assignment and the skill level of the worker. The manual operation time may be extended or shortened relative to the plan, in which case the proposed schedule is modified according to the constraints described below. • Machine change constraint: Even if work on a job is delayed, the job is not assigned to another machine and continues to work on the same machine. • Worker change constraint: Even if work on a job is delayed, the worker who processes the job is not changed and the job continues to be processed. • Processing order change constraint: If a preceding job on the same machine or by the same worker is delayed, the job is interrupted and work continues without interrupting the work of another subsequent job. • Start time change constraint: If the completion time of the preceding job of the same machine or the same worker is earlier, the start time of the subsequent job is brought forward as much as possible. 4.2 Experiment 1: Performance Comparison of Proposed Method and Conventional Deterministic Methods In Experiment 1, we compare the performance of the proposed method with considering uncertainty of the processing time and the deterministic method without considering uncertainty. The effects of changing the level of uncertainty are also compared. Experimental Conditions The experiments are performed with the following conditions. [a, b] means that it is an integer constant between a and b. • • • •

The number of jobs (J ) : 6 The number of processes for each job (Kj ): [1, 4] The number of machines (M ) : 6 The number of workers (I ) : 3 (A, B, C) = (1, 1, 1), (0, 3, 0)

768

D. Nagata et al.

The advanced level is denoted as A, the intermediate level as B, and the beginner level as C.   • Processing time for each PTjk : [5, 30]  process  • Due date of each job Dj : 1.2 ∗ k PTjk • Standard ratio of manual operation time to total processing time (γjk ) : 0.5 • Weight of objective functions (W1 , W2 ) :(100,1) • The range of scenario of uncertainties: [−2σ, +2σ ] • The step size of scenarios : 1σ • Number of simulations: 1000 In this experiment, the variance of the workers’ operation time for each process is varied in order to evaluate the influence of the uncertainty in the proposed method. Experiments are conducted on the 8 cases shown in Table 1 regarding the standard deviation of the processing time when workers of each skill level process the processes. Case 1 has the lowest level of uncertainty, and Case 8 has the highest level of uncertainty. Table 1. Conditions for standard deviation of processing time for each process at each skill level Advanced

Intermediate

Beginner

Case 1

0.00

0.00

0.00

Case 2

0.35

0.43

0.50

Case 3

0.71

0.87

1.00

Case 4

1.41

1.73

2.00

Case 5

2.83

3.46

4.00

Case 6

4.24

5.20

6.00

Case 7

5.66

6.93

8.00

Case 8

7.07

8.66

10.00

Results Table 2 shows the first term of the objective function, the total tardiness in the planning schedule (F1 ) and the total tardiness and the standard deviation in the operational phase simulation (TT (S) ) against the planned schedule obtained by the proposed method considering uncertainty and the conventional deterministic method without considering uncertainty. Note that the results in Table 2 are for the worker condition (A, B, C) = (1, 1, 1) and Case 8 is used for the mode and variance of operation time for each process. Table 2 also shows the t and p values obtained from the two-tailed t test (degree of freedom 1998) at 5% level of significance for the results of the proposed method and the deterministic method. From Table 2, it has been confirmed that the proposed method reduces the total tardiness more than the comparison method. The results of the p-values obtained from the two-tailed t-test at the 5% level of significance indicate that there is a significant difference. In the proposed method, it is possible to reduce the total tardiness in the operation phase compared to the conventional deterministic method by making a

A Proposal for Production Scheduling Optimization Method

769

production schedule that minimizes the total tardiness with considering the variation in the total processing time for each job. Table 2. Comparison of total tardiness between planning schedules and simulation results of the proposed and deterministic methods

Next, Table 3 shows a breakdown of the total tardiness obtained from simulations using the two methods. Figure 5 shows Gantt charts of the planning schedule developed using the proposed method and the deterministic method. The difference in Table 3 shows that the tardiness of Job1 was significantly reduced by the proposed method compared to the deterministic method. As shown in Table 3, Job1 has four processes, which is more than the other jobs, and has a large standard deviation in total processing time and a large uncertainty. Therefore, as shown in Fig. 5, the proposed method reduced the tardiness of Job1 compared to the deterministic method by deriving a schedule that advances the submission and completion times of Job1, which has a large uncertainty. Table 3. Tardiness for each job in the simulation results of the two methods of operation

Job tardiness

Total tardiness

Job index

Number of Processes

Deterministic

Proposed

Avg

Avg

Diff

Job1

4

9.67

2.15

7.52

Job2

1

0.16

0.15

0.01

Job3

2

4.68

4.52

0.16

Job4

2

19.05

19.18

-0.13

Job5

3

28.76

29.39

-0.63

Job6

1

0.86

0.38

0.48

65.23

54.98

10.25

Finally, we evaluate the influences of different levels of uncertainty. Figure 6 shows the simulation results of the schedules obtained using the proposed method and the deterministic method for 8 cases with different levels of processing time uncertainty. Note that for simplicity, all workers are assumed to be intermediate ((A, B, C) = (0, 3, 0)). In all 8 cases with different levels of processing time uncertainty, the proposed method reduced the total tardiness compared to the deterministic method, and also the difference

770

D. Nagata et al.

Fig. 5. Gantt charts of planning schedule for proposed and deterministic methods

increased as the uncertainty level increased. Therefore, the proposed method, which considers the uncertainty of job processing time, is more useful than the conventional deterministic method.

120 Total tardiness

100 80

Deterministic Proposed

60 40 20 0 Case1 Case2 Case3 Case4 Case5 Case6 Case7 Case8

Fig. 6. Comparison of total tardiness between the proposed and deterministic methods under different uncertainty level conditions

A Proposal for Production Scheduling Optimization Method

771

4.3 Experiment 2: Analysis of the Impact of Different Combinations of Workers of Various Skill Levels In Experiment 2, we evaluate the production schedule devised by the proposed method for a combination of workers of various skill levels who are available to work. As explained in Experiment 1, we evaluate the effectiveness of the proposed method and the relationship between the resource utilization ratio and the total tardiness in operation phase simulation. The experimental conditions for jobs in Experiment 2 use the same dataset as in Case 8 of Experiment 1. However, experiments are conducted using combinations of workers with the number of workers and their skill levels changed as shown in Table 4. The number of advanced workers is denoted as A, the number of intermediate workers as B, and the number of novice workers as C. Table 4. Combination of workers (A, B, C) Case1

(0, 2, 0)

Case2

(0, 3, 0)

Case3

(0, 4, 0)

Case4

(0, 5, 0)

Case5

(1, 1, 1)

Results Table 5. Optimization and simulation results comparison for different number of workers (A, B, C)

F1

F2

Total Tardiness Avg

S.D

C.V

Case1: (0, 2, 0)

74

39

93.04

32.66

0.35

Case2: (0, 3, 0)

64

39

63.27

20.80

0.33

Case3: (0, 4, 0)

60

39

56.30

17.31

0.31

Case4: (0, 5, 0)

60

39

56.95

18.93

0.33

Case5: (1, 1, 1)

45

33

55.77

20.11

0.36

Table 5 shows the bi-objective function values (F1 , F2 ) of each planned schedule under different workers conditions, the average, standard deviation, and coefficient of variation of the total tardiness obtained from the operation phase simulation using these obtained schedules. Table 6 shows the utilization rates of each machine and each worker for the planned schedule under different worker conditions. The utilization rate of each machine and worker is calculated by dividing each processing time by its makespan. In

772

D. Nagata et al.

Case5, each skill is Advanced for Worker1, Intermediate for Worker2, and Beginner for Worker3. First, we discuss the influence of the difference in the number of workers from the results of Cases 1 to 4 in Tables 5 and 6. Table 5 shows that the value of F1 , the expected total tardiness of the planned schedule, tended to decrease as the number of workers increases. The average of the total tardiness in the operation phase simulation was also reduced accordingly. However, when the number of workers was compared between the case with four (Case3) and the case with five (Case4), there was no change in both the objective function value at the time of planning and the simulation results. The coefficient of variation of the total tardiness in the operation phase simulation tended to decrease as the number of workers increases. Table 6. Utilization ratio of resources in planned schedule under the conditions in Table 5 Case1

Case2

Case3

Case4

Case5

(A, B, C)

(0, 2, 0)

(0, 3, 0)

(0, 4, 0)

(0, 5, 0)

(1, 1, 1)

Machine1

0.44

0.46

0.46

0.46

0.51

Machine2

0.20

0.21

0.21

0.21

0.22

Machine3

0.24

0.25

0.25

0.25

0.24

Machine4

0.45

0.47

0.47

0.47

0.49

Machine5

0.64

0.67

0.67

0.67

0.67

Machine6

0.71

0.74

0.74

0.74

0.71

Avg

0.45

0.47

0.47

0.47

0.47

Worker1

0.69

0.54

0.25

0.26

0.70

Worker2

0.65

0.37

0.34

0.45

0.49

0.46

0.45

0.33

0.16

0.34

0.23

Worker3 Worker4 Worker5 Avg

0.12 0.67

0.46

0.34

0.28

0.46

Table 6 shows that as the number of workers increases, the average utilization rate of workers decrease. This result suggests that the idle time of each worker increases as the number of workers increases. When workers have sufficient idle time, even if the operation time of a given process is longer than planned and delays occur, it is prevented from propagating delays to subsequent processes. As a result of this effect, the coefficient of variation of total delivery delays tended to decrease as the number of workers increases, as shown in Table 5. On the other hand, increasing the number of workers didn’t significantly change the utilization rate of each machine. This is because the machines that process each process of a job are predetermined, and increasing the number of workers does not improve the processing efficiency of the machines. Therefore, as shown in Table 5, there was no improvement in the average and variance of total tardiness when

A Proposal for Production Scheduling Optimization Method

773

the number of workers increases to four or more. In this experiment, all workers in Cases 1 ~ 4 are intermediates. In order to confirm the effect of the difference in the number of workers, it is necessary to conduct the same experiment for the cases where all workers are advanced and the case where all workers are beginners. Next, we discuss the influence of the difference in the skill of workers from the results of Cases 2 and 5 in Tables 5 and 6. Table 5 shows that the average of the total tardiness in Case 5 was smaller than that in Case 2, but the variation of the total tardiness (coefficient of variation) was larger. Advanced workers can process the work faster on average than other workers, and have less variation (standard deviation) in the operation time. Therefore, in Case 5, the proposed method reduced the value of the dual objective by considering the skill level of the workers and allocating a large number of processes to the advanced workers. About the average machine and worker utilization rates, the both cases were the same, and that the utilization rates of each machine were also the same as shown in Table 6. Regarding the utilization rate of workers by skill, when workers with different skills were included as in Case5, the utilization rate of advanced workers was high and that of beginner workers was low. On the other hand, in Case 2, although there was a difference in the utilization rate for each worker, the difference in the utilization rate of each worker was small. The reason for the large difference in utilization rates in Case 5 is that the proposed method assigns more processing to advanced workers with less variation in processing time in order to reduce the variation in working time. However, as shown in Table 5, the coefficient of variation of total tardiness was smaller in Case 2. This suggests that the load is disproportionately placed on specific workers with high work efficiency, which may affect the variation of total tardiness. These results suggest that in order to reduce the variation in total tardiness, planning that does not propagate process delays is required, taking into consideration about the load and idle time of each resource, in addition to worker allocation that consider the skill level of each worker.

5 Conclusion In this paper, we proposed a job shop scheduling optimization method with worker assignment for minimizing total tardiness for a production system with uncertain processing times. The proposed method minimized the linearly weighted sum of the expected total tardiness and the variance of the total processing time by considering the difference in the skill level of each worker and the probability distribution of the operation time. The computer experiments suggested that the proposed method can reduce the total latency time better than the deterministic method. The effectiveness of the proposed method became more pronounced as the level of uncertainty increased. The proposed method reduced the load of each worker, and then reduced the coefficient of variation of the total tardiness. In summary, we found that the proposed method can reduce the mean and variability of the total tardiness. We also found that reducing the workload of each worker and providing margin time can reduce delay propagation. Based on this finding, we would like to examine the number of workers and their capabilities required to reduce the risk of tardiness, and how much time should be set aside in the schedule. On the other hand, there are three points where the method needs to be improved. The

774

D. Nagata et al.

first is the problem of computation time. The proposed method solves the problem using the branch-and-bound method with a solver, but it can only solve small-scale problems due to the long computation time. Therefore, we plan to investigate methods such as meta-heuristics for large-scale experiments in the future. The second point concerns the accuracy of the solution. Since the probability distribution of the total tardiness is calculated by using intermediate workers in the proposed method, the distribution is different from the true value. Therefore, it is important to calculate a more accurate distribution of total tardiness. And as a third improvement, we will also consider extending the model to take into account of uncertainties other than the variability of operation time. Acknowledgement. We would like to thank Professor emeritus Susumu Fujii (Kobe university), Professor Nobutada Fujii (Kobe university), Dr. Harumi Haraguchi (Ibaraki university), Dr. Ruriko Watanabe (Waseda university) and Mr. Hideo Ikeda (Kobe Steel, Ltd) for providing appropriate advices.

References 1. Kanagachidambaresan, G.R., Anand, R., Balasubramanian, E., Mahima, V.: Internet of Things for Industry 4.0 Design, Challenges and Solutions. EAI/Springer Innovations in Communication and Computing (2020) 2. Daniel Alejandro, R., Fernando, T., Mariano, F.: A data-driven scheduling approach to smart manufacturing. J. Ind. Inf. Integr. 15, 69–79 (2019) 3. Ministry of internal affairs and communications.: WHITE PAPER information and communications, Part 1, Chapter 2, Section 4 (2020). https://www.soumu.go.jp/johotsusintokei/whi tepaper/ja/r02/pdf/n2400000.pdf 4. Marichevam, M.K., Geetha, M., Tosun, Ö.: An improved particle swarm optimization algorithm to solve hybrid flowshop scheduling problems with the effect of human factors – a case study. Comput. Oper. Res. 114, 104812 (2020) 5. Shahrul, K., Khan, Z.A., Noor Siddiquee, A., Wong, Y.S.: The impact of variety of orders and different number of workers on production scheduling performance: a simulation approach. J. Manuf. Technol. Manag. 24(8), 1123–1142 (2013) 6. Tarek, C., Sondes, C., Nassima, A., Damien, T.: Scheduling under uncertainty: survey and research directions. In: 2014 International Conference on Advanced Logistics and Transport, pp. 229–234 (2014) 7. Nagata, D., Kaihara, T., Fujii, N., Kokuryo, D., Umeda, T., Mizuhara, H.: Improvement of production scheduling method with robust optimization approach considering operation time variation and production efficiency. Proc. ICPE2022, C217 (2022) 8. Wojciech, B., Rajba, P., Uchro´nski, M., Wodecki, M.: A job shop scheduling problem with due dates under conditions of uncertainty. Comput. Sci. ICCS 2021, 198–205 (2021) 9. Qin, W., Zhang, J., Song, D.: An improved ant colony algorithm for dynamic hybrid flow shop scheduling with uncertain processing time. J. Intell. Manuf. 29, 891–904 (2018) 10. Dhiflaoui, M., Nouri, H.E., Driss, O.B.: Dual-resource constraints in classical and flexible job shop problems: a state-of-the-art review. Procedia Comput. Sci. 126, 1507–1515 (2018) 11. ILOG CPLEX. https://www.ibm.com/jp-ja/products/ilog-cplex-optimization-studio

The Impact of the Design Decisions of an Order Picking System on Human Factors Aspects of the Order Pickers Vivek Vijayakumar(B)

and Fabio Sgarbossa

Norwegian University of Science and Technology, 7491 Trondheim, Norway {vivek.vijayakumar,fabio.sgarbossa}@ntnu.no

Abstract. Warehouses are crucial for supply chain management and the success of businesses in production and logistics systems. The order picking (OP) system plays a vital role in achieving short lead times and high customer satisfaction and order pickers are essential to achieve flexibility in the OP system due to their cognitive and motor skills. However, manual order picking is time-consuming and accounts for approximately 50% of overall operating costs in warehousing. It is important to consider the human factors (HF) aspects of the order picker to avoid deviations from expected performance outcomes and reduce the risk of errors that could cause delays and financial losses. Negligence of HF could also lead to musculoskeletal disorders in the order picker. This study aims to empirically show the impact of such decisions on the HF aspects of order pickers. The study uses case studies and survey-based empirical data to analyze the impact of decisions on the HF aspects of order pickers. The findings suggest that consideration of HF is crucial for the success of the OP system and the wellbeing of order pickers. The study highlights the need for further research on HF aspects in the OP system and provides insights for decision-makers to optimize the performance of the system. Keywords: Order picking · Human Factors · Design decisions · Empirical study

1 Introduction Warehouses are considered to key players in supply chain management and are essential for the success for business in production and logistics systems [1]. Order picking (OP) systems are critical in warehousing to achieve short lead times and excellent customer satisfaction. The OP system consists different tasks such as setting up of picking list, travelling inside the warehouse, searching for the desired products, and picking the products from the desired locations in a warehouse to meet customer needs [1]. Fully automating order picking processes brings about various disadvantages, including high investment costs, the need for standardization, and reliance on computer systems, in addition to its lack of flexibility [2]. Despite the high labor costs, manual © IFIP International Federation for Information Processing 2023 Published by Springer Nature Switzerland AG 2023 E. Alfnes et al. (Eds.): APMS 2023, IFIP AICT 689, pp. 775–786, 2023. https://doi.org/10.1007/978-3-031-43662-8_55

776

V. Vijayakumar and F. Sgarbossa

operation persists in up to 80% of order picking warehouses [2]. Thus, order pickers are vital part of OP systems to achieve high amount of flexibility to the OP system due to combination of their cognitive and motor skills. Furthermore, humans can pick complex products from storage locations, something machines and automated systems cannot do reasonably [3]. These explanations demonstrate that the order picking system is a time-consuming operation that requires a significant amount of manual handling of products. As a result, the cost of process in warehousing accounts for approximately 50% of overall operating costs [1]. However, when manual order pickers are in the system, it is important to consider the HF aspects of the order picker. The four HF aspects are the physical, mental, perceptual, and psychosocial aspects. This is because if HF aspects are not considered then it could lead to deviation from the expected performance outcomes of an OP system [3]. This is because there could be high risk of errors in a manual OP system, as order pickers could pick wrong or incorrect number of items. As a result, these pick error could cause delay in delivering the products or financial losses [4]. The negligence of HF not only impacts the performance of the OP system, but also impacts the wellbeing of the order picker. This is because of handling heavy products in awkward body postures, which thereby increasing the chances of developing musculoskeletal disorders (MSDs), with low back disorders being the most common injury. Thus, the human factors in the OP system are one of an important player in the performance of the system [2]. Even tough, it is important to consider the HF aspects in an OP system, there has been negligence of HF in the decision making of setting up of the OP system or for the introduction of technologies into the system [5]. If these aspects are not considered, then as said before negative consequence could lead to the performance outcomes of an OP system [5]. According to [6], most OP research has concentrated on establishing a mathematical model and simulation model but has rarely included case studies and survey-based empirical data. Therefore, the aim of this study is to conduct an empirical study to show the impact of decision regarding the system settings and technologies in an order picking system on HF aspects of the order pickers. The remainder of the paper is organized as follows. Section 2 explains the literature review. Section 3 explains the methodology adopted in this paper. Section 4 presents the findings and analysis from the study. Section 5 summarizes the paper by highlighting key points, limitation of the study and the future research opportunities.

2 Literature Review The OP system could be classified into two method, picker to parts and parts to picker. In a picker to parts system, the order picker moves towards the desired parts which are the products remain stationary. On the other side, the order picker remains stationary, and the products moves towards the order picker. It is understood that for two different method of OP system, different system settings are required. According to [6], the system setting of an OP system could be classified into mechanization level, information availability and warehouse dimensionality. Few years ago, [7] have explained the system settings of an OP system could be categorized into layout and storage assignments. These factors

The Impact of the Design Decisions of an Order Picking System

777

are considered because they are related with the design of an OP system. The layout design is associated with the number, length, and width of aisles in the blocks and also, the shelf layout and configuration. The storage assignment defines the allocations of the products to the storage locations in the warehouse based on the product characteristics. Thus, it is important to understand the impact of human factors on the system settings. In the case of focusing human factors of order pickers with the design of the warehouse layout, [8] studied on an optimal layout problem as mixed-integer programming with the aim of minimizing the total ergonomics strain on the order pickers. When it comes to the focus on human factors aspects of order pickers in combination with system settings. [22] have represented a new ergonomic storage location assignment algorithm which reduces the mechanical load of the lumbar spine on the order pickers. [23] introduced a model for storage assignment problem using the integer linear optimization with consideration of the OP time, energy expenditure rate and health risk associated with the order pickers. [26] have studied on order pickers picking from different pallet rack layout considering the economic and ergonomics objectives. [8] with the help of mixed integer programming have achieved to minimize the ergonomic strain during the order picking. [9] have developed a mathematical model to address the storage assignment with respect to the fatigue level of the order pickers. [10] introduced a bi-objective approach considering the total order picking time and human energy expenditure into a storage assignment problem. [11] has talked on storage assignment decisions based on the learning and forgetting of order pickers. [12] introduced an algorithm to help in selecting the highly efficient storage locations thereby reducing the probability of mis-picks and improving the ergonomics for the order pickers. When it comes to the technologies in OP systems, [13] have described the transition of automated order picking systems and its impact on the order picker’s learning and work organization. [14] have proposed a model that considers the capacity, ergonomics, and cost of training an automated system. [15] have presented a simulation model to evaluate the fatigue on order pickers, who work in close collaboration with the picking robots. [16] has evaluated the workload and ergonomics design of workstations in picker to parts order picking system. [17] has compared the horizontal carousels with shelving systems with the consideration of space, time, and ergonomics. [18] studied the influence of paperless picking and the usage of forklifts for transportation on order picker’s wellbeing and productivity. [19] presents an empirical analysis of learning curves in pick by voice and semi-automated OP systems.

3 Methodology This section describes the research design, which contains a description of the study strategy, data gathering procedure, and analytic methodologies. The section begins with a thorough discussion of the research approach, followed by a description of the data gathering procedure, and continues with an overview of both descriptive and content analysis techniques (Table 1).

778

V. Vijayakumar and F. Sgarbossa Table 1. Research design

Method

Technique

Purpose

Case Study Questionnaire To understand the well-being of the order pickers (from order pickers perspective) Interview

To understand the design characteristics of the warehouse (from managers perspective) To understand the well-being of the order pickers (from managers perspective)

3.1 Research Method The research method adopted for this study is case study. The case study is conducted with the help of a Norwegian grocery distribution centre. The case study managed to provide a detailed investigation, with the empirical data collected to deliver an analysis and the process involved in the context [20]. The main objective of the case study is to do intensive research on a specific case to highlight the essential process and relationships [21]. In this study, the focus is to find the essential system design of the three different warehouses of the case company and to evaluate their relationship with the HF aspects of the order pickers. 3.2 Data Collection Technique Two types of data collection technique are used to support the case study research method: questionnaires and interviews. These data collecting techniques are described in further detail below. Questionnaire This study used the NASA TLX questionnaire, a multidimensional rating method that provides an overall workload score based on the weighted average of ratings on six subscales [24]. They are mental demands, physical demands, temporal demands, own performance, and frustration. Mental Demand assesses cognitive requirements and complexity. Physical Demand measures physical effort. Temporal Demand evaluates time pressure. Performance gauges perceived task success. Effort considers overall mental and physical exertion. Frustration captures emotional strain and dissatisfaction. These dimensions provide a comprehensive understanding of workload, enabling identification of high workload areas and informing task design and resource allocation improvements. The questionnaire is provided to the order picker to collect the information regarding the well-being of the order pickers. The duration of order picker to fill the questionnaire is approximately 30min. The questionnaire is distributed to all the order pickers in the warehouse. Interview A semi structured interview is conducted with the managers and order pickers [20]. The questions for the manager’s interview would address the data regarding the design of the

The Impact of the Design Decisions of an Order Picking System

779

order picking system. For e.g., layout, storage assignment, technology etc. As a result, an understanding of the warehouse’s system setting is provided. One manager, who is responsible for the order picking system, was selected for the interview. Secondly the questions for the order picker’s interview would evaluate HF aspects of the order picker. One experienced order picker from each warehouse who is comfortable to communicate in English were selected for the interview. The duration of the interview with the manager and order pickers were approximately 1 h each. 3.3 Data Analysis Triangulation is employed in the case study to enhance the validity of the acquired data by conducting two distinct types of data analysis, namely descriptive analysis and content analysis. This approach ensures a more robust and comprehensive examination of the findings. By combining these two approaches, the study benefits from a more holistic understanding of the phenomenon under investigation, strengthening the validity and reliability of the research outcomes. Descriptive Analysis In the initial phase of the study, a descriptive analysis is conducted using quantitative data collected through a questionnaire [24]. This analysis involves presenting the raw data without manipulation to understand the trends in the HF aspects when altering the system settings for the order picker. By analysing the unprocessed data, authors can observe patterns and changes in the level of workload on the order pickers. The descriptive findings derived from this analysis provide a comprehensive overview of the observed trends, serving as a foundation for further exploration of the relationship between system settings and HF aspects. This analysis aims to gain insights into the impact on the order picker’s working conditions and performance. Content Analysis After completing the descriptive analysis, content analysis is utilized to gain a more profound comprehension of the observed trends [25]. The coding process in content analysis involves identifying significant themes, concepts, or words within the qualitative data obtained from semi-structured interviews conducted with order pickers [25]. The constructs used in this analysis may focus on the order pickers’ perceptions, experiences, and attitudes towards the system settings and their impact on HF aspects. Evaluating the performance in content analysis involves examining the order pickers’ narratives and identifying any references to changes in their performance or productivity resulting from the system settings. The objective of this analysis is to assess the presence, meanings, and connections of specific words, themes, or concepts, shedding light on the reasons behind the alterations in the human factors aspects when the order picker’s system settings were modified.

780

V. Vijayakumar and F. Sgarbossa

4 Findings and Analysis 4.1 Case Company The case study focuses on three key warehouses within a major Norwegian grocery distribution centre. These warehouses, namely Warehouse A, Warehouse B, and Warehouse C, have been selected for in-depth analysis. Warehouse A is the most automated among the three, employing the parts to picker OP method with the assistance of an AS/RS (Automated Storage and Retrieval System). This means that the system retrieves the necessary parts and brings them directly to the order picker, streamlining the picking process. Warehouse B is a semi-automated warehouse that utilizes the picker to parts OP method. It incorporates pick by voice technology, where order pickers are guided by a microphone to perform various picking tasks. This technology assists them in setting up the picking list and locating and picking the required parts from designated locations. In contrast, Warehouse C is a fully manual warehouse, relying on the manual picker to parts OP method. Order pickers perform all the picking tasks manually, without the aid of automated systems or voice technology. It is worth noting that Warehouse A is the largest in size, while Warehouse C is the smallest among the three. The varying degrees of automation and manual involvement in these warehouses provide an opportunity to study and compare the effects of different system settings and technologies on the HF aspects of the order pickers. 4.2 Descriptive Findings The findings from the NASA TLX obtained from the three distinct warehouses are described in this section. Figure 1 in the study provides a visual representation of the data collected from the three warehouses (A, B, and C) and how it is classified and categorized into four primary order picking tasks: setup, travel, search, and pick. These tasks represent different stages or activities involved in the order picking process. To assess the workload associated with each task, the study utilizes the NASA Task Load Index (TLX), which consists of six dimensions: Mental, physical, temporal, performance, effort, and frustration. These dimensions capture different aspects of the workload experienced by order pickers during their tasks. By evaluating each dimension within the context of the three warehouses, a comprehensive analysis of the workload is conducted. This analysis allows for a deeper understanding of how workload factors vary across different aspects of order picking and between the warehouses (Figs. 2, 3 and 4).

The Impact of the Design Decisions of an Order Picking System

Fig. 1. Descriptive findings of the setup task

Fig. 2. Descriptive findings of the travel task

Fig. 3. Descriptive findings of the search task

781

782

V. Vijayakumar and F. Sgarbossa

Fig. 4. Descriptive findings of the pick task

4.3 Content Analysis Moreover, the content analysis provides a detailed exploration of the trends identified in the previous section, offering further explanations and insights specific to each order picking task. This analysis delves deeper into the findings, providing a more comprehensive understanding of the factors and patterns influencing performance within each task of the order picking process. Setup Of all the dimensions, the physical demand is reported more from all the warehouses and the least for mental demand. The mental demand is least in the setup task for the warehouse B because the search list for the picking is prepared by the pick by voice technology. This could reduce the mental demand for the operators to make decisions on picking the orders and warehouse A has the most mental fatigue because the order pickers have reported that the system crashes in between and this would annoy the order pickers in picking the orders and order pickers have also reported that they have to reboot the system to retrieve the picking list from the system. It is also more physically demanding for the warehouse A and least for warehouse C because order pickers from warehouse A notified that the standing in the same position for preparing the collection could case back pain to the operators and in warehouse C the order pickers feel less fatigue because order pickers have less parts to pick compared to other warehouses. The temporal demand for creating the setup list is equally same for all the three warehouses but bit high for the warehouse A and less for warehouse C because warehouse A being the biggest with huge movement of goods than warehouse C, the pressure on order pickers from warehouse A is more when compared with the warehouse C. When it comes to performance, the warehouse A is performing better than the remaining warehouses and the least performance was reported by warehouse C because the order pickers of warehouse A stated that if the system work without crashing then they could prepare more picking list faster than the conventional paper picking list of warehouse C. Also, in case of effort and frustration both the warehouse A and B has shown equal level and warehouse C has the highest because the warehouse C has conventional picking list,

The Impact of the Design Decisions of an Order Picking System

783

that causes order pickers more frustration and effort on cross checking the orders that they need to pick. Travel When it comes to the travel tasks, the mental demand is reported in warehouse A and the least for the warehouse C because the order pickers of warehouse C are used to follow the same route all the time which led them with least mental demand. On the other hand, the order pickers of warehouse A travel less often only when there is need of intervention with the AS/RS. Secondly, the physical workload is rated the most among all the dimensions. The physical demand is stated more for warehouse A and least for warehouse C because during the time of intervention for warehouse A, the order pickers has to climb to check the problem, which causes more fatigue to the them. Thereafter, the temporal demand is more for warehouse B and the least for warehouse C because warehouse B is bigger than warehouse C and the order pickers from warehouse B has to cover a greater distance to meet their picking target. However, when it comes to the performance, the warehouse A has better output and the least for warehouse C because if there is no intervention for AS/RS, there doesn’t exist a need for the order pickers to travel. Secondly, the effort is most for warehouse A and the least for the warehouse B only when there is an intervention in warehouse A and the order picker has to take actions to rectify it. Finally, the order pickers faced more frustration for warehouse C and the least for warehouse A because the order pickers of warehouse C states that they have to keep on travelling over the warehouse repeatedly causing them frustration. Search In the search task, the mental workload in stated most in warehouse A and least for warehouse B because if there is an intervention the order picker of warehouse A takes time to find the location of the item. But, in warehouse C since the order pickers are used to pick the parts, they remember the storage locations effecting in lower mental demand. Secondly, in case of physical demand and temporal demand is equally high for warehouse A and B and the least for warehouse C because both the order pickers of warehouse A and B has to pick more items in comparison with C. The performance is best reported for the warehouse A and least for warehouse C because order pickers from warehouse A has less travel to pick the items, but only have to travel during the need of intervention for robot. The effort is rated most for warehouse A and the least for warehouse C because during the intervention for robot in warehouse A, the order pickers has a greater effort to rectify and bring to work when compared with conventional warehouse C. As in the most cases the frustration is most of warehouse C and the least for the warehouse A because the order pickers of warehouse C has no assistive technology to help find the location of the item, which is not the case for warehouse A that provides the assistance to the order pickers to find the right item with its location. Pick Finally, in the case of pick task, the mental demand is rated most for the warehouse A and the least for the warehouse B because order pickers of warehouse A feels that

784

V. Vijayakumar and F. Sgarbossa

standing at a single place and picking the items from one point is boring. The physical demand was reported equally high in all the three warehouses, but a slightly higher for warehouse B because the order pickers of warehouse B has to pick more items from the shelf when compared to other warehouses. The temporal demand is high for both the warehouse B and C and the least for warehouse A because both the warehouse B and C are parts to picker OP method and the order pickers from both warehouses has reported that they have to pick the parts from the warehouses from different locations within the provided time span to meet the picking rate. The pick performance is high for warehouse A and the least for the warehouse C because the rate of picking is decided by the system and the order pickers has to perform the pick task based on the feedback from the system while in the case of warehouse C the rate of picking is entirely depended on the speed of the order picker. The effort for picking is rated equally in all the three warehouses because the picking task is same in all the three warehouses. Finally, the frustration is rated most for the warehouse C and the least for the warehouse A because the order pickers from warehouse C informed that the frustration for picking is due to the temporal demand of the work to keep up the picking rate.

5 Conclusion The study has evaluated the impact of the design decisions of an order picking system on human factors aspects of the order pickers. It was fascinating to see that parts to picker OP method has better performance but could lead to more effort to order pickers if there is a need for interventions for the AS/RS. Overall, order pickers in a manual warehouse reported more frustration doing OP tasks than order pickers in a warehouse that made use of assistive technology for the picker to part OP method. This study’s managerial implications emphasize that effective decision-making by managers plays a vital role in attaining the intended performance outcomes in order picking systems. Managers can optimize system performance and align it with planned goals by taking into account human factors and making well-informed design decisions. Additionally, fostering open communication channels and prioritizing the well-being of order pickers contribute significantly to achieving the desired performance outcomes. One significant drawback of this study is its limited scope, as it solely examines a grocery distribution centre located in Norway. By confining the research to this specific context, the findings may lack broader applicability and generalizability to other settings, limiting the validity of the findings to internal factors. However, by examining various sectors and conducting comparative analyses, researchers can enhance external validity and gain a deeper understanding of how different design choices influence order pickers’ performance and well-being. Another limitation of this study is its narrow focus on a limited set of technologies within the order picking system, specifically paperless picking technologies and AS/RS. While these technologies were thoroughly examined, other emerging or alternative technologies that could impact order pickers may not have been included. To address this limitation, future steps of this research could involve exploring and evaluating the effects of a broader spectrum of technologies. By considering a wider range of technologies, a more comprehensive understanding of their impact on order

The Impact of the Design Decisions of an Order Picking System

785

pickers could be achieved. This expanded analysis would provide managers and decisionmakers with a more holistic view when making informed choices about integrating different technologies into order picking systems. Additionally, it would ensure that the study’s conclusions remain relevant and applicable in a rapidly evolving technological landscape.

References 1. De Koster, R., Le-Duc, T., Roodbergen, K.J.: Design and control of warehouse order picking: A literature review. Eur. J. Oper. Res. 182(2), 481–501 (2007) 2. Grosse, E.H.: Application of supportive and substitutive technologies in manual warehouse order picking: a content analysis. Int. J. Prod. Res. 1–20 (2023) 3. Grosse, E.H., Glock, C.H., Neumann, W.P.: Human factors in order picking: a content analysis of the literature. Int. J. Prod. Res. 55(5), 1260–1276 (2017) 4. Setayesh, A., Grosse, E.H., Glock, C.H., Neumann, W.P.: Determining the source of humansystem errors in manual order picking with respect to human factors. Int. J. Prod. Res. 60(20), 6350–6372 (2022) 5. Vijayakumar, V., Sgarbossa, F., Neumann, W.P., Sobhani, A.: Framework for incorporating human factors into production and logistics systems. Int. J. Prod. Res. 60(2), 402–419 (2022) 6. Davarzani, H., Norrman, A.: Toward a relevant agenda for warehousing research: literature review and practitioners’ input. Logist. Res. 8, 1–18 (2015) 7. Boysen, N., De Koster, R., Weidinger, F.: Warehousing in the e-commerce era: a survey. Eur. J. Oper. Res. 277(2), 396–411 (2019) 8. Diefenbach, H., Glock, C.H.: Ergonomic and economic optimization of layout and item assignment of a U-shaped order picking zone. Comput. Ind. Eng. 138, 106094 (2019) 9. Zangaro, F., Battini, D., Calzavara, M., Persona, A., Sgarbossa, F.: A model to optimize the reference storage assignment in a supermarket to expedite the part feeding activities. IFAC-PapersOnLine 51(11), 1470–1475 (2018) 10. Battini, D., Glock, C.H., Grosse, E.H., Persona, A., Sgarbossa, F.: Human energy expenditure in order picking storage assignment: a bi-objective method. Comput. Ind. Eng. 94, 147–157 (2016) 11. Grosse, E.H., Glock, C.H., Jaber, M.Y.: The effect of worker learning and forgetting on storage reassignment decisions in order picking systems. Comput. Ind. Eng. 66(4), 653–662 (2013) 12. Marvel, J.H., Shell, R.L., Weckman, G.R.: An application of heuristic algorithms for determining inventory location in a distribution warehouse. Int. J. Ind. Eng. 8, 5–15 (2001) 13. Loske, D.: Empirical evidence on human learning and work characteristics in the transition to automated order picking. J. Bus. Logist. 43(3), 302–342 (2022) 14. Rieder, M., Bonini, M., Verbeet, R., Urru, A., Bartneck, N., Echelmeyer: Evaluation of humanrobot order picking systems considering the evolution of object detection. In 2021 IEEE International Conference on Industrial Engineering and Engineering Management (IEEM), pp. 1–8 (2016) 15. Zhang, M., Winkelhaus, S., Grosse, E.H., Glock, C.H.: A simulation model for evaluating the efficiency of robot-supported order picking warehouses. In: Symposium on Logistics (2021) 16. Wakula, J., Steinebach, T., Klaer, V., Rabenhaupt, W., Maier, G.: Analysis of the physical workload and ergonomic design of workstations for “goods-to-person” order picking. In: Black, N.L., Neumann, W.P., Noy, I. (eds.) Proceedings of the 21st Congress of the International Ergonomics Association (IEA 2021). IEA 2021. Lecture Notes in Networks and Systems, vol. 221, pp. 522–529. Springer, Cham (2021). https://doi.org/10.1007/978-3-03074608-7_64

786

V. Vijayakumar and F. Sgarbossa

17. Ðuki´c, G., Opetuk, T., Gajšek, B.: Space, time and ergonomic assessment of order picking using horizontal carousel. In: Sumpor, D., Jambroši´c, K., Jurˇcevi´c Luli´c, T., Milˇci´c, D., ˇ Salopek Cubri´ c, I., Šabari´c, I. (eds.) Proceedings of the 8th International Ergonomics Conference. ERGONOMICS 2020. Advances in Intelligent Systems and Computing, vol. 1313 pp. 73–83. Springer, Cham (2021). https://doi.org/10.1007/978-3-030-66937-9_9 18. Gajšek, B., Ðuki´c, G., Butlewski, M., Opetuk, T., Cajner, H., Kaˇc, S.M.: The impact of the applied technology on health and productivity in manual “picker-to-part” systems. Work 65(3), 525–536 (2020) 19. Loske, D., Klumpp, M.: Smart and efficient: Learning curves in manual and human-robot order picking systems. IFAC-PapersOnLine 53(2), 10255–10260 (2020) 20. Yin, R.K.: Discovering the future of the case study. Method in evaluation research. Eval. Pract. 15(3), 283–290 (1994) 21. Rashid, Y., Rashid, A., Warraich, M.A., Sabir, S.S., Waseem, A.: Case study method: a step-by-step guide for business researchers. Int. J. Qual. Methods 18, 1609406919862424 (2019) 22. Steinebach, T., Wakula, J., Mehmedovic, A.: The influence of an ergonomic storage location assignment on human strain in manual order picking. In: Black, N.L., Neumann, W.P., Noy, I. (eds.) Proceedings of the 21st Congress of the International Ergonomics Association (IEA 2021). IEA 2021. Lecture Notes in Networks and Systems, vol. 221, pp. 511–521. Springer, Cham (2021). https://doi.org/10.1007/978-3-030-74608-7_63 23. Gajšek, B., Šinko, S., Kramberger, T., Butlewski, M., Özceylan, E., Ðuki´c, G.: Towards productive and ergonomic order picking: multi-objective modeling approach. Appl. Sci. 11(9), 4179 (2021) 24. Hart, S.G., Staveland, L.E: Development of NASA-TLX (Task Load Index): results of empirical and theoretical research. In Advances in Psychology, vol. 52, pp. 139–183, North-Holland (1998) 25. Harwood, T.G., Garry, T.: An overview of content analysis. Mark. Rev. 3(4), 479–498 (2003) 26. Calzavara, M., Glock, C.H., Grosse, E.H., Sgarbossa, F.: An integrated storage assignment method for manual order picking warehouses considering cost, workload and posture. Int. J. Prod. Res. 57(8), 2392–2408 (2019)

SME 5.0: Exploring Pathways to the Next Level of Intelligent, Sustainable, and Human-Centred SMEs

From Surviving to Thriving: Industry 5.0 at SMEs Enhancing Production Flexibility Zuhara Zemke Chavez1(B) , Ala Arvidsson2 , Jannicke Baalsrud Hauge1 , Monica Bellgran1 , Seyoum Eshetu Birkie1 , Patrik Johnson2 , and Martin Kurdve2 1 KTH, Royal Institute of Technology, Stockholm, Sweden

[email protected] 2 Chalmers University of Technology, Gothenburg, Sweden

Abstract. This study explores how human-centered digitalization can contribute to the flexibility and adaptability of small and medium-sized enterprise (SME) production processes, resulting in more resilient systems. This study explains the relationship between digital technologies and production system features through progressively more human-centric stages of a digitalized manufacturing system. The authors present a case study of an SME that implemented a human-centric strategy, placing people’s needs and interests at the center of its processes, leading to more flexible and inclusive production processes and consistent with the goals of Industry 5.0. The results suggest that a digitalized working method that considers human capabilities and needs can enable a more diverse workforce and the rapid setup of new and additional production processes, thus helping SMEs respond to supply chain disruptions. The findings have implications for managers and practitioners interested in driving or supporting the transition of SMEs to humancentric, resilient, and sustainable businesses. Keywords: Adaptability · Flexibility · Human-centric production

1 Introduction Industry 5.0 (I5.0) is a concept that acknowledges the potential of industries to contribute to society beyond just creating jobs and generating economic growth. It aims to create a sustainable and prosperous future by ensuring that production processes align with our planet’s limits and prioritizing workers’ welfare within the industry [1]. At the same time, the I5.0 paradigm promotes systems’ agility and resiliency with the utilization of flexible and adaptable technologies. I5.0 is still a relatively new concept, and there are ongoing discussions about the extent and speed to which manufacturers will adopt it. However, many experts see the concept as a potential solution to challenges that Industry 4.0 (I4.0) does not address. It has become evident that the I4.0 concept is highly technology driven and doesn’t include sustainability and human orientation to the extent necessary when digitalizing the industry. Hence, there is a risk that I4.0 emphasizes digitalization as a goal rather than a mean. © IFIP International Federation for Information Processing 2023 Published by Springer Nature Switzerland AG 2023 E. Alfnes et al. (Eds.): APMS 2023, IFIP AICT 689, pp. 789–802, 2023. https://doi.org/10.1007/978-3-031-43662-8_56

790

Z. Z. Chavez et al.

During the COVID19 pandemic, organizations and companies in almost all industries realized digital technologies’ potential to address disruptions. Some went on board with digital transformations, while others prioritized digital initiatives to cope with supply chain disruptions [2]. The Ukrainian war and the energy crisis are critical drivers of digital and business model transformations [3]. The resilience of industrial organizations is one of the central themes of I5.0 to alleviate the challenges of uncertainties [4]. Emerging technologies play a key role in I5.0; with technologies such as AI, digital twins, and the big data analytics, the industry is already improving product design and manufacturing processes. In contrast to I4.0, under I5.0, the purpose of technology implementation is expanded to support sustainability and resilience. Technology is applied to societal problems triggered by high-impact shocks, and problems can be monitored and analyzed in real-time to prepare and apply early measures and to avoid shocks turning into crises [5]. Focusing on I5.0 could also contribute to resilience in small- and medium-sized enterprises (SMEs). SMEs are the backbone of the European economy, representing 99% of all businesses in the EU [6], and are seen as fundamental to the EU’s transition to a sustainable and digital economy. Research has shown that digitally transformed SMEs can capitalize on opportunities during a crisis and reorganize resources to cope with a crisis [3]. A critical factor in building the resilience of SMEs is the flexibility that fewer but key individuals add to the organizations. However, even though introducing new technologies and practices is always uncertain in SMEs, research on Industry 4.0 at SMEs has shown a limited but thus positive impact of new technologies on SMEs’ operational performance [7, 8]. Since I5.0 is a rather new topic recently studied by researchers and practitioners [1, 4, 9, 10], there are yet scarce reflections of I5.0 practices in general and even less at SMEs. SMEs’ business strategy is often based on adaptability, a reactive approach, and customer proximity. In this sense, SMEs need to be flexible by nature, and one can argue that the more flexibility an SME can demonstrate, the higher the chances of surviving the market and staying competitive. Thus, the promise of enabling greater flexibility and resilience at the industrial level through I5.0 motivates exploring the topic in the context of SMEs. This study focuses on a unique case of an SME utilizing the digitalization of work processes to increase the flexibility and transferability of the production work across individuals regardless of their cognitive and physical capabilities or skills. The case company has collaborated with academia previously around ecological and social sustainability [11–13], though the studies were not directed to resilience and flexibility. The digitalized work method was initially developed to integrate marginalized people into Sweden’s workforce and manufacturing. However, as the SME was hit by the shortages and lockdowns triggered by the COVID19 pandemic, this work process was adapted to the context during 2020–2022 by using voluntary workers without previous training. The method was deployed to quickly industrialize a production process for protective equipment for hospitals and care providers in Sweden. The case represents a unique application of agile digitalization for human-centric production that helped a Swedish SME survive and thrive.

From Surviving to Thriving: Industry 5.0 at SMEs

791

Consequently, the study presented in this paper explores a central research question: How can the human-centric digitalization of production work processes enhance SMEs’ flexibility? We consider this unique case and its digital implementations to demonstrate what I5.0-related practices in SMEs could look like. Here, we focus on process flexibility, i.e., the ability of a company to change or adapt its manufacturing setups according to demand and supply changes and place. The first section of the paper presents the aim of the research and its considerations. Section 2 presents the theoretical background regarding resilience and flexibility at SMEs and I5.0 as an approach from now on. In Sect. 3, the method followed is presented. Section 4 includes the case findings and analysis based on our conceptual framework for flexibility in a human-centric manufacturing system. Section 5 covers discussions about flexibility in human-centric digitalization of production processes at SMEs and further reflections on our proposed conceptual framework. Lastly, as we advance, conclusions and recommendations are provided in Sect. 6.

2 Theoretical Background 2.1 Resilience and Flexibility in SMEs SMEs are essential to Europe’s competitiveness and prosperity, industrial ecosystems, economic and technological autonomy, and resilience to external shocks. Large enterprises may have relatively more resources, such as capital and manpower, to absorb the shock of unexpected events. SMEs may face limitations such as resource scarcity, cash flow, and dependence on external systems [14]. Still, they may have more agility and flexibility due to the shorter chain of command, allowing them to pivot and adapt to changing circumstances more quickly [15–17], making the context arguably suitable for I5.0 practices. To manage supply disruptions, SMEs often rely on mobilizing resources within their supply network [18–20]. Flexibility refers to the ability of a system to change (adapt) in dynamic environments. It is often used interchangeably with adaptability [2]. Flexibility is about how the whole manufacturing system can change what it does and how individual parts can change [21]. It concerns both technical and human responsiveness to disturbances [22]. Moreover, an efficiently flexible organization is one that is adapted to what the environment needs [23]. Process flexibility examines the ability of a company to change manufacturing setups at flexible lines according to demand or supply changes [24]. It also refers to the company’s ability to provide goods or services using different facilities or resources, for instance, multiple facilities or lines [24, 25]. Flexibility has been extensively studied in manufacturing, services, and supply chain at different levels, i.e., strategic, tactical, and operative perspectives [21, 23, 25, 26]. In their study, [26] considered emerging issues in flexibility, such as risk and uncertainty management, environmental sustainability, optimal strategies under competition, and optimal operations with strategic consumer behaviors. Flexibility mechanisms comprise the tools, management practices, and systems that can be used to achieve a particular type of flexibility, i.e., product variety (modification flexibility), location of production (volume flexibility), and rapid introduction of new products (changeover flexibility), for instance, ERP system, design-for-manufacture

792

Z. Z. Chavez et al.

principles, multi-skilled workers and holding inventory [27]. In manual assembly, reducing complexity is a key factor for increased flexibility, whereas digital systems may increase or decrease complexity [22]. Risk mitigation is a crucial challenge for SMEs, especially in the manufacturing sector, as supply cost usually represents the most considerable budgetary portion [28]. Studies on risk mitigations classify solutions into either redundancy or flexibility approaches [29, 30]. Redundancy is a less common strategy in SMEs. It is an expensive strategy since many of the built redundancies will wait for use in an emergency case without creating any value in business-as-usual times [31]. Studies [30] have found that more efficient strategies for risk mitigation focus on flexibility rather than redundancy for supply chain failures. Ellegaard [32] also reported that SMEs do not implement bufferrelated, supplier development, or formal process-oriented measures due to their limited internal resources, weak position power in the supply chain, and informal management structure. Thus, the mindset and orientation of SMEs play an essential role, as many prefer to network with suppliers for risk mitigation [33]. 2.2 Early Insights on Industry 5.0 According to the European Commission, I5.0 demands technologies to address new industrial, societal, and environmental requirements. It means using technologies to increase production flexibility during disruption and making value chains more robust. It also means implementing technology that adapts to the worker rather than the other way around. It means using technology for circularity and sustainability [33], ultimately becoming a competitive industry that respects planetary boundaries and minimizes its negative environmental impact [34]. The EU Commission states that I5.0 has three characteristics: human-centricity, sustainability, and resiliency, which frame the vision for industry evolution [10]. This approach contributes to three of the commission’s priorities: i. an economy that works for people”, ii. “European Green Deal” and iii. “Europe fit for the digital age”. According to the EU Commission [34], policy brief report, the I5.0 characteristics are defined as: Human-centricity refers to the industry putting core human needs and interests at the heart of the production process. The vision is to use technology to adapt the production process to the worker’s needs, rather than the industry worker to adapt their skills to the rapidly evolving technology needs. It also includes securing industrial workers’ privacy, autonomy, and human dignity. Resilience refers to the need to develop higher robustness in industrial production, be prepared against disruptions, and ensure it can provide and support critical infrastructure in times of crisis. It includes developing suitable resilient strategic value chains, adaptable production capacity, and flexible business processes, primarily where value chains serve basic human needs, e.g., healthcare or security. Sustainability refers to industry respecting the planetary boundaries and includes developing circular processes that reuse, re-purpose and recycle natural resources, reduce waste, and have an environmental impact. It also means reducing energy consumption and greenhouse emissions to ensure the needs of today’s generations without risking the needs of future generations. Technologies play a prominent role in optimizing resource efficiency and minimizing waste.

From Surviving to Thriving: Industry 5.0 at SMEs

793

A challenge concerning adopting I5.0 in SMEs is the scalability in ensuring a broadscale implementation of technologies across value chains and ecosystems, including the small players. According to the European Commission 2020 report from technology leaders [35], the technologies under the umbrella of the I5.0 concept comprise: i. Human-centric solutions and human-machine-interaction technologies that interconnect and combine the strengths of humans and machines. ii. Bio-inspired technologies and smart materials that allow materials with embedded sensors and enhanced features while being recyclable. iii. Real-time-based digital twins and simulation to model entire systems. iv. Cyber-safe data transmission, storage, and analysis technologies that can handle data and system interoperability. v. Artificial Intelligence, e.g., to detect causalities in complex, dynamic systems, leading to actionable intelligence. vi. Technologies for energy efficiency and trustworthy autonomy, as the technologies mentioned earlier, will require large amounts of energy. Literature on how digital technologies supported industry to cope during the COVID19 pandemic [36] demonstrate that investing in long-term resilience and business sustainability over short-term optimization and improvements must be a priority for manufacturers and managers to be prepared for similar circumstances in the future. In addition, industries may face a higher demand for personalized products while having the requirement to ensure that the human needs of industrial workers and environmental impact are not compromised by the economic drive to satisfy product demand. In their study, Lu et al. [37] present a framework for a human-centric manufacturing system in an unstructured and distributed manufacturing environment to enable “ultra-flexible” (as described by the authors) manufacturing automation of personalized products. The framework contextualizes the human-centric manufacturing system towards mass personalization. Here, flexibility must include human wellness and working freedom to be optimized while ensuring good system productivity, which is different from traditional production scheduling and control. According to the authors, the highly dynamic team collaboration and contingencies resulting from maximum freedom given to industrial workers will directly cause traditional centralized production scheduling algorithms to fail. Learnings from Lu et al. [37] helps us understand the system characteristics concerning human and technologies of the I5.0 approach. Our study tries to understand how digitalization enables flexibility at the plant level at SMEs.

3 Methodology The research presented adopts a qualitative research design based on data collection through mainly semi-structured interviews and observations [38]. The study focused on a Swedish SME called the PS case, which has faced significant disruptions in its production system due to COVID19’s impact on its market and supply networks. The case strongly supports answering the research question, as it allowed us to investigate how, despite those disruptions, the SME implemented human-centric digitalization in their production work processes and how it enhanced the SMEs’ flexibility.

794

Z. Z. Chavez et al.

3.1 Data Collection and Analysis The study relies on two main types of primary data, i.e., phenomena in reality and people’s perceptions and experiences of reality [39]. Therefore, data was gathered from workshops, presentations, semi-structured interviews, and on-site observations at the SME firm. Data was collected from March 2021 to Dec 2022 (see Table 1). Our study focuses on the operations management and strategic aspects of enabling flexibility and resilience in SMEs’ production processes, with emphasis on inclusive workspace design; therefore, interviews were conducted with the CEO of the case company (also acting as COO), and production workers’ interviews were not included at this stage in this study (in a previous inclusive work study, operators have been interviewed [11, 12, 40]). Workshops included the PS case and two other SMEs in similar situations, focusing on resilience and flexibility while responding to disruptions. At one of the workshops, the CEO of the PS case gave a 60-min presentation on their transformation efforts/projects and their background, and an on-site visit of the transformation operations was carried out. The data collection targeted two main areas: first, to investigate the experienced unexpected events (with emphasis on COVID19), the production disturbances at the SME, and understanding the effects on the SME; and second, to capture the strategic actions, practices, and flexibility of the SME in response to the shocks. The CEO of the PS case described the internal operations and working methods in the initial interview. All interviews and workshops were recorded and transcribed. The transcripts and notes were then reviewed to 1) develop the overall story of the case and 2) extract the information according to the concepts and themes identified in the literature review. Triangulation improves the accuracy and completeness of a case study, strengthening the credibility of the research findings [38, 41, 42]. At least two researchers were part of all semi-structured interviews and presentations to ensure triangulation. One researcher complemented the process with input from previous studies [11, 12, 40]. Two leading researchers analyzed interview transcripts and notes from the workshops separately and discussed them afterward. Later, those were complemented by documentation of Table 1. Data collected throughout the research process at PS case Form of data collection

Description

Presentations

3 × 30–60 min presentations on the transformation by the CEO and project lead (March 2021, March 2022, Sept 2022)

Semi-structured interviews

4 interviews (60–90 min each) CEO and head of operations (May 2021, Sept 2022, Oct 2022)

On-site observations

1 × 120 min site visit (May 2022)

Workshops

4 × 240 min joint workshops focused on the transformation for resilience, challenges, and opportunities in connection to other industries (Dec 2021, March 2022, May 2022, Sept 2022)

From Surviving to Thriving: Industry 5.0 at SMEs

795

firms and on-site observation notes. The data analysis combined different techniques to get the most out of each data set. A summative content analysis of all transcripts and presentations was conducted to build a timeline of events that identified both occurrences of high-impact events and patterns in connection to our conceptualization.

4 Findings 4.1 Introduction and Development of the Digital Working Method The SME in this study was a Swedish firm that has divided its operations into two sub-firms, one focusing on packaging solutions (PS) and one on assembly systems for innovative construction solutions (CS). PS supplies packaging, preassembly, and kitting solutions to the manufacturing industry. They deliver manufacturing packaging solutions to the automotive, electronics, pharma, and food industries. CS is a manufacturing startup developing wooden housing modules composed of standardized components. The modules are built on standardized demountable, recyclable single-story units. The firm makes every effort to keep short lead times in high-quality products, accomplished by two key characteristics. First, it is an independent firm choosing which materials and subcontractors to work with. Second, sales associates’ function as project managers, meaning that the customers’ point of contact will be the same person from the idea definition to the delivery of the finished solution. Over the past six years since the initialization of the company, CS has developed a digital working method supporting the development of manufacturing stations for people with limited cognitive and physical capabilities, allowing flexibility in educational and language backgrounds [11, 12, 40]. Based on their developed concept of manufacturing their products, people without industrial experience can quickly learn and start productive work in assembly operations. The concept includes training individuals with no previous industrial experience in performing industrial work. It involves designing inclusive workstations with digitalized and animated visualizations of standardized work instructions, less language limitations, and assembly templates with a fault-proofing process. In addition, developing a learning culture and inclusive attitudes between colleagues is emphasized. The company provides individual coaching and feedback to encourage positive energy and learning, which has been seen as an essential enabler for training [13]. The digitalized method has enabled jobs to be suitable for a broader spectrum of individuals regardless of having special cognitive and physical capabilities, for instance, hearing impairment, low proficiency in the Swedish language, and lack of previous industrial experience. Hence, the SME has utilized digitalization as a means to develop a new and more inclusive training method. In March 2020, the pandemic hit the world, with most countries having compulsory measures to slow the spread of the virus, including limited access to or closure of borders and physical workplaces. Many industries (including the SME in our study) continued to deal with material shortages following this time. The case company was not an exception and was heavily impacted by the shortages in supply (e.g., glue) and lockdowns. Here, PS tapped into the experience gained from the digitalized working method; before COVID19, it was mainly used to build housing modules. The method was successfully deployed during COVID19 for building personal protective equipment

796

Z. Z. Chavez et al.

(PPE). The production was set up at different manufacturing facilities and public spaces that have shut down due to the national restrictions to battle the pandemic. The workers used for PPE production were all volunteers or part-time workers e.g. students. The digitalized production method helped fulfill demand from the healthcare sector equivalent to thousands of PPE -protective coats and visors. 4.2 Exemplary Leverage of Human-Centricity for Flexibility in Production Table 2 summarizes the types of flexibility and the human-centric characteristics displayed in the case. For instance, the digitalized working process method enabled the increase of product variety (modification flexibility), expanded the possibilities of production location (volume flexibility), and allowed for products to be rapidly introduced (change over flexibility) as no specific training per new product is required. Table 2. Flexibility mechanisms and human-centric characteristics in PS case Case

Mechanisms implemented

Type of displayed flexibility

Tools, methods

Management practices

Systems Humancentric

Level 1 Level 2

Level 3

Modification

Volume

Changeover

The digitalized working process method allowed the introduction of new products and processes N/A

The digitalized working process method was deployed at multiple locations to produce PPE

Products are rapidly introduced, and no specific training per new product is required

The workforce can The practice of emdo productive work ploying a workforce with different cogni- right from the start, and the learning tive and physical curve is shortened capabilities expands the production capacity by being able to employ a more diverse workforce The digitalized working process deployment to multiple locations and products is seen as a way of working and developing the production system The digitalized working process integrates ergonomics consideration in the development of fixtures and workforce capabilities Removing physical and cognitive capabilities limitations is embedded in the design of the method; the digitalized working process accommodates lowering complexity and supports workforce diversity PS case already working with a broad range of personalized products in combination with reaching workforce wellbeing and skills development

From Surviving to Thriving: Industry 5.0 at SMEs

797

By designing a digitalized work process system that could employ a workforce with diverse cognitive and physical capabilities, the firm managed to increase production capacity during disruptions by employing, e.g., volunteers. At the same time, the method allows for the rapid setup of new and additional production processes if needed. Under this method, the workforce can do productive work right from the start. Hence the learning curve is shortened. In this way, PS managed to secure supply and fast response to their customers; for instance, when a shortage of a particular material happened, they would propose alternative products as a substitute. For the supply of PPE, the fast production enabled by the digitalized working process method alleviated the healthcare shortage in the region, helping battle the crisis with COVID19. It is indicated that PS displays the human-centric characteristics envisioned for level 3 (see Fig. 1), where technologies allow for optimizing human well-being. The digital method considers human capabilities, needs, and skills development while responding to customer requirements. In alignment with the I5.0 human-centric vision, the PS case company puts human needs and interests at the heart of the production processes. By design, the digitalized working process method removes physical and cognitive restrictions to the workforce, enabling diversity and inclusiveness of the personnel at PS.

5 Discussions: Flexibility and Adaptability in Human-Centric Digitalization at SME Manufacturer The findings illustrate how human-centric digitalization may play a critical role in enabling SMEs to achieve process resilience and flexibility during high-impact events. In their study, Chou et al. [24] show that partial flexibility structures, properly designed, can already accrue most of the benefits of the full flexibility system. The case study company illustrates how adopting digital technologies that leverage human-centric characteristics, such as their digitalized working process method, can enhance the production process’s ability to adapt to the environment, increase agility and responsiveness, i.e., process flexibility, and enable SMEs to overcome crises more effectively. In Fig. 1, the link between flexibility mechanisms enabled by digital technologies, i.e., digital tools, systems, and management practices, and the production system characteristics is demonstrated through progressive levels of a human-centric manufacturing system. Under this framework, firms can leverage human-centric characteristics to increase process flexibility. On the initial level 1, the workforce preferences, capabilities, and ergonomic aspects are considered for task assignment. On level 2, a human’s physical, cognitive, and psychological capacities are considered, and the capability information determines the feasibility of assigning a manufacturing task to a human [37]. At the highest level 3, technologies allow for optimizing human well-being by considering human capabilities, needs, and skills development while also responding to customer requirements. The different mechanisms under a digital strategy can support the development of different types of flexibility by integrating human-centric characteristics. For instance, for the rapid introduction of new products, a company may need to rapidly train the workforce in new skills and mobilize the production location. Often these activities

798

Z. Z. Chavez et al.

Fig. 1. A conceptual framework for flexibility in a Human-centric manufacturing system inspired by Lu et al. (2022)

convey high stress to workers. In a human-centric manufacturing system, workforce needs are taken care of from the outset, and actions hindering human wellness would ideally be avoided. In our case study, it is important to highlight that digitalization aimed at maximum support and minimum complexity, as suggested in the previous literature [18]; the aim in the case firm was increasing the employability of the workers and designing more inclusive working environments. Further, in our case, the level of digitalization and visual support was high, but the level of automation was low. In this respect, small changes in work procedures at the case company helped accommodate people with functional variations, such as physical or hearing impairments. For example, when making protective clothing, adjustments were made to the worktable height to accommodate a volunteer who used a wheelchair, and materials were delivered to the work area to avoid contamination of the protective clothing. Another example showed how communication was adapted for a person with hearing impairment during the assembly of protective clothing, i.e., physical and cognitive variations when making changes in the workplace.

6 Conclusions and Contributions This study adds to the existing research on production flexibility, mainly focusing on the context of SMEs and digitalization. It advances our knowledge of how human-centered digitalization can increase the adaptability and flexibility of SME production processes,

From Surviving to Thriving: Industry 5.0 at SMEs

799

leading to resilient systems both during high-impact events and normal company operations. It advises businesses on improving the flow of their production processes and dealing more skillfully with upcoming difficulties. The study utilizes a conceptual framework for comprehending the relationship between flexibility and adaptability mechanisms made possible by digital technologies and the production system features through progressively integrating human-centric characteristics in a digitalized manufacturing system. The case study illustrates a human-centric strategy for designing a workspace where people’s capabilities and interests are placed at the center of the manufacturing processes, consistent with Industry 5.0’s goal. Our findings indicate that a digitalized working method considering human capabilities and needs can enable a more flexible and inclusive production process. It can help expand the production capacity by employing a more diverse workforce and enable a rapid setup of new and additional production processes. Such a method can be leveraged to quickly respond to supply chain disruptions caused by events like the pandemic, as demonstrated by the studied SME. Additionally, sales associates functioning as project managers can help provide better customer service, resulting in high-quality products with short lead times. Developing a learning culture and attitudes between colleagues can promote positive energy and skill development, resulting in better production processes and products. The findings of this study have implications for managers and practitioners interested in driving or supporting the transition of SMEs to Human-centric, resilient, and sustainable businesses. Future work of interest is to focus on identifying other examples of I5.0 and its implication on the resilience of SMEs. There is also a need to explore employee engagement in a given environment like the one in the presented case study and investigate concepts such as job autonomy, job demands, and social support to promote job satisfaction and organizational commitment. Acknowledgement. This work was partially supported by Production2030 and Sweden’s Government Agency for Innovation VINNOVA Programme, RESPIRE project [grant number 2021– 03685], and Marie Skłodowska-Curie Actions (MSCA), SME 5.0 project [grant agreement ID 101086487]. The authors thankfully acknowledge the support of the case company and project partners.

References 1. Xu, X., Lu, Y., Vogel-Heuser, B., Wang, L.: Industry 4.0 and Industry 5.0—inception, conception and perception. J. Manufact. Syst. 61(September), 530–535 (2021). https://doi.org/ 10.1016/j.jmsy.2021.10.006 2. H. D. Mohammadian and M. Castro, “The Development of a Readiness Assessment Framework for Tomorrow ’ s SMEs / SME 5 . 0 for Adopting the Educational Components of future of I4.0,” in 2022 IEEE Global Engineering Education Conference (EDUCON), 2022, pp. 1699–1708 3. Skare, M., de las Mercedes de Obesso, M., Ribeiro-Navarrete, S.: Digital transformation and European small and medium enterprises (SMEs): a comparative study using digital economy and society index data. Int. J. Inf. Manage. 68(October 2022), 102594 (2023). https://doi.org/ 10.1016/j.ijinfomgt.2022.102594

800

Z. Z. Chavez et al.

4. European Commission: Industry 5.0 A Transformative Vision for Europe. ESIR Policy Brief No.3, no. 3, p. 30 (2021). https://doi.org/10.2777/17322 5. Huang, S., Wang, B., Li, X., Zheng, P., Mourtzis, D., Wang, L.: Industry 5.0 and Society 5.0—comparison, complementation and co-evolution. J. Manufact. Syst. 64(July), 424–428 (2022). https://doi.org/10.1016/j.jmsy.2022.07.010 6. European Commission: What is an SME? (2023). https://single-market-economy.ec.europa. eu/smes/sme-definition_en 7. Müller, J.M., Buliga, O., Voigt, K.I.: Fortune favors the prepared: how SMEs approach business model innovations in Industry 4.0. Technol. Forecast. Soc. Change 132(December 2017), 2–17 (2018). https://doi.org/10.1016/j.techfore.2017.12.019 8. Mittal, S., Khan, M.A., Romero, D., Wuest, T.: A critical review of smart manufacturing & Industry 4.0 maturity models : implications for small and medium-sized enterprises (SMEs). J. Manuf. Syst. (November 2018). https://doi.org/10.1016/j.jmsy.2018.10.005 9. Mourtzis, D., Angelopoulos, J., Panopoulos, N.: A literature review of the challenges and opportunities of the transition from Industry 4.0 to Society 5.0. Energies 15(17) (2022). https://doi.org/10.3390/en15176276 10. Leng, J., et al.: Industry 5. 0: prospect and retrospect. J. Manufact. Syst. 65(September), 279–295 (2022). https://doi.org/10.1016/j.jmsy.2022.09.017 11. Chen, X., Kurdve, M., Johansson, B., Despeisse, M.: Enabling the twin transitions: Digital technologies support environmental sustainability through lean principles. Sustain. Prod. Consum. 38, 13–27 (2023). https://doi.org/10.1016/j.spc.2023.03.020 12. Kurdve, M., Hildenbrand, J., Jönsson, C.: Design for green lean building module production - case study. Procedia Manufact. 25, 594–601 (2018). https://doi.org/10.1016/j.promfg.2018. 06.096 13. Mattsson, S., Kurdve, M., Almström, P., Skagert, K.: Synthesis of universal workplace design in assembly-a case study. Adv. Transdiscipl. Eng. SPS 21, 184–196 (2022). https://doi.org/ 10.3233/ATDE220138 14. Polyviou, M., Croxton, K.L., Knemeyer, A.M.: Resilience of medium-sized firms to supply chain disruptions: the role of internal social capital. Int. J. Oper. Prod. Manage. 40(1), 68–91 (2020). https://doi.org/10.1108/IJOPM-09-2017-0530 15. Ambulkar, S., Blackhurst, J., Grawe, S.: Firm’s resilience to supply chain disruptions: scale development and empirical examination. J. Oper. Manage. 33–34, 111–122 (2015). https:// doi.org/10.1016/j.jom.2014.11.002 16. d’Amboise, G., Muldowney, M.: Management theory for small business: attempts and requirements. Acad. Manage. Rev. 13(2), 226–240 (1988). https://doi.org/10.5465/amr.1988.430 6873 17. Ramaswami, S.N., Srivastava, R.K., Bhargava, M.: Market-based capabilities and financial performance of firms: insights into marketing’s contribution to firm value. J. Acad. Mark. Sci. 37(2), 97–116 (2009). https://doi.org/10.1007/s11747-008-0120-2 18. Chowdhury, P., Lau, K.H., Pittayachawan, S.: Operational supply risk mitigation of SME and its impact on operational performance: a social capital perspective. Int. J. Oper. Prod. Manage. 39(4), 478–502 (2019). https://doi.org/10.1108/IJOPM-09-2017-0561/FULL/HTML 19. Harel, R., Kaufmann, D.: Financing innovative SMEs of traditional sectors: the supply side. EuroMed J. Bus. 11(1), 84–100 (2016). https://doi.org/10.1108/EMJB-02-2015-0007/FULL/ HTML 20. Prasad, S., Baltov, M., Neelakanteswara Rao, A., Lanka, K.: Interdependency analysis of lean manufacturing practices in case of Bulgarian SMEs: interpretive structural modelling and interpretive ranking modelling approach. Int. J. Lean Six Sigma 12(3), 503–535 (2020). https://doi.org/10.1108/IJLSS-09-2019-0100/FULL/HTML 21. Slack, N.: The flexibility of manufacturing systems. Int. J. Oper. Prod. Manage 35–45 (1987)

From Surviving to Thriving: Industry 5.0 at SMEs

801

22. Asadi, N., Jackson, M., Fundin, A.: Drivers of complexity in a flexible assembly system- a case study. Procedia CIRP 41(March 2016), 189–194 (2016). https://doi.org/10.1016/j.pro cir.2015.12.082 23. Molina, L.M., Verdu, A.J., Llore, F.J.: Flexibility of manufacturing systems, strategic change and performance. Int. J. Prod. Econ. 98, 273–289 (2005). https://doi.org/10.1016/j.ijpe.2004. 05.011 24. Chou, M.C., Chua, G.A., Teo, C.-P., Zheng, H.: Design for process flexibility: efficiency of the long chain and sparse structure. Oper. Res. 58(1), 43–58 (2010). https://doi.org/10.1287/ opre.1080.0664 25. Simchi-levi, D., Wei, Y.: Operations research jnfflflnfü worst-case analysis of process flexibility des. Oper. Res. 63(1), 166–185 (2015) 26. Ivanov, D., Das, A., Choi, T.M.: New flexibility drivers for manufacturing, supply chain and service operations. Int. J. Prod. Res. 56(10), 3359–3368 (2018). https://doi.org/10.1080/002 07543.2018.1457813 27. Schmenner, R.W., Tatikonda, M.V.: Manufacturing process flexibility revisited. Int. J. Oper. Prod. Manag. 25(12), 1183–1189 (2005). https://doi.org/10.1108/01443570510633585 28. Thakkar, J., Kanda, A., Deshmukh, S.G.: Interpretive structural modeling (ISM) of IT-enablers for Indian manufacturing SMEs. Inf. Manage. Comput. Secur. 16(2), 113–136 (2008). https:// doi.org/10.1108/09685220810879609/FULL/HTML 29. Chopra, S., Sodhi, M.S.: Managing risk to avoid supply-chain breakdown. MIT Sloan Manage. Rev. 46(2004) (2004) 30. Talluri, S., Kull, T.J., Yildiz, H., Yoon, J.: Assessing the efficiency of risk mitigation strategies in supply chains. J. Bus. Logist. 34(4), 253–269 (2013). https://doi.org/10.1111/jbl.12025 31. Ivanov, D.: The Industry 5 . 0 framework : viability-based integration of the resilience, sustainability , and human-centricity perspectives. Int. J. Prod. (2023). https://doi.org/10.1080/ 00207543.2022.2118892 32. Ellegaard, C., Normann, U., Lidegaard, N.: Intuitive global sourcing – a study of supplier selection decisions by apparel SMEs. Int. J. Oper. Prod. Manage. 42(2), 151–181 (2022). https://doi.org/10.1108/IJOPM-03-2021-0205/FULL/HTML 33. Ellegaard, C.: The purchasing orientation of small company owners. J. Bus. Ind. Mark. (2009) 34. European Commission: Industry 5.0 - towards a sustainable, human-centric and resilient European industry (2021). https://doi.org/10.2777/308407 35. European Commission: Enabling Technologies for Industry 5.0 (2020). https://doi.org/10. 2777/082634 36. Ardolino, M., Bacchetti, A., Dolgui, A., Franchini, G., Ivanov, D., Nair, A.: The Impacts of digital technologies on coping with the COVID-19 pandemic in the manufacturing industry: a systematic literature review. Int. J. Prod. Res. (2022). https://doi.org/10.1080/00207543. 2022.2127960 37. Lu, Y., et al.: Outlook on human-centric manufacturing towards Industry 5.0. J. Manufact. Syst. 62(February), 612–627 (2022). https://doi.org/10.1016/j.jmsy.2022.02.001 38. Yin, R.K.: Case study research: Design and methods, 4th edn. SAGE, London (2009) 39. Kristina Säfsten, M.G.: Research methodology - for engineers and other problem-solvers, 1st edn. Studentlitteratur AB (2020) 40. Kurdve, M., De Goey, H.: Can social sustainability values be incorporated in a product service system for temporary public building modules? Procedia CIRP 64, 193–198 (2017). https:// doi.org/10.1016/j.procir.2017.03.039

802

Z. Z. Chavez et al.

41. Stuart, I., McCutcheon, D., Handfield, R., McLachlin, R., Samson, D.: Effective case research in operations management: a process perspective. J. Oper. Manag. 20(5), 419–433 (2002). https://doi.org/10.1016/S0272-6963(02)00022-0 42. Morgan, S.J., Pullon, S.R.H., MacDonald, L.M., McKinlay, E.M., Gray, B.V.: Case study observational research: a framework for conducting case study research where observation data are the focus. Qual. Health Res. 27(7), 1060–1068 (2017). https://doi.org/10.1177/104 9732316649160

Challenges in Designing and Implementing Augmented Reality-Based Decision Support Systems for Intralogistics: A Multiple Case Study Moritz Quandt1,2(B) , Hendrik Stern2 , Markus Kreutz1 , and Michael Freitag1,2 1

BIBA – Bremer Institut f¨ ur Produktion und Logistik at the University of Bremen, Bremen, Germany [email protected] 2 Faculty of Production Engineering, University of Bremen, Bremen, Germany Abstract. Numerous Augmented Reality-based prototypes and proof of concepts of assistance systems have been developed for industrial application scenarios that have not yet reached operational use. Alongside remaining technical challenges, the human-centered development of the systems and the associated improvement of the human-machine interfaces are seen as critical challenges for successful development and introduction. Therefore, this contribution addresses the challenges of designing and implementing augmented reality-based decision support systems for intralogistics work processes focusing on system users. The authors compared the qualitative results of a multiple case study conducted in ten intralogistics companies with challenges for industrial AR systems based on a literature review and the resulting focus areas regarding a human-centered design process. The view of the experts from the companies confirms the literature-based challenges on many points. However, some aspects pose particular challenges in intralogistics, such as the high-efficiency requirements for work processes. Overall, the results underscore the importance of involving operators in system development and highlight factors like usability, workload, and user acceptance, which can be evaluated through user research methods. Keywords: Industrial Augmented Reality · Artificial Intelligence · Cognitve Assistance Systems · Human-centered design · Multiple Case Study · Intralogistics

1

Human-Centered Augmented Reality in Intralogistics

The digitalization of industrial work environments, including production environments under the term Industry 4.0, brings together information, technologies, and people in smart work environments [1]. The interaction of increasingly c IFIP International Federation for Information Processing 2023  Published by Springer Nature Switzerland AG 2023 E. Alfnes et al. (Eds.): APMS 2023, IFIP AICT 689, pp. 803–817, 2023. https://doi.org/10.1007/978-3-031-43662-8_57

804

M. Quandt et al.

autonomous systems with human intelligence in industrial work environments from a human-centric perspective is addressed by the term Industry 5.0 [2]. For operators, the work is increasingly shifting to knowledge-based activities, decentralized decision-making, and collaborative tasks, which translates into a higher cognitive load for operators [3]. Assistance technologies are coming into use to support people in increasingly complex and dynamic industrial work environments [4]. One of the assistance technologies suitable for operator decision support in complex work environments is augmented reality (AR) [5]. For the visual processing of information on mobile devices, AR promises natural interaction between humans and technical systems and context-based information provision [6]. However, AR-based assistance systems have so far predominantly corresponded to perceptual assistance systems. Artificial intelligence (AI) techniques can provide operators with decision support for different tasks and open up new ways of collaboration with the operators [7]. In recent years, the possibilities of developing AR solutions for mobile hardware, such as smartphones, tablets, or smart glasses, have led to a strong interest in AR technology and the development of numerous mobile AR applications. This development trend can also be observed in the development of industrial AR applications [8]. In the industrial application context, a rapidly increasing number of industrial AR developments and numerous potential applications are contrasted by the fact that AR solutions have not yet been widely deployed in industrial practice [9–11]. Due to the numerous possibilities of the technology, such as the visualization of virtual information, interaction in three-dimensional space, or the design of user interfaces, no standards for the design of human-machine interaction of industrial AR systems have yet been established [12]. Therefore, the human-centered view of the technology is seen as an essential research direction [13]. The concrete structure of the humancentered design process and the associated investigation of factors contributing to the successful development and introduction of AR-based assistance systems in industrial application contexts from the user perspective have not yet been comprehensively investigated [14]. Therefore, this paper examines the current challenges to the industrial deployment of AR-based decision assistance systems. The article relates this state-of-the-art research to the empirical results from a multiple case study. Based on the empirical results, we discuss the implications for a human-centered design process and derive factors that need to be considered for the humancentered development of AR-based decision assistance systems from the perspective of both science and practice. Intralogistics was considered as the industry in this case, as there are many potential applications for the use of AR-based systems for process support, especially in personnel-intensive areas such as picking [15–17]. In the multiple case study, we interviewed company representatives about previous experience with digital assistance systems for process support, the requirements for AR-based decision assistance systems, and the challenges and opportunities of such systems. In addition, concrete implementation potentials were jointly identified as to which processes are suitable for AI-based AR

AR-Based Decision Support Systems for Intralogistics

805

support. These were identified in incoming goods inspection, picking and warehouse management, and packaging. The paper is divided into the presentation of the research design in Sect. 2, the analysis of open challenges for industrial AR assistance systems from selected scientific review papers in Sect. 3, the description of the empirical results from the case study in Sect. 4, the discussion of the challenges from both science and practice and the resulting focus areas concerning a human-centered design process in Sect. 5.

2

Research Design - Literature Analysis and Multiple Case Study

As a first step, we analyzed the current challenges to industrial AR assistance systems based on 23 review papers in a scoping review following Arksey [18]. Therefore, we conducted a search Scopus database for review papers on the topic of industrial AR (76 search results). We selected the review papers according to the subject (content), the impact, and the inclusion of challenges of industrial AR. In each case, the subject is AR assistance systems for the industrial application context. Of the search results from Scopus, nine review papers met the defined criteria. We extended these results with a search in google scholar, from which an additional twelve relevant review papers were identified. By conducting forward and backward searches, we manually identified an additional two relevant review papers that also met our search criteria. Kim [19], Dey [20] and De Souza Cardoso [9] address industrial AR use independent from the application field. Other authors analyze the state of research for specific application fields: Manufacturing [6,8,21–26], maintenance [11,27], shipbuilding [28], automotive [29], engineering services [30], construction [31–35], and logistics [17,36] To assess the article’s impact on the scientific community, we determined the average number of citations using the Average Citation Score (ACC), according to [20]. The ACC is calculated by dividing the total number of citations by the time since publication in years. For this purpose, we used the current citation counts according to Google Scholar. Dey [20] set a threshold of 1.5 for the ACC in their review to ensure a moderate impact of the article on the scientific community. The reviews considered here have an ACC between 12 and 107. The consideration was limited to papers from 2012 onwards, as there have been significant technical advancements from this period ahead, especially in mobile AR. For the case study, we selected ten companies from the intralogistics environment. The companies interviewed included four manufacturing companies: mechanical engineering (1), medical technology (2), and construction (1). In addition, we interviewed two logistics service providers, each of which operates its own logistics infrastructure. Furthermore, we interviewed experts from AR and AI software development (2), one IT security company (1), and one logistics IT consulting company (1). We surveyed the companies’ previous experiences

806

M. Quandt et al.

with developing and introducing cognitive assistance systems for process support, the requirements for such assistance systems, and the existing challenges for AI-based AR support with regard to a human-centered design process. To ensure a systematic qualitative approach, we created a case study protocol, preparing the objectives and background, the data collection procedure, the questionnaire, and the case study report, according to Yin [37]. Based on a catalog of questions, we interviewed 15 experts in semi-structured interviews. In addition to the expert interviews, we conducted process inspections at five companies to provide a detailed insight into the companies’ intralogistics processes. All expert interviews were recorded and transcribed. The transcripts were supplemented by the records of the process inspections. A total of over 120 pages of text material were analyzed using qualitative content analysis according to [38], being an established evaluation method for expert interviews [39]. The qualitative content analysis derives categories from the texts and quantifies them according to the number of mentions in the text material [40]. To ensure scientific rigor, a systematic approach to data analysis is essential [41]. Therefore, we opted for inductive category formation, according to [38]. To ensure a systematic approach, we applied the web application QCAmap [42]. Initially, the first author of this paper conducted the text analysis for all three research questions under study and derived categories. These categories were aggregated in several iterations and assigned to each main category. The author team reviewed text coding and categorization of the text and jointly developed a summary in main categories across all three research questions.

3

Background - Current Challenges for Industrial AR Assistance Systems from a Scientific Perspective

We identified a total of 34 challenges from the analyzed 23 review papers. For structuring reasons, we first assigned the identified challenges to the categories user, technology, and interaction. Figure 4 shows this categorization, including the empirical results of the case study. Linked to interaction are challenges in designing intuitive user interfaces using different forms of interaction [6,17,21,23–25,27–33,36]. Other challenges are for example the development of guidelines for AR user interfaces [22] and the standardization of interaction patterns for different forms of interaction [6]. Regarding technology, we could identify numerous challenges that we divided into hardware- and software-related challenges. Hardware-related challenges include the lack of ergonomics of AR hardware [6,9,11,17,19,22,24,27,28,30,32– 34,36], or the so-far insufficient Field of View of head mounted displays (HMD) [6,8,9,11,19,22,28–30,36]. Furthermore, in the industrial work context the hardware must comply with occupational safety regulations [6,27,29,32,36]. In the software domain, tracking virtual objects in the real environment is one of the key challenges over the entire review period [6,8,9,11, 19,21,23,24,26– 28,30,32,33,35]. Furthermore, the challenges mentioned include improving rendering, i.e., embedding virtual objects into the real scene [6,9,11,19,24,26,28] or providing contextual data in real-time [22,23,25–27,31–33,36].

AR-Based Decision Support Systems for Intralogistics

807

Concerning users, we could identify challenges in increasing usability and user experience [6,11,20,23,25–27,32,33,35,36]. Furthermore, gaining knowledge through user-centered evaluation is an essential prerequisite for the practicality of industrial AR solutions [8,19,20,23,33]. Other reviews name the physical strains, e.g., dizziness, nausea, or headaches, and cognitive load of the users in using an AR assistance system as a challenge [9,11,21,23–26,28,32,33,36]. Boboc [29], Egger [6], and Rejeb [36], name user acceptance of industrial AR systems as a current challenge. The research needs we identified in analyzing the review papers concerning the users were the basis and motivation to conduct the case study. The results of the literature analysis served as a basis for the development of the case study protocol and for the classification of the case study results.

4

Empirical Results - A Practical Perspective from Intralogistics

For the case study, we interviewed the experts using an interview guide. Of the ten companies surveyed, seven each had prior experience with AR or AIbased approaches. Based on the inductive categorization, according to Mayring [38], four main categories could be derived in which we combined a total of 48 categories from the text analysis: user (17), system/technology (18), process (7), and system implementation (6). The findings for each category are described below. User. Concerning users, the experts emphasized the added value of an assistance system for processing the work task (see Fig. 1). Most experts stated that the introduced assistance system or an assistance function must represent added user value (13 mentions). Further, achieving user acceptance is a high priority for the companies (14 mentions). In this context, the companies mentioned the consideration of users’ individual motivation for using an assistance system and a higher general acceptance of new technologies for future introduction as particularly relevant topics. Other issues often mentioned by the experts were ensuring usability (7 mentions), mainly by involving operators in technology selection and system development. In addition, the experts named the avoidance of cognitive and physical burdens as essential for AR-based decision support systems (7 mentions). For example, a mobile assistance system must not restrict operators’ freedom of movement or cause dizziness or headaches during prolonged use. All identified user-related categories concerning the number of their occurrence and the context of the statement (experience, requirement, challenge, or opportunity) are depicted in Fig. 1. System/Technology. For most companies interviewed, the effort and costs were the most critical challenge associated with system development (10 mentions). The experts focus on ensuring a short payback period and assume that

808

M. Quandt et al.

Fig. 1. Identified categories and number of mentions from the case study concerning the users

developing an AR-based decision support system might be related with a higher development effort than other solutions. In connection with the introduction of AR solutions, many experts mentioned the hardware’s practicality (8 mentions). Particularly concerning the use of data glasses, the experts see a need for further development due to the current technological maturity of the hardware. For using AI-based methods for decision support in the work process, the companies see the associated effort with data preparation and maintenance as one of the most significant challenges (8). As a basis for AI-supported data evaluation, companies must digitize the work process, connect data streams, and record highly reliable data. As per previous experience, the companies stated about assistance systems and the use of new technologies that the systems must achieve a high level of reliability to be considered for practical use (6 mentions). In addition, the companies see further system-related challenges concerning data preparation for visualizing information on AR systems (5). Depending on the data basis, this involves a great deal of effort. For all other empirical results on system/technology-related categories, see Fig. 2.

Fig. 2. Identified categories and number of mentions from the case study concerning system/technology

AR-Based Decision Support Systems for Intralogistics

809

Process. Increasing process efficiency is essential for intralogistics processes (14 mentions). The experts are concerned with avoiding errors, increasing process quality, or increasing speed with system use. In this context, the high demands on the system’s speed are seen as a central challenge since assistance must not slow down the operators by data input or related tasks. Regarding the work process, the experts also see the need to align the design of assistance functions with the work process and to increase process efficiency through the use of the assistance system, e.g., that the use of the system reduces the workload of the operators or AI-based assistance functions provide targeted support (8 mentions). Implementation. The most important aspect mentioned from the experts concerning implementation is a gradual introduction of hardware and functionalities (6 mentions). In this context, the experts see positive effects on the acceptance of a cognitive assistance system and the requirement that assistance functions be designed to be expandable to ensure long-term usability. Analogous to the costs of developing an assistance system, companies consider it a requirement to keep the introduction costs as low as possible (3 mentions). The introduction of AI-based assistance functions, in particular, poses significant challenges in integrating the successfully tested processes into the companies’ existing IT structures. All results of the main categories process and implementation are summarized in Fig. 3.

Fig. 3. Identified categories and number of mentions from the case study concerning processes and implementation

5

Discussion

In this section, we match the challenges for industrial AR assistance systems with the empirical results from the case study, which relate to AI-based process support using AR technology for intralogistics processes. Figure 4 divides the categories from the literature review into user, technology, and interaction. From the case study, we added the main categories of process and implementation. To

810

M. Quandt et al.

enable a comparison between the scientific and practical views of the experts, we have expressed the individual categories in ideograms (Harvey Balls) according to the number of mentions in the literature or case study. The different scaling is explained by the number of review papers examined (23) and the aggregated mentions of the categories in the case studies. User. Concerning users, the empirical results confirm the consideration of userrelated factors, usability/user experience (UX), physical and cognitive loads, and user acceptance. Additional factors mentioned in both the review papers and the case studies are data security or the handling of personal data, and user participation in system development. The review papers additionally address the evaluation of the systems, which is expressed, e.g., in the requirement to evaluate with real users in the actual application environment [20] or the adaptation of user-oriented evaluation methods to the requirements of AR systems [19]. The companies did not explicitly name the involvement of users in the evaluation of AR assistance systems. In our opinion this is addressed by the objective to develop usable solutions and to involve users in system development. Most surveyed companies emphasized the added value of an assistance system for the users in the respective work session. This demand for efficient support of the work process is reflected, e.g., in the usability definition according to ISO 9241-11 (“extent to which specified users can use a system, product or service to achieve specified goals with effectiveness, efficiency and satisfaction in a specified context of use” [43, p. 9]) or the construct of usefulness from Davis’ TAM model (“The degree to which a person believes that using a particular system would enhance his or her job performance” [44, p. 320]). Another new factor, which was mentioned in particular by the companies from the software development sector, is the management of expectations concerning AR technology. From our point of view, the expectations towards the technical solution can be met by the users’ participation and the decision makers’ early involvement. Moreover, the companies expressed the need for different experience levels, which should offer the possibility of adjusting the level of support to operators with different qualification profiles. Technology. For companies and academia, the better adaptation of AR hardware to the industrial work environment is a crucial challenge, including the need to achieve better ergonomics of AR hardware or better protection against environmental influences. The effort associated with the development, is estimated to be an equally important factor. The companies did not mention the detailed technical challenges, e.g., the field of view and the quality of the displays of the HMD, or the improvement of tracking and rendering on the software side. In our opinion, this might relate to the lack of concretization of the AR application or the self-perception as a user and not as a developer of an AR solution. The concrete technical challenges should be considered in the context of the individual technical implementation and are addressed depending on the type of AR application. The field of view, for example, is only relevant when using an HMD;

Fig. 4. Comparing the identified challenges for industrial AR-based assistance systems from literature and the intralogistics case study

AR-Based Decision Support Systems for Intralogistics 811

812

M. Quandt et al.

the required computing power of mobile hardware also depends on the specific application. The availability and reliability of data are the foundation for the use of AI-based learning methods. In addition, data management is related to preparing data for visualization in AR systems that represented an essential part of the experience of AR-related developments of the companies. Data preparation for AR visualization can be a reason for deciding against an AR solution due to the high effort involved. Interaction. If a smartphone or tablet is selected for process support for procedural or organizational reasons, usually more familiar interaction forms, such as touch- or voice-based interaction, are used. When selecting an HMD, hand gesture-based interaction forms and accurate three-dimensional representations can be used. This confronts most users with new interaction patterns. The ARspecific possibilities of interaction in three-dimensional space or direct interaction with virtual objects pose new challenges for science. For companies, we believe that interaction issues closely link to usability and the practicality of the AR hardware used. Process. The review papers and the experts mentioned process-based support as an important challenge. The case study and scientific publications emphasized that an AR-based assistance system must be oriented to the work process, including hardware selection and assistance functions. In addition, many logistics processes considered place high demands on process efficiency, which is expressed in, among other things, low error rates and a high process speed. The specific requirements of the work processes in terms of speed and quality must be considered when designing assistance functions. For example, the experts expect to see an improvement in inspection results for incoming goods through image-based component inspection. Implementation. Some companies see a big difference between the development effort of an AR- or AI-based assistance function and the actual integration into operational processes. For example, from an executable prototype of a databased AI solution for evaluating operational enterprise data, there may still be a long way to go to a system integrated into enterprise IT. Another point for companies is the gradual introduction, both from an investment point of view and to gradually confront operators with new hardware and new software functions and thus not overstrain them. From the companies’ point of view, this increases acceptance of introducing new technologies. Limitations. Among the limitations of this paper, we count for the literature review that we limited ourselves to 23 review papers. Nevertheless, we believe that we were able to identify the fundamental challenges to industrial AR assistance systems, as the practical view from the case study confirmed many of the

AR-Based Decision Support Systems for Intralogistics

813

challenges mentioned there. The execution of the case study is limited to German companies. We assume that the results are more generalizable by interviewing other companies from different markets. The expert interviews were mainly conducted at the management level. In some cases, we could also interview operators during the process inspections. When it comes to the requirements for an AI-based AR assistance function, the requirements of operators should be considered. Despite the methodologically guided evaluation of the case study, we assume an inevitable subjectivity of the results and the category formation. We have tried to counteract this through discussions, joint revisions, and feedback within the author team. Considering the limitations mentioned above, we believe that the results from science and practice allow us to draw conclusions about which factors need special attention when designing, developing, and implementing AR decision support systems for intralogistics work processes.

6

Conclusion and Outlook

In this contribution, we investigated the challenges of designing, developing, and implementing AR-based decision support systems for intralogistics work processes from scientific and practical perspectives. We conducted a multiple case study in ten intralogistics companies based on a scoping review. We compared the empirical results of the case study with the challenges from science and discussed them. For the companies, costs and efforts related to system development, implementation, and data preparation are essential for AI-based AR assistance functions for process support. Logistics processes require high efficiency, affecting the type of assistance functions and the interaction between humans and the system. Companies focus on intuitive systems that guide operators through the work process without overwhelming them with numerous setting options and new interaction patterns. Nevertheless, there should be customization options for the assistance functions to suit users’ preferences and skill levels, which will determine the level of support the system provides. In terms of AR hardware, the focus for companies is on practicality. The numerous technical challenges from science, such as the field of view of HMD or improvement of the tracking of virtual objects, must be addressed in development depending on the type of assistance function. The increased involvement of users in system development, which is also intensively demanded by the scientific community, was present throughout the case study. The empirical results confirm the literature that users need to be more involved in the development process. For a human-centered design process of ARbased decision assistance systems, the work process and the work environment are crucial for selecting assistance functions and the hardware used. Compliance with occupational health and safety and data security guidelines are fundamental prerequisites for implementing process-related assistance functions. Involving users ensures their early contact with new technology. It enables drawing conclusions about the design of an assistance function and adapting the solution to the users’ requirements in the respective application context. In this context, the

814

M. Quandt et al.

needs of different user groups must also be taken into account, e.g. with regard to age or experience with new technologies. This was not explicitly addressed by the experts in the case study, but should be taken into account in the design and development of AR assistance functions. In addition to collecting quantitative data, such as throughput times or error rates, user-related factors can be evaluated, e.g., usability/UX, workload, or user acceptance. Questionnaire-based methods can be used for user-based evaluation, such as System Usability Scale (SUS) for usability assessment [45], the User Experience Questionnaire (UEQ) for UX assessment [46], NASA TLX for workload assessment [47], or the Technology Acceptance Model for user acceptance prognosis [44]. This contrasts with the efficiency requirements and high effort of user-based evaluation. Therefore, by prioritizing the factors for the specific use case of intralogistics, we strive to develop a requirements-based evaluation method that enables the assessment of the most critical factors while considering the limited resources. Therefore, the concrete method selection and adaptation for the human-centered design process will be the subject of our further research work. In this context, we aim to adapt user acceptance models for industrial AR assistance systems, which have not yet been investigated in detail [48]. Therefore, through further cooperation with experts from the application field, we aim to narrow down the factors to be investigated and adapt established evaluation methods to the needs. Acknowledgements. The authors would like to thank the German Federal Ministry for Economic Affairs and Climate Action (BMWK) and the German Federation of Industrial Research Associations (AiF) as part of the programme for promoting industrial cooperative research (IGF) for their support within the project “AR Improve Development of a guideline for the demand-oriented use of AR-based assistance systems in intralogistics” (grant number 22458 N) and the University of Bremen for funding the independent postdoc project “Human Factors in Hybrid Cyber-Physical Production Systems” by the University of Bremen’s Central Research Development Fund.

References 1. Romero, D., Stahre, J.: Towards the resilient operator 5.0: the future of work in smart resilient manufacturing systems. Procedia CIRP 104, 1089–1094 (2021). https://doi.org/10.1016/j.procir.2021.11.183 2. Nahavandi, S.: Industry 5.0 - a human-centric solution. Sustainability 11, 4371 (2019). https://doi.org/10.3390/su11164371 3. Peruzzini, M., Grandi, F., Pellicciari, M.: Exploring the potential of Operator 40 interface and monitoring. Comput. Ind. Eng. 139, 105600 (2020). https://doi.org/ 10.1016/j.cie.2018.12.047 4. Mark, B.G., Rauch, E., Matt, D.T.: Worker assistance systems in manufacturing: a review of the state of the art and future directions. J. Manuf. Syst. 59, 228–250 (2021). https://doi.org/10.1016/j.jmsy.2021.02.017 5. Danielsson, O., Holm, M., Syberfeldt, A.: Augmented reality smart glasses in industrial assembly: current status and future challenges. J. Ind. Inf. Integrat. 20, 100175 (2020). https://doi.org/10.1016/j.jii.2020.100175

AR-Based Decision Support Systems for Intralogistics

815

6. Egger, J., Masood, T.: Augmented reality in support of intelligent manufacturing - a systematic literature review. Comput. Ind. Eng. 140, 106195 (2020). https:// doi.org/10.1016/j.cie.2019.106195 7. Stern, H.; Freitag, M.: Human-centered design of hybrid cyber-physical production systems - use of human autonomy teaming as a future way of working. In: Schriftenreihe der Wissenschaftlichen Gesellschaft f¨ ur Arbeits- und Betriebsorganisation (WGAB) e. V. (2022), pp. 97–113. https://doi.org/10.30844/WGAB 2022 6 8. Bottani, E., Vignali, G.: Augmented reality technology in the manufacturing industry: a review of the last decade. IISE Trans. 51, 284–310 (2019). https://doi.org/ 10.1080/24725854.2018.1493244 9. Souza Cardoso, L.F., de Queiroz Mariano, F.C.M., Zorzal, E.R.: A survey of industrial augmented reality. Comput. Ind. Eng. 139, 106159 (2020). https://doi.org/ 10.1016/j.cie.2019.106159 10. Martinetti, A., Marques, H.C., Singh, S., van Dongen, L.: Reflections on the limited pervasiveness of augmented reality in industrial sectors. Appl. Sci. 9, 3382 (2019). https://doi.org/10.3390/app9163382 11. Palmarini, R., Erkoyuncu, J.A., Roy, R., Torabmostaedi, H.: A systematic review of augmented reality applications in maintenance. Rob. Comput.-Integrat. Manuf. 49, 215–228 (2018). https://doi.org/10.1016/j.rcim.2017.06.002 12. Gattullo, M., Evangelista, A., Uva, A.E., Fiorentino, M., Gabbard, J.: What, how, and why are visual assets used in industrial augmented reality? a systematic review and classification in maintenance, assembly, and training (from 1997 to 2019). IEEE Trans. Vis. Comput. Graph. 28, 1443–1456 (2020). https://doi.org/10.1109/ TVCG.2020.3014614 13. Masood, T., Egger, J.: Augmented reality in support of Industry 4.0 - implementation challenges and success factors. Rob. Comput.-Integrat. Manuf. 58, 181–195 (2019). https://doi.org/10.1016/j.rcim.2019.02.003 14. Quandt, M., Stern, H., Zeitler, W., Freitag, M.: Human-centered design of cognitive assistance systems for industrial work. Procedia CIRP 107, 233–238 (2022). https://doi.org/10.1016/j.procir.2022.04.039 15. Wang, W., Wang, F., Song, W., Su, S.: Application of augmented reality (AR) technologies in inhouse logistics. In: E3S Web of Conferences, vol. 145, p. 02018 (2020). https://doi.org/10.1051/e3sconf/202014502018 16. Kim, S., Nussbaum, M.A., Gabbard, J.L.: Influences of augmented reality headworn display type and user interface design on performance and usability in simulated warehouse order picking. Appl. Ergon. 74, 186–193 (2019). https://doi.org/ 10.1016/j.apergo.2018.08.026 17. Stoltz, M.-H., Giannikas, V., McFarlane, D., Strachan, J., Um, J., Srinivasan, R.: Augmented reality in warehouse operations: opportunities and barriers. IFACPapersOnLine 50, 12979–12984 (2017). https://doi.org/10.1016/j.ifacol.2017.08. 1807 18. Arksey, H., O’Malley, L.: Scoping studies: towards a methodological framework. Int. J. Social Res. Methodol. 8, 19–32 (2005). https://doi.org/10.1080/ 1364557032000119616 19. Kim, K., Billinghurst, M., Bruder, G., Duh, H.B.-L., Welch, G.F.: Revisiting trends in augmented reality research: a review of the 2nd decade of ISMAR (2008–2017). IEEE Trans. Vis. Comput. Graph. 24, 2947–2962 (2018). https://doi.org/10.1109/ TVCG.2018.2868591 20. Dey, A., Billinghurst, M., Lindeman, R.W., Swan, J.E.: A systematic review of 10 years of augmented reality usability studies. Front. Rob. AI 5, 1–28 (2018). https://doi.org/10.3389/frobt.2018.00037

816

M. Quandt et al.

21. Nee, A.Y.C., Ong, S.K., Chryssolouris, G., Mourtzis, D.: Augmented reality applications in design and manufacturing. CIRP Ann. 61, 657–679 (2012). https://doi. org/10.1016/j.cirp.2012.05.010 22. Syberfeldt, A., Danielsson, O., Gustavsson, P.: Augmented reality smart glasses in the smart factory: product evaluation guidelines and review of available products. IEEE Access 5, 9118–9130 (2017). https://doi.org/10.1109/ACCESS.2017.2703952 23. Wang, X., Ong, S.K., Nee, A.Y.C.: A comprehensive survey of augmented reality assembly research. Adv. Manuf. 4(1), 1–22 (2016). https://doi.org/10.1007/ s40436-015-0131-4 24. Nee, A.Y.C., Ong, S.K.: Virtual and augmented reality applications in manufacturing. IFAC Proc. Vol. 46, 15–26 (2013). https://doi.org/10.3182/20130619-3-RU3018.00637 25. Baroroh, D.K., Chu, C.-H., Wang, L.: Systematic literature review on augmented reality in smart manufacturing: collaboration between human and computational intelligence. J. Manuf. Syst. 61, 696–711 (2021). https://doi.org/10.1016/j.jmsy. 2020.10.017 26. Sahu, C.K., Young, C., Rai, R.: Artificial intelligence (AI) in augmented reality (AR)-assisted manufacturing applications: a review. Int. J. Prod. Res. 59, 4903– 4959 (2021). https://doi.org/10.1080/00207543.2020.1859636 27. Quandt, M., Knoke, B., Gorldt, C., Freitag, M., Thoben, K.-D.: General requirements for industrial augmented reality applications. Procedia CIRP 72, 1130–1135 (2018). https://doi.org/10.1016/j.procir.2018.03.061 28. Fraga-Lamas, P., Fernandez-Carames, T.M., Blanco-Novoa, O., Vilar-Montesinos, M.A.: A review on industrial augmented reality systems for the industry 4.0 shipyard. IEEE Access 6, 13358–13375 (2018). https://doi.org/10.1109/ACCESS.2018. 2808326 29. Boboc, R.G., Gˆırbacia, F., Butil˘ a, E.V.: The application of augmented reality in the automotive industry: a systematic literature review. Appl. Sci. 10, 4259 (2020). https://doi.org/10.3390/app10124259 30. Dini, G., Dalle Mura, M.: Application of augmented reality techniques in throughlife engineering services. Procedia CIRP 38, 14–23 (2015). https://doi.org/10. 1016/j.procir.2015.07.044 31. Rankohi, S., Waugh, L.: Review and analysis of augmented reality literature for construction industry. Vis. Eng. 1(1), 1–18 (2013). https://doi.org/10.1186/22137459-1-9 32. Chi, H.-L., Kang, S.-C., Wang, X.: Research trends and opportunities of augmented reality applications in architecture, engineering, and construction. Autom. Constr. 33, 116–122 (2013). https://doi.org/10.1016/j.autcon.2012.12.017 33. Wang, X., Kim, M.J., Love, P.E.D., Kang, S.-C.: Augmented Reality in built environment: classification and implications for future research. Autom. Constr. 32, 1–13 (2013). https://doi.org/10.1016/j.autcon.2012.11.021 34. Ahmed, S.: A review on using opportunities of augmented reality and virtual reality in construction project management. Organ. Technol. Manag. Constr. 11, 1839– 1852 (2019). https://doi.org/10.2478/otmcj-2018-0012 35. Sidani, A., et al.: Recent tools and techniques of BIM-based augmented reality: a systematic review. J. Build. Eng. 42, 102500 (2021). https://doi.org/10.1016/j. jobe.2021.102500 36. Rejeb, A., Keogh, J.G., Leong, G.K., Treiblmaier, H.: Potentials and challenges of augmented reality smart glasses in logistics and supply chain management: a systematic literature review. Int. J. Prod. Res. 59, 3747–3776 (2021). https://doi. org/10.1080/00207543.2021.1876942

AR-Based Decision Support Systems for Intralogistics

817

37. Yin, R.K.: Case Study Research and Applications - Design and Methods. SAGE, Los Angeles (2018) 38. Mayring, P.: Qualitative content analysis - theoretical background and procedures. In: Bikner-Ahsbahs, A., Knipping, C., Presmeg, N. (eds.) Approaches to Qualitative Research in Mathematics Education, pp. 365–380. Springer, Dordrecht (2015). https://doi.org/10.1007/978-94-017-9181-6 13 39. Kohlbacher, F.: The use of qualitative content analysis in case study research. Forum Qualitative Sozialforschung Forum: Qual. Social Res. 7 (2006). https://doi. org/10.17169/fqs-7.1.75 40. Meyer, M., Vetter, E., Jenner, B., Titscher, S., Wodak, R.: Methods of Text and Discourse Analysis. SAGE, Los Angeles (2007) 41. Gioia, D.A., Corley, K.G., Hamilton, A.L.: Seeking qualitative rigor in inductive research - notes on the gioia methodology. Organ. Res. Methods 16, 15–31 (2013). https://doi.org/10.1177/1094428112452151 42. QCAmap - a software for Qualitative Content Analysis. https://www.qcamap.org/ ui/en/home 43. ISO 9241–11:2018, Ergonomics of human system interaction - Part 11: Usability: Definitions and concepts. Geneva: International Organization for Standardization (2018) 44. Davis, F.D.: Perceived usefulness, perceived ease of use, and user acceptance of information technology. MIS Q. 13, 319–340 (1989). https://doi.org/10.2307/ 249008 45. Brooke, J.: SUS: a quick and dirty usability scale. In: Jordan, P.W., Thomas, B., McClelland, I.L., Weerdmeester, B. (eds.) Usability Evaluation in Industry, pp. 189–194. Taylor & Francis, London (1996) 46. Laugwitz, B., Held, T., Schrepp, M.: Construction and evaluation of a user experience questionnaire. In: Holzinger, A. (ed.) USAB 2008. LNCS, vol. 5298, pp. 63–76. Springer, Heidelberg (2008). https://doi.org/10.1007/978-3-540-89350-9 6 47. Hart, S.G.: Nasa-task load index (NASA-TLX); 20 years later. Proc. Human Fact. Ergon. Society Annual Meet. 50, 904–908 (2006). https://doi.org/10.1177/ 154193120605000909 48. Quandt, M., Freitag, M.: A systematic review of user acceptance in industrial augmented reality. Front. Educ. 6, 700760 (2021). https://doi.org/10.3389/feduc. 2021.700760

Data at the Heart of the Industry of the Future: New Information Issues from an Information and Communication Sciences Perspective Nathalie Pinède1

and Bruno Vallespir2(B)

1 MICA, Université Bordeaux Montaigne, Pessac, France

[email protected]

2 Univ. Bordeaux, CNRS, IMS, UMR 5218, 33405 Talence, France

[email protected]

Abstract. In this article, we propose to take a forward-looking look at the issue of data and information systems in industrial companies, in the context of the industry of the future. Indeed, if, as many claim, we are to take advantage of this industrial revolution to reaffirm the central role of human beings in industrial organizations, it is essential to involve the social and human sciences, and particularly the information and communication sciences, in current research on the subject. The aim of this paper is therefore to identify themes where the contribution of this discipline would be major. To this end, after tracing the major stages in the evolution of ERP (Enterprise Resources Planning) type information systems and positioning the major challenges of the industry of the future in this respect, we raise four main questions. The first concerns the need to develop a new enterprise information system model, taking into account global communication, IoT and CPS. The second concerns the coverage of this information system, and the need to take account of the company’s ecosystem. The third concerns the need to integrate the informal aspect of the organization into this model. Finally, the fourth question relates to the need for a new decision model in a context of distributed intelligence and decision-making. Ultimately, these questions lead to the definition of a roadmap for the evolution of information and decision systems in industrial companies, which deeply involves information and communication sciences. Keywords: Data · ERP · industrial engineering · industry of the future · information and communication sciences · information systems

1 Introduction and Purpose In this paper, we propose to set out a reflective framework on the issue of data, information and its systems in the context of Industry 4.0, also known as the industry of the future. This perspective of the industry of the future lies at the crossroads of reaffirmed human issues and strong technological innovations - new manufacturing processes (additive manufacturing, 3D printing), principles of collaborative robotics (cobotics), advanced simulations (digital twins), massive production of data (big data) and interconnection of these data (IOT - internet of things) (PWC 2016; Barneveld and Jansson 2017). © IFIP International Federation for Information Processing 2023 Published by Springer Nature Switzerland AG 2023 E. Alfnes et al. (Eds.): APMS 2023, IFIP AICT 689, pp. 818–830, 2023. https://doi.org/10.1007/978-3-031-43662-8_58

Data at the Heart of the Industry of the Future

819

The problem of multiple, heterogeneous data, coming from automatic sources (sensors) as well as from human or organisational sources, is therefore an important facet of these new industrial environments, marked by a strong digitalisation and technicalisation of processes and activities. The development of sensors, recording different types of human activities (production, design, manufacturing, etc.) within the company or in connection with external actors (suppliers, customers) produces a colossal quantity of data, requiring automatic (algorithmic) processing. These modalities, characteristic of the new industrialised environments, therefore coexist with more structured info-documentary modalities, linked to the information system of the organisation concerned. This quick overview shows that the industry of the future is an opportunity to rethink the place of people in industrial systems. Indeed, much of the literature on the industry of the future puts forward the principle of “people at the heart of the system”. If we now want this principle to go beyond the mere stage of argument to become a reality, we need to analyse how disciplines stamped as Social Sciences and Humanities (SSH) can fully participate in the debate. One of these disciplines, the Information and Communication Sciences, seems to be very close to the issues of the industry of the future (particularly the data and information issue). However, an analysis of this discipline’s fields of application shows that it generally focuses on areas with which it feels naturally close (education systems, libraries, etc.), and only very rarely takes an interest in industrial systems. Consequently, the aim of this paper is not to present precise results, but rather to launch a forward-looking reflection aimed at identifying scientific themes linked to the industry of the future, in which information and communication sciences can play a major role. In this sense, the proposals made in this paper contribute to the development of industry 5.0, as defined by (Müller 2020). To this end, based on problems identified by industrial engineering but viewed from an information and communication science perspective, this paper explores several questions. How do these different types of data make up information and can they be integrated into information systems (ERP type) but also into documented and structured systems (metadata, knowledge organisation systems)? What new modes of aid to action, decision-making and governance are emerging, in connection with artificial intelligence (algorithmic governance – Henman 2020; Rieder 2020)? What ethical questions arise in these working contexts? What new mediations, data literacies and information cultures should be developed in these so-called industries of the future? The paper is organized as follows. We will review the evolution of information systems in companies focusing on ERP (Enterprise Resources Planning) solutions. Secondly, we will focus on the key points of the so-called industry of the future, in order to confront them with ERP. From this confrontation, we will open up on four major industrial issues of interest to the information and communication sciences.

2 Enterprise Information Systems and ERP The first element of context concerns information systems in companies and ERP (Magal and Word 2012; Ganesh et al. 2014; Sagegg and Alfnes 2020). At the time when management information systems began to be used in industrial companies, their information systems were developed in line with their organisation. Thus, each function, each department in the company built its own information system to meet its own needs and uses.

820

N. Pinède and B. Vallespir

The result was to improve the functioning of these departments, but at the same time to increase the balkanisation of the company, which was already very often the rule. The overall functioning of the company gained nothing from this situation, which reflected a cruel lack of integration. Faced with this situation, new approaches were developed, all with the aim of breaking down the walls erected between departments. The first aspect concerned data. The aim was to bring them all together in a single system, thereby avoiding the same data being found in several places in the company under different names and with different values. This approach was based on the development of large databases and was described as integration by data. The second step was to build business processes which use and produce the data we have just mentioned and which have the characteristic of not being adjusted to the perimeter of the company’s services but of following a functional logic. For example, a process could be envisaged that goes from the customer (order) to the customer (delivery), crossing and relying on as many of the company’s departments as necessary. This approach, largely inspired by Business Process Reengineering (Hammer and Champy 1993), also ignores the boundaries between departments.

Fig. 1. Position of ERP in the CIM pyramid (Invilution)

This approach, with a single database to which business processes implemented on the basis of workflow technology are connected, has gradually become the standard and is the basic principle of ERP. In this sense, ERPs play a major role in the CIM (Computer Integrated Manufacturing) perspective (Waldner 1992), which was particularly popular in the 1980s, promoting the integration of the enterprise through computerisation. Figure 1 presents a revised view of this concept.

Data at the Heart of the Industry of the Future

821

Thus, an ERP has many advantages (integration of data, formalisation and transparency of business processes) which effectively contribute to the integration of the company as long as it covers all the functions of the company (Fig. 2). However, this integration is achieved at the cost of many defects and criticisms (Herr 2013; Dorobat and Nastase 2012): cumbersome implementation and use; complexity; monolithic nature; lack of ergonomics; rigidity and inability to integrate new uses. A survey carried out in 2013 (Cegid 2019) also showed that the main requests for change concerned mobility (working anywhere and on different media) and modularity (business-oriented solutions). The study concluded that: “The ERP of tomorrow […] will be social, will process external data and will be used on the move and in the cloud”. The ensuing evolution can be structured in two main stages. The first stage, which is still in progress, is based on migration to the cloud with several expected impacts (Dorobat and Nastase 2012): mobility; reduction of investments (CXP 2018), improvement of the user experience (ease of use and adaptation to new mobile and collaborative uses). The second, more forward-looking stage aims to move towards agile and scalable ERPs, less monolithic and rigid architectures with better handling and more reliable deployment projects (Dorobat and Nastase 2012), and finally, an evolution of the solution (thanks to a marketplace operation).

Fig. 2. Functional coverage of traditional ERP systems (I-scoop 2023)

As part of this trend, we can also note the I-scoop’s vision (2023) concerning intelligent ERPs (i-ERPs): “[…] you can see the intelligent ERP on the most basic level as the ERP becoming a system of engagement, system of insight and intelligence, system of decisions and system of integration and connectivity instead of just a system of records”. To conclude, while ERP is a salient variation of information systems in companies, it is important to emphasise that it does not encompass all information-related

822

N. Pinède and B. Vallespir

approaches. In particular, the design of the offer (products/services, research departments), the improvement and evolution of the company (organisational innovation) or data linked to social media and digital networks do not fall within the scope of ERP.

3 The Industry of the Future The very strong current developments in information and communication technologies are leading to a new technological offer which has many applications in the industrial sector. It is in this context that the concept of the industry of the future has been flourishing for the last ten years. Whatever the terms used (“factory of the future”, “industry 4.0”, “industry of the future”, etc.), it refers to the technological and organisational changes that will impact the world of work following the emergence and implementation of a new generation of IT solutions.

Fig. 3. The four industrial revolutions (translated from Geandarme, 2018)

It is clear that these components are only facilitators of transformation. Technological developments are only of interest if they lead to new services, new uses, new organisations or new business models. Therefore, talking about the industry of the future is not limited to considering these changes in terms of technology, but integrates the entire current transformation. This is why the factory of the future can be understood in a more general way by integrating all current trends. Beyond the trivial observation that things are changing and that trends are emerging, it is the speed of change that is taking place today that is most characteristic, a speed that leads to talk of an industrial revolution (hence the term Industry 4.0 in reference to the three previous industrial revolutions, Fig. 3).

Data at the Heart of the Industry of the Future

823

In a report presenting the main results of an international survey conducted in 26 countries with 2,000 respondents (PWC, 2016), Industry 4.0 is identified on the basis of a set of tools (mobile media, Internet of Things (IoT)-oriented platforms, intelligent sensors, 3D printing, Big Data-type approaches, augmented reality, cloud computing, advanced human-machine interfaces, etc.) but also on the basis of a number of transformations (improved performance through digitalisation, the crucial role of data, the changing role of the customer, who is more autonomous and more closely connected to the company, etc.). A particularly interesting element of the report is the observation that the pivotal point of the industry of the future is not, in the end, a technological one, but rather linked to the human element and a new digital culture in the company. The role of data, and in particular big data, in the industry of the future is nevertheless crucial. In this respect, we can recall the words of Menger and Paye (2017) concerning the five key developments superimposed through Big data: • • • •

scientific and technical innovations; multiplication of data-generating sources and connectivity; expansion of the market for tools and uses of mass data exploitation; placing of exploited informational imbalances in ethical and political controversy and legal and legislative regulation; • formation of new service markets transforming privacy into a gradable and negotiable utility or “disutility”. An analysis carried out in 2018 (Vallespir 2018) sought to define the content of the industry of the future field on the basis of sources of various kinds (industrial announcements, policy documents of public institutions involved in innovation - regional councils, technology transfer platforms, competitiveness clusters, etc.). This analysis is, of course, indicative, firstly because the set of sources compiled is small and, secondly, because this classification is necessarily debatable. if we now seek to isolate from this set of themes those that correspond to the subject of this paper, we can obtain the following classification: a. The data. This first set concerns the data itself, from its generation (sensors and actuators) to its storage and analysis (Analytics). b. Connectivity. Here we find everything that corresponds to the exchange of data between players, whether it be technical solutions (Internet of Things) or the possible perimeters and scales of exchange: between machines (connected machine), within the framework of collaboration between companies (digital supply chain) or with the customer (adaptation to demand, customisation, integration from order to delivery). An important point is the speed of connection (fast connection, adaptation to demand in real time). c. Management and Control. Here We Find Levels of Control Traditionally Devolved to Control Layers External to the Resources and Requiring Human Intervention Which Are Now Found Implanted into the Technical Resources. This Concerns Supervision (Automatic Control and Supervision), Optimisation (Self-Optimising Systems, Product/process Optimisation) or Quality (Automatic Quality Control). d. Intelligence. This Concerns the Intelligence and Learning Capabilities (Machine Learning) Now Attributed to Technical Resources (Decentralised Intelligence, Intelligent Mechatronic Systems).

824

N. Pinède and B. Vallespir

e. The Convergence of Information Technology and Operational Technology. All the Points We Have just Seen Can Be Encapsulated in a General Principle of Convergence Between Physical and Information Transformation Technology. This Principle is Embodied in the Concept of Cyber-Physical Systems (CPS) or, Even More Generally, Cyber-Physical and Human Systems (Poursoltan et al. 2021). If we compare the evolution of ERPs, as presented above, with the new paradigm represented by the industry of the future, understood as a human society, a place of behaviour and informal exchanges, it appears that this evolution of ERPs remains far from the proposals of the industry of the future (Haddara and Elragal 2015). This is why a number of scientific studies have focused on this subject (Majstorovic et al., 2020; Akyurt et al., 2020; Bytniewski et al., 2020). Anyway, a fundamental question emerges: how can these systems adapt to an organisation where all the players communicate, supply data, demonstrate intelligence and decision-making capabilities? Indeed, one may wonder about the capacity of these systems to meet the needs of complex industrial systems composed solely of CPSs.

4 Main Questioning From an Information and Communication Sciences Point of View On the basis of these two contextual elements (ERP-type information systems and the industry of the future), we can identify four main questions. these stem from infocommunication issues and are enriched here with industrial engineering issues. 4.1 Question 1: What is the Enterprise Information System Model in the Age of Global Communication, IoT and CPS? This questioning comes from the point of view of data: how can the integrative logic of current information systems (based on a single central data model) accommodate an organisation where all the players communicate and supply heterogeneous data, in mass and in real time (characteristic of big data)? More specifically, the problem concerns the vertical integration of data, i.e. across the hierarchical levels of the company. the central and unique model solves the problem de facto since only one level of data is proposed. however, if the human actor continues to be a producer and user of data, in relation to his activity, the production of automatic data via sensors is intensifying and will become dominant in the near future. Several types of data thus coexist at the heart of the organisation’s functioning: formal/informal data; raw/structured data; automatic/socio-personal data. It should be noted that the intensification of data production through sensors is not without risks, whether this concerns health (interweaving of the body and technology, implanted sensors), ethical issues surrounding the data collected or asymmetrical relations between users and technological systems (for example, forms of opacity concerning the conditions of capture, storage and exploitation). If we now want to take advantage of the emerging wealth of information, it is undoubtedly necessary to envisage local information systems, on the scale of autonomous

Data at the Heart of the Industry of the Future

825

resources (CPSs?) which would be responsible for their own data, in terms of updating values and model evolution. How then can we ensure that all these local information systems are interoperable and can be aggregated in order to ensure coherence and integration and thus avoid returning to the previous situation of silos? At the informationdocumentary level, the stakes and challenges are also high. They concern the processes of transforming these heterogeneous data into information / documents / knowledge and raise the question of metadata and knowledge organisation systems to be deployed. 4.2 Question 2. What is the Coverage of This Information System (Integration of Suppliers, Customers, Innovation Ecosystem, Etc.)? Even though enterprise information systems necessarily integrate information relating to their environment, they are still very much focused on their own company. However, the expected information enrichment suggests that this environment will be the source of information of major interest to the company in question. the injunction to refocus on one’s core business, which has been at work for several decades, also pushes in this direction: having outsourced essential functions, the company cannot function alone. This is all the especially true in the context of the industry of the future, where the boundaries of the company, between endogenous and exogenous, are becoming more porous, with the user/customer playing a growing role in industrial processes, whether through feedback and satisfaction, or through new forms of collaborative production/innovation (Morrar et al. 2017). Also, its performance is linked to the knowledge it has of its environment and the diversification of useful/usable data sources. The first point concerns the value chain and therefore relates to horizontal integration: the company’s information system must integrate its suppliers and customers. This raises many questions: what information is needed? what information is accessible? Should we stay at the first level or extend the approach to suppliers of suppliers and customers of customers? What level of integration of “external” social data (for example from digital social networks or CRM tools)? How should it be linked to more flexible working and data production tools, in connection with collaborative innovations (within or outside the company)? In other words, how can we articulate, within what could be considered as a meta-information system, data from substitutive (automation of a certain number of tasks), rationalising (ERP type) and enabling technologies (more flexible and cooperation-oriented technologies) (Bobillier Chaumon 2021, Martine et al. 2016). The second point concerns the rest of the environment: competitors, innovation ecosystems, etc. here too, questions arise about the need for and accessibility of information. The main challenge relating to this questioning therefore concerns the development of information ecosystems, horizontal and open to the environment, but according to perimeters and with data accessibility/utility/usability criteria to be defined. 4.3 Question 3. How Can the Informal Aspect of the Organisation Be Integrated into This Model? The shortcomings of ERPs with regard to “human” aspects were very often noted. In response, it has been proposed that digital social networks-type approaches be integrated,

826

N. Pinède and B. Vallespir

that ergonomics be worked on, etc. At the same time, one of the main slogans of the industry of the future is the repositioning of people at the centre of the system. If the industry of the future aims for this aspect to go beyond the rank of slogan, it must be an opportunity to remember that a company is above all a human society and that, beyond the engineering and managerial vision (of which the ERP is the result), a company is a place of behaviour and informal exchanges. This debate is not new and has already been discussed many times (Roethlisberger and Dickinson 1939; Maslow 1943; Mayo 1971). Also, this hidden side of the company’s information system contributes to its overall performance and must therefore be studied and taken into account: it is necessary to know how to analyse the two sides, formal and informal (Morega 2008). The problem is not easy to solve: we stand in an intermediate space between a mechanistic and explicit approach to information (data model, business process model) and an informal and therefore invisible part. What is the limit, what is the balance between, in one hand, a ERP-type extreme formalisation and, in another hand, the consideration of the generation of activities, exchanges and communication situations in a start-up, non-formalised way? The question is therefore how to represent what cannot be represented (because it is informal), or in other words, how to make the intangible tangible (informationas-thing, Buckland 1991). In this respect, knowledge management, the documentation of innovative projects, and investigative approaches are some of the methods that can be used to transform the informal into the formal, useful for the company. Of course, not everything has to be formalised, recorded and documented, and the public/private boundary within the company, based on the law and the protection of personal data, has to be drawn. 4.4 Question 4. What is the Decision Model for Analysing the Distribution of Intelligence and Decision Making? A simple way of representing the production activity is to consider an operational level and a management level. At the operational level, we find operators who interfere with technical resources which, in turn, act on products. At the management level, we find managers who rely on IT resources to make their decisions. If we now think in terms of intelligence and decision-making capacity, we can distinguish between the current situation (a) and that which can be envisaged in the context of the industry of the future (b). Current situation (a) Operational level: 1. The operator has intelligence and decision-making capacity, both of which can be described as operational, 2. The technical resources do not have these capabilities, they respond either to orders given by the operator or directly sent by the control system, 3. The products also have none of these capabilities and simply undergo the transformations. Management level: 4. The manager has intelligence and decision-making capability,

Data at the Heart of the Industry of the Future

827

5. The IT resources can be described as decision support, but at a sufficiently primitive level to consider that they do not have these capabilities. New situation (b). Operational level: 1. The operator keeps intelligence and decision-making capability, 2. The technical resources still respond to the orders given by the operator or by the control system but have new capabilities: supervision, optimisation, adaptation, diagnosis, learning, 3. Products continue to undergo transformations but get smart capabilities. From the simple data-carrying product to the product capable of deciding on its routing or triggering actions necessary to preserve its integrity or quality for example, the notion of “intelligent product” (Trentesaux et al. 2013) covers a wide range of situations. Management level: 4. The manager always has intelligence and decision-making capability, 5. IT resources now have data analysis and learning capabilities that allow them to become much higher-level decision aids with undoubted intelligence and decisionmaking capacity. Figure 4 summarises this evolution. The problem that then arises concerns the development of a new decision model that takes this evolution into account. Beyond this model, many questions arise: how to represent in this model all the possible solutions, from the traditional centralised system to the system where all the actors have a significant decision-making capability? How to choose the best configuration for each industrial case? Behind this question, another one emerges: how to evaluate these configurations in order to choose the best one according to the performances expected by the company? Finally, the last important

Fig. 4. Evolution of the distribution of intelligence and decision-making capabilities between actors

828

N. Pinède and B. Vallespir

question is: how does the place of the human actor evolve in these new configurations? In connection with artificial intelligence and big data, questions of governance and ‘algorithmic governmentality’ (Henman 2020) arise with acuity, between control and authority of digital devices and processes of negotiation/contestation by human actors, such as proximity managers (Jemine 2021). From the point of view of information and communication sciences, it therefore seems necessary to develop the issue of digital literacies and to bring a critical look at data culture (social, ethical, communicational, informational issues) in training courses (especially technological) and in companies. A data mediation perspective is also a promising area of development (see in particular the model of Human-Data mediation proposed in (Nesvijevskaia 2019). Finally, in connection with these emerging models of decision making, it is important to also alert to the new vulnerabilities and new techno-digital fractures that may emerge.

5 Conclusion We can attempt to summarise our thoughts in Fig. 5. Many questions are thus open, of a fundamentally interdisciplinary nature. In particular, what may interest information and communication sciences, in connection with industrial engineering issues, is the place of the user, of the informal, of ‘human’ data and activities as well as of document and knowledge processing operations in these new industrial company contexts and these informational environments tending to be dominated by massive, heterogeneous data, used independently of any human authority. To put it differently, it is a question, through the context of the industry of the future, of considering information systems, no longer from a strictly technological point of view but as situated anthropo-technical systems. To move in this direction, it is clear that SSH disciplines need to be involved at the forefront, particularly information and communication sciences. The aim of this paper was to identify the main themes that would benefit

Fig. 5. Synthesis

Data at the Heart of the Industry of the Future

829

from such involvement and it now remains to translate these themes into operational research projects. Acknowledgement. These initial reflections were carried out as part of the Réseau de Recherche Impulsion BEST - Usine du futur (Research network on the factory of the future) of the University of Bordeaux.

References Akyurt, B.Z., Kuvvetli, Y., Deveci, M.: Enterprise resource planning in the age of Industry 4.0 - a general overview. In: Paksoy, T., Kochan, C.G., Ali, S.S. (eds.) Logistics 4.0 - Digital Transformation of Supply Chain Management. CRC Press (2020) Barneveld van, J., Jansson, T.: Additive manufacturing: a layered revolution. Technopolis group (2017) Bobillier Chaumon, M.E. (ed.): Digital Transformations in the Challenge of Activity and Work: Understanding and Supporting Technological Changes. ISTE Wiley (2021) Buckland, M.K.: Information as thing. J. Am. Soc. Inf. Sci. 42(5), 351–360 (1991) Bytniewski, A., Matouk, K., Rot, A., Hernes, M., Kozina, A.: Towards Industry 4.0: functional and technological basis for ERP 4.0 systems. In: Hernes, M., Rot, A., Jelonek, D. (eds.) Towards Industry 4.0 — Current Challenges in Information Systems. Studies in Computational Intelligence. Springer, Heidelberg, vol. 887, pp. 3–19 (2020). https://doi.org/10.1007/978-3030-40417-8_1 Cegid: En quoi les ERP nouvelle génération seront-ils vraiment disruptifs ? CEGID, 3 June 2019 CXP: ERP Survey 2018. CXP (2018) Dorobat, I., Nastase, F.: Training issues in ERP implementations. Acc. Manage. Inf. Syst. 11(4), 621–636 (2012) Ganesh, K., Sanjay Mohapatra, Anbuudayasankar, S.P., Sivakumar, P.: Enterprise Resource Planning: Fundamentals of Design and Implementation, Springer, Cham (2014). https://doi.org/10. 1007/978-3-319-05927-3 Geandarme, F.: Industrie 4.0 : définition et mise en œuvre vers l’usine connectée. Visiativ (2018) Haddara, M., Elragal, A.: The readiness of ERP systems for the factory of the future. Procedia Comput. Sci. 64, 721–728 (2015) Hammer, M., Champy, J.: Reengineering the Corporation: Manifesto for Business Revolution. A. Zondervan (1993) Henman, P.: Governing by algorithms and algorithmic governmentality: towards machinic judgement. In: The Algorithmic Society. Routledge, pp. 19–34 (2020) Herr, B.: ERP à l’heure du big data. Solutions Logiciels, 42 (2013) Invilution (2023). http://www.invilution.com/comprehensive-solution I-scoop: From ERP to intelligent ERP in the smart factory and supply chain (2023). www.i-scoop. eu/erp-intelligent-erp-smart-factory-supply-chain Jemine, G.: Deconstructing new ways of working: a five-dimensional conceptualization proposal. In: Mitev, N., Aroles, J., Stephenson, K.A., Malaurent, J. (eds) New Ways of Working. Technology, Work and Globalization. Palgrave Macmillan, Cham, pp. 453–480 (2021). https://doi. org/10.1007/978-3-030-61687-8_18 Magal, S.R., Word, J.: Integrated Business Processes with ERP Systems. Wiley (2012) Majstorovic, V., Stojadinovic, S., Lalic, B., Marjanovic, U.: ERP in Industry 4.0 context. APMS 2020, August 30th – September 3rd, Novi Sad, Serbia. In: Lalic, B., Majstorovic, V., Marjanovic, U., von Cieminski, G., Romero, D. (eds.) Advances in Production Management Systems. The Path to Digital Transformation and Innovation of Production Management Systems.

830

N. Pinède and B. Vallespir

IFIP Advances in Information and Communication Technology, Springer, Cham, vol. 591, pp. 287–294 (2020). https://doi.org/10.1007/978-3-030-57993-7_33 Martine, T., Cooren, F., Bénel, A., Zacklad, M.: What does really matter in technology adoption and use? A CCO approach. Manage. Commun. Q. 30(2), 164–187 (2016) Maslow, A.H.: The theory of Human Motivation. Psychol. Rev. 50, 370–396 (1943) Mayo, E.: Hawthorne and the Western Electric Company. Organization Theory (1971) Menger, P.M., Paye, S.: Big data et traçabilité numérique : Les sciences sociales face à la quantification massive des individus. Collège de France, Paris (2017) Morega, C.E.: L’informel dans le cadre des organisations - Manière cachée du fonctionnement organisationnel. Master thesis, University of Bucarest (2008) Morrar, R., Arman, H., Mousa, S.: The fourth industrial revolution (Industry 4.0): a social innovation perspective. Technol. Innov. Manag. Rev. 7(11), 12–20 (2017) Müller, J.: Enabling Technologies for Industry 5.0 - results of a workshop with Europe’s technology. European Commission, Directorate-General for Research and Innovation, Directorate F — Prosperity, Unit F5 Industry 5.0 (2020) Nesvijevskaia, A.: Phénomène Big Data en entreprise : processus projet, génération de valeur et Médiation Homme-Données. Ph.D. thesis. CNAM (2019) Poursoltan, M., Pinède, N., Traore, M.K., Vallespir, B.: A new descriptive, theoretical framework for Cyber-physical and human systems based on Activity Theory. INCOM, 7–9 June, Budapest, Hungary (2021) PWC: Industry 4.0: Building the digital enterprise. PWC (2016) Rieder, B.: Engines of Order: A Mechanology of Algorithmic Techniques. Amsterdam University Press (2020) Roethlisberger, F.J., Dickinson, W.J.: Management and the Worker. Harvard University Press (1939) Sagegg, O.J., Alfnes, E.: ERP Systems for Manufacturing Supply Chains: Applications, Configuration, and Performance. CRC Press (2020) Trentesaux, D., Grabot, B., Sallez, Y.: Intelligent products: a spinal column to handle information exchange in supply chains. APMS, State College, PY, USA (2013) Vallespir, B.: L’usine du futur : tendances technologiques et organisationnelles. Université d’été pluridisciplinaire et internationale sur le travail – Travail et innovations technologiques. 2–6 july, Bordeaux, France (2018) Waldner, J-B.: Principles of Computer-Integrated Manufacturing. Wiley, Hoboken (1992)

Author Index

A Abonyi, János 543 Agbomemewa, Lorenzo 702 Agerskans, Natalie 311 Ahmadi, Alireza 716 Aitzhanova, Malika 297 Almström, Peter 228 Amrani, Anne Zouggar 3, 69 Arana-Landin, Germán 432 Arvidsson, Ala 789 Aschenbrenner, Doris 559 Ashjaei, Mohammad 311 Azamfirei, Victor 587 B Bareche, Imene 730 Bauernhansl, Thomas 184 Bellgran, Monica 213, 789 Bettoni, Andrea 702 Birkie, Seyoum Eshetu 789 Bortolotti, Thomas 109, 200 Boscari, Stefania 109, 200 Bousdekis, Alexandros 649 Braun, Greta 513 Brehm, Nico 417 Bruch, Jessica 311 Bugger, Mette Holmriis 97 C Cagliano, Anna Corinna 54 Chavez, Zuhara Zemke 213, 789 Chifu, Emil St. 386 Chirumalla, Koteshwar 311, 401 Cimini, Chiara 501 Clemens, Florian 528 Colloseus, Cecilia 559 Confalonieri, Matteo 702 Costa, Federica 716 Crnobrnja, Jelena 269 Cutrona, Vincenzo 702

D da Silva, Márcia Terra 487 Daniele, Fabio 702 de Camargo, Rogerio Queiroz 487 Demiralay, Yüksel De˘girmencio˘glu 691 Dietsch, Maximilian 417 Digiesi, Salvatore 745 Dikhanbayeva, Dinara 297 Dreyer, Heidi Carin 171 Drobnjakovic, Milos 458 E El-Thalji, Idriss 297 Emmanouilidis, Christos 617 Engelmann, Frank 417 Eren, Serkan 171 Ericsson, Kristian 139 Eriksson, Victor 171 Evans, Richard 372 F Facin, Ana Lucia Figueiredo 487 Ferrario, Andrea 702 Ferrazzi, Matteo 125, 242 Fiedler, Jannick 573 Florescu, Tiberiu 617 Foosherian, Mina 649 Franciosi, Chiara 633 Franke, Susanne 602 Frecassetti, Stefano 39, 157 Freitag, Michael 471, 803 G Gaiardelli, Paolo 27 Gerlach, Johanna 417 Gittler, Thomas 339 Gonçalves, Rodrigo Franco 487 Grimaldi, Sabrina 54 Grob, Willem 200

© IFIP International Federation for Information Processing 2023 Published by Springer Nature Switzerland AG 2023 E. Alfnes et al. (Eds.): APMS 2023, IFIP AICT 689, pp. 831–833, 2023. https://doi.org/10.1007/978-3-031-43662-8

832

H Haga, Yasunori 253 Halász, Gergely 543 Hall, Roland 184 Hämmerle, Oliver 471 Hatakeyama, Shintaro 253 Hauge, Jannicke Baalsrud 789 Hautzinger, Michael 184 Henriksen, Bjørnar 282 Holmemo, Marte Daae-Qvale 15, 171 Holmen, Elsebeth 171 I Iakymenko, Natalia 15, 97, 171 Islam, Mohammad Hasibul 213 Ivezic, Nenad 458 J Jarebrant, Caroline 228 Jarrar, Qussay 326 Javadi, Siavash 401 Jelisic, Elena 458 Johansson, Björn 678 Johnson, Patrik 789 K Kaihara, Toshiya 760 Kara, Yakup 691 Kato, Karimu 253 Khan, Prince Waqas 730 Kimura, Atsushi 253 Kokuryo, Daisuke 760 Korsen, Eirik Bådsvik Hamre 15 Kreutz, Markus 803 Kulvatunyou, Boonserm 458 Kurata, Takeshi 253 Kurdve, Martin 789 L Lagorio, Alexandra 501 Lahlou, Khadija 326 Lalic, Bojan 357 Lalic, Danijela Ciric 357 Landeta-Manzano, Beñat 432 Larsen, Sverre Sørbye 84 Lodgaard, Eirin 171 Löfving, Malin 228

Author Index

Lorenzi, Andrea 27 Lucchese, Andrea 745 Lukhmanov, Yevgeniy 297 Lutzer, Bernhard 649 M Macedo, Helena 69 Maffei, Antonio 139 Maghazei, Omid 339, 573 Mangano, Giulio 54 Marjanovic, Ugljesa 269 May, Gökan 587 Medini, Khaled 326 Medvegy, Tibor 543 Mentzas, Gregoris 649 Messerli, Marco 339 Miranda, Salvatore 633 Miura, Takahiro 253 Mizuhara, Houei 760 Montini, Elias 702 Mori, Ikue 253 Morton, Etta 109 Mummolo, Giovanni 745 Mütze-Niewöhner, Susanne 528 N Nagata, Daiki 760 Nakahira, Katsuko 253 Netland, Torbjørn 339, 573 Nwogu, Uchechukwu 372 Nyberg, Jarle 84 O Ogiso, Satoki 253 Oh, Hakju 458 Økland, Sunniva 15, 171 Oliveira, Eduardo e 444 P Papadimitropoulou, Christina 69 Paraskevopoulos, Fotis 649 Pedersen, Ann-Charlott 171 Pereira, Teresa 444 Pfeifroth, Tobias 417 Pinède, Nathalie 818 Popara, Jelena 357

Author Index

Portioli-Staudacher, Alberto 39, 125, 157, 242, 716 Powell, Daryl 15, 27, 69, 109, 171, 200 Presciuttini, Anna 157 Prezioso, Matteo 27 Psarommatis, Foivos 587 Q Quandt, Moritz 803 R Rabelo, Ricardo José 662 Rafele, Carlo 54 Reke, Eivind 15, 97 Riedel, Alexander 417 Riedel, Ralph 602 Riemma, Stefano 633 Rocco, Paolo 702 Romero, David 69, 269, 501, 662, 678 Rossini, Matteo 39, 157 Ruppert, Tamás 543 S Sagli, Signe 15, 171 Salunkhe, Omkar 678 Sand, Sigrid Eliassen 15, 171 Savkovic, Milena 357 Schöllmann, Jan 184 Schuh, Günther 528 Schumacher, Simon 184 Seeliger, Arne 573 Sengupta, Sourav 171 Sgarbossa, Fabio 775 Shehab, Essam 297 Sigüenza Tamayo, Waleska 432 Singer-Coudoux, Katrin 513 Softic, Selver 269, 386

833

Stahre, Johan 513, 678 Stefanovic, Darko 269 Stern, Hendrik 803 Syberfeldt, Anna 678 T Tay, Mayari Perez 213 Thomassen, Maria Kollberg Turcin, Ioan 386 Turkyilmaz, Ali 297 U Umeda, Toyohiro 760 Ungureanu, Radu 386 Uriarte-Gallastegi, Naiara Urquia, Ilse 3

282

432

V Vallespir, Bruno 3, 818 Veneroso, Ciele Resende 633 Verginadis, Yiannis 649 Vijayakumar, Vivek 775 W Waschull, Sabine 617 Watanabe, Rei 253 Wellsandt, Stefan 649 Widfeldt, Magnus 228 Willemsen, Fabian 528 Wuest, Thorsten 326, 730 Z Zambiasi, Lara Popov 662 Zambiasi, Saulo Popov 662 Zanchi, Matteo 27 Zenezini, Giovanni 54