Engineering Societies in the Agents World III: Third International Workshop, ESAW 2002, Madrid, Spain, September 16-17, 2002, Revised Papers (Lecture Notes in Computer Science, 2577) 3540140093, 9783540140092

Citation preview

Lecture Notes in Artificial Intelligence Subseries of Lecture Notes in Computer Science Edited by J. G. Carbonell and J. Siekmann

Lecture Notes in Computer Science Edited by G. Goos, J. Hartmanis, and J. van Leeuwen

2577

3

Berlin Heidelberg New York Barcelona Hong Kong London Milan Paris Tokyo

Paolo Petta Robert Tolksdorf Franco Zambonelli (Eds.)

Engineering Societies in the Agents World III Third International Workshop, ESAW 2002 Madrid, Spain, September 16-17, 2002 Revised Papers

13

Series Editors Jaime G. Carbonell, Carnegie Mellon University, Pittsburgh, PA, USA J¨org Siekmann, University of Saarland, Saarbr¨ucken, Germany Volume Editors Paolo Petta Austrian Research Institute for Artificial Intelligence Intelligent Agents and New Media Group Schottengasse 3, 1010 Vienna, Austria E-mail: [email protected] Robert Tolksdorf Freie Universität Berlin Institut für Informatik, AG Netzbasierte Informationssysteme Takustr. 9, 14195 Berlin, Germany E-mail: [email protected] Franco Zambonelli Università di Modena e Reggio Emilia Dipartimento di Scienze e Metodi dell’Ingegneria Via Allegri 13, 42100 Reggio Emilia, Italy E-mail: [email protected] Cataloging-in-Publication Data applied for A catalog record for this book is available from the Library of Congress. Bibliographic information published by Die Deutsche Bibliothek Die Deutsche Bibliothek lists this publication in the Deutsche Nationalbibliografie; detailed bibliographic data is available in the Internet at . CR Subject Classification (1998): I.2.11, I.2, C.2.4, D.1.3, D.2.2, D.2.7, D.2.11 ISSN 0302-9743 ISBN 3-540-14009-3 Springer-Verlag Berlin Heidelberg New York This work is subject to copyright. All rights are reserved, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, re-use of illustrations, recitation, broadcasting, reproduction on microfilms or in any other way, and storage in data banks. Duplication of this publication or parts thereof is permitted only under the provisions of the German Copyright Law of September 9, 1965, in its current version, and permission for use must always be obtained from Springer-Verlag. Violations are liable for prosecution under the German Copyright Law. Springer-Verlag Berlin Heidelberg New York a member of BertelsmannSpringer Science+Business Media GmbH http://www.springer.de © Springer-Verlag Berlin Heidelberg 2003 Printed in Germany Typesetting: Camera-ready by author, data conversion by DA-TeX Gerd Blumenstein Printed on acid-free paper SPIN: 10872344 06/3142 543210

Preface

The characteristics of software systems are undergoing dramatic changes. We are moving rapidly into the age of ubiquitous information services. Persistent computing systems are being embedded in everyday objects. They interact in an autonomous way with each other to provide us with increasingly complex services and functionalities that we can access at any time from anywhere. As a consequence, not only do the numbers of components of software systems increase; there is also a strong qualitative impact. Software systems are increasingly made up of autonomous, proactive, networked components. These interact with each other in patterns and via mechanisms that can hardly be modeled in terms of classical models of interaction or service-oriented coordination. To some extent, future software systems will exhibit characteristics making them more resemblant of natural systems and societies than of mechanical systems and software architectures. This situation poses exciting challenges to computer scientists and software engineers. Already, software agents and multi-agent systems are recognized as both useful abstractions and effective technologies for the modeling and building of complex distributed applications. However, little is done with regard to effective and methodic development of complex software systems in terms of multiagent societies. An urgent need exists for novel approaches to software modeling and software engineering that enable the successful deployment of software systems made up of a massive number of autonomous components, and that allow us to control and predict their behaviour. It is very likely that such innovations will exploit lessons from a variety of different scientific disciplines, such as sociology, economics, organization science, modern thermodynamics, and biology. The sequel to successful editions in 2000 in Berlin and 2001 in Prague, ESAW 2002 remained committed to the use of the notion of multi-agent systems as the seed for animated, constructive, and highly interdisciplinary discussions about technologies, methodologies, and tools for the engineering of complex distributed applications. While the workshop placed an emphasis on practical engineering issues, it also welcomed theoretical, philosophical, and empirical contributions, provided that they clearly documented their connection to the core applied issues. The organizers received 35 papers that underwent a strict scientific peer review for quality, with three reviews per paper. This process led to the selection of 20 papers for presentation at the workshop. The accepted papers were reviewed again for this volume to ensure the integration of reviewers’ comments and issues that arose from discussions at the workshop. The discussions at the workshop were stimulated by an invited presentation of Marco Dorigo, Institut de Recherches Interdisciplinaires et de D´eveloppements en Intelligence Artificielle Universit´e, IRIDIA, Universit´e Libre de Bruxelles on “Ant Algorithms, Self-organization and Stigmergy”.

VI

Preface

The organizers want to thank Europe’s IST-funded Network of Excellence for agent-based computing, AgentLink II, the Austrian Society for Artificial ¨ Intelligence (OGAI) and the Universidad Rey Juan Carlos, Madrid for financial support in realizing the workshop.

Madrid, September 2002

Paolo Petta Robert Tolksdorf Franco Zambonelli

Organization Workshop Organizers Paolo Petta Robert Tolksdorf Franco Zambonelli

Austrian Research Institute for Artificial Intelligence, Vienna, Austria Institut f¨ ur Informatik, Freie Universit¨at Berlin, Germany Dip. di Elettronica, Informatica e Sistemistica, Univ. Bologna, Reggio Emilia/Modena, Italy

Local Organizing Chair Sascha Ossowski

Grupo de Investigaci´on en Inteligencia Artificial, Univ. Rey Juan Carlos, Madrid, Spain

Programme Committee Federico Bergenti

Dip. di Ingegneria dell’Informazione, Univ. degli Studi di Parma, Italy Jeffrey Bradshaw Inst. for Human and Machine Cognition, Univ. of West Florida, USA Cristiano Castelfranchi Inst. of Psychology, CNR, Italy Paolo Ciancarini Dip. di Scienze dell’Informazione, Univ. di Bologna, Italy Helder Coelho Dept. of Informatics of the Faculty of Sciences, Univ. of Lisbon, Portugal Keith Decker Dept. of Computer and Information Sciences, Univ. of Delaware, Newark, DE, USA Paul Davidsson Dept. of Software Engineering and Computer Science, Blekinge Inst. of Technology, Karlskrona, Sweden Bruce Edmonds Centre for Policy Modelling, Manchester Metropolitan Univ., Manchester, UK Rino Falcone Inst. of Psychology, CNR, Italy Tim Finin Computer Science and Electrical Engineering Dept., Univ. of Maryland, Baltimore County, MD, USA Stephan Flake Cooperative Computing & Communication Lab, Innovation Center of Siemens and Univ. Paderborn, Germany Martin Fredriksson Dept. of Software Engineering and Computer Science, Blekinge Inst. of Technology, Karlskrona, Sweden Marie-Pierre Gleizes Inst. de Recherche en Informatique de Toulouse, Univ. Paul Sabatier, Toulouse, France Rune Gustavsson Dept. of Software Engineering and Computer Science, Blekinge Inst. of Technology, Karlskrona, Sweden

VIII

Organization

Nicholas Jennings

Intelligence, Agents and Multimedia Group, Dept. of Electronics and Computer Science, Univ. of Southampton, UK Paul Kearney Intelligent Agents, BT exact, UK Matthias Klusch German Research Center for Artificial Intelligence – DFKI GmbH, Saarbr¨ ucken, Germany Yannis Labrou Powermarket, Inc., Belmont, CA, USA Lyndon C. Lee Intelligent Agents, BT exact, UK Michael Luck Intelligence, Agents and Multimedia Group, Dept. of Electronics and Computer Science, Univ. of Southampton, UK Scott Moss Centre for Policy Modelling, Manchester Metropolitan Univ., Manchester, UK Pablo Noriega Laboratorio Nacional de Inform´ atica Avanzada, A.C, Xalpa, Vera Cruz, M´exico Andrea Omicini Dip. di Elettronica, Informatica e Sistemistica, Univ. di Bologna, Cesena, Italy Sascha Ossowski Grupo de Investigaci´on en Inteligencia Artificial, Univ. Rey Juan Carlos, Madrid, Spain H. Van Dyke Parunak Altarum Inst., Ann Arbor, MI, USA Michal Pechoucek Dept. of Cybernetics, Faculty of Electrical Engineering, Czech Technical Univ., Prague, Czech Republic Jeremy Pitt Intelligent and Interactive Systems, Electrical and Electronic Engineering Dept., Imperial College of Science, Technology and Medicine, London, UK Agostino Poggi Dip. di Ingegneria dell’Informazione, Univ. degli Studi di Parma, Italy John R. Rose Dept. of Computer Science and Engineering, College of Engineering, Univ. of South Carolina, Columbia, USA Onn Shehory IBM Haifa Research Laboratories, Haifa, Israel Christophe Inst. de Recherche en Informatique de Toulouse, Sibertin-Blanc Univ. Paul Sabatier, Toulouse. France Jos´e M. Vidal Dept. of Computer Science and Engineering, College of Engineering, Univ. of South Carolina, Columbia, USA Gerhard Weiß Theoretische Informatik und Grundlagen der K¨ unstlichen Intelligenz, Inst. f¨ ur Informatik, Technische Univ. M¨ unchen,Germany Bin Yu Information Technology and Engineering, College of Engineering, Carolina State Univ., Raleigh, USA

Table of Contents

Views A Discussion of Two Major Benefits of Using Agents in Software Development . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 Federico Bergenti Signs of a Revolution in Computer Science and Software Engineering . . . . . . .13 F. Zambonelli and H. Van Dyke Parunak Models Rationality, Autonomy and Coordination: The Sunk Costs Perspective . . . . . 29 Matteo Bonifacio, Paolo Bouquet, Roberta Ferrario, and Diego Ponte A Normative and Intentional Agent Model for Organisation Modelling . . . . . 39 Joaquim Filipe Simulating Computational Societies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53 Lloyd Kamara, Alexander Artikis, Brendan Neville, and Jeremy Pitt Co-Fields: Towards a Unifying Approach to the Engineering of Swarm Intelligent Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . 68 Marco Mamei, Franco Zambonelli, and Letizia Leonardi A Schema for Specifying Computational Autonomy . . . . . . . . . . . . . . . . . . . . . . . . 82 Matthias Nickles, Michael Rovatsos, and Gerhard Weiß Activity Theory as a Framework for MAS Coordination . . . . . . . . . . . . . . . . . . . . 96 Alessandro Ricci, Andrea Omicini, and Enrico Denti An Operational Framework for the Semantics of Agent Communication Languages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111 Giovanni Rimassa and Mirko Viroli Access-as-you-need: A Computational Logic Framework for Accessing Resources in Artificial Societies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126 Francesca Toni and Kostas Stathis Motivating Participation in Peer to Peer Communities . . . . . . . . . . . . . . . . . . . . 141 Julita Vassileva Engineering ADELFE: A Methodology for Adaptive Multi-agent Systems Engineering . . . . . . . . . . . . . . . . . . . . . . . . . . . . 156 Carole Bernon, Marie-Pierre Gleizes, Sylvain Peyruqueou, and Gauthier Picard

X

Table of Contents

Evaluating Multi-agent System Architectures: A Case Study Concerning Dynamic Resource Allocation . . . . . . . . . . . . . . . . . . 170 Paul Davidsson and Stefan Johansson Engineering Agent Systems for Decision Support . . . . . . . . . . . . . . . . . . . . . . . . . .184 Sascha Ossowski, Josefa Z. Hern´ andez, Carlos A. Iglesias, and Alberto Fern´ andez Co-ordinating Heterogeneous Interactions in Systems Composed of Active Human and Agent Societies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 199 Konstantinos Prouskas and Jeremy Pitt Modelling and Design SABPO: A Standards Based and Pattern Oriented Multi-agent Development Methodology . . . . . . . . . . . . . 213 O˘guz Dikenelli and Riza Cenk Erdur Modelling a Multi-agent System Environment . . . . . . . . . . . . . . . . . . . . . . . . . . . . .227 Sehl Mellouli, Guy W. Mineau, and Daniel Pascot Towards a Methodology for Coordination Mechanism Selection in Open Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 241 Simon Miles, Mike Joy, and Michael Luck Specification by Refinement and Agreement: Designing Agent Interaction Using Landmarks and Contracts . . . . . . . . . . . . . 257 Hans Weigand, Virginia Dignum, John-Jules Meyer, and Frank Dignum An Agent and Goal-Oriented Approach for Virtual Enterprise Modelling: A Case Study . . . . . . . . . . . . . . . . . . . . . . . . . . . 270 Zhi Liu and Lin Liu Author Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .285

A Discussion of Two Major Benefits of Using Agents in Software Development )HGHULFR%HUJHQWL 

$27/DE 3DUFR$UHDGHOOH6FLHQ]H$3DUPD,WDO\ [email protected] http://aot.ce.unipr.it  )5$0H7HFK 9LD6DQ/HRQDUGR3DUPD,WDO\ [email protected] http://www.frame-tech.it

Abstract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

1

Introduction

7KH XOWLPDWH REMHFWLYH RI DJHQWRULHQWHG VRIWZDUH HQJLQHHULQJ $26(  LV WR SURYLGH DOO QHFHVVDU\ WRROV WR HQJLQHHU DJHQWEDVHG V\VWHPV 6XFK WRROV LQFOXGH EXW DUH QRW UHVWULFWHG WR PHWKRGRORJLHV DQG GHYHORSPHQW IUDPHZRUNV 1RZDGD\V $26( KDV SDUWLDOO\ DFFRPSOLVKHG LWV WDVN ZH KDYH VRPH JRRG PHWKRGRORJLHV DQG VRPH JRRG GHYHORSPHQWIUDPHZRUNVDOVR@ZDVLQWURGXFHGLQPLGt VDQGWKH\NHHSVD\LQJWKDWRQH 33HWWDHWDO(GV (6$:/1$,SS ©6SULQJHU9HUODJ%HUOLQ+HLGHOEHUJ

)HGHULFR%HUJHQWL

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

2

A Model of Agentware

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

$'LVFXVVLRQRI7ZR0DMRU%HQHILWVRI8VLQJ$JHQWVLQ6RIWZDUH'HYHORSPHQW

KROGVIRUWKHVHQGHUDQGWKDWWKHVHQGHULVWU\LQJWRDFKLHYHWKHFRUUHVSRQGLQJUDWLRQDO HIIHFW7KLVLVEDVLFDOO\DUDWKHUNQRWW\ZD\WROHWWKHUHFHLYHUNQRZZKDWLVJRLQJRQ LQ WKH VHQGHUt V PHQWDO VWDWH 7KH DGYDQWDJH RI XVLQJ D VWUXFWXUHG $&/ LQVWHDG RI D PRUHQDWXUDOH[FKDQJHRIUHSUHVHQWDWLRQVRIEHOLHIVLQWHQWLRQVDQGFDSDELOLWLHVVLPSOL ILHVWKHGHYHORSPHQWRIUHDFWLYHDJHQWVFDSDEOHRIFRPSOH[LQWHUDFWLRQV 7KH VHPDQWLFV WKDW ZH QRUPDOO\ XVH ZLWK PRVW SRSXODU $&/V UHO\ RQ KHDY\ DV VXPSWLRQVRQWKHFDSDELOLWLHVRIDJHQWV)RUH[DPSOHWKH),3$$&/UHTXLUHVDQDJHQW WRUHDVRQRQWKHUHFHLYHUt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able 1.6HPDQWLFVRIVRPHSHUIRUPDWLYHVRIWKH3DU$'($&/

0HVVDJH ,QIRUPVUS $FKLHYHVUS 4XHU\5HIVUS[ 5HTXHVWVUD $JUHHVUD 5HIXVHVUD

6HPDQWLFV ,V%US ,V,US ,V%V>%UT∧XQLILHVS[ T @ __DFKLHYHVUGRQHD __ __LQIRUPVU,VGRQHD __ __LQIRUPVU¬,VGRQHD __

)HGHULFR%HUJHQWL

%V,UGRQHD ĸ,UGRQHD %V¬,UGRQHD ĸ¬,UGRQHD %VGRQHD ĸ%UGRQHD



7KHILUVWUXOHVWDWHVWKDWLIDJHQW LQWHQGVWRSHUIRUPWKHQPXVWNQRZLW7KH VHFRQGUXOHFRYHUVWKHRSSRVLWHFDVHLI GHFLGHVQRWWRSHUIRUPWKHQPXVWNQRZ LW7KHWKLUGUXOHVD\VWKDWDVVRRQDV FRPHVWRNQRZWKDWKDVEHHQGRQHWKHQDOVR PXVWNQRZLW $QLQWHUDFWLRQODZKDVDSUHFRQGLWLRQDQGDWHUPLQDWLRQFRQGLWLRQ,QWKHH[DPSOH DERYHWKHSUHFRQGLWLRQIRU LV LH UHFHLYHGWKHUHTXHVW7KHWHUPLQDWLRQ FRQGLWLRQLV  ∨¬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

3

Comparison with Software Components

7KH FRPSDULVRQ EHWZHHQ DJHQWV DQG FRPSRQHQWV VWDUWV IURP WDEOH ,W VKRZV VRPH LPSRUWDQW FRPSRQHQWRULHQWHG DEVWUDFWLRQV DQG DVVRFLDWHV WKHP ZLWK DJHQWRULHQWHG FRXQWHUSDUWV 7DEOHFRPSDUHVVRPHRIWKH DEVWUDFWLRQV WKDW IRUP WKH FRPSRQHQWVt  PHWDPRGHO ZLWK WKH FRUUHVSRQGLQJ DEVWUDFWLRQV RI WKH DJHQWVt  PHWDPRGHO 7KLV VXJJHVWV WKDW D FRPSOHWHFRPSDULVRQVKRXOGWDNHLQWRDFFRXQWDOVRWKHDEVWUDFWLRQVWKDWDUHPLVVLQJLQ WDEOH )ROORZLQJ WKLV DSSURDFK , VKRXOG WDNH LQWR FRQVLGHUDWLRQ HJ WKH 1(7 PHWDPRGHO>@DQGFRPSDUHLWZLWKHJWKH60$57IUDPHZRUN>@7KLVDSSURDFK VHHPVWRRYHUNLOOWKHSUREOHPEHFDXVHPDQ\DEVWUDFWLRQVLQWKHVHPHWDPRGHOVDUHWRR GLIIHUHQWDQGWKH\DUHKDUGO\FRPSDUDEOH Table 2.&RPSDULVRQEHWZHHQFRPSRQHQWRULHQWHGDQGDJHQWRULHQWHGDEVWUDFWLRQV

$EVWUDFWLRQ &RPPXQLFDWLRQ 0HVVDJH ,QWHUDFWLRQ ZLWK WKHHQYLURQPHQW 6WDWH ,QWHUDFWLRQV EHWZHHQSDUWLHV 5XQWLPH

&RPSRQHQWRULHQWHG 7DVNGHOHJDWLRQ 5HTXHVWVIRUDFWLRQV (YHQWV

$JHQWRULHQWHG 7DVNDQGJRDOGHOHJDWLRQ $&/PHVVDJHV 8SGDWHVRIEHOLHIV

3URSHUWLHVDQGUHODWLRQV ,QWHUIDFHV ,QWHUIDFHUHSRVLWRU\ $SSOLFDWLRQVHUYHU

0HQWDODWWLWXGHV &DSDELOLWLHV 6HPDQWLFPDWFKPDNHU ),3$SODWIRUP

$'LVFXVVLRQRI7ZR0DMRU%HQHILWVRI8VLQJ$JHQWVLQ6RIWZDUH'HYHORSPHQW

3.1

Comparing Abstractions

,QWKHIROORZLQJ,WDNHWKHFRPSRQHQWRULHQWHGDEVWUDFWLRQVLGHQWLILHGLQWDEOHDQG FRPSDUH WKHP ZLWK WKHLU DJHQWRULHQWHG FRXQWHUSDUWV , GLG QRW FKRRVH DQ\ UDQNLQJ FULWHULD\HWDQGWKHUHIRUH,FDQQRWVD\ZKLFKRIWKHWZRDSSURDFKHVLVEHWWHU Communication Model. 7KH PDLQ GLIIHUHQFH EHWZHHQ DJHQWV DQG FRPSRQHQWV LV LQ WKHPHFKDQLVPWKH\XVHWRFRPPXQLFDWH$JHQWVXVH$&/VOLNHWKHRQHWKDW,LQWUR GXFHGLQWKHSUHYLRXVVHFWLRQZKLOHFRPSRQHQWVXVHDPHWDREMHFWSURWRFRO>@,QWKH DJHQWRULHQWHGDSSURDFKDPHVVDJHLVVHQWWRWUDQVIHUSDUWRIWKHVHQGHUt VPHQWDOVWDWH WR WKH UHFHLYHU $ VSHFLDO FDVH RI WKLV PHFKDQLVP LV ZKHQ WKH VHQGHU HQGHDYRUV WR GHOHJDWH D JRDO WR WKH UHFHLYHU HJ WKURXJK WKH    SHUIRUPDWLYH 7KLV VSHFLDO FDVHNQRZQDVJRDOGHOHJDWLRQ>@LVWKHEDVLFPHFKDQLVPWKDWDJHQWVXVHWRGHOHJDWH UHVSRQVLELOLWLHV 7KH FRPSRQHQWVt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

WR DQGWKHQ ZRXOGGHFLGHDXWRQRPRXVO\WKHZD\WRJRHJLWZRXOGGHFLGHWR SHUIRUPDQG7KLVDSSURDFKFRXSOHVDQG WKURXJK WKXVHQDEOLQJ WRSHU IRUPFURVVRSWLPL]DWLRQVEHWZHHQDQG Interaction with the Environment. 7KH HQYLURQPHQW LV D VWUXFWXUDO SDUW RI WKH DJHQWVt  PHWDPRGHO $JHQWV H[HFXWH LQ DQ HQYLURQPHQW WKDW WKH\ FDQ XVH WR DFTXLUH NQRZOHGJH$JHQWVFDQPHDVXUHWKHHQYLURQPHQWDQGWKH\FDQUHFHLYHHYHQWVIURPLW ,QERWKFDVHVDJHQWVUHDFWWRDQ\FKDQJHLQWKHHQYLURQPHQWWKURXJKFKDQJHVLQWKHLU PHQWDOVWDWH7KLVLVUDGLFDOO\GLIIHUHQWIURPWKHFRPSRQHQWRULHQWHGDSSURDFKZKHUH WKHHQYLURQPHQWFRPPXQLFDWHVZLWKFRPSRQHQWVWKURXJKUHLILHGHYHQWV&RPSRQHQWV UHDFWWRUHLILHGHYHQWVDIWHUFRQVWUXFWLQJDUHODWLRQZLWKWKHP 7KH FRPSRQHQWRULHQWHG DSSURDFK 

 WR EHWWHU UHVSHFW HQFDSVXODWLRQ WKDQ WKH DJHQWRULHQWHGDSSURDFKWKHVWDWHRIWKHFRPSRQHQWLVFKDQJHGRQO\ZKHQWKHFRPSR QHQWLWVHOIGHFLGHVWRFKDQJHLWLQUHDFWLRQWRDQHYHQW,IZHJRPRUHLQGHWDLOZHVHH WKDW DOVR WKH DJHQWRULHQWHG DSSURDFK UHVSHFWV HQFDSVXODWLRQ $JHQWV KDYH UHDVRQLQJ FDSDELOLWLHVWKDWDUHWKHDFWXDOUHVSRQVLEOHRIDQ\FKDQJHLQWKHPHQWDOVWDWH$Q\GL

)HGHULFR%HUJHQWL

UHFWSXVKRINQRZOHGJHIURPWKHVHQVRUVWRWKHPHQWDOVWDWHLVUXOHGWKURXJKUHDVRQLQJ DQGWKHPHQWDOVWDWHUHPDLQVHQFDSVXODWHG State Representation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

$JHQWVKDYHDQH[SOLFLWUHSUHVHQWDWLRQRIWKHLUJRDOV $JHQWV KDYH H[SOLFLW NQRZOHGJH RI WKHLU HQYLURQPHQW DQG QRW RQO\ RI RWKHU DJHQWV ([FHSWIRUDXQLTXHLGHQWLILHUDJHQWVGRQRWKDYHSURSHUWLHVWKH\RQO\KDYHUHOD WLRQVZLWKRWKHUDJHQWVDQGZLWKHQWLWLHVLQWKHHQYLURQPHQW $JHQWV PD\ XVH GHGXFWLRQ WR FRPH WR NQRZ PRUH WKDQ ZKDW RWKHU DJHQWV WROG WKHPDQGPRUHWKDQZKDWWKH\PHDVXUHG

,WLVZRUWKGLVFXVVLQJWKHODVWSRLQWDELW&RPSRQHQWVFDQGHGXFHWKHYDOXHRIDSURS HUW\XVLQJDSSOLFDWLRQVSHFLILFPHFKDQLVPV7KHGLIIHUHQFHLVWKDWSURSHUWLHVDQGUHOD WLRQVDUHQRWHDVLO\VWUXFWXUHGLQDORJLFIUDPHZRUNDQGLWLVGLIILFXOWWRXVHJHQHUDO SXUSRVHUHDVRQLQJWHFKQLTXHVZLWKWKHPDQ\UHDVRQLQJLVKDUGFRGHGLQWKHFRPSRQHQW LWVHOI Interaction between Parties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

+LJKOHYHOPHQWDOLVWLFDEVWUDFWLRQVLHEHOLHIVDQGLQWHQWLRQV 7KHPRGHORIWKHHQYLURQPHQWLHHQWLWLHVDQGWKHLUUHODWLRQV

Runtime Environment. ),3$ VWDUWHG LWV ZRUN LQ  WKURXJK WKH GHILQLWLRQ RI D UXQWLPHHQYLURQPHQWWRVXSSRUWLQWHURSHUDEOHPXOWLDJHQWV\VWHPV7KLVZRUNLVYDOX DEOHDQGWRGD\WKHDJHQWRULHQWHGFRXQWHUSDUWRIDQDSSOLFDWLRQVHUYHULVD),3$SODW IRUP OLNH -$'(>@ DQG /($3>@ $FWXDOO\ DSSOLFDWLRQ VHUYHUV DUH KXJH SLHFHV RI VRIWZDUH WKDW RIIHU HQWHUSULVH IHDWXUHV WKDW DYDLODEOH ),3$ SODWIRUPV GR QRW SURYLGH

$'LVFXVVLRQRI7ZR0DMRU%HQHILWVRI8VLQJ$JHQWVLQ6RIWZDUH'HYHORSPHQW

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

Agents Improve Reusability and Composability

7KHFRPSDULVRQWKDW,VNHWFKHGDERYHHQXPHUDWHVGLIIHUHQFHVDQGVLPLODULWLHVEHWZHHQ FRPSRQHQWV DQG DJHQW EXW LW GRHV QRW VWDWH ZKHWKHU DQ DSSURDFK LV EHWWHU WKDQ WKH RWKHUEHFDXVH,GLGQRWDGRSWDQ\UDQNLQJFULWHULD,QWKLVVHFWLRQ,FRQVLGHUWZRRIWKH PRVW LPSRUWDQW IHDWXUHV RI D GHYHORSPHQW WHFKQRORJ\ LH UHXVDELOLW\ DQG FRPSRV DELOLW\ DQG , HYDOXDWH KRZ DQG ZKHQ DJHQWV DUH EHWWHU WKDQ FRPSRQHQWV IURP WKHVH SRLQWVRIYLHZ$VQRWHGDERYHDJHQWV   

&DQXVHJRDOGHOHJDWLRQLQVWHDGRIWDVNGHOHJDWLRQRQO\ 8VH$&/VWKDWH[SORLWWKHDJHQWPRGHOWRJLYHDVHPDQWLFVWRLQWHUDFWLRQV +DYHDPHQWDOVWDWHDQGFDQXVHUHDVRQLQJWRJHQHUDWHNQRZOHGJHIURPHYHQWVDQG PHVVDJHV

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

)HGHULFR%HUJHQWL

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

4

Introducing the Agent Level

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t V ZRUGV LQWR DFFRXQW !  

               

        "#$%& 7KHPHDQLQJIXOFROOHFWLRQVRIDJHQWRULHQWHG HQWLWLHVFRPSULVHEHOLHIVJRDOVDQGFDSDELOLWLHVDQGWKH\DUHREYLRXVO\FORVHUWRKX PDQ LQWXLWLRQ WKDQ FRPSRQHQWRULHQWHG FRXQWHUSDUWV HJ SURSHUW\ HYHQW DQG PHV VDJH )ROORZLQJ WKH VWDQGDUG DSSURDFK , IRUPDOL]H WKLV LGHD GHILQLQJ D QHZ V\VWHP OHYHO  

0HVVDJHVGHSHQGDOVRRQWKHFKRVHQFRQWHQWODQJXDJHEXWWKLVLVQRWDUHDOLVVXH

$'LVFXVVLRQRI7ZR0DMRU%HQHILWVRI8VLQJ$JHQWVLQ6RIWZDUH'HYHORSPHQW

$V\VWHPOHYHO>@LVDVHWRIFRQFHSWVWKDWSURYLGHVDPHDQVIRUPRGHOLQJLPSOH PHQWWDEOHV\VWHPV6\VWHPOHYHOVDUHOD\HUHGLQDVWDFNDQGHDFKOHYHOLVPDSSHGRQ RQH ORZHU OHYHO 6\VWHP OHYHOV DEVWUDFW DZD\ IRU LPSOHPHQWDWLRQ GHWDLOV DQG KLJKHU OHYHOV SURYLGH FRQFHSWV WKDW DUH FORVHU WR KXPDQ LQWXLWLRQ DQG IDU DZD\ IURP LPSOH PHQWDWLRQ 7KLV LV ZK\ ZH VD\ WKDW V\VWHP OHYHOV DUH OHYHOV RI DEVWUDFWLRQ 6\VWHP OHYHOVDUHVWUXFWXUHGLQWHUPVRIWKHIROORZLQJHOHPHQWV    

&RPSRQHQWVDWRPLFEXLOGLQJEORFNVIRUEXLOGLQJV\VWHPVDWWKDWOHYHO /DZV RI FRPSRVLWLRQV ODZV WKDW UXOH KRZ FRPSRQHQWV FDQ EH DVVHPEOHG LQWR D V\VWHP $PHGLXPDVHWRIDWRPLFFRQFHSWVWKDWWKHV\VWHPOHYHOSURFHVVHV /DZVRIEHKDYLRUODZVWKDWGHWHUPLQHKRZWKHEHKDYLRURIWKHV\VWHPGHSHQGVRQ WKHEHKDYLRURIHDFKFRPSRQHQWDQGRQWKHVWUXFWXUHRIWKHV\VWHP

1HZHOOt VNQRZOHGJHOHYHO>@SURYLGHVDEHWWHUSHUVSHFWLYHIRUFKDUDFWHUL]LQJDJHQWV WKDQ WKDW RI DQ\ RSHUDWLRQDO FKDUDFWHUL]DWLRQ %HVLGHV 1HZHOOt V ZRUN GDWHV EDFN WR DQGLWGLGQRWWDNHLQWRDFFRXQWPDQ\RIWKHIHDWXUHVWKDWZHZRXOGOLNHDJHQWVWR KDYH WRGD\ ,Q SDUWLFXODU 1HZHOO ZDV YHU\ PXFK FRQFHQWUDWHG RQ WKH UHDOL]DWLRQ RI VLQJOHDJHQWVFDSDEOHRIH[KLELWLQJLQWHOOLJHQWEHKDYLRU1RZDGD\VZHDUHQRORQJHU LQWHUHVWHGLQUHDOL]LQJDV\VWHPLQWHUPVRIDVLQJOHPRQROLWKLFDJHQWDQGZHSUHIHUWR XVHDJHQWVDVEXLOGLQJEORFNVRIPXOWLDJHQWV\VWHPV7KHLQWHOOLJHQFHUHVXOWVIURPWKH LQWHUDFWLRQRIDJHQWVDQGLWLVDQHPHUJLQJSURSHUW\RIWKHZKROHV\VWHP,Q:HJQHUt V ZRUGV!    GXPEDOJRULWKPV   VPDUWDJHQWV%& -HQQLQJVt  SURSRVDO RI WKH VRFLDO OHYHO>@ JRHV LQ WKLV GLUHFWLRQ WU\LQJ WR UDLVH WKH OHYHO RI DEVWUDFWLRQ RI WKH NQRZOHGJH OHYHO 7KLV DSSURDFK KDV WKH DGYDQWDJH RI UH TXLULQJPLQLPDOFDSDELOLWLHVWRDJHQWVWKXVDFFRPPRGDWLQJDQ\VRUWRIDJHQWDYDLODEOH LQ WKH OLWHUDWXUH 1HYHUWKHOHVV LW PD\ UHPDLQ WRR JHQHULF DQG LWV XWLOLW\ LQ EXLOGLQJ FRPSOH[V\VWHPVVHHPVSRRU 7DNLQJ LQWR DFFRXQW WKH OLPLWDWLRQV RI ERWK 1HZHOOt V DQG -HQQLQJVt  DSSURDFKHV , SURSRVHDQRYHOV\VWHPOHYHOFDOOHGDQG,ZLOOMXVWLI\ODWHUZK\   7KHDJHQW OHYHO LV WKRXJKW DV D UHDVRQDEOH FRPSURPLVH EHWZHHQ 1HZHOOt V SRVLWLRQ WKDW DYRLGV GHDOLQJ VWUXFWXUDOO\ ZLWK LQWHUDFWLRQV DQG -HQQLQJVt  YLHZ RI LQWHUDFWLRQ DV WKH RQO\ UHOHYDQW FRPSRQHQW RI D V\VWHP $W WKH DJHQW OHYHO VLQJOH DJHQWV H[KLELW D UDWLRQDO EHKDYLRU DQG WKH\ DUH GHVFULEHG LQ WHUPV RI PHQWDO DWWLWXGHV WKH\ DUH DOVR SDUW RI D VRFLHW\DQGWKH\LQWHUDFWWKURXJKDQ$&/7KHDJHQWOHYHOH[SORLWVWKHSRVVLELOLWLHVRI UDWLRQDOLW\DQGLWSURYLGHVDZD\IRUGLVWULEXWLQJFRPSOH[LW\DFURVVDJHQWVUDWKHUWKDQ FHQWUDOL]LQJLQWHOOLJHQFHLQWRDVLQJOHPRQROLWKLFDJHQW ,Q RUGHU WR GHILQH WKH DJHQW OHYHO , ILUVW HQXPHUDWH WKH FRPSRQHQWV WKDW LW FRP SULVHV    

$JHQWVWKDWLQWHUDFWWRLPSOHPHQWWKHIXQFWLRQDOLW\RIWKHPXOWLDJHQWV\VWHP %HOLHIVWKDWDQDJHQWKDV ,QWHQWLRQVWKDWDQDJHQWLVWU\LQJWRDFKLHYH $FWLRQVWKDWDQDJHQWFDQSHUIRUPRQWKHHQYLURQPHQWZKHUHLWH[HFXWHV

 

1HZHOOH[SOLFLWO\VWDWHGWKHFRQWUDU\EXWLWVHHPVWKDWKHXVHGWKHSKUDVHu OHYHORIDEVWUDF WLRQv LQDGLIIHUHQWZD\IURPKRZZHXVHLWLQVRIWZDUHHQJLQHHULQJWRGD\

)HGHULFR%HUJHQWL

 

5ROHVWKDWDJHQWVFDQSOD\LQWKHV\VWHP ,QWHUDFWLRQODZVWKDWUXOHKRZDJHQWVSOD\LQJGLIIHUHQWUROHVFDQLQWHUDFW

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

5HVSRQVLELOLW\DQLQWHQWLRQWKDWDQDJHQWLVFRQVWDQWO\EULQJLQJDERXWDQGWKDWLW GRHVQRWGURSHYHQLILWKDVDOUHDG\EHHQDFKLHYHG &DSDELOLW\WKHSRVWFRQGLWLRQVRIDQDJHQWt VDFWLRQLHZKDWWKHDJHQWLVFDSDEOH WRDFKLHYHRQLWVRZQDQGWKHLQWHUDFWLRQODZVLWVXSSRUWVLHKRZLWFDQLQWHUDFW ZLWKRWKHUDJHQWVWRDFKLHYHLWVLQWHQWLRQV

7KHPHGLXPRIWKHDJHQWOHYHOLVWKHUHSUHVHQWDWLRQWKDWHDFKDJHQWKDVDERXWLWVDQG RWKHU DJHQWVt  EHOLHIV LQWHQWLRQV DQG FDSDELOLWLHV $Q DJHQW SURFHVVHV WKH PHGLXP RI WKHDJHQWOHYHOLHFRPHVWRNQRZRWKHUDJHQWVt EHOLHIVLQWHQWLRQVDQGFDSDELOLWLHV WKURXJKLQWHUDFWLRQLQRUGHUWRDFKLHYHLWVJRDOV,QP\SURSRVDOWKHSULQFLSDOPHDQV RILQWHUDFWLRQLVFRPPXQLFDWLRQDQGDJHQWVXVHDQ$&/WRH[FKDQJHUHSUHVHQWDWLRQVRI WKHLUVDQGRWKHUVEHOLHIVLQWHQWLRQVDQGFDSDELOLWLHV2QO\DJHQWVSURGXFHDQGFRQVXPH PHVVDJHVDQGWKHFRPSXWDWLRQRIWKHPXOWLDJHQWV\VWHPUHVXOWVIURPWKHPHVVDJHVWKDW DJHQWVPDQLSXODWH 7KHODZRIEHKDYLRULVDQDGDSWDWLRQRI1HZHOOt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t VEHKDYLRUFDQQRWEHGHVLJQHGZLWKRXWJRLQJWRD ORZHUOHYHO Table 3.(OHPHQWVRIWKHDJHQWOHYHO

(OHPHQW 6\VWHP &RPSRQHQWV 0HGLXP %HKDYLRU

(OHPHQWRIWKH$JHQW/HYHO 0XOWLDJHQWV\VWHP %HOLHIVLQWHQWLRQVDFWLRQVUROHVDQGLQWHUDFWLRQODZV 5HSUHVHQWDWLRQVRIEHOLHIVLQWHQWLRQVDQGFDSDELOLWLHV 3ULQFLSOHRIUDWLRQDOLW\

$'LVFXVVLRQRI7ZR0DMRU%HQHILWVRI8VLQJ$JHQWVLQ6RIWZDUH'HYHORSPHQW

$QDJHQWOHYHOPRGHOLVQRWXVHIXOLIZHGRQRWSURYLGHDPHDQVIRUFRQFUHWHO\LP SOHPHQWLQJWKHV\VWHP,IZHDJUHHWKDWWKHDJHQWOHYHOFDQVLWRQWRSRIWKHFRPSRQHQW OHYHOWKHQZHQHHGDPDSSLQJEHWZHHQHOHPHQWVRIWKHDJHQWOHYHODQGV\VWHPVDWWKH FRPSRQHQWOHYHO*HQHUDOO\VSHDNLQJDQ\GHYHORSPHQWWRROWKDWVXSSRUWVDJHQWOHYHO DEVWUDFWLRQVOHYHODFWXDOO\LPSOHPHQWVWKLVPDSSLQJ-XVWWRFLWHRQHH[DPSOH3DU$'( DOORZV XVLQJ 80/ GLDJUDPV WR PRGHO D V\VWHP DW WKH DJHQW OHYHO WKHQ LW JHQHUDWHV -DYDVNHOHWRQVWKDWDUHDWWKHVDPHOHYHORIDEVWUDFWLRQRIFRPSRQHQWV

5

Conclusions

7KHZKROHOHQJWKRIWKLVSDSHULVGHYRWHGWRJLYHUHDVRQDEOHDUJXPHQWDWLRQVRQZK\ GHYHORSHUVVKRXOGVHULRXVO\WDNHLQWRDFFRXQWWKHSRVVLELOLW\RIXVLQJDJHQWVIRUWKHLU SURGXFWV7KHVHFWLRQRQWKHDJHQWOHYHOLVWKHUHWRSURYLGHDVFLHQWLILFIRXQGDWLRQWR VXFKDUJXPHQWDWLRQV7KHWZRPRVWLPSRUWDQWRXWFRPHVRIWKHVHGLVFXVVLRQVDUHWKDW DJHQWV  

3URYLGHKLJKOHYHODEVWUDFWLRQVWRGHYHORSHUV +DYHVRPHDGYDQWDJHVRYHUFRPSRQHQWVLQWHUPVRIUHXVDELOLW\DQGFRPSRVDELO LW\

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

$UHSDUWLFXODUO\FRPSOH[DQGFRXOGEHQHILWIURPKLJKOHYHODEVWUDFWLRQV :HZDQWWRUHXVHDQGFRPSRVHIUHHO\LQSRVVLEO\GLIIHUHQWSURMHFWV

2QWKHFRQWUDU\ZHFDQIDOOEDFNRQUHDFWLYHDJHQWVRUFRPSRQHQWVZKHQZHKDYHDQ XUJH IRU VSHHG 7KLV GHFLVLRQ FULWHULRQ VHHPV VRXQG EHFDXVH WKH PRUH DQ DJHQW LV FRPSOH[ DQG YDOXHDGGHG WKH PRUH ZH ZDQW WR UHXVH LW DQG FRPSRVH LW ZLWK RWKHU DJHQWV0RUHRYHUUHDFWLYHDJHQWVDUHSHUIHFWO\HTXLYDOHQWWRFRPSRQHQWV DQG ZH GR QRW ORRVH DQ\WKLQJ XVLQJ WKH DJHQWRULHQWHG DSSURDFK LQVWHDG RI WKH FRPSRQHQW RULHQWHGDSSURDFK

)HGHULFR%HUJHQWL

References >@ >@ >@ >@ >@ >@ >@ >@ >@ >@ >@ >@ >@ >@ >@ >@ >@ >@

%HOOLIHPLQH)3RJJL$5LPDVVD*'HYHORSLQJ0XOWLDJHQW6\VWHPVZLWK D ),3$&RPSOLDQW $JHQW )UDPHZRUN 6RIWZDUH 3UDFWLFH DQG ([SHULHQFH   x  %HUJHQWL ) 3RJJL $ $ 'HYHORSPHQW 7RRONLW WR 5HDOL]H $XWRQRPRXV DQG ,QWHURSHUDEOH $JHQWV 3URFV WK ,QWt O &RQIHUHQFH RQ $XWRQRPRXV $JHQWV  x  %HUJHQWL)3RJJL$%XUJ % &DLUH * 'HSOR\LQJ ),3$&RPSOLDQW 6\V WHPVRQ+DQG+HOG'HYLFHV,(((,QWHUQHW&RPSXWLQJ  x  %RRFK*2EMHFW2ULHQWHG$QDO\VLVDQG'HVLJQZLWK$SSOLFDWLRQV$GGLVRQ :HVOH\ &DVWHOIUDQFKL & 0RGHOOLQJ 6RFLDO $FWLRQ IRU $, $JHQWV $UWLILFLDO ,QWHOOL JHQFH  (XURSHDQ&RPSXWHU 0DQXIDFWXUHUt V $VVRFLDWLRQ 6WDQGDUG (&0$ 3DUWL WLRQ,,0HWDGDWD'HILQLWLRQDQG6HPDQWLFV$YDLODEOHDW http://www.ecma.ch )RXQGDWLRQIRU,QWHOOLJHQW3K\VLFDO$JHQWV6SHFLILFDWLRQV$YDLODEOHDW http://www.fipa.org -HQQLQJV152Q$JHQW%DVHG6RIWZDUH(QJLQHHULQJ$UWLILFLDO,QWHOOLJHQFH  x  .LF]DOHV*GHV5LYL UHV-%REURZ'*7KH$UWRIWKH0HWDREMHFW3URWR FRO0,73UHVV /XFN 0 Gt ,QYHUQR 0 $ &RQFHSWXDO )UDPHZRUN IRU $JHQW 'HILQLWLRQ DQG 'HYHORSPHQW7KH&RPSXWHU-  x  0H\HU%2EMHFW2ULHQWHG6RIWZDUH&RQVWUXFWLRQ3UHQWLFH+DOO 1HZHOO$7KH.QRZOHGJH/HYHO$UWLILFLDO,QWHOOLJHQFH x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evolutionRIFXUUHQW PRGHOVDQGPHWKRGRORJLHVEXWDrevolutionDUDGLFDO FKDQJH RI SDUDGLJP >@ SLR QHHUHGE\WKH0$6FRPPXQLW\7KLVUHYROXWLRQZLOOLPSDFWPRVWUHVHDUFKFRPPXQL WLHVDQGZLOOGUDPDWLFDOO\FKDQJHKRZZHFRQFHLYHPRGHODQGEXLOGVRIWZDUHFRP SRQHQWVDQGVRIWZDUHV\VWHPV1H[WJHQHUDWLRQVRIWZDUHV\VWHPVZLOOQRORQJHUEH PRGHOHG DQG GHVLJQHG DV PHFKDQLFDO RU DUFKLWHFWXUDO V\VWHPV EXW UDWKHU DV SK\VLFDORULQWHQWLRQDOV\VWHPV:HWU\WRLGHQWLI\WKHLPSDFWRIVXFKDFKDQJHRI SDUDGLJPLQFRPSXWHUVFLHQFHDQGVRIWZDUHHQJLQHHULQJSUDFWLFHV



 ! " #

7KH IRXU FKDUDFWHULVWLFV LGHQWLILHG LQ WKH LQWURGXFWLRQ VLWXDWHGQHVV RSHQQHVV ORFDO FRQWUROORFDOLQWHUDFWLRQV DIIHFWPRVWRIWRGD\t VVRIWZDUHV\VWHPV $

   

7RGD\t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u ZUDSv WKH HQYLURQPHQW DQG PRGHO LW LQ WHUPV RI D

6LJQVRID5HYROXWLRQLQ&RPSXWHU6FLHQFHDQG6RIWZDUH(QJLQHHULQJ

QRUPDOFRPSRQHQWVRWKDWWKHHQYLURQPHQWLQLWVHOIGRHVQRWH[LVWDVDSULPDU\DE VWUDFWLRQ 8QIRUWXQDWHO\ WKH HQYLURQPHQW LQ ZKLFK FRPSRQHQWV OLYH DQG ZLWK ZKLFK WKH\LQWHUDFWGRHVH[LVWDQGPD\LPSDFWH[HFXWLRQDQGPRGHOLQJ •

6HYHUDO HQWLWLHV ZLWK ZKLFK VRIWZDUH FRPSRQHQWV PD\ QHHG WR LQWHUDFW DUH WRR FRPSOH[LQWKHLUVWUXFWXUHDQGEHKDYLRUWRHQDEOHDWULYLDOZUDSSLQJ

•

)RUDV\VWHPZKRVHJRDOLVWRPRQLWRUVHQVH DQGFRQWURODIIHFW DSK\VLFDORU ORJLFDO FRPSXWDWLRQDO  HQYLURQPHQW PDVNLQJ WKH HQYLURQPHQW LV QRW QDWXUDO ,QVWHDGSURYLGLQJDQH[SOLFLWFRQVFLRXVQHVVPD\EHDSULPDU\DSSOLFDWLRQJRDO

•

7KH HQYLURQPHQW FDQ KDYH LWV RZQ G\QDPLFV LQGHSHQGHQW RI D VRIWZDUH V\V WHPt V LQWULQVLF G\QDPLFV :UDSSLQJ WKH HQYLURQPHQW ZLOO LQWURGXFH XQSUHGLFW DEOHQRQGHWHUPLQLVPLQWKHEHKDYLRURIVRPHSDUWVRIRXUDSSOLFDWLRQV

)RUWKHVHUHDVRQVERWKFRPSXWHUVFLHQFHDQGVRIWZDUHHQJLQHHULQJQRZWHQGWRGH ILQH D V\VWHPt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xamples &RQWUROV\VWHPVIRUSK\VLFDOGRPDLQVHJPDQXIDFWXULQJWUDIILFFRQWUROKRPHFDUH KHDOWK FDUH  DUH RIWHQ EXLOW E\ H[SOLFLWO\ WDNLQJ LQWR DFFRXQW XQSUHGLFWDEOH HQYLURQ PHQWDOG\QDPLFVYLDVSHFLILFHYHQWKDQGOLQJRUVLPLODU SROLFLHV>@6LPLODUSURE OHPVDULVHLQVHQVRUDQGURERWQHWZRUNV>@ZKHUHPDQ\FRPSRQHQWVDUHVSUHDGLQ DQXQNQRZQHQYLURQPHQWWRH[SORUHDQGPRQLWRULW 0RELOHDQGSHUYDVLYHFRPSXWLQJUHFRJQL]HVWKHLPSRUWDQFHRIFRQWH[WDZDUHQHVV HJDSSOLFDWLRQVt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

)UDQFR=DPERQHOOLDQG+9DQ'\NH3DUXQDN

)LQDOO\VHYHUDOSURPLVLQJSURSRVDOVLQ GLVWULEXWHG SUREOHP VROYLQJ DQG RSWLPL]D WLRQ LH ZRUNV RQ DQWEDVHG FRORQLHV > @  H[SORLW D G\QDPLF YLUWXDO HQYLURQ PHQWLQIOXHQFLQJWKHDFWLYLWLHVRIGLVWULEXWHGSUREOHPVROYHUSURFHVVHV $

%  

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•

6LPSO\ HQDEOLQJ LQWHURSHUDELOLW\ LV QRW HQRXJK ZKHQ VRIWZDUH V\VWHPV PD\ FRPHWROLIHDQGGLHLQDQHQYLURQPHQWLQGHSHQGHQWO\RIHDFKRWKHURUZKHQD VRIWZDUHV\VWHPRULWVFRPSRQHQWV FDQH[SOLFLWO\PRYHDFURVVGLIIHUHQWHQYL URQPHQWVGXULQJLWVOLIH7KHVHFKDUDFWHULVWLFVLQWURGXFHDGGLWLRQDOSUREOHPV

•

:KHQFRPSRQHQWLQGLIIHUHQWVRIWZDUHV\VWHPVLQWHUDFWDQGZKHQFRPSRQHQWV PRYH DFURVV GLIIHUHQW HQYLURQPHQW LW PD\ EHFRPH KDUG LI QRW LPSRVVLEOH WR LGHQWLI\FOHDUO\WKHV\VWHPWRZKLFKDFRPSRQHQWEHORQJV,QRWKHUZRUGVLWEH FRPHVGLIILFXOWWRXQGHUVWDQGFOHDUO\WKHERXQGDULHVRIVRIWZDUHV\VWHPV

•

:KHQDFRPSRQHQWFRPHVWROLIHLQDQHQYLURQPHQWRULVVRPHKRZPRYHGWRD VSHFLILFHQYLURQPHQW LWPXVWVRPHKRZEHPDGHDZDUHRIZKDWLVDURXQGLQWKH HQYLURQPHQW ZKDW RWKHU FRPSRQHQWV DUH WKHUH IRU LQWHUDFWLRQ DQG KRZ LW FDQ LQWHUDFWZLWKQHZO\HQWHULQJV\VWHPVVDIHO\IRUERWKWKHV\VWHPVDQGLWVHOI

•

(QDEOLQJFRPSRQHQWVWRHQWHUDQGOHDYHDQHQYLURQPHQWLQDIXOO\IUHHZD\DQG LQWHUDFW ZLWK HDFK RWKHU PD\ PDNH LW YHU\ KDUG WR XQGHUVWDQG DQG FRQWURO WKH RYHUDOOEHKDYLRURIWKHVHVRIWZDUHV\VWHPVDQGHYHQRIDVLQJOHVRIWZDUHV\VWHP

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

6LJQVRID5HYROXWLRQLQ&RPSXWHU6FLHQFHDQG6RIWZDUH(QJLQHHULQJ

Examples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t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

' 

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

)UDQFR=DPERQHOOLDQG+9DQ'\NH3DUXQDN

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•

,QDQRSHQZRUOGDXWRQRP\RIH[HFXWLRQHQDEOHVDFRPSRQHQWWRPRYHDFURVV V\VWHPVDQGHQYLURQPHQWVZLWKRXWKDYLQJWRUHSRUWEDFNWRRUZDLWIRUDFNE\ LWVRULJLQDODSSOLFDWLRQ

•

:KHQFRPSRQHQWVDQGV\VWHPVDUHLPPHUVHGLQDKLJKO\G\QDPLFHQYLURQPHQW ZKRVHHYROXWLRQPXVWEHPRQLWRUHGDQGFRQWUROOHGDQDXWRQRPRXVFRPSRQHQW FDQEHHIIHFWLYHO\GHOHJDWHGWRWDNHFDUHRIDSRUWLRQRI WKHHQYLURQPHQWLQGH SHQGHQWO\RIWKHJOREDOIORZRIFRQWURO

•

6HYHUDO VRIWZDUH V\VWHPV DUH QRW RQO\ PDGH XS RI VRIWZDUH FRPSRQHQWV EXW DOVRLQWHJUDWHFRPSXWHUEDVHGV\VWHPVZKLFKDUHE\WKHLU YHU\ QDWXUH DXWRQR PRXVV\VWHPVDQGFDQEHPRGHOHGDFFRUGLQJO\

•

$V WKH VL]H RI D VRIWZDUH V\VWHP LQFUHDVHV ERWK SHUIRUPDQFH DQG FRQFHSWXDO VLPSOLFLW\UHTXLUHGHOHJDWLQJFRQWUROWRFRPSRQHQWV&RRUGLQDWLQJDJOREDOIORZ RIFRQWURODPRQJDYHU\ODUJHQXPEHURIFRPSRQHQWVPD\EHFRPHXQIHDVLEOH $XWRQRP\WKHQEHFRPHVDQDGGLWLRQDOGLPHQVLRQRIPRGXODULW\>@

2XUFRQFHSWRIDXWRQRP\UHIHUVQRWRQO\WRWKRVHVRIWZDUHFRPSRQHQWVWKDWDUHGH VLJQHGDVDXWRQRPRXVEXWDOVRWRWKRVHWKDWFDQEHSHUFHLYHGDVDXWRQRPRXV Examples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

6LJQVRID5HYROXWLRQLQ&RPSXWHU6FLHQFHDQG6RIWZDUH(QJLQHHULQJ

$(

'   

'LUHFWO\GHULYLQJIURPWKHVHWKUHHLVVXHVWKHFRQFHSWRIORFDOLQWHUDFWLRQVLQDJOREDO ZRUOGLVPRUHDQGPRUHSHUYDGLQJWRGD\ VVRIWZDUHV\VWHPV %\QRZZHKDYHGHOLQHDWHGDVFHQDULRLQZKLFKVRIWZDUHV\VWHPVFRPSRQHQWVDUH LPPHUVHGLQDVSHFLILFHQYLURQPHQWH[HFXWHLQWKHFRQWH[WRIDVSHFLILFVXE V\VWHP DQGDUHGHOHJDWHGWRSHUIRUPVRPHWDVNLQDXWRQRP\7DNHQDOOWRJHWKHUWKHVHDVSHFWV QDWXUDOO\OHDGWRDVWULFWHQIRUFHPHQWRIORFDOLW\LQLQWHUDFWLRQV,QIDFW •

$XWRQRPRXVFRPSRQHQWVFDQLQWHUDFWZLWKWKHHQYLURQPHQWLQZKLFKWKH\DUH LPPHUVHG E\ VHQVLQJ DQG DIIHFWLQJ LW ,I WKH HQYLURQPHQW LV SK\VLFDO WKH DPRXQWRIWKHSK\VLFDOZRUOG D VLQJOH FRPSRQHQW FDQ VHQVH DQG DIIHFW LV OR FDOO\ERXQGHGE\SK\VLFDOODZV,IWKHHQYLURQPHQWLVORJLFDOPLQLPL]DWLRQRI FRQFHSWXDODQGPDQDJHPHQWFRPSOH[LW\VWLOOIDYRUPRGHOLQJLWLQORFDOWHUPV DQGOLPLWLQJWKHHIIHFWRIDVLQJOHFRPSRQHQWRQWKHHQYLURQPHQW

•

&RPSRQHQWVFDQQRUPDOO\LQWHUDFWZLWKHDFKRWKHULQWKHFRQWH[WRIWKHVRIW ZDUH V\VWHP WR ZKLFK WKH\ EHORQJ WKDW LV ORFDOO\ WR WKHLU V\VWHP ,Q RSHQ ZRUNKRZHYHUDFRPSRQHQWRIDV\VWHPFDQDOVRLQWHUDFWZLWKFRPSRQHQWV RI RWKHUV\VWHPV,QWKHVHFDVHVPLQLPL]DWLRQRIFRQFHSWXDOFRPSOH[LW\VXJ JHVWV PRGHOLQJ WKH FRPSRQHQW LQ WHUPV RI D FRPSRQHQW WKDW KDV WHPSRUDULO\ PRYHGWRDQRWKHUFRQWH[WDQGWKDWLQWHUDFWVORFDOO\LQWKHQHZFRQWH[W>@

•

,QDQRSHQZRUOGFRPSRQHQWVQHHGVRPHIRUPRI FRQWH[WDZDUHQHVV WR H[H FXWHDQGLQWHUDFWHIIHFWLYHO\)RUDFRPSRQHQWWREHPDGHDZDUHRILWVFRQWH[W HIIHFWLYHO\DQGHIILFLHQWO\ WKLVFRQWH[WPXVWQHFHVVDULO\EHORFDOO\FRQILQHG

/RFDOLW\LQLQWHUDFWLRQVLVDVWURQJUHTXLUHPHQWZKHQWKHQXPEHURIFRPSRQHQWVLQ DV\VWHPLQFUHDVHVRUDVWKHGLPHQVLRQRIWKHGLVWULEXWLRQVFDOHLQFUHDVHV,QDQ\FDVH WUDFNLQJDQGFRQWUROOLQJFRQFXUUHQWDXWRQRPRXVDQGDXWRQRPRXVO\LQLWLDWHGLQWHUDF WLRQVLVPXFKPRUHGLIILFXOWWKDQLQREMHFWEDVHGDQG FRPSRQHQWEDVHG DSSOLFDWLRQV HYHQLIWKHVHLQWHUDFWLRQVDUHVWULFWO\ORFDO Examples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x DWDJLYHQWLPHx RQO\ZLWKDOLPLWHGSRUWLRQRIWKHZRUOG>@ $SSOLFDWLRQVGLVWULEXWHGLQWKH,QWHUQHWKDYHWRWDNHLQWRDFFRXQWWKHVSHFLILFFKDU DFWHULVWLFV RI WKH ORFDO DGPLQLVWUDWLYH GRPDLQ LQ ZKLFK LWV FRPSRQHQWV H[HFXWH DQG KDYHWRLQWHUDFWDQGFRPSRQHQWVDUHXVXDOO\DOORFDWHGLQ,QWHUQHWQRGHVVRDVWRHQ IRUFHDVPXFKDVSRVVLEOHORFDOLW\LQLQWHUDFWLRQV>@

)UDQFR=DPERQHOOLDQG+9DQ'\NH3DUXQDN

$)

*   + 

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

,$$7KH6FHQDULRRI0RGHUQ6RIWZDUH6\VWHPV

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

&

 % 

7UDGLWLRQDOO\VRIWZDUHV\VWHPVDUHPRGHOHGIURPDPHFKDQLFDOVWDQFHDQGHQJLQHHUHG IURP D GHVLJQ VWDQFH &RPSXWHU VFLHQWLVWV DUH ERWK EXUGHQHG DQG IDVFLQDWHG E\ WKH

6LJQVRID5HYROXWLRQLQ&RPSXWHU6FLHQFHDQG6RIWZDUH(QJLQHHULQJ

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

-     #

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

)UDQFR=DPERQHOOLDQG+9DQ'\NH3DUXQDN

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

6LJQVRID5HYROXWLRQLQ&RPSXWHU6FLHQFHDQG6RIWZDUH(QJLQHHULQJ

&$

-      #

7KHFKDQJHLQWKHPRGHOLQJDQGXQGHUVWDQGLQJRIFRPSOH[VRIWZDUHV\VWHPVZLOODOVR LPSDFWKRZVXFKV\VWHPVDUHGHVLJQHGPDLQWDLQHGDQGWHVWHG 7RGD\VRIWZDUHV\VWHPVDUHGHVLJQHGWRH[KLELWVSHFLILFSUHGLFWDEOHDQGGHWHUPL QLVWLFEHKDYLRUDWDQ\OHYHOIURPVLQJOHXQLWVXSWRWKHZKROHV\VWHP+RZHYHULQWKH SUHVHQFH RI DXWRQRPRXV FRPSRQHQWV VLWXDWHG LQ DQ RSHQ DQG G\QDPLF HQYLURQPHQW REWDLQLQJSUHGLFWDEOHEHKDYLRURIDPXOWLFRPSRQHQWV\VWHPLQPHFKDQLFDODQGGHWHU PLQLVWLFWHUPVLVQRWIHDVLEOH7KHQH[WFKDOOHQJHIRUWKHHIIHFWLYHFRQVWUXFWLRQRIODUJH VRIWZDUHV\VWHPVRYHUFRPLQJWKHLPSRVVLELOLW\RIFRQWUROOLQJWKHEHKDYLRURIHDFKRI LWVVLQJOHFRPSRQHQWVDQGWKHLULQWHUDFWLRQVLVWREXLOGLWVRDVWRJXDUDQWHHWKDWWKH V\VWHPDVDZKROHZLOOEHKDYHUHDVRQDEO\FORVH WR DQ LGHDO GHVLUHG EHKDYLRU GHVSLWH WKH ODFN RI H[DFW NQRZOHGJH DERXW LWV PLFUREHKDYLRU )RU LQVWDQFH E\ DGRSWLQJ D SK\VLFDO DWWLWXGH WRZDUG VRIWZDUH GHVLJQ D SRVVLEOH DSSURDFK FRXOG EH WR EXLOG D V\VWHPWKDWGHVSLWHXQFHUWDLQW\RQWKHLQLWLDOFRQGLWLRQVLVDEOHWRUHDFKDJLYHQVWDEOH EDVLQRIDWWUDFWLRQ%\DGRSWLQJDWHOHRORJLFDODWWLWXGHRQHFRXOGEXLOG DQ HFRV\V WHP RU D VRFLHW\ RI FRPSRQHQWV DEOH WR EHKDYH LQ DQ LQWHQWLRQDO ZD\ DQG UREXVW HQRXJKWRGLUHFWLWVJOREDODFWLYLWLHVWRZDUGWKHDFKLHYHPHQWRIWKHUHTXLUHGJRDO 7KHGHVLJQRIDV\VWHPPXVWDOVRH[SOLFLWO\WDNHLQWRDFFRXQWWKHIDFWWKDWWKHV\V WHPZLOOEHLPPHUVHGLQDQRSHQDQGG\QDPLFHQYLURQPHQWDQGWKDWWKHEHKDYLRURI WKHV\VWHPFDQQRWEHGHVLJQHGLQLVRODWLRQ5DWKHUWKHHQYLURQPHQWDQGLWVG\QDPLFV DVZHOODVWKRVHVRIWZDUHFRPSRQHQWVWKDWFDQHQWHUDQGOHDYHWKHV\VWHPVPXVWEH FRPHSULPDU\FRPSRQHQWVRIWKHGHVLJQDEVWUDFWLRQV$defensiveGHVLJQWUHDWVWKHVH IDFWRUVDVVRXUFHVRIXQFHUWDLQW\WKDWFDQVRPHKRZEHGDPDJLQJWRWKHJOREDOEHKDYLRU RIDV\VWHPDQGWKDWWKHV\VWHPVKRXOGEHSUHSDUHGWRIDFH$Qoffensive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

)UDQFR=DPERQHOOLDQG+9DQ'\NH3DUXQDN

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u JRRGv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

6LJQVRID5HYROXWLRQLQ&RPSXWHU6FLHQFHDQG6RIWZDUH(QJLQHHULQJ

6RIWZDUHPDLQWHQDQFHZLOOFKDQJHWRR$XWRQRP\DQGRSHQQHVVPHDQWKDWQRVRIW ZDUHV\VWHPZLOOEHVWDEOHEXWWRVRPHH[WHQWQHHGVPDLQWHQDQFHFRQWLQXDOO\GXHWR FKDQJHGFRQGLWLRQV:KHQDODUJHVRIWZDUHV\VWHPQRORQJHUEHKDYHVDVQHHGHGHJ ZKHQH[WHUQDOFRQGLWLRQVUHTXLUHFKDQJHV XSGDWHZLOOQRORQJHULPSO\VWRSSLQJWKH V\VWHPUHEXLOGLQJLWDQGUHWHVWLQJLW,QVWHDGLWZLOOLPSO\LQWHUYHQLQJH[WHUQDOO\E\ DGGLQJQHZFRPSRQHQWVZLWKGLIIHUHQWDWWLWXGHVDQGE\UHPRYLQJVRPHRILWVH[LVWLQJ FRPSRQHQWVVRDVWRFKDQJHDVQHHGHGWKHRYHUDOOEHKDYLRURIWKHV\VWHP7KHRSHQ QHVVDQGVLWXDWHGQHVVRIWKHV\VWHPDQGWKHDXWRQRP\RILWVFRPSRQHQWVFDQDOVRPDNH PDLQWHQDQFHDQGXSGDWLQJVPRRWKHUDQGOHVVH[SHQVLYHLQWKDWWKHV\VWHPLVDOUHDG\ GHVLJQHGWRWROHUDWHDQGVXSSRUWG\QDPLFFKDQJHVLQLWVVWUXFWXUH &$&

/

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u E\GHVLJQv DVRIWZDUHV\VWHPPD\VLPSO\PDNHLWORVHWKHSURSHUWLHVQHHGHG WRVXSSRUWRSHQQHVVDQGWRWROHUDWHHQYLURQPHQWDOG\QDPLFV ,QDQ\FDVHZHDUHDZDUHWKDWIRURXUYLVLRQRQVRIWZDUHHQJLQHHULQJWR EHFRPH SUDFWLFDO DQG XVHG PXFK WKHRUHWLFDO PHWKRGRORJLFDO DQG H[SHULPHQWDO ZRUN LV UH TXLUHGSDYLQJWKHZD\IRUDVXLWDEOHVHWRIFRQFHSWXDOWRROVDQGIUDPHZRUNVWREHH[ SORLWHGE\WKHHQJLQHHUVRIQH[WJHQHUDWLRQVRIWZDUHV\VWHPV

(



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

)UDQFR=DPERQHOOLDQG+9DQ'\NH3DUXQDN

WKHGRRUWRQHZLQWHUHVWLQJUHVHDUFKDQGHQJLQHHULQJFKDOOHQJHVLQZKLFKKRSHIXOO\ WKHPXOWLDJHQWUHVHDUFKFRPPXQLW\ZLOOKROGDOHDGLQJSRVLWLRQ

    >@ >@ >@ >@ >@ >@ >@ >@ >@ >@ >@ >@ >@ >@ >@

+$EHOVRQ'$OOHQ'&RRUH&+DQVRQ*+RPV\7.QLJKW51DJSDO (5DXFK*6XVVPDQDQG5 :HLVV u $PRUSKRXV &RPSXWLQJv  &RPPXQLFD WLRQVRIWKH$&0 0D\ * ' $ERZG ( ' 0\QDWW u &KDUWLQJ 3DVW 3UHVHQW DQG )XWXUH 5HVHDUFK LQ 8ELTXLWRXV&RPSXWLQJv $&07UDQVDFWLRQVRQ&RPSXWHU+XPDQ,QWHUDFWLRQ  0DUFK :$GMLH:LQRWR(6FKZDUW]+%DODNULVKQD-/LOOH\u 7KH'HVLJQDQG,P SOHPHQWDWLRQ RI DQ ,QWHQWLRQDO 1DPLQJ 6\VWHPVv  WK $&0 6\PSRVLXP RQ 2SHUDWLQJ6\VWHPV3ULQFLSOHV6263  $&0 5$OEHUW+-HRQJ$%DUDEDVLu 'LDPHWHURIWKH:RUOG:LGH:HEv 1DWXUH 6HSW 5 $OEHUW + -HRQJ $ %DUDEDVL u (UURU DQG $WWDFN 7ROHUDQFH RI &RPSOH[ 1HWZRUNVv 1DWXUH-XO\ -%URFK'$0DOW]'%-RKQVRQ@ >@ >@ >@

>@ >@ >@ >@ >@ >@ >@ >@ >@ >@ >@ >@ >@ >@

$+RZDUG0-0DWDULFu &RYHU0H$6HOI'HSOR\PHQW$OJRULWKPIRU0R ELOH6HQVRU1HWZRUNVv ,QWHUQDWLRQDO&RQIHUHQFHRQ5RERWLFVDQG$XWRPDWLRQ WRDSSHDU %$+XEHUPDQ7+RJJ7KH (PHUJHQFH RI &RPSXWDWLRQDO (FRORJLHV LQ /HFWXUHVLQ&RPSOH[6\VWHPV$GGLVRQ:HVOH\ 0+XKQV,QWHUDFWLRQ2ULHQWHG3URJUDPPLQJVW,QWHUQDWLRQDO:RUNVKRSRQ $JHQW2ULHQWHG6RIWZDUH(QJLQHHULQJ/1&61R-DQ & ,QWDQDJRQZLZD\ 5 *RYLQGDP ' (VWULQ u 'LUHFWHG 'LIIXVLRQ D 6FDODEOH DQG 5REXVW &RPPXQLFDWLRQ 3DUDGLJP IRU 6HQVRU 1HWZRUNVv  WK $&0,((( &RQIHUHQFHRQ0RELOH&RPSXWLQJDQG1HWZRUNLQJ%RVWRQ0$ $XJ SS 1 5 -HQQLQJV $Q $JHQW%DVHG $SSURDFK IRU %XLOGLQJ &RPSOH[ 6RIWZDUH 6\VWHP&RPPXQLFDWLRQVRIWKH$&0  7.XKQ7KH6WUXFWXUHRI6FLHQWLILF5HYROXWLRQV8QLYHUVLW\RI&KLFDJR3UHVV UG(GLWLRQ1RY 00DPHL//HRQDUGL)=DPERQHOOLu $3K\VLFDOO\*URXQGHG$SSURDFKWR &RRUGLQDWH 0RYHPHQWV LQ D 7HDPv  VW ,QWHUQDWLRQDO :RUNVKRS RQ 0RELOH 7HDPZRUNDW,&'&6,(((&63UHVV-XO\ :*0LWFKHQHU$9DKGDWu (SLGHPLF5RXWLQJIRU3DUWLDOO\&RQQHFWHG$G +RF1HWZRUNVv 'XNH7HFKQLFDO5HSRUW1R&6-XO\ @

$QWXQHV / DQG + &RHOKR  5DWLRQDOLVLQJ GHFLVLRQV XVLQJ PXOWLSOH YDOXHV0 

  1  2  2  3  6LHQD ,WDO\ >@ $QWXQHV / DQG + &RHOKR  'HFLVLRQV EDVHG XSRQ PXOWLSOH YDOXHV WKH %9*DJHQWDUFKLWHFWXUH0 

  0    42  10455 6SULQJHU9HUODJ /1$, /HFWXUH 1RWHV LQ $UWLILFLDO ,QWHOOLJHQFH  Qœ   SS  YRUD3RUWXJDO >@ %HOQDS 1  %DFNZDUGV DQG )RUZDUGV LQ WKH 0RGDO /RJLF RI $JHQF\ 3KLORVRSK\DQG3KHQRPHQRORJLFDO5HVHDUFKYRO >@ &DVWHOIUDQFKL &  &RPPLWPHQWV IURP ,QGLYLGXDO ,QWHQWLRQV WR *URXSV DQG2UJDQLVDWLRQV,Q:RUNLQJ1RWHVRI$$$,t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

-RDTXLP)LOLSH

>@ 5DR $ DQG 0 *HRUJHII  0RGHOLQJ 5DWLRQDO $JHQWV ZLWKLQ D %', DUFKLWHFWXUH0 

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

Simulating Computational Societies Lloyd Kamara, Alexander Artikis, Brendan Neville, and Jeremy Pitt Intelligent & Interactive Systems Group, Imperial College Electrical Engineering Dept., London, UK, SW7 2BT Tel.: +44 (0)20 7594 6221 {l.kamara,a.artikis,brendan.neville,j.pitt}@ic.ac.uk

Abstract. Multi-agent systems can be considered from a variety of perspectives. One such perspective arises from considering the architecture of an agent itself. Another is that of an instantiated agent architecture and its interaction with its peers in a MAS. A third perspective is that of an external observer. These three perspectives cover a potentially overlapping but essentially distinct set of issues concerning MAS simulation and modelling. In this paper, we consider each of these perspectives in turn and demonstrate how a simulation framework can support a collective treatment of such concepts. We discuss the implications for agent development and agent society design arising from the results and analysis of our simulation approach.

1

Introduction

Multi-agent systems (MAS) can be considered from a variety of perspectives. One such perspective arises from considering the architecture of an agent itself (e.g. [1]). This is relevant to agent designers and implementors, as it concerns the internal operation of an agent. The basic architectural properties largely determine how the agent will interact with its environment. Another perspective to consider is that of an instantiated agent architecture and its interaction with its peers in a MAS. This covers both communicative and socio-cognitive aspects of agent interaction (e.g. [2, 3, 4]). In addition to being relevant to agent designers and implementors, this can be used to model, analyse and explain the behaviour of the agents in terms of social theories (e.g. [3, 4]). The introduction of sociocognitive elements to agent reasoning allows interactions to be characterised in increasingly anthropomorphic terms. A third perspective is that of an external observer (e.g. [5, 6, 7, 8]). Under such circumstances, we adopt a bird’s eye view of computational systems — we are not concerned with the internal architecture of the participating agents. We can specify the legal and social aspects of these systems, such as the normative positions and institutional powers of the agents, without making any assumptions about mentalistic concepts like beliefs, desires or intentions. We aim to enable society designers and agent developers to view simulated societies both at a micro (i.e. from an intensional perspective) and a macro

P. Petta et al. (Eds.): ESAW 2002, LNAI 2577, pp. 53–67, 2003. c Springer-Verlag Berlin Heidelberg 2003 

54

Lloyd Kamara et al.

(i.e. from an extensional perspective) level (as is the case with [9] and [7])1 . To this end, we propose simulation tools which address concerns and requirements identified at both (and intermediate) levels. We refer to these tools collectively as our simulation framework. Pitt et al. [11] describe an abstract producer/consumer (APC) scenario where producers sell information to consumers. In this scenario the producers are explorer agents that map out the distribution of oil in their environment and consumers are cartographer agents that initiate contract-net protocols [12] (CNP) to acquire the maps from the explorers. We have used variants of the CNP in Section 3 and Section 4 as vehicles for analysing trust, experience and reputation from the socio-cognitive (micro) perspective; and power, permission and obligation from the external (macro) perspective. The rest of this paper is organised as follows. In Section 2, we present MAS and agent simulation architectures. In Section 3, we describe an approach to trust and reputation modelling in MAS. In Section 4, we consider executable specifications of norm-governed computational societies in a manner independent of agent internals. In Section 5, we relate our findings to agent development and agent society design endeavours, concluding that they offer some preliminary guidelines to agent implementors and society developers for scoping social MAS.

2

Agent and MAS Simulation Architectures

The basis of our MAS simulation architecture is a collection of agents whose interaction is governed by a communications interface and an accompanying set of protocols. The MAS is neutral with respect to the architecture of participating agents: no assumptions are made about internal operation and motivations of an agent by its peers. Interaction between two agents takes place through their communications interfaces, either creating or extending a communicative context, or conversation. 2.1

Communications Interface

The communications interface consists of a higher and lower-level component pair: the Agent Communication Language (ACL) and a socket-based module for TCP/IP transmission of messages. The former defines (amongst other things) the supported syntax of messages while the latter defines how messages are exchanged. Following [2], we define an ACL in terms of three components: a content language, which defines the syntax of messages, a set of protocols which define patterns of interaction and a reply function which provides an external semantics for the ACL. The content language is the set of performatives that agents use in communication, while the protocols identify sequences of performatives 1

We also note that these intensional and extensional perspectives are respectively related to the subjective and objective approaches to multi-agent co-ordination dicussed in [10].

Simulating Computational Societies

55

that constitute meta-level interaction (such as auction and query/response protocols). Given an input performative, protocol and current conversation state, the reply function returns the set of ‘acceptable’ responses. These take the form of performative-protocol pairs, which, depending on the options available, may enable a responding agent to continue the conversation using the same protocol, or to initiate an auxiliary conversation using a different protocol. The relationship between content language, protocols and reply function is specified in a separate (Prolog) module, making it straightforward to refine and enhance the communicative behaviour of agents. Pitt and Mamdani [2] justify the use of a protocol-based semantics to cover the external aspects of agent interaction in MAS featuring agents with heterogeneous architectures and distinct reasoning and motivational traits. A similar justification is used here, with the introduction of a socio-cognitive perspective arguably serving to re-inforce the need for such an approach. The additional subjective treatment of concepts including trust and reputation makes each agent even more distinct and thus external characterisations of interactions become increasingly relevant. In the following section, we describe an agent architecture that can be parameterised with different reasoning and socio-cognitive behaviour, allowing us to represent heterogeneous MAS in our experiments while retaining a compact simulation base. 2.2

Agent Architecture

In our proposed experimental configuration, a generic BDI-based [1] architecture forms the computational base for each agent. We use the same architecture for reasons of simplicity: the principle of heterogeneity outlined earlier has not been abandoned. It would be equally possible to use agents with different architectures as long as they possess compatible interfaces (like in [13, 14]). Specific behaviour is determined by parameterisation of the respective agent instances. This approach can be used, for example, to pre-assign the roles of auctioneer and bidders in a group of agents enacting an electronic auction. Figure 1 gives a diagrammatic overview of the agent architecture. The control module contains the agent interface and interpreter, in addition to housing the protocols sub-module. The agent interface facilitates and manages conversations (as mentioned previously), representations of which are to be found in the conversational component of the agent’s mental state. The interpreter serves as the animating core of the agent architecture, operating in the form of a typical BDI cycle. It processes incoming messages at the content level, consulting and updating mental state representation while executing communicative actions appropriate to the agent’s recent perception and interpretation of events. The mental state representation consists of several elements (see Figure 1). The intentional state holds the agent’s belief and motivational state alongside the agent’s current intentions. These are expressed as time-stamped Prolog terms with assertion being the means of updating state content. Records of conversations — again, represented through Prolog terms — are maintained in the

56

Lloyd Kamara et al.

server socket

intentional state socio−cognitive characteristics

control module get msgs reply

interpreter

agent interface

bi−directional connections to other agent processes

mental state

protocols

conversational state

user interface

initial behaviour

user

trust model graphing utility

Fig. 1. Generic agent architecture for experiments

conversational state component. Each conversation is recorded as a four-tuple denoting the correspondent (the other agent in the conversation), local and remote conversation identifiers [2] and the current state of the conversation (from a local perspective). A sub-set of the contained data reflects proceedings in current, or active, conversations. The user interface provides a means for the experimenter to initialise the agent and to monitor and control its behaviour. The architecture does not feature a fixed procedure for belief update at present, although the way has been prepared for the introduction of such a mechanism. In particular, with each Prolog term used to represent a belief there is an associated credence measure (in the range of 0–1) which reflects the level of certainty the agent has in the content.

3

Socio-cognitive Modelling

Our investigation of socially motivated behaviour is based on formal models of anthropomorphic socio-cognitive relations. Accordingly, our socio-cognitive model defines three component beliefs of each agent about its peers in the society: trust, direct experience and reputation. This section defines the sociocognitive model and the software developed to facilitate experimentation with agent societies whose members are implemented with a (parameterised) version of the model. Our computational representation of trust is based on the formal model of Castelfranchi and Falcone [3, 4]. The essential conceptualisation is as follows: the degree to which Agent A trusts Agent B about task τ in Ω (state of the world) is a subjective probability; this is the basis of agent A’s decision to rely upon B for τ . Our method incoporates this stance, defining trust as the resultant belief of one agent about another, borne out of direct experience of that other party and/or from the testimonies of peers (i.e. reputation).

Simulating Computational Societies

57

We define direct experience as the belief one agent has about the trustworthiness of another, based on first-hand interactions. Having ‘trusted’ another agent to perform task τ and assessed the outcome, an agent will update its direct experience beliefs concerning the delegated agent accordingly. The outcome of delegating a task may be either successful or unsuccessful: this categorisation is the basis of an experience update rule [15] that calculates the revised level of trustworthiness to associate with the agent in question. Our motivation in investigating reputation mechanisms by simulation is provided by Conte [16], who outlines the need for ‘decentralised mechanisms of enforcement of social order’ noting that ‘reputation plays a crucial role in distributed social control’. In an agent society, reputation is the collectively informed opinion held by a group of agents about the performance of a peer agent within a specific social context. By consulting its peers, an agent can discover the individual reputation of an agent. However, the received testimonies may be affected by existing relationships and attitudes (for example, agent C may be willing to divulge reputation information to agent A but not to agent B). Thus, reputation is also a subjective concept which we define as a belief held/derived by one agent. The Subjective Reputation Evaluation Function (SREF) formulates subjectively — that is, from the perspective of an agent A — the reputation of an agent B, based on information from the peers of A and B. SREF may take several forms; examples are a weighted sum or a fuzzy relation. In our current work, we use the former approach; a summing function in which n assertions received from peers regarding agent B are weighted by the agent’s trust (confidence) in each peer to make accurate recommendations. This approach incorporates the credibility of each assertion’s source (the credibility of a belief being dependent upon the credibility of its sources, evidences and supports [3]). We briefly address a proposed method of combining direct experience with reputation to formulate the mental state of trust. Each belief has associated with it a degree of confidence signifying an agent’s trust in the belief’s accuracy. This confidence measure is essential to what trust will be based on, be it one or indeed both of the beliefs, experience or reputation. An agent with strong confidence in its experience beliefs and little confidence in the accuracy of its reputation beliefs should rationally choose to calculate its degree of trust (DoT) primarily from its experiences. An agent recently introduced to the system should have little confidence in its own (in)experience and should thus base its trust solely on reputation. The number of direct experiences is therefore significant in evaluating confidence in a belief about direct experience. It is similarly the case for reputation, where the number of agreeing testimonies is a decisive factor. Experience and reputation thus become influences of trust, the weighting assigned to them dependent upon an agent’s confidence in their respective accuracies. 3.1

The Trading Economy

We simulate our commerce society as trading agents executing a variant of the CNP [12]. In this variant, the corresponding communicative acts for informing of the successful completion of a task and making payment are sent without any

58

Lloyd Kamara et al.

intermediate evaluation — that is, the cartographer (manager) does not know how well (if at all) a task has been performed before it sends a payment in response to a contracted explorer’s (bidder) task completed message. In effect, we have a CNP concluding with a prisoner’s dilemma [17] (PD). Each agent’s actions at any stage will be influenced by expectations about the actions of its trading counterpart. The agents are able to execute the CNP repeatedly, forming the basis for an iterated PD. This differs from the game theoretic studies in [17], however, in that agents choose the peers with which they are willing to enter the PD on a trust/cost trade-off basis. The reasoning behind which action to take once a contract is established is similarly distinguished. An explorer agent is able to vary bidding price from one CNP instance to another. This could be used to increase the desirability of an agent’s bid, by under-cutting the bids of other explorers held in higher esteem. Agents not being awarded contracts may lower their prices in an attempt to generate business. Once a loyal trade relationship has been established, an explorer agent can gradually increase its bidding price in order to increase profitability. When a cartographer awards a contract to a specific explorer, a PD-style dependence is obtained. The explorer is relying upon the cartographer to make prompt payment; the cartographer is relying upon the explorer to provide the required information. Our system however, does not guarantee that agent actions will always have the intended effect. Agents may fail to complete a task for reasons other than conscious deception. The inclusion of this fallibility factor is to account for unpredictability and unreliability within MAS environments and real-world applications. 3.2

Trust-Based Prediction

Here, we discuss the use of a trust mechanism as a predictor for future peer behaviour. In this context, we interpret trust as a subjective probability that a peer will take a certain action, the outcome of which will characterise the corresponding interaction context. For each outcome, the agent will experience profit or loss to a varying extent — we refer to this as the outcome’s pay-off. Knowing the pay-offs of each outcome in advance and having an estimate of their probabilities of occurrence (trust), the agent and calculate the expected pay-off of each of its options. How the agent determines trading partners and what action to take (co-operate or defect ) once a contract is established depends upon its character type and the expected pay-offs of each possible action. We currently simulate three basic character types. Co-operators only choose co-operation strategies; if the expected outcome for co-operation is not greater than zero, a co-operator will not enter into a contract. Defectors will pursue a policy of co-operation or defection purely based upon maximising the expected pay-off of an interaction. Reciprocators have probabilistic intentions; the probability that they will co-operate (their trustworthiness) is matched to their evaluation of a trading partner’s behaviour (DoT in the peer). As a result, their trustworthiness displayed towards competent co-operators is high, whereas they are likely to defect against defectors and incompetent agents in general.

Simulating Computational Societies

3.3

59

Agent Configuration and Monitoring

The Multiple Agent Launcher (MAL) is a software tool that allows experimenters to create and launch an arbitrary number of agents with distinct configuration parameters. Agents may be launched individually or in groups. The configuration parameters are expressed as arguments to the agent architecture presented in Section 2 and range from low-level (e.g. agent cycle delay time) to high-level (e.g. file-name of Prolog source for run-time behaviour). The motivation behind the MAL is to provide the user with a single interface from which to configure and monitor agent experiments, as well as to visualise real-time agent interactions. It is important to note that the agents thus launched remain semi-autonomous processes; the MAL is not a centralised control mechanism, but behaves as a convenient abstraction of one. The MAL supports configuration of agents with heterogeneous socio-cognitive characteristics and behavioural traits, by allowing the experimenter to adjust the parameters of the trust and reputation update functions (for example). The output of these functions is logged at various intervals during a simulation run and can also be graphed in real-time. 3.4

Simulation and Results

We now discuss an example simulation for trust update based on direct experience. The experiment consists of twelve agents, split evenly between managers and bidders. All of the agents are of reciprocator interaction type, differing only in their ability to complete the tasks relied upon by their respective contract partners. This ability is manifested through a probability parameter which determines the rate of successful task completion per agent. In this experiment, three ability levels were specified: 95%, 75% and 50%. Two cartographers and two explorers were all assigned the same probability parameter for each of the three levels, resulting in a tripartite ability configuration. When an agent fails to successfully perform its part of the contract, its partner perceives this and consequently reacts as if it were a conscious act of defection. In the graphs of Figure 2 and Figure 3, each line reflects the mean data values belonging to a pair of equally skilled agents of the same type. The mean data readings for two explorer agents whose degree of ability is 95%, for example, is labelled ‘Explorers (95%)’. Time-points occur at ten second intervals; the current trust beliefs and monetary worth of the agents are logged at each point. Figure 2 plots the level of trust directed towards different members, over the time of the simulation. The featured DoT is simply the mean of the opinions of all the members in the society who know the agent in question. Agents are configured to initially blindly trust their peers (DoT = 1.0). As the simulation progresses their opinions change dramatically — in particular, there is a drop in communal trust for the capable agents. This is because the incompetent agents do not trust the competent agents, while the capable agents reciprocate for the incompetents’ failures by defecting against them. The competent agents have a very strong trust relationship among themselves but a very poor one with others — the average is

Average Degree of Trust (Direct Experience)

60

Lloyd Kamara et al. 1

0.8 Explorers (95%) 0.6

Cartographers (95%) Explorers (75%)

0.4

Cartographers (75%) Explorers (50%)

0.2 Cartographers (50%) 0 0

250

500

750

1000 1250 1500 1750 2000 2250 2500 2750 3000 Timepoint (10×Seconds)

Fig. 2. Experimental results - Trust evolution

therefore lowered. Our next graph (Figure 3) shows the performance measure, the amount of assets accrued or lost. From this it can be seen that a society of reciprocators is a meritorious one. The highly capable agents (95%) are able to succeed; agents of intermediate skill (75%) break roughly even while the least skilled agents (50%) are punished. Although our presented experiments have not covered reputation, we intend to address this in future work, as we believe it is an important aspect of agent interaction. In particular, we note how reputation is related to the notion of scalability: in a system with a large number of agents, the likelihood of any individual agent having interacted with a specific peer is less. This leads to an overall reduction in the amount of direct experience available to an agent during its decision-making. We are investigating when and how agents (in their capacity as third parties for reputation information transmission) can effectively

2500

Explorers (95%) Cartographers (95%)

Assets Value

2000

1500

1000 Cartographers (75%) 500

Explorers (75%) Cartographers (50%) Explorers (50%)

0 0

250

500

750

1000 1250 1500 1750 2000 2250 2500 2750 3000 Timepoint (10×Seconds)

Fig. 3. Experimental results - Monetary performance

Simulating Computational Societies

61

propagate information [16] under such circumstances. It is expected that the results of further experiments will establish what parameters and socio-cognitive characteristics lead to ‘stable’ trust relationships; ones that are resistant to minor aberrations in agent behaviour due to circumstances beyond their control.

4

Executing the Specifications of Computational Societies

Artikis et al. [5] present a theoretical framework for providing executable specifications of particular kinds of multi-agent systems, called open computational societies. Open computational societies (as defined in [5]) have the following characteristics: first, the internal architecture of the agents is not publicly known, i.e. there is no direct access to an agent’s mental state. Second, the members of the society do not necessarily share a notion of global utility. Members may fail to, or choose not to, conform to specifications in order to achieve their individual goals. In addition to these properties, in open societies ‘the behaviour of the members and their interactions cannot be predicted in advance’ [18]. In this framework the computational systems are viewed from an external perspective, that is to say, the internal architecture of the agents is not considered. Three key components of computational systems are specified, namely the social constraints, social roles and social states. The specification of these concepts is based on and motivated by the formal study of legal and social systems (a theory of institutionalised power [19] and a theory of normative positions [20]) and traditional distributed computing techniques (state transition systems). The social constraints and roles have been specified with two different formalisms from AI: the Event Calculus [21] (see [5]) and the C + language [22] (see [6]). Next, we briefly discuss the way social constraints are specified in [5, 6]. Then we present a computational framework that executes the specifications (i.e. the social constraints) of the agent societies during their execution. 4.1

Social Constraints

Artikis et al. [5, 6] follow Jones and Sergot [19] and maintain the standard, in the study of social and legal systems, long established distinction between permission, physical capability and institutionalised power. The social constraints specify: – What kind of actions ‘count as’ [19] valid (‘effective’) actions. Distinguishing between valid and invalid actions enables the separation of meaningful from meaningless activities. – What kind of actions (valid, invalid) are permitted. Determining the permitted, prohibited, obligatory actions enables the classification of the agent behaviour as ‘legal’ or ‘illegal’, ‘social’ or ‘anti-social’, etc. – What are the sanctions and enforcement policies that deal with ‘illegal’, ‘anti-social’ behaviour.

Lloyd Kamara et al.

agent

agent

observer

62

NARRATIVE: t:action1. t:action2. …

SOCIETY VISUALISER

CURRENT SOCIAL STATE: t+1:-pow(agentA, action1). t+1:pow(agentB, action3). t+1:permitted(agentB,action3). …

SOCIAL CONSTRAINTS: i:pow(Agent, Act) => i:permitted(Agent, Act) i:action1=> i+1:pow(agentB,action3) …

Fig. 4. The Society Visualiser

Valid actions are specified as follows: An action ‘counts as’ a valid action at a point in time if and only if the agent that performed that action had the institutionalised power [19] to perform it at that point in time. Differentiating between valid (‘meaningful’) and invalid (‘meaningless’) actions is of great importance in the analysis of agent systems. For example, in an auction, the auctioneer has to determine which bids are valid and therefore, which bids are eligible for winning the auction. 4.2

The Society Visualiser

In this section we describe a software platform called Society Visualiser (SV), that builds upon the theoretical framework for the specification of open systems [5] and that compiles (produces) and updates the global state of an agent society. The global state includes information like the institutional powers, permissions, obligations, sanctions and social roles of the members. Figure 4 describes the architecture of the SV in terms of inputs/outputs. In order to produce the global state of the societies, the SV has two inputs: a narrative and the social constraints 2 . The narrative is a description of the externally observable events that take place in a computational society. For example, t : action1 states that action1 took place at time point t (Figure 4). The narrative is produced in the following manner: When a member of the society sends a message to a peer, he also sends that message to an observer agent for monitoring purposes3 . These messages are called report messages and are necessary for the compilation of the social states. The report messages (and all other monitored events) consist of the narrative. In other words, the SV observes (without intervening) the interactions of the members of the computational societies in order to produce the social states of the societies. The social constraints specify the valid actions, institutional powers, permissions and so on. of the members of the society. Consider 2

3

The social constraints and the narrative are expressed in terms of either the Event Calculus or the C+ language. However, in order to simplify our analysis, we illustrate the social constraints and narrative in Figure 4 with the use of a generic notation. The observer agents monitors the execution of the members for the benefit of the Society Visualiser.

Simulating Computational Societies

Starting/ stopping the compilation of social states

Agent description in terms of: roles, powers, permissions, obligations, sanctions, valid actions. If the environment were selected at the list of agents then this table would show a description of the environment

Viewing either the social states or the narrative of the current time-point Current timepoint of the simulation. The user may scroll to any time-point in the simulation.

63

List of the members of the society and the environment.

Agent (or environment) description in the form of Prolog lists.

Fig. 5. The main GUI of the Society Visualiser

the following constraint (expressed in a generic notation): i : pow(Agent, Act) ⇒ i : permitted(Agent, Act) The above constraint states that if at time point i an agent is ‘empowered’ [19] to perform an action (represented by pow (Agent , Act )) then this agent is also ‘permitted’ to perform the same action (represented by permitted (Agent , Act )). The remaining constraints are specified in a similar fashion. As mentioned earlier, the output of the SV is the global state of the society at a point in time. In other words, given the narrative of time point t (say), the SV will produce the states of affairs that hold at time t + 1 . Global states are stored in a database and are displayed in a graphical interface for the benefit of the society designer. This graphical display the represents the following information about each agent: (i) institutional powers, (ii) permissions, (iii) obligations, (iv) sanctions and (v) valid actions. Figure 5 demonstrates the GUI of the SV during the simulations of a variation of the CNP (see [5] for more details). Global states are produced for the benefit of the members of the societies as well. Agents can send query messages to the SV in order to find out the social state of the group or of their peers. Apart from producing the global state of each time point, the SV includes a number of additional capabilities. One of these capabilities is planning. Given an initial state and a goal state (and the set of social constraints), the SV can determine if there exists a sequence of actions that will lead from the initial to the goal state. Due to the high complexity of such tasks, computation of planning queries can be mainly performed in an off-line phase, that is before the commencement of the interactions of the agent societies. Such a functionality enables the society designer to prove various properties of the agent systems and, therefore, evaluate the design of the social constraints. For example, having specified the constraints of an auction protocol, the designer can determine with the use of planning queries if it is possible to reach a state where

64

Lloyd Kamara et al.

two different agents have won the auction. Currently, this functionality is only provided if the social constraints are expressed by means of the C+ language4 .

5

Concluding Remarks

We presented a simulation framework that addresses the internal architecture of the agents in addition to accounts of the trust relationships of the agents and of the institutional and social aspects of agent systems. We intentionally did not address/fully investigate issues like low-level details of the simulation framework, experimentation, scalability and other criteria discussed in the community for the experimental study of MAS (see for example [13]). Our aim is to stress the need for the representation of conceptually different perspectives in the modelling and simulation of computational societies. The issues that were not addressed in this paper will be investigated in future work5 . Several simulation platforms have similar objectives to the framework presented in this paper (e.g. Mace3j [13], MadKit, [14], FishMarket [23] and MyWorld [24]). These platforms do not formally address all the issues raised in this paper. For example, in Fishmarket there is a lack of formalisation of concepts such as rights and obligations as well as a lack of an explicit representation of them during simulations. The MadKit platform is based on the organisation metaphor AGR (agent/group/role) developed in the context of the Aalaadin project and integrates heterogeneous agent systems. As pointed out in [8], the AGR model views organisations simply as collections of roles and does not incorporate the necessary notion of organisational rules (i.e. social constraints). There exist several approaches that modify the traditional BDI cycle in order to increase the ‘social awareness’ of agents (e.g. [25]). Here, we have selected trust as the crucial modifier. The results of our experiments (reported in Section 3) support the notion of trust as an important and beneficial input to agent-based reasoning in MAS. We believe that a similar view can be taken of the normative and institutional concepts discussed in Section 4. We have not fully integrated the trust and deontic concepts of our framework at this stage, but one of our objectives is to make the information compiled by the Society Visualizer available to all agents, either through communication (via query messages) or via the addition of a module equivalent to the Society Visualiser to each agent. Reasoning about the institutional and social properties of the agents (as produced by the Society Visualiser) is also a crucial part of trust relationships. Members of a society, governed by a set of rules, must consider when deciding to delegate a task to a peer what are its associated institutional powers, permissions, obligations, sanctions and social roles. For example, it would not be ‘rational’ to delegate a task to an agent if that agent does not have the institutional power or is forbidden to perform that task. Future work includes modifying the process of the 4 5

In this case the SV is an extended version of the Causal Calculator software tool (see [6]). Limited space is another reason why we do not fully investigate these issues here. However, details on experimental results can be found in [5, 6].

Simulating Computational Societies

65

trust update mechanism by considering attributes such as powers (institutional and physical), permissions, obligations and roles of the peers. In conclusion, we believe the combination of architectural, socio-cognitive and external observational perspectives presented in this paper affords a comprehensive approach to modelling MAS. A key aspect of future work will be to refine this combination into a socio-technical substrate defining and supporting the activities of agents within MAS as well. We believe that building a sociotechnical layer of computation, complementary to (and exploiting functionality of) a lower level of security, is technically feasible, and indeed offers a more tractable solution to certain security problems in e-commerce [26, 27] and digital rights management [28] than, say, ever more sophisticated encryption algorithms or encoding technology. For example, advocates of the Extensible rights Mark-up Language (XrML) position XrML as the foundation of trusted system development [29]: a higher-level approach is to use trust as the foundation of trusted system development. We plan to fully validate the presented concepts and arguments through continued design, simulation and analysis.

Acknowledgements This work has been undertaken in the context of the EU-funded ALFEBIITE Project (IST-1999-10298). We have also benefitted from Marek Sergot’s contributions on the specification of agent societies and on the development of the Society Visualiser (Section 4). The authors are especially grateful for the comments received both from reviewers and during the ESAW 2002 workshop.

References [1] A. Rao and M. Georgeff. BDI agents: from theory to practice. In Victor Lesser, editor, Proceedings of ICMAS, pages 312–319, San Francisco, CA, 1995. MIT Press. 53, 55 [2] J. Pitt and A. Mamdani. Designing agent communication languages for multiagent systems. In F. Garijo and M. Boman, editors, Proceedings of MAAMAW Workshop, number 1647 in LNAI, pages 102–114. Springer-Verlag, 1999. 53, 54, 55, 56 [3] C. Castelfranchi and R. Falcone. Social trust: A cognitive approach. In C. Castelfranchi and Y.-H. Tan, editors, Trust and Deception in Virtual Societies, pages 55–90. Kluwer Academic Press, 2000. 53, 56, 57 [4] R. Falcone and C. Castelfranchi. The socio-cognitive dynamics of trust: Does trust create trust? In R. Falcone, Y.-H. Tan, and M. Singh, editors, Trust in Cyber-Societies, number 2246 in LNAI, 2001. 53, 56 [5] A. Artikis, J. Pitt, and M. Sergot. Animated specifications of computational societies. In C. Castelfranchi and L. Johnson, editors, Proceedings of AAMAS, pages 1053–1062. ACM, 2002. 53, 61, 62, 63, 64 [6] A. Artikis, M. Sergot, and J. Pitt. Specifying electronic societies with the causal calculator. In Proceedings of AOSE, pages 75–86, 2002. 53, 61, 64 [7] M. Wooldridge, N. Jennings, and D. Kinny. The Gaia methodology for agentoriented analysis and design. JAAMAS, 3(3):285–312, 2000. 53, 54

66

Lloyd Kamara et al.

[8] F. Zambonelli, N. Jennings, and M. Wooldridge. Organisational rules as an abstraction for the analysis and design of multi-agent systems. International Journal of Software Engineering and Knowledge Engineering, 11(3):303–328, 2001. 53, 64 [9] J. Filipe. A normative and intentional agent model for organisation modelling. In P. Petta, R. Tolksdorf, and F. Zambonelli, editors, Proceedings of ESAW, LNCS. Springer, 2002. 54 [10] A. Ricci, A. Omicini, and E. Denti. Activity theory as a framework for MAS coordination. In P. Petta, R. Tolksdorf, and F. Zambonelli, editors, Proceedings of ESAW, LNCS. Springer, 2002. 54 [11] J. Pitt, L. Kamara, and A. Artikis. Interaction patterns and observable commitments in multi-agent trading scenario. In Proceedings of Autonomous Agents, pages 481–489. ACM Press, 2001. 54 [12] R. Smith and R. Davis. Distributed problem solving: The contract-net approach. In Proceedins of the 2nd Conference of Canadian Society for CSI, 1978. 54, 57 [13] L. Gasser and K. Kakugawa. MACE3J: Fast, flexible distributed simulation of large, large-grain multi-agent systems. In C. Castelfranchi and L. Johnson, editors, Proceedings of AAMAS, pages 745–852, 2002. 55, 64 [14] O. Gutknecht, J. Ferber, and F. Michel. Integrating tools and infrastructures for generic multi-agent systems. In Proceedings of Autonomous agents, pages 441–448. ACM Press, 2001. 55, 64 [15] M. Witkowski, A. Artikis, and J. Pitt. Experiments in building experiential trust in a society of objective-trust based agents. In R. Falcone, M. Singh, and Y.-H. Tan, editors, Trust in Cyber Societies, LNAI 2246, pages 110–132. Springer, 2001. 57 [16] R. Conte. A cognitive memetic analysis of reputation. Alfebiite project deliverable, 2002. http://alfebiite.ee.ic.ac.uk/docs/Deliverables/D5D6.zip. 57, 61 [17] R. Axelrod. The Evolution of Cooperation. Basic Books, New York, 1984. 58 [18] C. Hewitt. Open information systems semantics for distributed artificial intelligence. Artificial Intelligence, 47:76–106, 1991. 61 [19] A. Jones and M. Sergot. A formal characterisation of institutionalised power. Journal of the IGPL, 4(3):429–445, 1996. 61, 62, 63 [20] M. Sergot. A computational theory of normative positions. ACM Transactions on Computational Logic, 2(4):522–581, 2001. 61 [21] M. Shanahan. The event calculus explained. In M. Wooldridge and M. Veloso, editors, Artificial Intelligence Today, LNAI 1600, pages 409–430. Springer, 1999. 61 [22] E. Giunchiglia, J. Lee, V. Lifschitz, N. McCain, and H. Turner. Nonmonotonic causal theories. To appear in the Journal of Artificial Intelligence, 2003. 61 [23] J. Rodriguez-Aguilar, F. Martin, P. Noriega, P. Garcia, and C. Sierra. Towards a test-bed for trading agents in electronic auction markets. In AI Communications, pages 5–19, 1998. 64 [24] M. Wooldridge. This is MyWorld: The logic of an agent-oriented DAI testbed. In M. Wooldridge and N. Jennings, editors, Intelligent Agents: Proceedings of the 1994 Workshop on ATAL. Springer-Verlag, 1995. 64 [25] J. Broersen, M. Dastani, Z. Huang, J. Hulstijn, and L. Van der Torre. The BOID architecture: Conflicts between beliefs, obligations, intentions and desires. In Proceedings of Autonomous Agents, pages 9–16. ACM Press, 2001. 64 [26] M. Blaze, J. Feigenbaum, J. Ioannidis, and A. D. Keromytis. The role of trust management in distributed systems security. In Secure Internet Programming: Security Issues for Mobile and Distributed Objects, number 1603 in LNCS, pages 185–210. Springer, 1999. 65

Simulating Computational Societies

67

[27] V. Swarup and J. T. Fabr´ega. Trust: Benefits, models and mechanisms. In Secure Internet Programming: Security Issues for Mobile and Distributed Objects, number 1603 in LNCS, pages 3–18. Springer, 1999. 65 [28] J. Bing. Managing copyright in a digital environment. In I. Butterworth, editor, The Impact of Electronic Publishing on the Academic Community, pages 52–62. Portland Press, 1998. 65 [29] XrML. Extensible rights mark-up language. http://www.xrml.org, 2002. 65

Co-Fields: Towards a Unifying Approach to the Engineering of Swarm Intelligent Systems 0DUFR0DPHL)UDQFR=DPERQHOOLDQG/HWL]LD/HRQDUGL 

'LSDUWLPHQWRGL,QJHJQHULDGHOOt ,QIRUPD]LRQH 8QLYGL0RGHQDH5HJJLR(PLOLD 9LD9LJQROHVH0RGHQD,WDO\  'LSDUWLPHQWRGL6FLHQ]HH0HWRGLGHOOt ,QJHJQHULD 8QLYGL0RGHQDH5HJJLR(PLOLD 9LD$OOHJUL0RGHQD,WDO\ {mamei.marco,franco.zambonelli}@unimo.it [email protected]

Abstract. 6ZDUP LQWHOOLJHQW V\VWHPV LQ ZKLFK WKH SDWKV WR SUREOHP VROYLQJHPHUJHDVWKHUHVXOWRILQWHUDFWLRQVEHWZHHQVLPSOHDXWRQRPRXV FRPSRQHQWVDJHQWVRUDQWV DQGEHWZHHQWKHPDQGWKHLUHQYLURQPHQW DSSHDU YHU\ SURPLVLQJ WR GHYHORS UREXVW DQG IOH[LEOH VRIWZDUH DSSOLFDWLRQ+RZHYHUWKHYDULHW\RIVZDUPEDVHGDSSURDFKHVWKDWKDYH EHHQ SURSRVHG VR IDU VWLOO ODFNV D FRPPRQ PRGHOLQJ DQG HQJLQHHULQJ PHWKRGRORJ\ ,Q WKH DWWHPSW WR RYHUFRPH WKLV SUREOHP WKLV SDSHU SUHVHQWV D JHQHUDO FRRUGLQDWLRQ PHWKRGRORJ\ LQ ZKLFK VZDUPt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

1

Introduction

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©6SULQJHU9HUODJ%HUOLQ+HLGHOEHUJ

&R)LHOGV7RZDUGVD8QLI\LQJ$SSURDFKWRWKH(QJLQHHULQJRI6ZDUP

FDVH RI SHUYDVLYH >@ PRELOH >@ DQG ZLGHDUHD ,QWHUQHW DSSOLFDWLRQV >@  7KHUH G\QDPLF DQG RSHQQHVV FDOO IRU V\VWHPV DUFKLWHFWXUHV FDSDEOH RI VXSSRUWLQJ ERWK SDUWLDOIDLOXUHVDQGG\QDPLFXQVXSHUYLVHGUHRUJDQL]DWLRQRILQWHUDFWLRQSDWWHUQV 7R IDFH WKLV VLWXDWLRQ VHYHUDO UHVHDUFKHUV WXUQHG WKHLU DWWHQWLRQ WR HQJLQHHULQJ DSSURDFKHVWKDWWDNHLQVSLUDWLRQIURP WKH FROOHFWLYH EHKDYLRU RI VRFLDO DQLPDOV HJ DQWV)URPWKHHQJLQHHULQJSRLQWRIYLHZLWLVLPSRUWDQWWRREVHUYHWKDWDQLPDOVVROYH DZLGHYDULHW\RIGLIIHUHQWSUREOHPVHJILQGLQJIRRGXVHVKRUWHVWSDWKVWRUHDFKWKH QHVWRUJDQL]HWDVNGLYLVLRQ LQDYHU\UREXVWDQGIOH[LEOHZD\5REXVWQHVVDULVHVIURP WKHIDFWWKDWWKHFRORQ\KDVWKHDELOLW\WRDFKLHYHLWVJRDOHYHQZKHQVRPHLQGLYLGXDOV GLH RU IDLO WR SHUIRUP WKHLU WDVN IOH[LELOLW\ DULVHV IURP WKH IDFW WKDW WKH SDWWHUQV RI LQWHUDFWLRQV EHWZHHQ WKH LQGLYLGXDOV DUH QRW IL[HG E\ FRQWUDFW DQG FDQ EH LQVWHDG G\QDPLFDOO\ UHVKDSHG WR VHOIDGDSW WR FKDQJLQJ HQYLURQPHQWV )ROORZLQJ RWKHU DXWKRUV>@ZHUHIHUWRWKRVHNLQGRIV\VWHPVDVVZDUPLQWHOOLJHQWV\VWHPVWRVWUHVV WKHIDFWWKDWWKHLUIHDWXUHVDQGFDSDELOLWLHVDUHQRWHPEHGGHGLQWKHVLQJOHFRPSRQHQWV RIWKHV\VWHPEXWHPHUJHE\WKHFRRUGLQDWHGDFWLYLWLHVRIDVZDUPRILQGLYLGXDOV 7KHVH IHDWXUHV RI VZDUP LQWHOOLJHQW V\VWHPV DWWUDFW PRUH DQG PRUH VRIWZDUH HQJLQHHULQJUHVHDUFKHUV>@GXHWRWKHIDFWWKDWWKHGDLO\SUREOHPVVROYHGE\JURXSV RIDQLPDOVKDYHGLUHFWFRXQWHUSDUWVLQHQJLQHHULQJDQGFRPSXWHUVFLHQFHILQGLQJIRRG VWUDWHJLHVFDQEHH[SORLWHGLQDQLQIRUPDWLRQUHWULHYDOFRQWH[WWKHVWUDWHJLHV XVHGWR ILQGWKHVKRUWHVWSDWKFRQQHFWLQJWZRORFDWLRQVFDQEHH[SORLWHGLQURXWLQJDOJRULWKPV IRU WHOHFRPPXQLFDWLRQ DQG FRPSXWHU QHWZRUNV PHFKDQLVPV IRU DQWt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t  DFWLRQV VR DV WR LPSOLFLWO\ HQFRGH FRRUGLQDWLRQ WDVNV 7KH NH\ DGYDQWDJH RI &R)LHOGVLVWKDW LW HQDEOHV WR UHSUHVHQW LQ D VLPSOH \HW HIIHFWLYH ZD\ ERWK GLIIHUHQW W\SHV RI u HQYLURQPHQWVv  LQ ZKLFK DJHQWV FDQ OLYH DQG GLIIHUHQW FRRUGLQDWLRQ WDVNV WKH\ KDYH WR DFKLHYH 7KLV PDNHV LW SRVVLEOH WR DGRSW WKH VDPH EDVLF DSSURDFK WR PRGHO GLIIHUHQW FODVVHV RI VZDUP LQWHOOLJHQW V\VWHPV ,Q DGGLWLRQ WKH SK\VLFDO LQVSLUDWLRQDOVRDOORZVWRIRUPDOO\WUHDWLQJ&R)LHOGVFRRUGLQDWHGV\VWHPVLQWHUPVRI G\QDPLFDO V\VWHPV > @ DQG WR GHULYH WKH GLIIHUHQWLDO HTXDWLRQV JRYHUQLQJ WKHLU

0DUFR0DPHLHWDO

EHKDYLRU ,Q WKH IXWXUH ZH WKLQN WKDW YHU\ LQWHUHVWLQJ UHVXOWV FRXOG EH REWDLQHG E\ DSSO\LQJ FODVVLF G\QDPLFDO V\VWHP WKHRUHPV DQG UHVXOWV WR WKLV PDWKHPDWLFDO GHVFULSWLRQ OLNH IRU H[DPSOH GHWHFWLQJ WKH V\VWHPt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t IRUPDOLVP6HFWLRQFRQFOXGHVWKHSDSHUDQGSUHVHQWVRXUIXWXUH ZRUN

2

The Co-Fields Model

7KH&R)LHOGVPRGHOSURYLGHVDPRGHOLQJIUDPHZRUNWRGHVLJQDPXOWLDJHQWV\VWHP 0$6 ,WVSULPDU\IRFXVLVWRFRQVLGHUDQGPRGHOWKH0$6DVDu ZKROHv LQZKLFK DJHQWVDFKLHYHWKHLUJRDOQRWEHFDXVHRIWKHLUVLQJOHFDSDELOLWLHVEXWEHFDXVHWKH\DUH SDUWRIDQVHOIRUJDQL]HGV\VWHPWKDWOHDGVWKHPWRWKHJRDODFKLHYHPHQW>@WKHIDFW WKDWWKHJRDOVDUHDFFRPSOLVKHGLVQRWDPHULWRIWKHVLQJOHDJHQWVEXWRIWKHV\VWHPDV DZKROH 7KH&R)LHOGVPRGHOFDQEHVFKHPDWL]HGLQWKHIROORZLQJIRXUSRLQWV 

  

7KHHQYLURQPHQWLVUHSUHVHQWHGDQGDEVWUDFWHGE\ILHOGVVSUHDGE\DJHQWVDQG E\WKHHQYLURQPHQWLWVHOI7KHVHILHOGVFRQYH\VRPHXVHIXOLQIRUPDWLRQIRUWKH DJHQWVt  FRRUGLQDWLRQ WDVNV DQG SURYLGH DJHQWV ZLWK VWURQJ FRRUGLQDWLRQWDVN WDLORUHGFRQWH[WDZDUHQHVV 7KHFRRUGLQDWLRQSROLF\LVUHDOL]HGE\OHWWLQJWKHDJHQWVWRPRYHIROORZLQJWKH u ZDYHIRUPv RIWKHVHILHOGV (QYLURQPHQW G\QDPLFV DQG DJHQWVt  PRYHPHQWV LQGXFH FKDQJHV LQ WKH ILHOGVt VXUIDFH FRPSRVLQJ D IHHGEDFN F\FOH WKDW LQIOXHQFHV DJHQWVt  PRYHPHQW SRLQW   7KLVIHHGEDFNF\FOHOHWVWKHV\VWHPDJHQWVDQGHQYLURQPHQW WRDXWRRUJDQL]H VRWKDWWKHFRRUGLQDWLRQWDVNLVILQDOO\DFKLHYHG

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t V    

&R)LHOGV7RZDUGVD8QLI\LQJ$SSURDFKWRWKH(QJLQHHULQJRI6ZDUP

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t  PRYHPHQW DQG WKH JUDGLHQW RI WKH ILHOGt V VXUIDFHDQGWKXVJRYHUQDQGGHVFULEHWKHV\VWHPt VEHKDYLRU $OPRVW DOO WKH FRRUGLQDWLRQ WDVNV FRQVLGHUHG LQ WKLV SDSHU GHDO ZLWK DJHQWVt PRYHPHQWV DQG VR WKH     FDQ EH FRQVLGHUHG VSUHDG LQ WKH SK\VLFDO VSDFHDQGIROORZLQJWKH  VXUIDFH PHDQV PRYLQJ IURPRQHSODFHWR DQRWKHU+RZHYHUWKHDSSURDFKLVPRUHJHQHUDODQGD  VSUHDGLQD PRUH DEVWUDFW VSDFH FRXOG HQFRGH FRRUGLQDWLRQ WDVNV WKDW GR QRW GHDO ZLWK VRPH SK\VLFDO PRYHPHQWV EXW ZLWK RWKHU NLQG RI DFWLRQV ,Q WKHVH FDVHV DJHQWV ZRXOG IROORZLWV  QRWE\PRYLQJIURPRQHSODFHWRDQRWKHUEXWE\PDNLQJ RWKHUNLQGRIDFWLRQV$QH[DPSOHRIWKLVPRUHDEVWUDFWHQFRGLQJZLOOEHSUHVHQWHGDW WKHHQGRIWKHQH[WVHFWLRQ

3

Classic Swarm Intelligence Examples

,QWKLVVHFWLRQZHDUHJRLQJWREULHIO\VXUYH\GLIIHUHQWVZDUPLQWHOOLJHQWFRRUGLQDWLRQ SROLFLHVPDLQO\WDNHQIURP>@DQGZHZLOOVKRZKRZFDQWKH\EHPRGHOHGLQ WHUPV RI WKH &R)LHOGV PRGHO 7KLV ZLOO VHUYH ERWK WR MXVWLI\ WKH FODLP RI XQLI\LQJ DSSURDFKRIWKLVSDSHUDQGWRFODULI\WKH&R)LHOGVPRGHOLWVHOI 

B

B A

A

A B

B A

A

A

B

Fig 1. $JHQW % IROORZLQJ WKH VXUIDFH RI LWV     ZKLFK FRLQFLGHV ZLWK WKH ILHOG JHQHUDWHGE\DJHQW$left WKHDJHQW%IROORZVGRZQKLOOWKHGHFUHDVHRILWV   center  WKH DJHQW % IROORZV XSKLOOWKH LQFUHDVH RI LWV    right  WKH DJHQW % IROORZVDQHTXLSRWHQWLDOOLQHRILWV  

0DUFR0DPHLHWDO 

. 

>

. −



− . − 



− . − 

. −

 >





−. 





Fig. 2.0RRVHt Vleft DQGZROIt Vright JHQHUDWHGILHOG

3.1

Wolves: Surrounding a Prey

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t PRYHPHQWV D PRRVH DW WKH FRRUGLQDWHV  @   ?  JHQHUDWHV D  ,    GHVFULEHGE\WKHIROORZLQJHTXDWLRQVHH)LJXUHOHIW      A   = − .  −  ((A − @  ) + (  −? ) )

.    >   $QDORJRXVO\ WKH ZROI  DW WKH FRRUGLQDWHV  @   ?   JHQHUDWHV D  ,    ZKRVHHTXDWLRQLVVHH)LJXUHULJKW 

(

)

    A   = .  −  (A − @  ) + ( −? ) .    >  7KHFRQWULEXWLRQRIWKLVDQDO\WLFDOGHVFULSWLRQZLOOEHFOHDULQVHFWLRQZKHQZH ZLOOGHULYHWKHGLIIHUHQWLDOHTXDWLRQVGHVFULELQJWKH&R)LHOGV PRGHOLQJHQHUDODQG ZHZLOODSSO\WKRVHHTXDWLRQVWRWKLVPDWKHPDWLFDOGHVFULSWLRQRIWKHILHOGVt VXUIDFHV WRGHULYHPRRVHt VDQGZROYHVt EHKDYLRU 1RZ LI ZH FRQVLGHU WKH PRRVHt V     FRQVLVWLQJ LQ WKH OLQHDU FRPELQDWLRQRIDOOWKHZROYHVt ILHOGV      A   =     A  



 

 

 = E\IROORZLQJWKHGHFUHDVHRIWKHUHVXOWLQJILHOGWKHPRRVHUXQVDZD\IURPZROYHV 6LPLODUO\ LI ZH FRQVLGHU HDFK ZROIt V     FRQVLVWLQJ LQ WKH OLQHDU FRPELQDWLRQEHWZHHQWKHPRRVHt VILHOGDQGDOOWKHRWKHUZROYHVt ILHOGV

&R)LHOGV7RZDUGVD8QLI\LQJ$SSURDFKWRWKH(QJLQHHULQJRI6ZDUP 

      A   =   A   +

   -  A  

- = - ≠

D ZROI LV GLUHFWHG WRZDUGV WKH PRRVH EXW VWD\LQJ DZD\ IURP RWKHU ZROYHV ,W LV FOHDU WKDW WKLV VLPSOH GHVFULSWLRQ LV SHUIHFWO\ DQDORJRXV WR WKH GHVFULSWLRQ EDVHG RQ GLVWDQFHV DQG E\ D SURSHU WXQLQJ RI WKH ILHOGVt  FRHIILFLHQWV FDQ OHDG WR WKH VDPH UHVXOWV 3.2

Birds Flocking

)ORFNVRIELUGVVWD\WRJHWKHUFRRUGLQDWHWXUQVDQGDYRLGHDFKRWKHUE\IROORZLQJD YHU\VLPSOHVZDUPDOJRULWKP>@6LPLODUSUREOHPVKDSSHQLQDLUWUDIILFFRQWURO DQGFRQYR\VRIVKLSVDQGWKH\FRXOGEHSRVVLEO\DGGUHVVHGE\XVLQJVLPLODUPHWKRGV 7KHFRRUGLQDWHGEHKDYLRURIIORFNVFDQEHH[SODLQHGE\DVVXPLQJWKDWHDFKELUGWULHV WRPDLQWDLQDVSHFLILHGVHSDUDWLRQIURPWKHQHDUHVWELUGVDQGWR PDWFK QHDUE\ELUGVs YHORFLW\ 7KH IORFN LV D VHOIFRQVWUDLQLQJ VWUXFWXUH LQ ZKLFK HDFK HQWLW\t V LQGLYLGXDO DFWLRQ VLPXOWDQHRXVO\ UHVSRQGV WR DQG FKDQJHV WKH RYHUDOO VWUXFWXUH RI WKH IORFN $OWKRXJKHDFKELUGVHQVHV RQO\ WKH PRYHPHQWV RI LWV QHDUHVW SHHUV LWV UHVSRQVHV WR WKHVH PRYHPHQWV SURSDJDWH WR RWKHUV VR WKDW WKH V\VWHP DV D ZKROH H[KLELWV JOREDO FRRUGLQDWLRQ 7RPRGHOWKLVVWUDWHJ\XQGHUWKH&R)LHOGVPRGHOLQJIUDPHZRUNZHFDQLPDJLQH WKDWHDFKELUGJHQHUDWHVD=:(2;ILHOGOLNHWKHRQHVKRZHGLQILJXUHOHIW ZKLFK FDQEHGHVFULEHGE\WKHIROORZLQJVLPSOHHTXDWLRQ  =  A − @ 0  +   − ?0  =:(2;   A   =   −   ⋅   7KLVILHOGLVXSGDWHGLQUHDOWLPHWRPDWFKWKHELUGt VPRYHPHQWV1RZVD\LQJWKDW ELUGV KDYH WR VWD\ D VSHFLILHG GLVWDQFHV IURP HDFK RWKHU LV HTXLYDOHQW WR VD\LQJ WKDW ELUGVKDVWRIROORZWKHGHFUHDVHRIRWKHUELUGVt JHQHUDWHGILHOGVVHH)LJXUHULJKW ,Q IDFWWKHVKDSHRIWKHILHOGFRQVWUDLQVELUGVWRVWD\FORVHHDFKRWKHULQDQDOPRVWUHJXODU JULGWKHPRYHPHQWRIDELUGDQGWKHFRQVHTXHQWFKDQJHLQWKHILHOGLWJHQHUDWHVIRUFH RWKHUELUGVWRPRYHDVZHOO 



Fig. 3.)ORFNLQJILHOGleft IORFNLQJELUGVright

0DUFR0DPHLHWDO

3.3

Ant Foraging

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t IRUDJLQJVWUDWHJ\XQGHUWKH&R)LHOGVPRGHOZHFDQLPDJLQH WKDWWKHHQYLURQPHQWLVDEVWUDFWHGLHSHUFHLYHGE\DQWV E\PHDQVRIWZRLQLWLDOO\ IODW  ILHOGV 7KHVH WZR ILHOGV ZKLFK ZH ZLOO FDOO WKH KRPH DQG WKH IRRG ILHOGV DUH JHQHUDWHG DQG VSUHDG E\ WKH HQYLURQPHQW LWVHOI 7KH HQYLURQPHQW UHDFWV WR DQWVt PRYHPHQWV E\ ZULQNOLQJ WKH ILHOGVt  VXUIDFH ZKLOH DQWVt  PRYHPHQWV DUH DIIHFWHG E\ WKHu ZDYHIRUPv RIWKHILHOG7KLVIHHGEDFNF\FOHFRQVWLWXWHVWKHNH\WROHWWKHV\VWHP DXWRRUJDQL]H 7KH DOJRULWKP IROORZHG E\ DQWV FDQ LQ IDFW EH UHVWDWHG E\ VXSSRVLQJ WKDW HDFK DQW ZDQGHUV DYRLGLQJ REVWDFOHV IROORZLQJ SUREDELOLVWLFDOO\  WKH GHFUHDVH RIWKHIRRGILHOGZKHQLWLVORRNLQJIRUIRRGIROORZLQJSUREDELOLVWLFDOO\ WKHGHFUHDVH RI WKH KRPH ILHOG ZKHQ LW LV EULQJLQJ IRRG EDFN WR WKH DQWKLOO 7KHQ WR FORVH VXFFHVVIXOO\WKHIHHGEDFNF\FOHZHZLOOVLPSO\LPDJLQHWKDWWKHHQYLURQPHQWUHDFWVWR WKHDQWVt SUHVHQFHE\ORFDOO\ZULQNOLQJWKHKRPHILHOGVXUIDFHLQFRUUHVSRQGHQFHRI WKH SRLQWV LQ ZKLFK DQWV ORRNLQJ IRU IRRG DUH ORFDWHG DQG LW ZULQNOHV WKH IRRG ILHOG VXUIDFHLQWKHSRLQWVLQZKLFKDQWVFDUU\LQJIRRGDUHORFDWHG7KHIRUPRIWKHZULQNOH LVGHSLFWHGLQILJXUHOHIW 0RUHRYHUZHZLOOVXSSRVHWKDWWKHHQYLURQPHQWLVDEOHWR FRQWUROWKHGHHSQHVVRIWKHZULQNOHVRWKDWHDFKQHZRUUHQHZHG ZULQNOHLVGHHSHU WKDWDOOWKHZULQNOHVLQLWVQHLJKERUKRRG)ROORZLQJWKHVHSULQFLSOHVDQWVt PRYHPHQW FUHDWHV D QHWZRUN RI FKDQQHOV LQ WKH ILHOGVt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

&R)LHOGV7RZDUGVD8QLI\LQJ$SSURDFKWRWKH(QJLQHHULQJRI6ZDUP 







Fig. 4.$ZULQNOHLQGXFHGRQDILHOGt VVXUIDFHE\DQDQWSUHVHQFHleft 7KHFKDQQHOLQWKHILHOG VXUIDFHFUHDWHGE\WKHDQWPRYHPHQWright

)LQDOO\WRDFFXVWRPIRUSKHURPRQHHYDSRUDWLRQZHZLOOLPDJLQHWKDWHDFKRIWKH ILHOGVt VXUIDFHVKDVVRPHIRUPRIPHPRU\DQGLWUHVKDSHVWRLWVRULJLQDOIODWIRUPLI XQWRXFKHG,QSDUWLFXODUWKLVFDQEHIRUPDOO\VWDWHGE\VD\LQJWKDWDILHOGt VZULQNOHDW FRRUGLQDWHV  @   ?   FUHDWHG D WLPH    FDQ EH PDWKHPDWLFDOO\ GHVFULEHG E\ WKH IROORZLQJIXQFWLRQ    .  A   = − .  −   −  (A − @  ) + (  − ? ) .    >  :KHUH.LVDIXQFWLRQWKDWJRHVWR]HURDVLQFUHDVHVZKLOHLVDFRQVWDQWWKDW H[SUHVVHVWKHIDFWWKHZULQNOHLVRQO\DORFDOGHIRUPDWLRQ ,Q JHQHUDO WKLV NLQG RI DQDO\WLFDO GHVFULSWLRQ LV QRW RQO\ XVHIXO WR VWDWH WKHVH FRQFHSWVIRUPDOO\EXWLWFDQEHDSSOLHGWRSURYLGHDG\QDPLFDOV\VWHPGHVFULSWLRQRI WKH PRGHO ,Q VHFWLRQ  WKLV SURFHGXUH ZLOO EH DSSOLHG WR WKH H[DPSOH GLVFXVVHG LQ VHFWLRQKRZHYHUWKHUHVXOWVFDQEHHDVLO\DSSOLHGDOVRWRWKLVFDVH

(

3.4

)

Ant Labor Division and Task Succession

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t V QHHGV 0RUHRYHU E\ SHUIRUPLQJ D FHUWDLQ WDVN LQGLYLGXDOVt  WDVN DVVRFLDWHG WKUHVKROG GHFUHDVH 7KLV VLPSOH VWUDWHJ\ LV WKH NH\ IRU VSHFLDOL]DWLRQ WKH PRUH RQH LQGLYLGXDO SHUIRUPHG D WDVN LQ WKH SDVW WKH PRUH OLNHO\ WKH VDPH LQGLYLGXDO ZLOO SHUIRUPWKHWDVNLQWKHIXWXUH 7KLV H[DPSOH LV SDUWLFXODUO\ LQWHUHVWLQJ WR EH PRGHOHG ZLWK &R)LHOGV EHFDXVH LW LQYROYHV ILHOGVSURSDJDWHGLQDVSDFH ZKLFKLVQRWWKHSK\VLFDO VSDFHLQ ZKLFKDQWV DUHHPEHGGHG

&R)LHOGV7RZDUGVD8QLI\LQJ$SSURDFKWRWKH(QJLQHHULQJRI6ZDUP

HQFRXUDJHVDQWVLQSHUIRUPLQJDWDVN%DVLFDOO\WKHVKDSHRIWKHVHILHOGVLVDOPRVWOLNH DVWHHSIODWVXUIDFHGHFUHDVLQJWRZDUGVWKHGLUHFWLRQRIWKHDVVRFLDWHGD[LV7KHPRUH WKHWDVNDFKLHYHPHQWLV XUJHQWWKH PRUHWKHHQYLURQPHQWLQFUHDVHV WKH VWHHSQHVV RI WKHILHOGVXUIDFH,QILJXUHWKUHHGLIIHUHQW7DVN$t VDVVRFLDWHGILHOGVDUHVKRZQ7KH PRUHWKHVXUIDFHJHWVWHHSWKHPRUH7DVN$t VDFKLHYHPHQWEHFRPHVXUJHQW $QWVt DOJRULWKPFDQWKXVEHUHVWDWHGE\ PHDQV RI WKH IROORZLQJ DFWLRQV HDFK DQW HYDOXDWHVDFRPELQDWLRQRIWKHVHQVHGILHOGVE\FRQVLGHULQJRQO\WKRVHILHOGVZKRVH VWHHSQHVVZKHUHWKHDQWLVORFDWHGRYHUFRPHVDFHUWDLQWKUHVKROG7KHQWKHDQWIROORZV GRZQKLOOWKHILHOGREWDLQHG$VWKHDQWVPRYHDORQJWKH7DVNVSDFHWKHILHOGt VVXUIDFH WHQG WR JHW IODWWHQHG EHFDXVH WKH DVVRFLDWHG WDVNV DUH DFKLHYHG $QWVt  GXWLHV JHW VWDELOL]HG LQ D VXLWDEOH FRQILJXUDWLRQ XQWLO QHZ VWLPXOL DQG WKXV ILHOGVt  VXUIDFHVt GHIRUPDWLRQVDSSHDU$VGHSLFWHGLQILJXUHDWDVNDVVRFLDWHGILHOGt VVXUIDFHLVQRW DFWXDOO\ D VWHHS IODW VXUIDFH EXW LWV VWHHSQHVV LQFUHDVHV ERWK LQ WKH GLUHFWLRQ RI WKH DVVRFLDWHGD[LVDQGERWKLQWKHSUR[LPLW\RIWKHDVVRFLDWHGD[LVLWVHOI7KHUHDVRQIRU WKLV QRQOLQHDU VKDSH LV WR HQIRUFH VSHFLDOL]DWLRQ ,Q IDFW RQ WKH RQH KDQG E\ LQFUHDVLQJWKHVWHHSQHVVWRZDUGVWKHGLUHFWLRQRIWKHDVVRFLDWHGD[LVWKHPRUHDQDQW LVSODFHGWRZDUGVWKDWGLUHFWLRQLHWKHPRUHLWLVSHUIRUPLQJWKDWWDVN WKHPRUHLW ZLOOHDV\WKDWWKHDQWZLOOSHUIRUPWKHWDVNLQWKHIXWXUHEHFDXVHWKHILHOGt VVWHHSQHVV LVWKHUHSDUWLFXODUO\KLJK2QWKHRWKHUKDQGWKHLQFUHDVHRIWKHVWHHSQHVVWRZDUGVWKH D[LV LWVHOI LH WRZDUGV WKH ]HUR  UHQGV HDVLHU WKDW u XQHPSOR\HGv  DQWV UDWKHU WKDQ DOUHDG\IXOO\FRPPLWWHGDQWVHQJDJHWKHWDVN

4

A Dynamical Description of Swarm Intelligence

,Q WKLV VHFWLRQ ZH DUH JRLQJ WR SUHVHQW D G\QDPLFDO V\VWHP GHVFULSWLRQ RI WKH &R )LHOGVPRGHOEDVHGRQWKHGLIIHUHQWLDOHTXDWLRQVWKDWJRYHUQWKHV\VWHPEHKDYLRU 7KH &R)LHOGV PRGHO VXLWV UHDOO\ ZHOO WR WKLV NLQG RI GHVFULSWLRQ EHFDXVH LQGLYLGXDOVt EHKDYLRUFDQEHDVVLPLODWHGWRWKHRQHRIDEDOOUROOLQJRQDVXUIDFHWKH     VXUIDFH  ,W LV WKXV UDWKHU HDV\ WR VHH WKDW LI ZH FRQVLGHU WKH LQGLYLGXDO  DQG ZH GHQRWH LWV FRRUGLQDWHV LQ D SDUWLFXODU VSDFH DV  A   A    A   DQG LWV     DV     @   @   @     7KHQ WKHGLIIHUHQWLDOHTXDWLRQVWKDWJRYHUQEHKDYLRUDUHLQWKHIRUP A ∂    @   @   @    = ± ⋅ - =    ∂@ :KHUH  H[SUHVVHV WKH u VSHHGv  RI LQGLYLGXDO  ZKLOH WKH VLJQ DW WKH ULJKW RI WKH HTXDOVLVGHFLGHGE\ZHDWKHUIROORZVWKHLQFUHDVHRUWKHGHFUHDVHRIWKHFRRUGLQDWLRQ ILHOG 7KLV LV EHFDXVH WKH GLUHFWLRQ RI WKH JUDGLHQW RI WKH LQGLYLGXDOt V     HYDOXDWHG WRZDUGV WKH VSDWLDO FRRUGLQDWHV SRLQW WR WKH GLUHFWLRQ LQ ZKLFK WKH   LQFUHDVH6RWKHLQGLYLGXDOZLOOIROORZWKLVJUDGLHQWRUZLOOJRLQ WKHRSSRVLWHGLUHFWLRQGHSHQGLQJRQLILWZDQWVWRIROORZWKHLQFUHDVHRUWKHGHFUHDVH RILWV  ,QDVLPLODUZD\LWLVSRVVLEOHWRPRGHOWKRVHFDVHVLQZKLFK IROORZVWKHHTXLSRWHQWLDOOLQHVRIWKHILHOGV ,W LV ZRUWK QRWLQJ WKDW WKH DERYH GHVFULSWLRQ  LV FRPSOHWHO\ GHWHUPLQLVWLF ZKLOH PRVWRIWKHVZDUPLQWHOOLJHQWV\VWHPVKDYHUDQGRPGHFLVLRQSURFHVVDWWKHLUFRUHHJ

0DUFR0DPHLHWDO

IRUDJLQJDQWVIROORZSKHURPRQHWUDLOVRQO\ZLWKVRPHSUREDELOLW\ :HWKLQNWKDWWKLV PLVPDWFKLVQRWUHOHYDQWIRUWKHFODLPPDGHLQWKHSDSHUWKDW&R)LHOGVFDQSURYLGHD XQLI\LQJDEVWUDFWLRQ IRUVZDUP LQWHOOLJHQWV\VWHPV,QIDFWUDQGRPHOHPHQWVDUH QRW H[SOLFLWO\QHFHVVDU\WRJHQHUDWHVZDUPV\VWHPVt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t ILHOGV OHWt VVXSSRVHQZROYHV       A   =     A    = :KLOHWKHZROI  LVJLYHQE\WKHVXPOLQHDUFRPELQDWLRQZLWKDOO WKH FRHIILFLHQWV HTXDO WR   RI WKH PRRVHt V ILHOG WKDW DWWUDFWV WKH ZROI  DQG DOO WKH RWKHUVZROYHVt ILHOGVWKDWUHSXOVHWKHZROI 





      A   =   A   +

   -  A  

- = - ≠

1RZ ZULWLQJ WKH GLIIHUHQWLDO HTXDWLRQV LV MXVW D PDWWHU RI VXEVWLWXWLQJ WKHVH     ZLWK WKH ILHOGVt  HTXDWLRQV GHVFULEHG LQ  DQG WR XVH WKH GLIIHUHQWLDO HTXDWLRQV DW WKH EHJLQQLQJ RI WKLV VHFWLRQ LQ WKH FDVH RI LQGLYLGXDOV IROORZLQJWKHGHFUHDVHRIWKH  %HFDXVHRIVSDFHOLPLWZHZLOOQRW UHSRUWWKHUHVXOWRIWKLVVXEVWLWXWLRQKRZHYHUWKHV\VWHPWXUQVRXWWREHGHVFULEHGE\D VHWRIVWULFWO\FRXSOHGQRQOLQHDUGLIIHUHQWLDOHTXDWLRQVEHFDXVHRIWKHH[SRQHQWLDOVLQ WKH ILHOGVt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t DQDO\WLFVROXWLRQ WRWKLV PDWKHPDWLFDOGHVFULSWLRQOLNHIRUH[DPSOHGHWHFWLQJWKHV\VWHPt VDWWUDFWRUVDQGWKHLU EDVLQ>@ 7KH UHVXOWV RI D QXPHULFDO LQWHJUDWLRQ RI WKH H[DPSOHt V GLIIHUHQWLDO HTXDWLRQV LQ WKHFDVHRIDV\VWHPFRPSRVHGE\RQHPRRVHDQGWKUHHZROYHVDUHGLVSOD\HGLQWKH IROORZLQJ ILJXUH VHH )LJXUH   ZKLFK VKRZV D FRPPRQ [\ SODQH ZLWK WKH WUDMHFWRULHVRIWKHHOHPHQWVRIWKHV\VWHPLHWKHVROXWLRQVRI  A      HYDOXDWHG IRUDFHUWDLQWLPHLQWHUYDO ,WLVZRUWKQRWLQJWKDWVROYLQJWKHVHHTXDWLRQVQXPHULFDOO\ LVDYHU\HIIHFWLYHZD\WRVLPXODWHWKHV\VWHPDQGLWFDQEHUHJDUGHGDVDYHU\HDV\ DQGSRZHUIXOWRROWRPDNHH[SHULPHQWVDQGWXQLQJFRHIILFLHQWV

&R)LHOGV7RZDUGVD8QLI\LQJ$SSURDFKWRWKH(QJLQHHULQJRI6ZDUP  :2/)

:2/) :2/)

:2/)

:2/)

0226( 0226(

:2/)

Fig. 7. left  :ROYHV GR QRW FDUH DERXW RWKHU ZROYHVt  ILHOGV DQG WKXV WKH\ DUH QRW DEOH WR VXUURXQG WKH PRRVH right  :ROYHV VHQVLQJ RWKHU ZROYHVt  ILHOGV DUH DEOH WR VXUURXQG WKH PRRVH

5

Conclusions, Current and Future Works

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



 Fig. 8.from left to right 'LIIHUHQWVWDJHVLQWKH&R)LHOGVJXLGHGPRYHPHQWRIDJURXSRI DJHQWVWKURXJKDYLUWXDOEXLOGLQJ$JHQWVPRYHZKLOHUHPDLQLQJLQDIORFN

0DUFR0DPHLHWDO

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t VEXLOGLQJ>@,Q)LJXUHZHVNHWFKWKHUHVXOWVRI VXFK D VLPXODWLRQ ZKHQ DSSOLHG WR WKH IORFNLQJ SUREOHP DQG YLVXDOL]LQJ WKH FRRUGLQDWHGPRYHPHQWVRIDVHWRIVHFXULW\JXDUGVZLWKLQWKHPXVHXP :H DOVR HPSKDVL]H WKDW LQ &R)LHOGV DV LQ RWKHU VZDUP V\VWHPV DGDSWLYLW\ DQG VHOIRUJDQL]DWLRQLVJXDUDQWHHGZLWKRXWDQ\FKDQJHWRWKHDJHQWVt FRGHWKHVHFXULW\ JXDUGVFDQFRQWLQXHPRYHLQDFRRUGLQDWHGZD\HYHQZKHQQHZVHFXULW\JXDUGVDUULYH RUZKHQVHFXULW\JXDUGVKDYHWRWDNHFDUHRIDWRWDOO\GLIIHUHQWPXVHXP:LWKUHJDUG WR WKLV DUJXPHQW ZH SODQ WR H[WHQG WKH &R)LHOGV PRGHO ZLWK UDQGRP HOHPHQWV LQ DJHQWVt  EHKDYLRU UDQGRP HOHPHQWV DUH SUHVHQW LQ PRVW VZDUP LQWHOOLJHQW V\VWHP WR DGGUREXVWQHVVDQGIOH[LELOLW\ 7KHVHUDQGRPHOHPHQWVFRXOGEHLQWURGXFHGHLWKHUE\ DVVXPLQJWKDWDJHQWVEHKDYLRULVQRWGHWHUPLQLVWLFEXWUDQGRPLHWKHDJHQWIROORZV GRZQKLOOWKHILHOGt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t VGLYLVLRQRIODERU,QWKLVGLUHFWLRQ ZH ZLOO H[SORUH RWKHU DSSURDFKHV OLNH $PRUSKRXV &RPSXWLQJ >@ 'LVVLSDWLYH &HOOXODU$XWRPDWD'&$>@DQG$UWLILFLDO1HXUDO1HWZRUNV>@WRVHHLIDQGKRZWKH &R)LHOGVPRGHOFDQEHFRPELQHGZLWKWKHP 2QWKHRWKHUKDQGZHZRXOGOLNHWRGHYHORSDSURSHULQIUDVWUXFWXUHPLGGOHZDUHWR VXSSRUW WKH &R)LHOGV PRGHO LQ D UHDO FRPSXWLQJ VFHQDULR 7KH UROH RI WKLV LQIUDVWUXFWXUHZRXOGEHWRPDQDJHWKHGLVWULEXWHGGDWDVWUXFWXUHVDVVRFLDWHGWRILHOGV E\OHWWLQJDJHQWVWRJHQHUDWHDQGVHQVHILHOGV

References >@ >@

>@

+$EHOVRQ'$OOHQ'&RRUH&+DQVRQ*+RPV\7.QLJKW51DJSDO ( 5DXFK * 6XVVPDQ DQG 5 :HLVV u $PRUSKRXV &RPSXWLQJv  &RPPXQLFDWLRQVRIWKH$&0 0D\ 2 %DEDRJOX + 0HOLQJ $ 0RQWUHVRU u $QWKLOO $ )UDPHZRUN IRU WKH 'HYHORSPHQW RI $JHQW%DVHG 3HHUWR3HHU 6\VWHPVv  QG ,QWHUQDWLRQDO &RQIHUHQFH RQ 'LVWULEXWHG &RPSXWLQJ 6\VWHPV ,&'&6  9LHQQD $  -XO\  ( %RQDEHDX 0 'RULJR * 7KHUDXOD] u 6ZDUP ,QWHOOLJHQFHv  2[IRUG 8QLYHUVLW\3UHVV

&R)LHOGV7RZDUGVD8QLI\LQJ$SSURDFKWRWKH(QJLQHHULQJRI6ZDUP

>@ >@ >@ >@

>@ >@ >@ >@ >@ >@ >@ >@ >@ >@ >@ >@ >@ >@

( %RQDEHDX * 7KHUDXOD] u 6ZDUP VPDUWVv  6FLHQWLILF $PHULFDQ  0DUFK * &DEUL / /HRQDUGL ) =DPERQHOOL u (QJLQHHULQJ 0RELOH $JHQW $SSOLFDWLRQV YLD &RQWH[W'HSHQGHQW &RRUGLQDWLRQv  ,((( 7UDQVDFWLRQV RQ 6RIWZDUH(QJLQHHULQJ 6HSWHPEHU '(VWULQ'&XOOHU.3LVWHU*6XNMDWPHu &RQQHFWLQJWKH3K\VLFDO:RUOG ZLWK3HUYDVLYH1HWZRUNVv ,(((3HUYDVLYH&RPSXWLQJ -DQ 0 *DWWL 0 6SDOODQ]DQL $ 5ROL ) =DPERQHOOL u :KDW &DQ &HOOXODU $XWRPDWD7HOO8V$ERXWWKH%HKDYLRXURI /DUJH0XOWL$JHQW6\VWHPV"v VW ,QWHUQDWLRQDO :RUNVKRS RQ 6RIWZDUH (QJLQHHULQJ IRU /DUJH $JHQW 6\VWHPV 2UODQGR)/ 0D\ 6 -RKDQVVRQ - -RKDQVVRQ $ 6DIILRWWL u $Q HOHFWULF ILHOG DSSURDFK WR DXWRQRPRXV URERW FRQWUROv  7HFKQLFDO 5HSRUW  %OHNLQJH ,QVWLWXWH RI 7HFKQRORJ\'HF 6-RKQVRQu (PHUJHQFHv 6FULEQHU - 0 .DKQ 5 + .DW] . 6 - 3LVWHU u (PHUJLQJ &KDOOHQJHV 0RELOH 1HWZRUNLQJ IRU 6PDUW 'XVWv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

Creating and Updating the User Model

$VLPSOHPHWKRGUHLQIRUFHPHQWOHDUQLQJLVVHOHFWHGDVDEDVLVIRUPDLQWDLQLQJDOO SDUWVRIWKHXVHUPRGHO 3.2.1 Modelling User Personality 7KH OHYHO RI XVHU V VHOILVKQHVVDOWUXLVP LV XSGDWHG EDVHG RQ XVHU DFWLRQV WKDW KDYH LPSDFWRQWKHFRPPXQLW\HJSURYLGLQJQHZILOHVWREHVKDUHGHQDEOLQJILOHVKDULQJ IRUIROGHUVRQWKHXVHU VKDUGGLVNWKHSUHIHUHQFHVWKHXVHUKDVFKRVHQZLWKUHVSHFWWR ILOHWUDQVIHUVIURPRWKHUSHRSOHVHH)LJ )RUH[DPSOHLIWKHXVHUZDQWVWREULQJD QHZUHVRXUFHSDSHU LQWKHV\VWHPVKHKDVWRVSHQGVRPHHIIRUWWRDQQRWDWHWKHILOH ZLWKUHVSHFWWRFDWHJRU\VRWKDWLWLVVHDUFKDEOH,IWKHXVHULVZLOOLQJWRGRWKLVWKH OHYHORIDOWUXLVPRIWKHXVHULVLQFUHDVHG,IWKHXVHULQWHUUXSWVDRQJRLQJWUDQVIHUWKLV LV D VLJQ RI D VHOILVK EHKDYLRXU DQG WKH OHYHO RI DOWUXLVP LV GHFUHDVHG 2WKHU SDUDPHWHUVFRQVLGHUHGLQWKHFRPSXWDWLRQRIWKHSHUVRQDOLW\FKDUDFWHULVWLFRIWKHXVHU DUH WKH QXPEHU RI ILOHV VKDUHG E\ WKH XVHU WKH UHODWLYH GXUDWLRQ LQ ZKLFK WKH XVHU V VHUYHQW LV DFWLYH DQG WKH QXPEHU RI XVHU DFWLRQV WKDW DUH GHHPHG DV XQFRRSHUDWLYH VXFKDVUHPRYLQJGRZQORDGHGILOHVIURPWKHVKDUHGIROGHURUGLVDOORZLQJVKDULQJRI IROGHUV 3.2.2 Modelling User Interests (DFKVHUYHQWNHHSVWUDFNRIWKHDUHDVHQWHUHGE\WKH XVHUIRUVHDUFKLQWKH PRGHORI XVHULQWHUHVWV7KHDUHDVUHIOHFWWKHRQWRORJ\RIWKHGRPDLQ,Q&207(//$WKHVHDUH DVXEVHWRIWKH$&0VHWRIVXEMHFWFDWHJRULHV7KHVWUHQJWKRIXVHULQWHUHVW6aDWWLPHt LQHDFKVXEDUHDa LQWKHRQWRORJLFDOKHWHUDUFK\WKDWLVRQWKHSDWKOHDGLQJWRWKHVXE DUHDUHODWHGWRWKHZRUGVRITXHU\LVXSGDWHGDFFRUGLQJWRWKHUHLQIRUFHPHQWOHDUQLQJ IRUPXOD 6DHWW  L 6DHWW L  HW ZKHUHWKH QHZDWWLPH t HYLGHQFHRILQWHUHVWet ∈ > @ LV FDOFXODWHG DV et = 1/ d ZKHUH d = 1 + the distance EHWZHHQ WKH OHYHO RI WKH FXUUHQW VHDUFK DUHD LQ WKH RQWRORJ\WUHHDQGWKHOHYHORIWKHDUHDa

0RWLYDWLQJ3DUWLFLSDWLRQLQ3HHUWR3HHU&RPPXQLWLHV

Fig. 2. 8SGDWLQJ WKH PRGHO RI XVHU LQWHUHVWV 7KH QXPEHUV DVVRFLDWHG ZLWK DUFV RQ WKH ULJKW LQGLFDWH WKH RUGHU RI XSGDWLQJ ILUVW HYLGHQFH DERXW XVHU LQWHUHVW LQ FKDQWV LV SURSDJDWHG QH[WHYLGHQFHDERXWXVHULQWHUHVWLQIRONSRSLVSURSDJDWHG

7KH SDUDPHWHU i ∈ > @ LV DQ LQIODWLRQ UDWH XVHG WR PRGHO WKH IDFW WKDW ROGHU H[SHULHQFHV EHFRPH OHVV LPSRUWDQW RYHU WLPH t ZKLOH WKH PRVW UHFHQW H[SHULHQFH LV WKHPRVWUHOHYDQWVLQFHWKHXVHU VSUHIHUHQFHVPD\FKDQJHRYHUWLPH ,WFDQEHIL[HG DW D JLYHQ YDOXH VD\  JLYLQJ HTXDO ZHLJKWV WR ROG DQG QHZ HYLGHQFH 7KH SDUDPHWHUi FDQDOVREHDYDULDEOHZKLFKGHSHQGVRQWKHWLPHHODSVHGVLQFHWKHODVW HYLGHQFHRILQWHUHVWLQWKLVDUHDZKLFKDOORZVFDSWXULQJEHWWHUWKHFXUUHQWWHQGHQF\LQ XVHULQWHUHVWV$QH[DPSOHLVVKRZQLQ)LJXUH 1.1.3 Modelling User Relationships 7RPRGHOWKHUHODWLRQVKLSVRIWKHXVHUWKHVHUYHQWNHHSVWUDFNRIWKHIROORZLQJ     

ZKLFKVHUYHQWVUHVSRQGZLWKUHODWLYHO\PDQ\KLWVWRWKHXVHU VTXHU\ IURPZKLFKVHUYHQWVWKHXVHUFKRRVHVWRGRZQORDGILOHVRUWRUHTXHVWVHUYLFH WKHVXFFHVVRIHDFKGRZQORDGVHUYLFH ZKRLVVXHVTXHULHVWKDWUHVXOWLQKLWVLQWKHOLVWRIUHVRXUFHVRIWKHXVHU ZKRGRZQORDGVILOHVIURPWKHXVHU

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

0RWLYDWLQJ3DUWLFLSDWLRQLQ3HHUWR3HHU&RPPXQLWLHV

LQWHUIDFHDVWRZKDWLVWKHUHDVRQIRUWKHLQFUHDVHGHFUHDVHLQWKH4R6DQGZKDWWKH XVHU VKRXOG GR LQ RUGHU WR LPSURYH LW  7KH 4R6 VKRXOG QRW GHWHULRUDWH FRPSOHWHO\ HYHQIRUXQFRRSHUDWLYHIUHHULGHUVVLQFHWKHUHLVDOZD\VDKRSHWKDWWKH\PD\EHFRPH FRRSHUDWLYHLIWKH\ILQGWKHULJKWLQWHUHVWJURXS $Q HFRQRPLF PRGHO LQ D 33 V\VWHP KRZHYHU EULQJV DQ RYHUKHDG LW UHTXLUHV DGGLWLRQDO UHDVRQLQJ FDSDELOLWLHV RQ EHKDOI RI WKH VHUYHQWV VHUYHQWV UHGXFHV WKH DQRQ\PLW\LQWKHV\VWHP>@DQGUHTXLUHVFHQWUDOL]HGFRPSRQHQWVWREHLQWURGXFHGWR EHUHVSRQVLEOHIRUWKHFXUUHQF\WUDQVDFWLRQVDVLQ0RMRQDWLRQ 

5

Conclusions

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

-XOLWD9DVVLOHYD

2XUIXWXUHVWHSVDUHHYDOXDWLQJWKHDGYDQWDJHVDQGGLVDGYDQWDJHVRIWKHSURSRVHG GHVLJQZLWKUHVSHFWWRXVHUVDWLVIDFWLRQDQGSHUIRUPDQFH

Acknowledgements 7KH DXWKRU WKDQNV +HOHQ %UHW]NH DQ 16(5&  &5$: VXPPHU VWXGHQW ZKR LQYHVWLJDWHG YDULRXV SHUVXDVLRQ VWUDWHJLHV DQG GHVLJQHG WKH &RPXWHOOD V LQWHUIDFH 7KDQNV JR DOVR WR &KULVWRSKHU &R[ DQ 16(5& VXPPHU VWXGHQW IRU WKH LPSOHPHQWDWLRQ RI WKH &RPXWHOOD VHUYHQW DQG WR @ >@ >@ >@ >@

>@ >@ >@ >@

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

0RWLYDWLQJ3DUWLFLSDWLRQLQ3HHUWR3HHU&RPPXQLWLHV

>@

>@

>@ >@ >@

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u VHOI DGDSWLYH VRIWZDUH HYDOXDWHV LWV RZQ EHKDYLRU DQG FKDQJHV EHKDYLRU ZKHQ WKH HYDOXDWLRQLQGLFDWHVWKDWLWLVQRWDFFRPSOLVKLQJZKDWWKHVRIWZDUHLVLQWHQGHGWRGR RUZKHQEHWWHUIXQFWLRQDOLW\RUSHUIRUPDQFHLVSRVVLEOHv 7KHPXOWLDJHQWFRPPXQLW\  

7KH$0$6WKHRU\LVGHYHORSHGDQGDSSOLHG IRU WKH ODVW \HDUV DW WKH 5HVHDUFK ,QVWLWXWH LQ &RPSXWHU 6FLHQFHLQ7RXORXVH,5,7 6HHKWWSZZZLULWIU60$&

33HWWDHWDO(GV (6$:/1$,SS ©6SULQJHU9HUODJ%HUOLQ+HLGHOEHUJ

$'(/)($0HWKRGRORJ\IRU$GDSWLYH0XOWLDJHQW6\VWHPV(QJLQHHULQJ

JHQHUDOO\FRQVLGHUVWKH0XOWL$JHQW6\VWHPV0$6 DVEHLQJDGDSWLYHEHFDXVHDJHQWV DUHDXWRQRPRXVVLWXDWHGSURDFWLYHVRFLDOg 7KHFRQGLWLRQIRUWKHV\VWHPWRDGDSWLV WREHFRPSRVHGRIDXWRQRPRXVDJHQWV,QRXUSRLQWRIYLHZWKLVNLQGRIDGDSWDWLRQLV ZHDNDGDSWDWLRQDQGDOOH[LVWLQJPXOWLDJHQWV\VWHPVDUHDGDSWLYHLQWKLVVHQVH :H FDOO u VWURQJ DGDSWDWLRQv  RI D PXOWLDJHQW V\VWHP WKH DELOLW\ WKLV V\VWHP SRVVHVVHVWRWDNHLQWRDFFRXQWXQSUHGLFWDEOHHYHQWVLQ RUGHU WR UHDFW WR HYROXWLRQDU\ HQYLURQPHQWV LQ RUGHU WR UHDOL]H LWV u ULJKWv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

2

The AMAS Theory

7KH ILUVW DLP RI WKLV WKHRU\ LV WR UHDOL]H 0$6 KDYLQJ WKH u FODVVLFDOv  FKDUDFWHULVWLFV JLYHQLQ>@WREXLOGDVRFLHW\RIVLWXDWHGDJHQWV%XWIURPRXUSRLQWRIYLHZVXFKD 0$6LVDOVRSOXQJHGLQWRDQHQYLURQPHQWDQGPXVWUHDFKDEHKDYLRUDODGHTXDF\E\ UHSURGXFLQJWKHEHKDYLRURIDVLPXODWHGVRFLHW\ RUDIXQFWLRQDORQHE\SHUIRUPLQJ WKH ULJKW WDVN WKH WDVN IRU ZKLFK WKH V\VWHP KDV EHHQ EXLOW  ,Q RXU YLVLRQ WKH LPSRUWDQW QRWLRQ LV WKH FROOHFWLYH WKH $0$6 WKHRU\ PXVW WKHQ OHDG WR D FRKHUHQW FROOHFWLYHDFWLYLW\WKDWUHDOL]HVWKHULJKWWDVN :HWKHQSURYHGWKHIROORZLQJWKHRUHP>@u )RUDQ\IXQFWLRQDOO\DGHTXDWHV\VWHP WKHUH LV DW OHDVW D FRRSHUDWLYH LQWHULRU PHGLXP V\VWHP ZKLFK IXOILOOV DQ HTXLYDOHQW IXQFWLRQ LQ WKH VDPH HQYLURQPHQWv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u FRRSHUDWLYHv WKLVGRHVQRWPHDQWKDWLWLV DOZD\VKHOSLQJWKHRWKHURQHVRUWKDWLWLVDOWUXLVWLFEXWRQO\WKDWLWLVDEOHWRUHFRJQL]H FRRSHUDWLRQ IDLOXUHV FDOOHG u 1RQ &RRSHUDWLYH 6LWXDWLRQVv  1&6 ZKLFK FRXOG EH UHODWHGWRH[FHSWLRQVLQFODVVLFDOSURJUDPV DQGWRWUHDWWKHP7KHORFDOWUHDWPHQWRI 1&6 LV D PHDQV WR EXLOG D V\VWHP WKDW GRHV WKH EHVW LW FDQ ZKHQ D GLIILFXOW\ LV

&DUROH%HUQRQHWDO

HQFRXQWHUHG 6XFK D GLIILFXOW\ LV SULPDULO\ GXH WR WKH G\QDPLFDO QDWXUH RI WKH HQYLURQPHQW RI WKH V\VWHP DV ZHOO DV WKH G\QDPLFV RI WKH LQWHUDFWLRQV EHWZHHQ DJHQWV0RUHSUHFLVHO\DQDJHQWFDQGHWHFWWKUHHNLQGVRI1&6 • • •

ZKHQ D VLJQDO SHUFHLYHG IURP LWV HQYLURQPHQW LV QRW XQGHUVWRRG DQG QRW UHDG ZLWKRXWDPELJXLW\ ZKHQ WKH LQIRUPDWLRQ SHUFHLYHG GRHV QRW LQGXFH WKH DJHQW WR DQ DFWLYLW\ SURFHVV ZKHQFRQFOXGLQJUHVXOWVOHDGWRDFWLQDXVHOHVVZD\LQWKHHQYLURQPHQW

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

3

A Short Overview of the ADELFE Methodology

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requirements workflow ZKLFK LV DOVR WDNHQ LQWR DFFRXQW LQ 752326 LV D IXQGDPHQWDO VWHS LQ VRIWZDUH HQJLQHHULQJ ,Q WKH $0$6 WKHRU\ WKH DGDSWDWLRQ SURFHVV VWDUWV IURP WKH LQWHUDFWLRQV EHWZHHQ WKH V\VWHP DQG LWV HQYLURQPHQW 7KHUHIRUHLWLVLPSRUWDQWWRJLYHD PRGHORIWKH HQYLURQPHQW GXULQJ WKLV ZRUNIORZ  

$'(/)( LV D QDWLRQDO SURMHFW VWDUWHG LQ 'HFHPEHU  DQG VXSSRUWHG E\ WKH )UHQFK 0LQLVWU\ RI WKH (FRQRP\ )LQDQFH DQG ,QGXVWU\ 7KH GLIIHUHQW SDUWQHUV RI WKLV SURMHFW DUH ,5,7 8QLYHUVLW\ RI 7RXORXVH ,,,  DQG /, 8QLYHUVLW\ RI /D 5RFKHOOH  IURP DFDGHPLD DQG DUH$57$/DQG71,IURPLQGXVWU\

$'(/)($0HWKRGRORJ\IRU$GDSWLYH0XOWLDJHQW6\VWHPV(QJLQHHULQJ

7KHenvironment modelFRQVLVWVLQWKUHHVWHSVGHWHUPLQLQJWKHDFWRUVGHILQLQJWKH FRQWH[WDQGFKDUDFWHULVLQJWKHHQYLURQPHQW ,QWKHanalysis workflowZHDGGHGWZRVWHSVWRWKH583WKHLGHQWLILFDWLRQRIWKH DJHQWVDQGWKHDGHTXDF\RIWKH$0$6WKHRU\$'(/)(IRFXVHVRQWKHidentification of the agents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the AMAS theory is adequate WR LPSOHPHQW KLV DSSOLFDWLRQ )RU H[DPSOH LI WKH DOJRULWKPUHTXLUHGWRUHVROYHWKHWDVNLVDOUHDG\NQRZQLIWKHWDVNLVQRWFRPSOH[LI WKHV\VWHPLVFORVHGRULIQRXQH[SHFWHGHYHQWVFDQRFFXUWKLVNLQGRISURJUDPPLQJLV XVHOHVV ,Q WKH design workflow WKH agent model DQG WKH NCS model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

Fig. 1.2YHUYLHZRIWKHWKUHHILUVWFRUHZRUNIORZVRI$'(/)(

&DUROH%HUQRQHWDO

4

Step-by-Step Details of the Methodology

,Q WKLV VHFWLRQ ZH SUHVHQW WKH GLIIHUHQW VWHSV RI WKH $'(/)( PHWKRGRORJ\ $ FDVH VWXG\DERXWWLPHWDEOLQJSUREOHPLOOXVWUDWHVWKHPHWKRGRORJ\DWHDFKVWHS7KHPDLQ DFWRUVRIWKLVFDVHVWXG\DUHWHDFKHUVDQGVWXGHQWVZKRQHHGWRERRNVHYHUDOURRPVWR DFKLHYH WKHLUV WDVNV 7KLV FDVH VWXG\ KDV EHHQ FKRVHQ EHFDXVH LW LV EDVHG RQ D QRQ SUHGLFWDEOH HQYLURQPHQW DQG LW LV DQ RSHQ SUREOHP $FWXDOO\ VRPH URRPV FRXOG EHFRPHXQDYDLODEOHEHFDXVHRIVRPHNLQGRISUREOHPRUDWHDFKHUt VDYDLODELOLW\FRXOG FKDQJHDQGJLYLQJDVROXWLRQWRWKLVSUREOHPPHDQVDGDSWLQJLQRUGHUWRUHDFWWRWKHVH G\QDPLFFKDQJHVRIWKHHQYLURQPHQW)XUWKHUPRUHWKLVSUREOHPFDQEHH[WHQGHGWRD GLDU\ PDQDJHPHQW IRU H[DPSOH LQ ZKLFK QHZ SHRSOH DQG VR QHZ DJHQWV PXVW EH DGGHGLQWKHUXQQLQJV\VWHPWKLVSUREOHPFDQEHFRQVLGHUHGDUHEHLQJRSHQ 4.1

Requirements Workflow

7KHDLPVRIWKLVVWDJHDUHWRGHILQHWKHV\VWHPWREHWRWUDQVIRUPWKLVYLHZLQDXVH FDVHPRGHODQGWRRUJDQL]HDQGWRPDQDJHWKHUHTXLUHPHQWVIXQFWLRQDORUQRW DQG WKHLU SULRULWLHV ,Q IDFW DW WKLV VWDJH WKH GHVLJQHU KDV WR GHILQH WKH IXQFWLRQ RI WKH VWXGLHGV\VWHPDQGWRPRGHOLWVHQYLURQPHQW Definition of the studied system.7KLVGHILQLWLRQLVWKHUHVXOWRIWKHDQDO\VLVRIWKH UHTXLUHPHQWVVHWDUWLIDFW7KHRXWSXWRIWKLVVWHSLVDNH\ZRUGVHWGHILQLQJWKHV\VWHP )URPWKHXVHUUHTXLUHPHQWVIRUWKHWLPHWDEOLQJSUREOHPWKHIROORZLQJNH\ZRUGVDQG FRQFHSWV DUH KLJKOLJKWHG SODQQLQJ URRPV WHDFKHUV VWXGHQWV FRQVWUDLQWV RUJDQL]DWLRQDQGFRQVWUDLQWPDQDJLQJ$VDQRXWORRNWKHSUREOHPFDQEHYLHZHGDV WKHRUJDQL]DWLRQRIWHDFKHUVDQGVWXGHQWVLQURRPVDQGDOOWKHSDUWLFLSDQWVDUHVXEMHFW WRFRQVWUDLQWV 7KLVSUREOHPEHORQJVWRWKHFRQVWUDLQWVDWLVIDFWLRQSUREOHPVFODVV Definition and Environment Modeling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



:HHODERUDWHGWKLVH[DPSOHDVDFDVHVWXG\WRFRPSDUHDQGGLVFXVVGLIIHUHQWPHWKRGRORJLHV DQG PXOWLDJHQW SODWIRUPV IRU WKH $6$ *URXS RI WKH $UWLILFLDO ,QWHOOLJHQFH )UHQFK $VVRFLDWLRQ

$'(/)($0HWKRGRORJ\IRU$GDSWLYH0XOWLDJHQW6\VWHPV(QJLQHHULQJ

V\VWHP LW PD\ RQO\ KDYH WR FKDQJH FRQVWUDLQWV RZQHU FRQVWUDLQWV RU URRP FRQVWUDLQWV  ,Q WKH RWKHU VHQVH WKH V\VWHP LQWHUDFWV ZLWK WKH DFWLYH HQWLW\ E\ GLVSOD\LQJWKHSODQQLQJ

Fig. 2.7KHXVHFDVHVDQGWKHLUUHODWHGDFWRUVIRUWKHWLPHWDEOLQJSUREOHP

7KH WKLUG VWUHVV RI WKLV VWHS LV WR FKDUDFWHUL]H WKH HQYLURQPHQW DFFRUGLQJ WR WKH FODVVLILFDWLRQJLYHQE\5XVVHOOLQ>@7KLVFKDUDFWHUL]DWLRQPD\HQDEOHWKHGHVLJQHU WRGHWHFWVRPHVSHFLDOXVHFDVHVWKDWDLPWRUHVSRQGWRWKHHQYLURQPHQWEHKDYLRU,Q WKHFDVHVWXG\ZHFDQFKDUDFWHUL]HWKHHQYLURQPHQWRIWKHV\VWHPDVIROORZLQJ • • • •

'\QDPLFWKHHYROXWLRQRIWKHDFWLYHHQWLWLHVGRHVQRWGHSHQGRQWKHV\VWHPLW LVXQSUHGLFWDEOHIURPWKHSRLQWRIYLHZRIWKHV\VWHP $FFHVVLEOHWKHV\VWHPFDQREWDLQLQIRUPDWLRQRQWKHVWDWHRIWKHHQYLURQPHQW 1RQGHWHUPLQLVWLFWKHV\VWHPLVQRWDEOHWRNQRZZKDWFRXOGEHWKHHIIHFWVRI LWVDFWLRQVRQWKHDFWLYHHQWLWLHV &RQWLQXRXVWKH QXPEHURILQWHUDFWLRQVEHWZHHQWKHV\VWHPDQGWKH HQWLWLHV LV LQILQLWH

Determination of the Use Cases. 7KH PDLQ REMHFWLYH RI WKLV VWHS ZKLFK HQGV WKH UHTXLUHPHQWV ZRUNIORZ LV WR FODULI\ WKH GLIIHUHQW IXQFWLRQDOLWLHV WKH V\VWHP KDV WR UHVSRQG WR 2QO\ WKH DFWLYH HQWLWLHV DUH LPSOLHG LQ WKHVH XVH FDVHV ZKLFK DUH WKH UHVXOWVRIDIXQFWLRQDOUHTXLUHPHQWVVHW7KHXVHFDVHVIRUWKHWLPHWDEOLQJSUREOHPDUH VKRZQLQILJXUH 4.2

Analysis Workflow

)URPDPXOWLDJHQWSRLQWRIYLHZWKHLGHQWLILFDWLRQRIWKHDJHQWVPXVWWDNHSODFHLQ WKLVZRUNIORZ7KHDQDO\VLVZRUNIORZKDVWRGHYHORSDQXQGHUVWDQGLQJRIWKHV\VWHP LWVVWUXFWXUHLQWHUPVRIFRPSRQHQWVDQGWRNQRZLIWKH$0$6WKHRU\LVUHTXLUHG Domain Analysis and Architecture Study.'RPDLQDQDO\VLVLVDVWDWLFYLHZDQGDQ DEVWUDFWLRQRIWKHUHDOZRUOGDQGWKHOLQNHGHQWLWLHV&RQVLGHULQJVHSDUDWHO\HDFKXVH FDVH E\ GHILQLQJ VFHQDULRV WKH GHVLJQHU KDV WR GLYLGH WKH V\VWHP LQWR HQWLWLHV 7KH UHVXOW RI WKLV VWHS LV D VHW RI HQWLWLHV LQ SUHOLPLQDU\ FODVV GLDJUDPV Teacher

&DUROH%HUQRQHWDO

CourseManager StudentGroup Room RoomManager DQG PPN FODVVHV DSSHDUQDWXUDOO\DVUHDOZRUOGHQWLWLHV,QDVHFRQGWLPH ZHWULHGWRGHWHUPLQH ZKDW HQWLWLHV FRXOG EH XVHIXO IRU RXU V\VWHP :H SURSRVH D ERDUG WR YLVXDOL]H WKH RUJDQL]DWLRQWKHGridDQGCellFODVVHV DQGWKHConstraintManagerFODVVWR FRQWURO FRQVWUDLQWV IRU HDFK HQWLW\ WKDW RZQV D Constraint FODVV LQVWDQFH &HOOV UHSUHVHQWLQWHUVHFWLRQVRIGLIIHUHQWGLPHQVLRQVGD\VURRPVHWF  Adequacy of the AMAS Theory.7KLVVWHSDLPVWRKHOSWKHGHVLJQHUWRGHFLGHLIWKH $0$6WKHRU\LVDGHTXDWHWRVROYHKLVSUREOHPEHFDXVHIRUFHUWDLQDSSOLFDWLRQVWKLV NLQG RI SURJUDPPLQJ FDQ EH XVHOHVV $ VRIWZDUH KDV EHHQ GHYHORSHG ZLWK VHYHUDO FULWHULDWRVWXG\WKHDGHTXDF\DWWZROHYHOV • •

$W WKH JOREDO OHYHO WR DQVZHU WKH TXHVWLRQ u LV D V\VWHP LPSOHPHQWDWLRQ XVLQJ $0$6QHHGHG"v $WWKHORFDOOHYHOWRWU\WRDQVZHUWKHTXHVWLRQu GRVRPHFRPSRQHQWVQHHGWREH LPSOHPHQWHGDV$0$6"v LHLVVRPHGHFRPSRVLWLRQRUUHFXUVLRQXVHIXOGXULQJ GHVLJQ"

)RUWKHFDVHVWXG\WKHGHFLVLRQWRROFOHDUO\ VXJJHVWVWRXVHWKH $0$6 WR GHVLJQ WKHJOREDOOHYHO0RUHRYHUWKHWRROLQGLFDWHVWKDWVRPHHQWLWLHVFRXOGEHGHFRPSRVHG DV$0$66RRQFHWKHDJHQWVDUHLGHQWLILHGWKHGHVLJQHUKDVWRUHXVHWKHPHWKRGRQ WKHPDVGHYHORSHGEHORZ Agent Identification. ,Q WKLV VWHS ZH DUH RQO\ LQWHUHVWHG LQ DJHQWV WKDW HQDEOH D GHVLJQHUWREXLOGRXUVRUWRI$0$67KHGHVLJQHUKDVWRGHWHUPLQHZKLFKHQWLWLHVILW ZLWKWKLVDJHQWW\SH7KLVLGHQWLILFDWLRQLVGRQHFRQVLGHULQJWKHQRWLRQRIFRRSHUDWLYH DJHQW DQG WKH DJHQWt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dequacy of the AMAS Theory at the Local Level.,IWKHILUVWVWHSRIDGHTXDF\WR WKH$0$6WKHRU\LQGLFDWHVDSRVVLEOHGHFRPSRVLWLRQHDFKDJHQWKDVWREHDQDO\]HG DV D V\VWHP 7KH JRDOV RI DQ DJHQW Teacher RU StudentGroup DUH WR ILQG GLIIHUHQWSODFHVDQGSDUWQHUVWR IROORZ RU WR JLYH HDFK FRXUVH 7KHVH JRDOV UDLVH WKH SUREOHP RI XELTXLW\ $JHQWV FDQQRW EH DW GLIIHUHQW SODFH DW GLIIHUHQW PRPHQWV 7KHUHIRUH ZH SURSRVH WR FUHDWH RQH DJHQW SHU FRXUVH IRU HDFK WHDFKHU RU VWXGHQW JURXS7ZRDJHQWOHYHOVDUHGLVWLQJXLVKHG • •

RepresentativeAgent5$ DWWKHKLJKHVWOHYHOLWUHSUHVHQWVDWHDFKHURU DVWXGHQWJURXSZLWKLQWKHV\VWHP BookingAgent %$  DW WKH ORZHVW OHYHO LW LV UHVSRQVLEOH IRU ILQGLQJ SDUWQHUV DQG ERRNLQJ URRPV IRU D RepresentativeAgent 7KHUH DUH DV PDQ\%$VDVWKHQXPEHURIFRXUVHVDWHDFKHUKDVWRJLYHRUDVWXGHQWJURXSKDV WRIROORZ

$'(/)($0HWKRGRORJ\IRU$GDSWLYH0XOWLDJHQW6\VWHPV(QJLQHHULQJ

7KHLGHQWLILHGDJHQWVKDYHWREHDGGHGWRWKHSUHOLPLQDU\FODVVGLDJUDPDVVKRZQ LQILJ Study of Interactions between the Different Entities.7KHUHVXOWRIWKLVVWHSLVDVHW RIVHTXHQFHGLDJUDPVDQGDFWLYLW\GLDJUDPVZKLFKH[SODLQVWKHSRVVLEOHLQWHUDFWLRQV EHWZHHQWKHGLIIHUHQWHQWLWLHVZLWKLQWKHV\VWHPDQGDWHDFKOHYHO$VWKH583LVXVH FDVHJXLGHGIRUHDFKXVHFDVH ZKLFK KDV EHHQ GHILQHG EHWZHHQ WKH V\VWHP DQG WKH HQYLURQPHQWDVHTXHQFHGLDJUDPKDVWREHGHILQHGWRVKRZWKHLQWHUQDOYLHZDQGWKH LQWHUDFWLRQV ZLWKLQ WKH V\VWHP 7KH LQWHUDFWLRQ EHWZHHQ DJHQWV FDQ EH ZULWWHQ ZLWK $80/ 4.3

Design Workflow

,Q WKLV VHFWLRQ ZH GHWDLO WKH GHVLJQ ZRUNIORZ LQ IRXU VWHSV %HFDXVH WKH FRPSOHWH GHVLJQ FDQQRW EH GHVFULEHG LQ WKLV SDSHU ZH RQO\ GHWDLO WKH VWHSV ZKLFK DUH QRW H[LVWLQJ LQ RWKHU PHWKRGRORJLHV VXFK DV WKH DJHQW PRGHO DQG WKH PRGHOLQJ RI 1RQ &RRSHUDWLYH6LWXDWLRQV Detailed Architecture and Agent Model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g ,Q$'(/)(ZHSURSRVHDVSHFLILFGHVLJQSDWWHUQQDPHGFRRSHUDWLYHDJHQW DUFKLWHFWXUH,QRXUFDVHVWXG\IRXUSDFNDJHVDSSHDU • • • •

$JHQWSDFNDJHWRPDQDJH%$VDQG5$V *ULGSDFNDJHWRPDQDJHWKHGLIIHUHQWGLPHQVLRQVRIDJULGDQGLWVFHOOV &RQVWUDLQWSDFNDJHZKLFKKDVWREHDFFHVVLEOHWRURRPVDQGDJHQWV ,QWHUIDFHSDFNDJHWRHQDEOHDXVHUWRLQWHUDFWZLWKWKHV\VWHP

Fig. 3.7KHUHYLVHGSUHOLPLQDU\FODVVGLDJUDPIRUWKHWLPHWDEOLQJSUREOHPJLYHVXVDQLGHDRI WKH0$6PRGHO

&DUROH%HUQRQHWDO Table 1. BookingAgent VSHFLILFDWLRQ VNLOOV DSWLWXGHV DQG LQWHUDFWLRQ ODQJXDJHV DUH PHWKRGVUHSUHVHQWDWLRQVDUHDWWULEXWHV BookingAgent Representations •

Constraints

•

partners

•

bookState

•

recentlyMetAgents

•

partnershipState

•

RAFather

•

brotherConstraints Skills

•

currentCell

•

moveInTheGrid

•

managePartnership

•

manageConstraints

•

manageMessages

•

manageBooking Aptitudes

•

bookARoom

•

cancelPartnership

•

cancelBooking

•

negotiatePartnership

•

negotiateBooking

•

SendMessage

•

establishPartnership • Interaction Language

InterpretMessage

•

messageInteraction

•

contactInteraction

Cooperative Agent Architecture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ookingAgentLVJLYHQLQWKHWDEOH$%$ LV DEOH WR PRYH LQ WKH JULG WR PHHW RWKHU %$V DQG WR QHJRWLDWH SDUWQHUVKLSV ,W FDQ DOVRVHQGDQGUHFHLYHPHVVDJHVDQGLQWHUSUHWWKHP7RDFKLHYHLWVWDVNLWPXVWNQRZ WKHFRQVWUDLQWVRIDOOWKH%$VZRUNLQJIRULWVRepresentativeAgent&RQVWUDLQWV FDQEHUHOD[HGGXULQJDQHJRWLDWLRQ

$'(/)($0HWKRGRORJ\IRU$GDSWLYH0XOWLDJHQW6\VWHPV(QJLQHHULQJ Table 2.7KH%RRNLQJ&RQIOLFW1&6WDEOHRIDBookingAgent Booking Conflict State Any Description The BA is in a cell that is interesting to book but this cell is already booked Conditions The BA is in a cell AND this latter is already booked AND yet the cell would be suitable if not booked Actions IF the cost of the new booking is less than the older one THEN the BA books the cell ELSE the BA moves elsewhere

Non Cooperative Situation Model. 7KLV VWHS UHSUHVHQWV WKH PDLQ FRQWULEXWLRQ RI $'(/)(WRWKLVZRUNIORZ'XULQJLWWKHGHVLJQHUPXVWILOOXSDWDEOHGHVFULELQJHDFK 1&6HQFRXQWHUHGE\HDFKSUHYLRXVO\LGHQWLILHGDJHQW7KLVWDEOHFRQWDLQVWKHQDPHRI WKH1&6WKHVWDWHLQZKLFKWKHDJHQWLV ZKHQGHWHFWLQJLWWKHWH[WXDOGHVFULSWLRQRI WKH 1&6 DQG WKH FRQGLWLRQV DQG DFWLRQV OLQNHG WR LW 7KH FRQGLWLRQV GHVFULEH WKH GLIIHUHQW HOHPHQWV WKDW HQDEOH WKH DJHQW WR ORFDOO\ GHWHFW WKLV 1&6 7KH DFWLRQV GHVFULEH ZKDW WKH DJHQW KDV WR GR WR UHPRYH LW )RU LQVWDQFH WKH 1&6 IRU D BookingAgentDUH • • • • • •

0       WKH %$ PHHWV DQRWKHU %$ WKDW PD\ EH DQ XQLQWHUHVWLQJSDUWQHU .    WKH%$LVLQDFHOOWKDWLVXQLQWHUHVWLQJWRERRN J       WKH %$ UHFHLYHV D PHVVDJH WKDW LV QRW FRUUHFWO\ DGGUHVVHG 0   WKH%$PHHWVDQRWKHU%$WKDWLVLQWHUHVWLQJEXWWKLVRWKHU %$KDVDOUHDG\DSDUWQHU .  WKH%$LVLQDFHOOWKDWLVLQWHUHVWLQJWRERRNEXWWKLVFHOOLV DOUHDG\ERRNHGVKRZQLQWDEOH  .     WKH %$ PHHWV LWV SDUWQHU WKH\ PXVW VHSDUDWH WR H[SORUH PRUHHIILFLHQWO\WKHJULG

(YHQ LI DOO WKH EHKDYLRUV RI WKH FRRSHUDWLYH DJHQWV DUH JLYHQ DJHQW PRGHO  1&6 PRGHO WKH0$6DVDZKROHFDQDGDSWLWVHOIEHFDXVHWKHLQWHUDFWLRQVDUHQRWDSULRUL FRGHG &KDQJLQJ WKH LQWHUDFWLRQV EHWZHHQ DJHQWV VHOIRUJDQL]DWLRQ  FKDQJHV WKH JOREDOIXQFWLRQRIWKHV\VWHPDQGWKHQDOORZVWKHu VWURQJDGDSWDWLRQv RIWKH0$6 Class Diagrams.7KLVVWHSSURYLGHVWKHGLIIHUHQWFODVVGLDJUDPV IRUWDNLQJWKH*8, DQGGDWDEDVHGHVLJQVLQWRDFFRXQW7KHVHFODVVGLDJUDPVDUHDOVR UHILQHPHQWV RI WKH SUHYLRXVFODVVGLDJUDPV$'(/)(GRHVQRWQHHGWRSURYLGHDGGLWLRQDOIXQFWLRQDOLWLHV WR EXLOG WKHP EHFDXVH WKH GHVLJQHU IROORZV WKH VDPH PHWKRG DV WKH RQH XVHG LQ DQ REMHFWRULHQWHG PHWKRGRORJ\ 7KH ILJXUH  VKRZV WKH ILQDO FODVV GLDJUDP IRU WKH V\VWHP

&DUROH%HUQRQHWDO

Fig. 4.7KH ILQDO FODVV GLDJUDP IRU WKH WLPHWDEOLQJ SUREOHP DQG WKH VWDWLF LQWHU SDFNDJH OLQNV EHWZHHQDJHQWVDQGREMHFWV

4.4

Implementation and Tests Workflows

7KLVZRUNIORZLVVLPLODUWRZKDWLVGRQHLQWKH583+RZHYHUWRIDFLOLWDWHWKHZRUN RIWKHGHYHORSHUZHSODQWRSURYLGHDUDSLGSURWRW\SLQJWRROWKDWZLOOEHLQWHJUDWHGLQ WKH2SHQ7RRO‹VRIWZDUHSURYLGHGE\RXUSDUWQHU71,KWWSZZZWQLIU 2SHQ7RRO LVDJUDSKLFDOWRROOLNH5DWLRQDO5RVHZKLFKVXSSRUWVWKH80/QRWDWLRQ &RQFHUQLQJ WKH WLPHWDEOLQJ SUREOHP ZH KDYH LPSOHPHQWHG WKH SURSRVHG DUFKLWHFWXUH$JHQWVILQGDGHTXDWHRUJDQL]DWLRQVDQGSURYLGHUHOHYDQWUHVXOWV $WWKH PRPHQW ZH RQO\ KDYH WHVWHG WKH V\VWHP RQ VLPSOH H[DPSOHV WKDW DSSHDU LQ WKH UHTXLUHPHQWVEXWZHKRSHWRHQODUJHLWVDSSOLFDWLRQ

5

Contributions, Strengths and Limitations of ADELFE

752326 >@ H[SUHVVHV WKH G\QDPLF DQG RSHQQHVV RI WKH DSSOLFDWLRQ LQ WKH UHTXLUHPHQWVSKDVHVZLWKWKHPRGHORIWKHHQYLURQPHQWDQGZLWKSDUWLFXODUVRIWJRDOV +RZHYHULWGRHVQRWJLYHJXLGHOLQHVWRGHVLJQWKHULJKWDJHQWVt EHKDYLRUDOORZLQJWKH

$'(/)($0HWKRGRORJ\IRU$GDSWLYH0XOWLDJHQW6\VWHPV(QJLQHHULQJ

DGDSWDELOLW\ RI WKH V\VWHP *$,$ LQGLFDWHV WKDW WKH GRPDLQ FRYHUHG E\ WKH PHWKRGRORJ\LVVWDWLFDQGWKDWWKHPHWKRGRORJ\LVGHGLFDWHGWRFORVHGGRPDLQZKHUH DJHQWVt VNLOOVDQGEHOLHIVDUHVWDWLFDWUXQWLPH>@+RZHYHUVRPHLPSURYHPHQWVWR *$,$DUHVXJJHVWHGLQ>@WRVXSSRUWDSSOLFDWLRQVLQG\QDPLFGRPDLQV0DQ\RWKHU PHWKRGRORJLHV OLNH $$,, >@ 0D6( >@ RU 0(66$*( >@ GR QRW IRFXV RQ WKH G\QDPLF DVSHFW RI WKH VRIWZDUH HQYLURQPHQW DQG RQ WKH DGDSWDWLRQ DELOLWLHV RI WKH VRIWZDUH 7KHVSHFLILFLW\RI$'(/)(LVWRSURYLGHDPHWKRGRORJ\WRGHVLJQDGDSWLYHPXOWL DJHQW V\VWHPV FRXSOHG ZLWK D WKHRU\ RQ WKRVH V\VWHPV 7KH SURYLGHG PHWKRGRORJ\ FRYHUVWKHZKROHu FODVVLFDOv OLIHF\FOHRIVRIWZDUHDQGUHXVHV80/QRWDWLRQV7KHVH QRWLRQV DUH DOUHDG\ IDPLOLDU WR PRVW RI GHVLJQHUV PDNLQJ WKLV WRRO HDVLO\ XVDEOH E\ GHYHORSHUVQRWVSHFLDOL]HGLQGHVLJQLQJ$0$6 ,Q DGDSWLYH PXOWLDJHQW V\VWHPV WKH HQYLURQPHQW LQ ZKLFK WKH V\VWHP LV RSHUDWLQJ LVDNH\QRWLRQEXWLQD JHQHUDO ZD\ WKH HQYLURQPHQW PRGHOLQJ LV QRW D FHQWUDOSRLQWLQH[LVWLQJPHWKRGRORJLHV,Q'(6,5(>@WKHHQYLURQPHQWLVWDNHQLQWR DFFRXQW DW WKH DJHQW OHYHO LQ WKH u ZRUOG LQWHUDFWLRQ PDQDJHPHQW PRGXOHv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

&DUROH%HUQRQHWDO

6

Conclusion and Future

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

>@

&%HUQRQ03*OHL]HV*3LFDUG 3*OL]Hx 7KH$'(/)(0HWKRGRORJ\ IRU DQ ,QWUDQHW 6\VWHP x  4 (      3  4(3, >  2431, >7RURQWR0D\ )07%UD]LHU3$79DQ(FN -7UHXU0RGHOOLQJDVRFLHW\RIVLPSOH DJHQWV)URPFRQFHSWXDOVSHFLILFDWLRQWRH[SHULPHQWDWLRQK  4      x  3 %XUUDIDWR  0 &RVVHQWLQR x  'HVLJQLQJ D 0XOWL$JHQW 6ROXWLRQ IRU D %RRNVWRUH ZLWK WKH 3$66, 0HWKRGRORJ\ x  4(3, >  2431, > 7RURQWR 0D\ * &DLUH ) /HDO 3 &KDLQKR 5 (YDQV ) *DULMR - *RPH] * 3DYRQ 3.HDUQH\ - 6WDUN  3 0DVVRQHW  $JHQW 2ULHQWHG $QDO\VLV XVLQJ 0(66$*(80/4(31>> 9 &DPSV 03 *OHL]HV 6 7URXLOKHW  3URSHUWLHV $QDO\VLV RI D /HDUQLQJ $OJRULWKP IRU $GDSWLYH 6\VWHPV x  ,Q    K    2 4  3  (GLWLRQV &KDRV /L JH %HOJLXP  $YDLODEOH DW KWWSZZZLULWIU60$& - &DVWUR 0 .ROS  - 0\ORSRXORV x  $ 5HTXLUHPHQWVGULYHQ 'HYHORSPHQW 0HWKRGRORJ\ x  ,Q 0 

     L    2    4  3 1 

 2431, >6WDIIRUG8.x -XQH 

$'(/)($0HWKRGRORJ\IRU$GDSWLYH0XOWLDJHQW6\VWHPV(QJLQHHULQJ

>@ >@

>@ >@ >@

>@ >@ >@ >@ >@ >@

>@

- )HUEHU  /HV 6\W PHV 0XOWL$JHQWV 9HUV XQH ,QWHOOLJHQFH &ROOHFWLYH  ,QWHU(GLWLRQV 03*OHL]HV9&DPSV3*OL]H$7KHRU\RI(PHUJHQW&RPSXWDWLRQ%DVHG RQ &RRSHUDWLYH 6HOI2UJDQL]DWLRQ IRU $GDSWLYH $UWLILFLDO 6\VWHPV  #1   2   3  3   9DOHQFLD  $YDLODEOH DW KWWSZZZLULWIU60$& &0 ,JOHVLDV 0 *DULMR  - & *RQ]DOH]  $ 6XUYH\ RI $JHQW2ULHQWHG 0HWKRGRORJLHV  ,Q    4  M 464:N5O /1$,  6SULQJHU 9HUODJ -DFREVRQ * %RRFK  - 5XPEDXJK x  6  P  3  *    0  x $GGLVRQ:HVOH\ '.LQQ\0*HRUJHII $5DR$0HWKRGRORJ\DQG0RGHOLQJ7HFKQLTXH IRU 6\VWHPV RI %', $JHQWV  ,Q : 9DQ GH 9HOGH DQG -: 3HUUDP HGLWRUV 4   .  47 0 

     Q 1   ) .  J  44 J 4 ) /1$, SS 6SULQJHU9HUODJ - 2GHOO +9 3DUXQDN  % %DXHU  ([WHQGLQJ 80/ IRU $JHQWV  ,Q 3URFHHGLQJVRIWKH$JHQW2ULHQWHG,QIRUPDWLRQ6\VWHPV$2,6 :RUNVKRSDW WKHWK1DWLRQDO&RQIHUHQFHRQ$UWLILFLDO,QWHOOLJHQFH$$$,  35REHUWVRQ 5 /DGGDJD  + 6KUREH  ,QWURGXFWLRQ WKH )LUVW ,QWHUQDWLRQDO :RUNVKRSRQ6HOI$GDSWLYH6RIWZDUH,Q0 

  )343HGLWHG E\35REHUWVRQ5/DGGDJD +6KUREHLQ/1&6SS 65XVVHO 31RUYLJ$UWLILFLDO,QWHOOLJHQFHD0RGHUQ$SSURDFK0   <  & + 6SDUNPDQ 6 $ 'HORDFK $ / 6HOI x  $XWRPDWHG 'HULYDWLRQ RI &RPSOH[ $JHQW $UFKLWHFWXUHV IURP $QDO\VLV 6SHFLILFDWLRQV x  4(31>> 0RQWUHDO&DQDGD0D\WK 0 :RROGULGJH  3 &LDQFDULQL  $JHQW2ULHQWHG 6RIWZDUH (QJLQHHULQJ WKH 6WDWH RI WKH $UW  ,Q 3 &LDQFDULQL  0 :RROGULGJH HG 4 (    3 1 

 6SULQJHU9HUODJ/1$,-DQXDU\ 0 :RROGULGJH 1 5 -HQQLQJV  ' .LQQ\  $ 0HWKRGRORJ\ IRU $JHQW 2ULHQWHG $QDO\VLV DQG 'HVLJQ  ,Q 0 

     L    2  44 4 55SS6HDWWOH :$ 0D\  ) =DPERQHOOL 1 5 -HQQLQJV DQG 0 :RROGULGJH  2UJDQLVDWLRQDO DEVWUDFWLRQV IRU WKH DQDO\VLV DQG GHVLJQ RI PXOWLDJHQW V\VWHPV  ,Q 3 &LDQFDULQLDQG0:RROGULGJHHGV4(31, >>/1&66SULQJHU9HUODJ

Evaluating Multi-agent System Architectures: A Case Study Concerning Dynamic Resource Allocation Paul Davidsson and Stefan Johansson Department of Software Engineering and Computer Science, Blekinge Institute of Technology, Soft Center, 372 25 Ronneby, Sweden {pdv,sja}@bth.se

Abstract. Much effort has been spent on suggesting and implementing new architectures of Multi-Agent Systems. However, we believe the time has come to compare and evaluate these architectures in a more systematic way. Rather than just studying a particular application, we suggest that more general problem domains corresponding to sets of applications should be studied. Similarly, we argue that it is more useful to study the properties of classes of multi-agent system architectures than particular architectures. Also, it is important to evaluate the architectures in several dimensions, both different performance-related attributes, which are domain dependent and more general quality attributes, such as, robustness, modifiability, and scalability. As a case study we investigate the general problem of ”dynamic resource allocation” and present four classes of multi-agent system architectures that solve this problem. These classes are discriminated by their degree of distribution of control and degree of synchronization. Finally, we instantiate each of these architecture classes and evaluate, through simulation experiments, how they solve a concrete dynamic resource allocation problem, namely load balancing and overload control of Intelligent Networks.

1

Introduction

Much effort has been spent on suggesting and implementing new architectures of Multi-Agent Systems (mas). Unfortunately, this work has been carried out in a quite unstructured way where a (group of) researcher(s) invents a new architecture and applies it to a particular domain and conclude that it seems to be appropriate for this domain. Often this new architecture is not even compared to any existing architecture. We believe that this area now has reached the level of maturity when it is appropriate to compare and evaluate these architectures in a more systematic way. 1.1

Applications

Of course, there is no single mas architecture that is the most suitable for all applications. On the other hand, to find out whether one architecture performs P. Petta et al. (Eds.): ESAW 2002, LNAI 2577, pp. 170–183, 2003. c Springer-Verlag Berlin Heidelberg 2003 

Evaluating Multi-agent System Architectures

171

better than another for a particular application is usually of limited scientific interest. (Although this information may be very useful to solve that particular problem.) Instead, we suggest the study of more general problem domains corresponding to sets of applications with common characteristics. In this paper we will exemplify this approach by investigating the general problem of dynamic resource allocation. Such studies tend to be quite abstract and are for that reason often of a theoretical and qualitative nature. They therefore should be supplemented with and validated by quantitative empirical studies in one or more concrete applications corresponding to instances of the general domain, in this paper exemplified by load balancing and overload control in Intelligent Networks, a type of telecommunication system. 1.2

Architectures

Just as it is useful to study classes of applications rather than particular applications, we argue that it is useful to study classes of mas architectures in addition to particular architectures. To make such studies possible, we need to describe mas architectures in a way that abstracts the particularities of the individual architectures but still captures their relevant characteristics. To develop a general way of characterizing mas architectures is in itself a major research task. In fact, it may be necessary to use several views to capture all relevant aspects of an architecture cf. Kruchten [9]. In this work we will categorize mas architectures according to two properties: the type of control used (from fully centralized to fully distributed), the type of coordination (synchronous vs. asynchronous). As with classes of applications, it is mostly theoretical studies that can be performed on classes of mas architectures. Therefore, they often need to be supplemented with empirical studies using instantiations of these architectures. Below we will present four concrete architectures corresponding to different combinations of the two architectural properties. 1.3

Quality Attributes

It is possible to evaluate mas architectures with respect to several different quality attributes, both different performance-related attributes and more general quality attributes, such as, robustness, modifiability, and scalability. Some of these attributes are domain independent and some are specific for each set of applications, e.g., performance-related attributes. As mentioned earlier, we do not think it is possible to find a mas architecture that is optimal with respect to all relevant attributes. Rather, there is an inherent trade-off between these attributes and different architectures balance this trade-off in various ways. The various applications, on the other hand, require different balances of this tradeoff. Thus, in order to choose the right architecture for a particular application, knowledge about relevant attributes and how different mas architectures support them is essential.

172

1.4

Paul Davidsson and Stefan Johansson

Evaluation Framework

To evaluate a set of architectures in a systematic way, we suggest an approach that can be described in terms of the following three-dimensional space: – the set of possible applications, – the set of possible mas architectures, and – the set of attributes used to evaluate the architectures. The suggested approach is to investigate substantial parts of this space rather than just single points. We believe that this approach, besides of enabling a more systematic investigation of the space, will lead to a deeper understanding of mass and their applications, which, in turn, will contribute to reach the long-term goal of obtaining general design principles of mas. In this paper we will apply this approach to the general problem of dynamic resource allocation and present four abstract mas architectures with different characteristics that solves the problem. These will then be compared with respect to a number attributes, e.g., reactivity, ability to balance the loads, fairness, utilization of resources, responsiveness, amount of communication overhead, robustness, modifiability, and scalability. Finally, we evaluate concrete instantiations of these abstract architectures in a concrete dynamic resource allocation problem, namely load balancing and overload control in Intelligent Networks.

2

Abstract Domain: Dynamic Resource Allocation

Multi-agent technology has proved to be successful for dynamic resource allocation, e.g. power load management [11] and cellular phone bandwidth allocation [3]. Basically, this problem concerns the allocating of resources between a number of customers, given a number of providers. The dynamics of the problem lies in that the needs of the customers, as well as the amount of resources made available by the providers, vary over time. The needs and available resources not only vary on an individual level, but also the total needs and available resources within the system. We will here assume that the resources cannot be buffered, i.e., they have to be consumed immediately, and that the cost of communication (and transportation of resources) between any customer-provider pair is equal. 2.1

Abstract Multi-agent Architectures

There are many ways of dividing the set of possible mas architectures into different subsets based on their characteristics, e.g.: – – – –

the the the the

topography of the system, degree of mobility and dynamics of the communications, degree of distribution of control, and degree of synchronization of interaction.

Evaluating Multi-agent System Architectures

173

We have chosen to focus the two last properties. By degree of distribution we mean to what degree the control of the system is distributed. The degree of synchronization is a measure of how the execution of the agents interrelate with each other. We may have agents that are highly sophisticated, but who only interacts at special slots in time, and thus have a high degree of synchronization. There are also systems in which the agents may interact continuously, independently of when other agents interact, which we will refer to as asynchronous. To sum up, we will compare the following four abstract classes of mas architectures for dynamic resource allocation: centralized synchronous architectures, centralized asynchronous architectures, distributed synchronous architectures, and distributed asynchronous architectures. 2.2

Abstract Attributes

We have identified the following important performance-related attributes to dynamic resource allocation: – Reactivity: How fast are resources re-allocated when there are changes in demand? – Load balancing: How evenly is the load balanced between the resource providers? – Fairness: Are the customers treated equally? – Utilization of resources: Are the available resources utilized as much as is possible? – Responsiveness: How long does it take for the customers to get response to an individual request? – Communication overhead: How much extra communication is needed for the resource allocation? In addition, there are a number of more general software architecture quality attributes [6] that should be addressed, e.g.: – Robustness: How vulnerable is the system to node or link failures? – Modifiability: How easy is it to change the system after it is implemented (and often deployed)? – Scalability: How good is the system at handling large numbers of users (providers and customers)? 2.3

Theoretical/Qualitative Evaluation

We will now make a brief theoretical, or qualitative, analysis of how the degree distribution and synchronization of the multi-agent system architecture influence the attributes identified in the last section. – Reactivity should be promoted by asynchronous architectures since there is no need to await any synchronization event before i) an agent can notify other agents about changes in demand and ii) other agents can take the appropriate actions to adapt to these changes.

174

Paul Davidsson and Stefan Johansson

– Load balancing should be favored, or at least not disfavored, by centralized control since it is possible to take advantage of the global view of the state of the system, e.g., the current load at the providers and the current demand of the customers. – Similarly should fairness be easier to achieve for architectures with centralized control since they have information about the global state of the system. – It is not clear from a strictly theoretical analysis if there is any correlation between the ability to utilize the resources and the architectural properties. Empirical studies are probably necessary. – Also, it is not clear from a strictly theoretical analysis if there is any correlation between responsiveness and the architecture properties. – Communication overhead can be measured either by the number of messages sent, or by the bandwidth required for the allocation. Synchronous architectures tend to concentrate the message sending to short time intervals, and thus requiring a large bandwidth, whereas asynchronous architectures tend to be better at utilizing a given bandwidth over the time. Also, communication in distributed architectures has a tendency to be more local than in centralized architecture, using smaller parts of the network. – Regarding robustness we conclude that the more centralized the control is, the more vulnerable the system gets. Basically, the reason for this is that the system cannot function, i.e., perform reallocation, if the agents that are responsible for the control fail. In more distributed systems, the reallocation may function partially even though some agents have failed. – The modifiability, to add or remove a provider or customer, seems to be better in centralized architectures. For instance, changes may only be necessary in one part of the system. – Scalability seems to be better supported by distributed architectures than centralized architectures. Firstly, the computational load for the resource allocation is divided between a number of computers, and secondly, the risk for communication bottlenecks is smaller.

3

Concrete Domain: Load Balancing in Intelligent Networks

One important area in which the dynamic resource allocation problem is present is telecommunications. The Intelligent Network (in) concept was developed in order to enable operators of telecommunication networks to create and maintain new types of services [10]. Two important entities of an in are the Service Switching Points (ssps) and the Service Control Points (scps). The ssps continuously receive requests of services which they cannot serve without the help of the scps where all service software resides. Thus, the scps are providers and the ssps are customers. The ssps and scps communicate via a signaling network which we here will represent as a cloud rather than a specific topology of signaling links and nodes. (See Figure 1.) It is assumed that a small part of the available bandwidth of this network is reserved for the resource allocation, i.e.,

Evaluating Multi-agent System Architectures

175

SSP1 SCP1

SSP2

SCP2

......

SS7 Signalling Network "cloud"

......

SCPm

SSPn

Fig. 1. A simplified view of an Intelligent Network (in) with m scps and n ssps (typically n > m)

the communication overhead caused by agent communication (and transportation). It is assumed that all scps support the same set of service classes and that all service requests can be directed by a ssp to any scp. 3.1

Concrete Multi-agent System Architectures

We have chosen one architecture of each of the abstract classes mentioned earlier (see table 1). Common for these architectures are the use of three different types of agents: quantifiers, allocators, and distributors. A quantifier acts on behalf of a provider of the resources, an allocator acts on behalf of a customer, and a distributor decides the allocation of some (or even all) available resources. Although these three types of agents have similar roles in all the four multiagent system architectures, the actual implementation may be rather different (in particular this hold for the distributors). The reason, of course, is that different system architectures may put different demands on the agents.

Table 1. The four different multi-agent system architectures classified in terms of distributedness and synchronicity centralized distributed synchronous Centralized auctions (ca) Hierarchical auctions (ha) asynchronous Centralized leaky bucket (clb) Mobile brokers (mb)

176

Paul Davidsson and Stefan Johansson

The Centralized Auction Architecture The Centralized Auction (ca) architecture is an example of a synchronous, centralized architecture. Arvidsson et al. [1] suggested an approach where the resource allocation is carried out by means of tokens (cf. market-based control [5]). Each token represents a service request and is consumed when the request is accepted by a provider. The three types of agents have the following functionality: – The quantifiers try to sell the amount of tokens that corresponds to the load that the provider is able serve between two auctions. – The allocators try to buy the amount of token corresponding to the resources it predicts their customer will receive during the time to the next auction. – The distributor receives bids from the quantifiers corresponding the available capacity at their provider (and the prices), and bids from the allocators containing the expected need for resources. The distributor then carries out the auction so that the common good is maximized and sends messages about the result to the involved agents. An allocator maintains a pool of tokens for each provider and type of resource. Each time the allocator feeds a provider with a request for a particular type of resource, one token is removed from the associated pool. If all pools associated with a particular resource type are empty, the customer cannot accept more requests. The pools are refilled at the auctions that take place at fixed time intervals. In order to avoid spending all tokens immediately during high loads (which would lead to excessive delays caused by long queues at the providers), percentage thinning is used so that the probability of buying a certain type of resource is never higher than the number of remaining tokens over the number of expected needs during the reminder of the interval. For more details we refer to Arvidsson et al. [1]. The Hierarchical Auction-Based Architecture One possible implementation of a distributed, synchronous system is the hierarchical auction (ha) architecture [11]. The idea is to partition the set of allocators and to use one distributor for aggregating bids and holding auctions for each partition. These distributors then connect to higher order distributors in a hierarchical manner until the total demand can be matched against the amount of available resources offered by the quantifiers. The Centralized Leaky Bucket Architecture The centralized asynchronous architecture we have chosen is based upon an asynchronous approach called Leaky bucket [2]. The basic idea is that each provider is equipped with a Leaky bucket that feeds requests to the provider at an even and optimal rate. This is done by inserting the incoming requests from the customers in a queue in the Leaky bucket. These requests are then dequeued at a rate corresponding to the maximum capacity of that provider. If the queue is full, the requests are rejected. To get a centralized architecture, we introduce a centralized leaky bucket ( clb) architecture, in which there is just one central distributor, common for

Evaluating Multi-agent System Architectures

177

all allocators and quantifiers. The allocators send all requests immediately to this distributor, which consists of a common leaky bucket for queuing the requests. It also has a router that continuously dequeues requests at a rate corresponding to the total capacity of the providers and then forwards the requests evenly to the providers in proportion to their capacity. If the bucket is full, the request is returned to the allocator where it is rejected. The Mobile Broker Architecture As an example of a distributed, asynchronous system, we choose a mobile broker (mb) architecture [4]. In this architecture, the distributors are implemented as mobile brokers (one for each provider) that sequentially visit each (or a subset) of the allocators offering the resources currently available at the corresponding provider The allocator then requests the resources it needs for the moment (or rather, predicts it will need in the near future). If possible, the broker gives this amount of resources to the allocator. Otherwise, it gives as much as is currently available at the provider. However, there are two problems with this naive approach: – If an allocator demands all the available resources, the broker will give them to that allocator. Thus, the broker will not be able to hand out any more resources for a while, which would not be fair. – If the overall load is low or moderate, the allocators are given just as much resources as they demand. However, if an allocator need slightly more resources than it asked for (predicted), it will have to turn down request, even though the provider has lots of surplus capacity. In order to solve these problems, we use a broker mechanism that strive to give out all the available resources and give each allocator resources in proportion to their part of the total current demand (of the allocators in the route). For the details of this approach we refer to Johansson et al. [7]. It should be noted that in case of a sudden increase in demand, the resources given out may momentarily exceed the available resources, which in the worst case will lead to a transient overload situation. However, the mechanism is self-stabilizing, and will find an equilibrium within one route (given that the demands are relatively stable). The mechanism ensures that the allocators are given resources that (relatively) correspond to their share of the total demand handled by the broker, thus solving both problems in the naive solution above. If an allocator are visited by several brokers it may happen that some of the brokers’ scp are carrying a higher load than the others. To deal with this problem an additional balancing function is used, making the allocators try to move load from those scps with relatively high load to those with relatively low load. The allocator calculates the load of a broker from the quotient between what it asked for and what is was given by the broker. 3.2

Concrete Attributes

We now operationalize the abstract attributes presented earlier. Thus, in the domain of in load management the attributes are defined as follows:

178

Paul Davidsson and Stefan Johansson

– Reactivity is measured by how fast the mas is able to re-allocate the available scp processing time when there are sudden changes of offered loads by the ssps. – Load balancing is measured by the standard deviation between the carried load of the scps. – Fairness is measured by the standard deviation of rejected calls divided by the accepted calls between the ssps, i.e., the rejection rate. – The utilization of resources is measured by how close the carried load is to the target load, or offered load, if the offered load is less than the target load. scp load levels should be as close to the target load (e.g., 0.9 Erlangs, corresponding to 90% of its capacity) as possible but not exceed it. If an overload situation is approaching, the ssps should throttle new requests. – Responsiveness is measured by the time it takes for the ssps to get response from an scp. – Communication overhead is measured by the bandwidth necessary for the mas to perform the reallocation. – Robustness is measured in terms of the consequences of a distributor agent failure. – Modifiability is measured in terms of how easy it is to add a new or remove an existing ssp or scp. – Scalability is measured in terms of how the number of ssps and scps influence the performance-related attributes. 3.3

Experimental Evaluation

The four concrete architectures have been evaluated in simulation experiments. For these experiments we used the same simulation model as Arvidsson et al. [1]. The Intelligent Network modeled has 8 scps and 32 ssps, which are assumed to be statistically identical respectively. The processors at the scps are engineered to operate at 90 percent of the maximum capacity (target load is 0.9 Erlang). All messages passing through the network experience a constant delay of 5 ms, and it is assumed that no messages are lost. Detailed descriptions the simulation results can be found in [8]. Since we use these experiments as a case study, we only summarize the results here: – As the theoretical evaluation predicted, reactivity was promoted by the two asynchronous architectures. For instance, we simulated a scenario in which half of the 32 ssps experience an offered load corresponding to 0.2 Erlang, and the other half a load corresponding to 1.4 Erlang. The whole system thus is offered a total average load of 0.8 Erlang, which is below the target load, and therefore possible for the system to carry. At time 400, the levels of loads are shifted from high to low and vice versa and 10 seconds later they are swapped back again, see Fig. 2. As can be seen in Fig. 3, the clb manage this difficult situation perfectly and the other asynchronous architecture mb does almost as well. The synchronous auction-based architectures, on the other hand, have apparent difficulties to adapt to the changes.

Evaluating Multi-agent System Architectures

179

– All the architectures were able to balance the load well. There were just small differences in the standard deviations (of the measured carried load of the scps), varying between 1 and 3 mErlang depending on the offered load. When the offered load was below target load the centralized architectures performed slightly better, whereas there were no measurable differences in overload situations. – With respect to fairness, all architectures performed well when the offered load was below target load. In overload situations, mb was significantly less fair than the other architectures. However, it may be the case that an improvement of the design of the broker routes may reduce these differences. – It was not clear from theoretical analysis if there is any correlation between the ability to utilize the resources and the architectural properties. The simulation results presented to the left in Fig. 4 indicates that the centralized asynchronous architecture performs best in this respect. – Also, it was not clear the theoretical analysis if there is any correlation between responsiveness and the architecture properties. But in this case the simulation results indicate that the centralized asynchronous architecture has the worst performance. See Fig. 5. Note that responsiveness is measured in terms of the average delay for the individual requests, whereas reactivity is a measure of the delay of reallocations which typically concerns a large number of requests. – We tuned the parameters in the simulation experiments so that both ca, ha and mb needed approximately the same bandwidth for communication overhead, about 40 messages/second irrespective of the amount of offered load. The bandwidth needed by clb, however, is proportional to the number of requests and is considerably higher than for the other architectures. For instance, when the offered load is 0.70 Erlang, 2150 messages/second are sent, and when the offered load is 2.0 Erlang, 9365, i.e., more than 200 times as much bandwidth as the other architectures need. For the general software architecture attributes (robustness, modifiability, and scalability), we refer to the theoretical analysis in section 2.3.

4

Conclusions and Future Work

We have described a systematic way of evaluating different aspects of different mas architectures. This approach was applied to an abstract domain, namely dynamic resource allocation, and a theoretical evaluation of abstract architectures was made. This was supplemented by an experimental study of an implementation of this abstract domain, load balancing and overload control in Intelligent Networks. The experimental evaluation confirmed the conclusions of the theoretical analysis, e.g., that asynchronous architectures are able to react faster than synchronous. In addition, it gave insights concerning attributes for which no clear conclusions could be achieved from the theoretical analysis, e.g., that centralized asynchronous architectures are able to utilize the available resources better than

180

Paul Davidsson and Stefan Johansson

2 SSPs 1-16 SSPs 17-32 1.8 1.6

Offered load in Erlangs

1.4 1.2 1 0.8 0.6 0.4 0.2 0 400

450 Time in seconds

Fig. 2. Reactivity: The expected offered load

1

Average carried load in Erlangs

0.8

0.6

0.4

0.2 Centralized auction Distributed auction Mobile brokers Centralized leaky bucket 0 400 Time in seconds

Fig. 3. Reactivity: The average carried load

Evaluating Multi-agent System Architectures

181

1 0.95 0.9 0.85

Carried load in Erlangs

0.8 0.75 0.7 0.65 0.6 0.55 0.5 Centralized auction Hierarcical auction Mobile broker Centralized leaky bucket Min(Target load, offered load)

0.45 0.4 0.35 0.4

0.6

0.8

1

1.2

1.4

1.6

1.8

2

Offered load in Erlangs

Fig. 4. Utilization of resources: The ability to carry different levels of offered load 0.2 Centralized auction Hierarcical auction Mobile broker Centralized leaky bucket

0.18

Average time to connect (s)

0.16 0.14 0.12 0.1 0.08 0.06 0.04 0.02 0 0.4

0.6

0.8

1

1.2

1.4

1.6

1.8

2

Offered load in Erlangs

Fig. 5. Responsiveness: The average response time for different levels of offered load

182

Paul Davidsson and Stefan Johansson

the other architectures, but have larger delays and need more bandwidth when the load is high. The results of the case study were, not very surprisingly, that different architectures excel in different dimensions. The choice of mas architecture for a particular application should be guided by the balance of the trade-off between these dimensions that is optimal for that application. We believe that if the systematic approach suggested here is widely adopted, such choices can be more informed than is currently practice. Our plans for future work includes: – Further experimental validation in the in domain of the theoretical results regarding, e.g., scalability. – Experimental validation in another dynamic resource allocation domain. – Use the outlined approach to investigate other domains than dynamic resource allocation. – Investigating to what extent the implementations of the individual agents influence system performance. – Develop a more refined method for characterizing mas architectures.

Acknowledgments The authors would like to thank Martin Kristell for assistance during the simulation experiments and the Swedish Knowledge Foundation for funding parts of this work.

References ˚. Arvidsson, B. Jennings, L. Angelin, and M. Svensson. On the use of agent [1] A technology for in load control. In Proceedings of the 16th International Teletraffic Congress. Elsevier Science, 1999. 176, 178 [2] A. W. Berger. Comparison of call gapping and percent blocking for overload control in distributed switching systems and telecommunication networks. IEEE Trans. Commun., 39:574–580, 1991. 176 [3] E. Bodanese and L. Cuthbert. An intelligent channel allocation scheme for mobile networks: An application of agent technology. In Proceedings of the 2nd International Conference on Intelligent Agent Technology, pages 322–333. World Scientific Press, 2001. 172 [4] B. Carlsson, P. Davidsson, S. J. Johansson, and M. Ohlin. Using mobile agents for in load control. In Proceedings of Intelligent Networks ’2000. IEEE, 2000. 177 [5] S. H. Clearwater, editor. Market-Based Control: Some early lessons. World Scientific, 1996. 176 [6] P. Clements, R. Kazman, and M. Klein. Evaluating Software Architectures. Addison Wesley, 2002. 173 [7] S. J. Johansson, P. Davidsson, and B. Carlsson. Coordination models for dynamic resource allocation. In A. Porto and G.-C. Roman, editors, Coordination Languages and Models, volume 1906 of Lecture notes in computer science, pages 182–197. Springer Verlag, 2000. Proceedings of the 4th International Conference on Coordination. 177

Evaluating Multi-agent System Architectures

183

[8] S. J. Johansson, P. Davidsson, and M. Kristell. Four architectures for dynamic resource allocation. In Proceedings of Fourth International Workshop on Mobile Agents for Telecommunication Applications, 2002. Springer Verlag, LNCS. 178 [9] P. B. Kruchten. The 4+1 view model of architecture. IEEE Software, pages 42–50, July 1995. 171 [10] T. Magedanz and R. Popescu-Zeletin. Intelligent Networks. International Thomson Computer Press, 1996. 174 [11] F. Ygge. Market-Oriented Programming and its Application to Power Load Management. PhD thesis, Lund University, Sweden, 1998. 172, 176

Engineering Agent Systems for Decision Support* 6DVFKD2VVRZVNL-RVHID=+HUQ QGH] &DUORV$,JOHVLDVDQG$OEHUWR)HUQ QGH] 

$,*URXS8QLYHUVLW\5H\-XDQ&DUORV &DPSXVGH0 VWROHVVQ(0DGULG {s.ossowski,al.fernandez}@escet.urjc.es  $,'SW7HFK8QLYHUVLW\RI0DGULG &DPSXVGH0RQWHJDQFHGRVQ(0DGULG [email protected]  7HFKQLFDO,QQRYDWLRQ'SW *HUPLQXV;;,*UDQ9 Dœ ,]T(0DGULG [email protected]

Abstract.7KLVSDSHUGLVFXVVHVKRZDJHQWWHFKQRORJ\FDQEHDSSOLHGWR WKH GHVLJQ RI DGYDQFHG ,QIRUPDWLRQ 6\VWHPV IRU 'HFLVLRQ 6XSSRUW ,Q SDUWLFXODULWGHVFULEHVWKHGLIIHUHQWVWHSVDQGPRGHOVWKDWDUHQHFHVVDU\ WR HQJLQHHU 'HFLVLRQ 6XSSRUW 6\VWHPV EDVHG RQ D PXOWLDJHQW DUFKLWHF WXUH7KHDSSURDFKLVLOOXVWUDWHGE\DFDVHVWXG\LQWKHWUDIILFPDQDJH PHQWGRPDLQ

1

Introduction

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

 :RUN VXSSRUWHG E\ WKH 6SDQLVK 0LQLVWU\ RI 6FLHQFH DQG 7HFKQRORJ\ 0&\7  XQGHU JUDQW 7,&&

33HWWDHWDO(GV (6$:/1$,SS ©6SULQJHU9HUODJ%HUOLQ+HLGHOEHUJ

(QJLQHHULQJ$JHQW6\VWHPVIRU'HFLVLRQ6XSSRUW

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

2

Engineering IDEAs

&XUUHQW UHVHDUFK LQ $JHQW 2ULHQWHG 6RIWZDUH (QJLQHHULQJ H[DPLQHV SULQFLSOHG ZD\V IRU FRQVWUXFWLQJ FRPSOH[ VRIWZDUH V\VWHPV EDVHG RQ WKH DJHQW PHWDSKRU >@>@>@2IWHQWUDGLWLRQDO -        VXFKDV80/>@DUH H[WHQGHGWRDFFRXQWIRUWKHFKDUDFWHULVWLFV RI DJHQW V\VWHPV SD\LQJ SDUWLFXODU DWWHQ WLRQWRWKHPRGHOOLQJRIDJHQWLQWHUDFWLRQVDQGDJHQWDUFKLWHFWXUH6XFKPHWKRGRORJLHV DOVR DSSHDU WR EH DQ DGHTXDWH FKRLFH IRU WKH FRQVWUXFWLRQ RI ,'($V EXW QHHG WR EH FRPSOHPHQWHG E\ WHFKQLTXHV IURP WKH .    

  ILHOG DV GHFLVLRQ VXSSRUWDJHQWVGUDZKHDYLO\XSRQFRPSOH[NQRZOHGJHEDVHGUHDVRQLQJSURFHVVHV7R

 6DVFKD2VVRZVNLHWDO

WKLV UHVSHFW WKH ZHOONQRZQ 2;4*3 DSSURDFK >@ WKDW VXSSRUWV DGYDQFHG NQRZOHGJHPRGHOOLQJLVSDUWLFXODUO\UHOHYDQW ,QWKHVHTXHOZHGHVFULEHKRZWRDSSO\DJHQWRULHQWHG VRIWZDUH HQJLQHHULQJ WHFK QLTXHVIURPERWKILHOGVLQRUGHUWRHQJLQHHULQWHOOLJHQWDVVLVWDQWVIRUGHFLVLRQVXSSRUW )LUVWZHXVHWKH8(5PRGHOOLQJWHFKQLTXHDQH[WHQVLRQRI80/IRUDQDO\VLV7KHQ D2;4*3OLNHDSSURDFKLVXVHGIRUWKHLGHQWLILFDWLRQRIEDVLFUHDVRQLQJWDVNV RI,'($V)LQDOO\ZHVKRZKRZWRREWDLQDQGVWUXFWXUHUHDVRQLQJPHWKRGVEDVHGRQ VXFKDNQRZOHGJHRULHQWHGDSSURDFK>@ 2.1

Analysis

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

Fig. 1.8(50RGHORI,'($6

(QJLQHHULQJ$JHQW6\VWHPVIRU'HFLVLRQ6XSSRUW

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• •

2 7KHXVHUUHTXHVWVVRPHH[SODQDWLRQIURPWKHV\VWHP 1A    DQRWKHU '66 LQWHUFKDQJHV SODQV ZLWK WKH V\VWHPV WR JHW D EURDGHUXQGHUVWDQGLQJRIWKHVLWXDWLRQ 7KHXVHFDVHFRQYHUVDWLRQFDQEHGHWDLOHGDV

•

)WKHGHFLVLRQPDNHU UHTXHVWVDQH[SODQDWLRQRILQIRUPDWLRQSURYLGHGE\WKH '66 )WKHGHFLVLRQPDNHU UHTXHVWVUHOHYDQWGDWDWRWDNHDGHFLVLRQ ) WKH GHFLVLRQPDNHU DVNV WR HYDOXDWH WKH UHVXOWV RI D GHFLVLRQ DQG LWV FRQVHTXHQFHV 3

WKHGHFLVLRQPDNHU UHTXHVWVSRVVLEOHDFWLRQVWREHWDNHQ

• • •

'HSHQGLQJRQWKHSDUWLFXODUGRPDLQDERXWZKLFKWKH,'($LVNQRZOHGJHDEOHWKH

       LGHQWLILHV UHOHYDQW HQYLURQPHQW REMHFWV DQG HYHU\ SRVVLEOHHYHQWJHQHUDWHGIURPWKHVHREMHFWVDQGSRVVLEOHDFWLRQVFDUULHGRXWRQWKHP )RUH[DPSOHLQDILQDQFLDO'66ZHFRXOGLGHQWLI\DQREMHFW 7KHSRVVLEOHLQSXW HYHQWVDUHQHZYDOXHVRIWKLVVKDUHDQGWKHSRVVLEOHDFWLRQVWREX\RUWRVHOO )LQDOO\LQWKH     ZHQHHGWRLGHQWLI\WKHDXWRQRPRXVEHKDYLRXUV RIDQ,'H$,WVPRVWLPSRUWDQWUHVSRQVLELOLWLHVXVXDOO\DUH • • •

2.2

J  REVHUYHWKHHQYLURQPHQWDQGGHWHFWSUREOHPDWLFEHKDYLRXUV 4    UDLVHDODUPVLIWKHUHLVDFULWLFDOVLWXDWLRQ )   ZDUQLQJ UHVSHFWLQJ XQGHVLUHG FRQVHTXHQFHV RI u EDGv  DFWLRQV DQG SRWHQWLDOO\VXJJHVWLQJEHWWHURQHV Task Design

,Q PHWKRGRORJLHV WKDW JR EDFN WR WKH NQRZOHGJH HQJLQHHULQJ ILHOG D WDVN LV XVXDOO\ FRQFHLYHGDVDQDEVWUDFWGHVFULSWLRQRIKRZWKHZRUOGRUDQDJHQWtVu PHQWDOPRGHOv RILW QHHGVWREHWUDQVIRUPHGLQRUGHUWRDFKLHYHDGHVLUHGEHKDYLRXURUIXQFWLRQDOLW\ ,QWKLVVHFWLRQZHLGHQWLI\WKHGLIIHUHQWWDVNVWKDWDQ,'($QHHGVWRFRSHZLWKVRDVWR EHDEOHWREHKDYHVXFFHVVIXOO\LQWKHDIRUHPHQWLRQHGFRQYHUVDWLRQFDVHV 7R JHQHUDWH DQVZHUV IRU WKH GLIIHUHQW FODVVHV RI TXHVWLRQV LQ RXU FRQYHUVDWLRQ IUDPHZRUNZHKDYHLGHQWLILHGHVVHQWLDOO\IRXUWDVNV

 6DVFKD2VVRZVNLHWDO

• • • •

0   %)URPWKHDQDO\VLVRIWKHLQIRUPDWLRQUHFHLYHGIURPD FRPPXQLFDWLRQLQIUDVWUXFWXUHRUGLUHFWO\IURPWKHRSHUDWRUDQ,'($FODVVLILHV WKHVWDWHRIWKHPRQLWRUHGV\VWHP * 7KHSUHVHQFHRIXQDFFHSWDEOHHYHQWVRUVLWXDWLRQVUHTXLUHVDQH[ SODQDWLRQLQWHUPVRIFDXVDOIHDWXUHVRIWKHVLWXDWLRQ 4  2QFHDSUREOHPKDVEHHQLGHQWLILHGDSRVVLEOHVHTXHQFHRI DFWLRQVDSSOLFDEOHRQWKHFDXVHVPD\EHHVWDEOLVKHG 0 WKHFRQVHTXHQFHVRIHYHQWVDQGRSHUDWRUDFWLRQVDUHVLPXODWHG

6XSSRVHVRPHV\VWHPFRPSRQHQWV3VRPHH[WHUQDOHYHQWV1DQGRSHUDWRUDFWLRQV 4%\FRPELQLQJWKHDERYHWDVNVLQGLIIHUHQWPDQQHUVVHYHUDOTXHVWLRQVWKDWDGHFL VLRQPDNHUW\SLFDOO\IDFHVFDQEHDQVZHUHG)RULQVWDQFH • • • •

u :KDWLVKDSSHQLQJLQ6"v SUREOHPLGHQWLILFDWLRQGLDJQRVLV $GLDJQRVLV*IRUVRPHSRWHQWLDOPDOIXQFWLRQLVSURGXFHG u :KDWWRGRRQ'LQ6"v DFWLRQSODQQLQJSUHGLFWLRQ 'HFLVLRQRSWLRQVDUHVKDSHGDQGWKHLUSRWHQWLDOHIIHFWVHYDOXDWHG u :KDWPD\KDSSHQLI1LQ3"v       3RWHQWLDOIXWXUHSUREOHPVLQHYROXWLRQRIWKHV\VWHPDUHLGHQWLILHG u :KDWWRGRLI(LQ6"v SUHGLFWLRQSUREOHPLGHQWGLDJQRVLVSODQQLQJ

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

(QJLQHHULQJ$JHQW6\VWHPVIRU'HFLVLRQ6XSSRUW

2.3

Method Design

0RVWNQRZOHGJHRULHQWHGPHWKRGRORJLHVPDNHXVHRIWKHFRQFHSWRISUREOHPVROYLQJ PHWKRGVLQRUGHUWRFRSHZLWKWDVNV,QSDUWLFXODUVXFKPHWKRGVLQGLFDWHDWDVNLV DFKLHYHGE\GHVFULELQJWKHGLIIHUHQWVWHSVE\ZKLFKLWVLQSXWVDUHWUDQVIRUPHGLQWRLWV RXWSXWV 7KH SUREOHPVROYLQJ SURFHVV DVVRFLDWHG WR D WDVN LV VWUXFWXUHG HDFK RI LWV VWHSVPD\VHWXSVHYHUDOVXEWDVNVZKLFKDJDLQDUHWR EH VROYHG E\ VLPSOHU PHWKRGV HWFXQWLO VRPH HOHPHQWDU\ WDVNV FDQ EH DFKLHYHG GLUHFWO\ ,Q WKH VHTXHO ZH LGHQWLI\ GLIIHUHQWVXFKu UHDVRQLQJVNHOHWRQVv WKDWRXU,'($VZLOOQHHGWRDSSO\VRDVWRFRSH HIIHFWLYHO\ZLWKWKHWDVNVLGHQWLILHGLQWKHSUHYLRXVVHFWLRQ Problem Identification Methods $FODVVLILFDWLRQPHWKRGZLWKWZRRSWLRQVPD\EHDSSOLHG • •

,GHQWLILFDWLRQRIDUHIHUHQFHVLWXDWLRQDQGFODVVLILFDWLRQRIWKHGLIIHUHQFHVEH WZHHQWKHUHIHUHQFHDQGWKHFXUUHQWVLWXDWLRQ 'LUHFWFODVVLILFDWLRQRIWKHFXUUHQWVLWXDWLRQEDVHGRQDSUHGHILQHGWD[RQRP\ ZKHUHSUREOHPVRIGLIIHUHQWW\SHVDUHGHVFULEHG

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

 R  PRGHO>@7KHVHFRQGDSSURDFKLV VLPLODU WR WKH ILUVW RQH EXW LQ WKLVFDVHDFRPSOHWHGHVFULSWLRQRIWKHVLWXDWLRQLVUHTXLUHGQRWRQO\WKHGLIIHUHQFHV ZLWKWKHUHIHUHQFHVLWXDWLRQ Diagnosis Methods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

 6DVFKD2VVRZVNLHWDO

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ction Planning Methods $IWHU WKH SUREOHP LGHQWLILFDWLRQ DQG GLDJQRVLV WDVNV VRPH VFHQDULRV RI FDXVHV RI SUREOHPVKDYHEHHQGHGXFHGWRJHWKHUZLWKLWVLPSDFWV7KH  WDVNPXVW JHQHUDWH D FRQVLVWHQW VHW RI DFWLRQV RULHQWHG WRZDUG WKH UHGXFWLRQ RU HOLPLQDWLRQ RI FDXVHV DQGRU WRZDUG WKH UHGXFWLRQ RI LPSDFW GDPDJHV ZKHUH QR SRVVLEOH FDXVH UHGXFWLRQPD\EHSURGXFHG 6SHFLI\LQJWKLVWDVNLQDJHQHUDOZD\UHTXLUHVGHILQLQJWKHHOHPHQWDU\DFWLRQVWKDW ZLOOEHWKHEDVLVIRUGHILQLWLRQRIDFFHSWDEOHGHFLVLRQSODQVWRJHWKHUZLWKWKHLUPRGHOV 7KH DFWLRQ SODQQLQJ WDVN PD\ EH SHUIRUPHG E\ D PHWKRG LQWHJUDWLQJ FRQVLVWHQW VH TXHQFHVRIDFWLRQVWRWUDQVLWWRDVLWXDWLRQZKHUHPRVWGDPDJHVKDYHEHHQDOOHYLDWHG DQGDOOWKHSUREOHPFDXVHVKDYHEHHQFDQFHOOHG 7KHJHQHUDOUHDVRQLQJPHWKRGWRGHDOZLWKWKLVIXQFWLRQDOLW\LVDSODQQHU+RZHYHU LWLVDVVXPHGWKDWWKHDUHDRIH[SHUWLVHWREHPRGHOOHGHPERGLHVVWUXFWXUHGNQRZOHGJH FULWHULD WKDW DUH VXIILFLHQWO\ SUHFLVH WR JHQHUDWH D SODQ LQ D PRUH VLPSOH ZD\ XVLQJ FODVVLILFDWLRQUHDVRQLQJRQSUHGHILQHGSODQVDQGVXESODQV$FFRUGLQJWRWKLVDVVXPS WLRQLWLVUHDVRQDEOH WR DSSO\ D VWHSZLVH UHDVRQLQJ PHWKRG GHULYHG IURP WKH URXWLQH GHVLJQPHWKRGSURSRVHGE\>@ 7KHGHFODUDWLYHNQRZOHGJHPD\EHRUJDQLVHGZLWKFROOHFWLRQVRISODQVFDSDEOHWR SHUIRUPVXEWDVNVWKDWPD\EHLQWHJUDWHGWREXLOGSODQVDLPLQJWRSHUIRUPPRUHFRP SOH[WDVNV$SODQPD\EHGHILQHGE\VHTXHQFHVRIHOHPHQWDU\DFWLRQVRUPD\LQFOXGH RWKHUVXEWDVNVVXSSRUWHGE\RWKHUFROOHFWLRQVRISODQV7KHPHWKRGXVLQJWKLVGRPDLQ NQRZOHGJHLVDW\SHRI.      VXSSRUWLQJDSURJUHVVLYHUHILQHPHQWRI WKHVXEWDVNVIURPLQWHUPHGLDWHSDUWLDOYHUVLRQVRIWKHSODQLQWKHIROORZLQJVWHSV 

$EDVLFPHWKRGJLVDSSOLHGZKLFKLQWURGXFHVDFROOHFWLRQRIFRQVLVWHQWDFWLRQV WRJHWKHU ZLWK D GHFRPSRVLWLRQ RI WKH JHQHUDO WDVN LQ SODQV RI DFWLRQV  LQWH

(QJLQHHULQJ$JHQW6\VWHPVIRU'HFLVLRQ6XSSRUW

JUDWHGZLWKVRPHVXEWDVNVQRW\HWIRUPXODWHGDVSODQV <   - %%%6 6  %%%  > 



ZKHUH6 6DUHVXEWDVNVZLWKQRSODQRIHOHPHQWDU\DFWLRQVVSHFLILHG )RUHYHU\VXEWDVNQRW\HWGHVFULEHGLQWHUPVRIEDVLFDFWLRQV-%%%VXFKDV 6 RU6DQLQIHUHQFHVWHSRIWKHVDPHW\SHLVDSSOLHGE\XVLQJDQLQVWDQWLDWLRQRI JZLWKRWKHUSUHPLVHVDQGGRPDLQNQRZOHGJH$WUHHH[SORULQJGLIIHUHQWVWHSV RISODQGHWDLOLVSURGXFHGXQWLOGHWDLOHGSODQVDUHREWDLQHGRUDQDSSOLFDWLRQRI JIDLOVDQGDEDFNWUDFNLQJLVSURGXFHGWRDOWHUQDWLYHWDVNGHFRPSRVLWLRQV     .  0

%%%

%%%

0.

%%%

0          .

0

0.     . 4%%%@ 0HWKRGV WKDW SHUIRUP WKLV W\SH RI PDQDJHPHQW XVXDOO\ FRPSULVH WKUHHVWHSV 





3

*     XVLQJGRPDLQNQRZOHGJHDERXWWKHGLIIHUHQWGHSHQGHQ FLHVWKDWPD\RFFXUSURGXFHUFRQVXPHUUHODWLRQVKLSVUHVRXUFHOLPLWDWLRQVHWF >@ SRVLWLYHDQGQHJDWLYHUHODWLRQVKLSVEHWZHHQWKHGLIIHUHQWORFDOWDVNVRIWKH DJHQWVDUHGHWHFWHG)RULQVWDQFHWZRORFDOWUDIILFDJHQWVPD\ZDQWWRGLVSOD\ GLIIHUHQWZDUQLQJPHVVDJHVRQWKHVDPHWUDIILFPHVVDJHSDQHO (   IRU HYHU\ GHSHQGHQF\ WKH VHW RI SRVVLEOH PDQDJHPHQW DFWLRQV LV JHQHUDWHG ,Q RXU WUDIILF H[DPSOH DQ\ RI WKH LQYROYHG DJHQWV PD\ FKDQJHLWVORFDODFWLRQSODQRUZHPD\PHUJHWKHLQFRPSDWLEOHPHVVDJHVu LQFL GHQW DW 4v  DQG u FRQJHVWLRQ DW v  WR EH GLVSOD\HG RQ WKH VDPH SDQHO LQWR WKH PHVVDJHu WUDIILFSUREOHPVDW4DQGv  J   ILQDOO\DGHFLVLRQPXVWEHWDNHQUHVSHFWLQJWKHGHSHQG HQF\PDQDJHPHQWDFWLRQWREHDSSOLHG,QWKHWUDIILFH[DPSOHRQHSRVVLEOHFUL WHULRQ IRU WDNLQJ WKLV GHFLVLRQ LV WKH DLP RI GLVWULEXWLQJ WUDIILF ORDG HTXDOO\ DPRQJWKHGLIIHUHQWSUREOHPDUHDV

An Example: IDEAs for Traffic Control

:H QRZ LOOXVWUDWH WKH DERYH IUDPHZRUN IRU HQJLQHHULQJ ,QWHOOLJHQW 'HFLVLRQPDNLQJ $VVLVWDQWVE\DQH[DPSOH:HZLOOGHVFULEHWKHFRQVWUXFWLRQRI,'($6IRUURDGWUDIILF PDQDJHPHQWDUHDOZRUOGGRPDLQDQGDQH[DPSOHRIWKHKLJKGHJUHHRIFRPSOH[LW\ WKDWGHFLVLRQVXSSRUWDSSOLFDWLRQVKDYHWRGHDOZLWK)RUWKLVSXUSRVHZHVHWRQIURP WKH75@ DJHQWEDVHG'66WKDWKDYHEHHQEXLOGDQG XVHGH[SHULPHQWDOO\LQGLIIHUHQW6SDQLVKWRZQV,QWKLV VHFWLRQ ZH SURYLGH D SULQFL SOHGu UHGHVLJQv RI WKH DUFKLWHFWXUH RI WKHVH V\VWHPV DORQJ WKH OLQHV SURSRVHG LQ WKH SUHYLRXVVHFWLRQ)LUVWRXUSDUWLFXODUWUDIILFPDQDJHPHQWGRPDLQLVVNHWFKHGDQGWKH UHTXLUHPHQWVIRUWUDIILFPDQDJHPHQW,'($VDUHDQDO\VHGLQWHUPVRIDQ8(5PRGHO 7KHQWKHGHVLJQRIVXFK,'($VIURPDNQRZOHGJHRULHQWHGSRLQWRIYLHZLVGHVFULEHG VRDVWRILQDOO\VNHWFKVRPHLPSOHPHQWDWLRQLVVXHV 3.1

The Traffic Problem: Analysis

,Q%DUFHORQDWKHORFDOWUDIILFFRQWUROFHQWUH-37LVLQFKDUJHRIPDQDJLQJXUEDQURDG WUDQVSRUW VR DV WR PDLQWDLQ DQG UHVWRUH WKH u VPRRWKv  IORZ RI YHKLFOHV 7UDIILF HQJL QHHUV FRQWLQXRXVO\ UHFHLYH LQIRUPDWLRQ DERXW WKH WUDIILF VWDWH LGHQWLI\ SRWHQWLDO SUREOHPVDQGDFWXSRQFRQWUROGHYLFHVWRRYHUFRPHWKHP,WKDVEHFRPHSDUWLFXODUO\

(QJLQHHULQJ$JHQW6\VWHPVIRU'HFLVLRQ6XSSRUW

GLIILFXOWIRUWKH-37HQJLQHHUVWRSHUIRUPWKLVMRELQUHDOWLPHDVLQWKHIROORZXSRI WKH  2O\PSLF *DPHV WKH WUDIILF PDQDJHPHQW LQIUDVWUXFWXUH LQ %DUFHORQD KDV EHFRPH LQFUHDVLQJO\ FRPSOH[ 1RZDGD\V LQIRUPDWLRQ DERXW WKH WUDIILF VWDWH RI WKH XUEDQPRWRUZD\QHWZRUNFRQVLVWLQJRIRQHULQJURDGDQGVHYHQDGMDFHQWPRWRUZD\V LVSURYLGHGE\RYHUWHOHPHWHUHGVHQVRUVu ORRSGHWHFWRUVv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

1 2 = 3   WKH RSHUDWRU UHFHLYHV GDWD IURP WKH VHQVRUV IURP DQ H[WHUQDOV\VWHPDQGKHVKHLQWURGXFHVWKHQHZFDVHPDQXDOO\ 1 4  = 0  WKHRSHUDWRUUHFHLYHVDQDODUPIURPWKHSROLFH 6LJQDOSODQVJHQHUDWHGE\RSHUDWRUVFRPSULVHWZRFODVVHVRIDFWLRQV

 

04J (0 VHWDZDUQLQJPHVVDJHRQDSDQHOHJu FRQJHVWLRQDW@v J8

  PRGLI\DUHJXODWRULQRUGHUWRSUHYHQWFRQJHVWLRQ

Fig. 3.7UDIILFLQIUDVWUXFWXUH>@

 6DVFKD2VVRZVNLHWDO

Fig. 4.8(5'LDJUDPIRUWKHWUDIILFGRPDLQ

:KHQFRPPXQLFDWLQJZLWKWKH'66WKHRSHUDWRULVLQYROYHGLQIOH[LEOHGLDORJXHV DVNLQJ IRU WKH UHDVRQ RI D GLDJQRVLV   ZKDW LV KDSSHQLQJ   RU D VSHFLILF VLPXODWLRQ 7KHIROORZLQJJRDOFDVHVDUHDVVLJQHGWRWKH'66  

*  0 2  GHWHFW ZKHQ WZR DFWLRQV IURP DGMDFHQW DUHDV JHQHUDWH GLIIHUHQWDFWLRQVRYHUWKHVDPHDUHDHOHPHQWV *  :  2  GHWHFW LI DFWLRQV IURP DGMDFHQW DUHDV JHQHUDWH D EDG RYHUDOOVROXWLRQ

7KH '66 FDQ PRGLI\ DQ HQYLURQPHQW SUREOHP  40  8

   7KHUHDUHQRFXUUHQWUHTXLUHPHQWVIRUUHDOUHDFWLYLW\EXWRQHPLJKWFRQVLGHUUHDFWLYH H[WHQVLRQVOLNHVSHHGOLPLWDWLRQVLQFDVHDVHYHUHLQFLGHQWLVGHWHFWHG2QFHWKH8(5 FDVHVKDYHEHHQGHILQHGDILUVWLGHQWLILFDWLRQRIWKHDJHQWVFDQEHGRQH,QWKLVFDVH VLQFHZHKDGGLVWULEXWHGH[SHUWLVHLQWKHSUREOHPHDFKRQHDJHQWKDVEHHQDVVLJQHG HDFKSUREOHPDUHDRIWKHURDGQHWZRUN 3.2

IDEAS for Traffic Management: Design

7KH PDJQLWXGH RI D WUDIILF SUREOHP LQ FHUWDLQ SDUW RI WKH URDG QHWZRUN FDQ EH H[ SUHVVHG E\ WKH DPRXQW RI WUDIILF GHPDQG WKDW H[FHHGV WKH FDSDFLW\ RI D FHUWDLQ URDG VHJPHQWLQYHKLFOHVSHUKRXU 7KLVLVFDOOHGWKHVHJPHQWtV  A 7KHTXDOLW\ RIDWUDIILFPDQDJHPHQWDFWLRQFDQEHPHDVXUHGE\WKHUHGXFWLRQRIWUDIILFH[FHVVWKDW WKH\ DUH H[SHFWHG WR SURGXFH ,Q FRQVHTXHQFH WUDIILF PDQDJHPHQW DJHQWV XVH WKH RYHUDOOUHGXFWLRQRIH[FHVVLQWKHLUSUREOHPDUHDVLHWKHVXPRYHUDOOURDGVHJPHQWV WKDWEHORQJWRWKHLUDUHD DVDORFDOXWLOLW\PHDVXUH6HWWLQJRXWIURPWKLVQRWLRQLQWKH VHTXHOZHVKRZKRZWRGHVLJQPHWKRGVWKDWFRSHZLWKDQ,'($tVWDVNV Problem Identification and Diagnosis (YHU\ FRXSOH RI PLQXWHV D WUDIILF PDQDJHPHQW DJHQW UHFHLYHV WHPSRUDO VHULHV RI PDJQLWXGHV VXFK DV WUDIILF VSHHG IORZ DQG RFFXSDQF\ IURP WKH URDG VHQVRUV RI LWV DUHD7KLVUDZGDWDLVLQLWLDOO\SUHSURFHVVHGLQRUGHUWRILOWHURXWQRLV\DQGHUURQHRXV GDWD 6XEVHTXHQWO\ IX]]\ GDWD DEVWUDFWLRQ LV SHUIRUPHG VHH )LJ   DQG DJJUHJDWH

(QJLQHHULQJ$JHQW6\VWHPVIRU'HFLVLRQ6XSSRUW

PDJQLWXGHV VXFK DV WHPSRUDO DQG VSDWLDO JUDGLHQWV DUH FDOFXODWHG IRU WKH GLIIHUHQW VHFWLRQV 7KH DFWXDO SUREOHP LGHQWLILFDWLRQ LV SHUIRUPHG E\ PDWFKLQJ WKH DEVWUDFWHG WUDIILF GDWD DJDLQVW D NQRZOHGJH EDVH RI IUDPHV ZKLFK PRGHO SUREOHP VFHQDULRV )LJ  VKRZVRQHVXFKIUDPHWKDWPDWFKHVWKHDEVWUDFWHGWUDIILFGDWD6XSSRVHWKDWDVDUHVXOW RIGDWDDEVWUDFWLRQORZVSHHGDQGKLJKRFFXSDQF\DUHLGHQWLILHGLQ5RQGDGH'DOWHQ 'LDJRQDO DQG PHGLXP WR KLJK VSHHG DQG ORZ RFFXSDQF\ LQ 5RQGD HQ Gt(VOXJXHV 7KHVHIDFWVPDWFKWKHIUDPHVKRZQLQ)LJVRWKDWDQLQFLGHQWLQWKHFHQWUDOODQHRI 'LDJRQDOURDGLVLGHQWLILHGZKLFKPDQLIHVWVLWVHOIDVDWUDIILFH[FHVVZLWKUHVSHFWWR WKHURDGtVFDSDFLW\ RIYHKKEHWZHHQ'LDJRQDODQG/OREUHJDWLQWKH'DOWULQJ URDG 7UDIILF IURP &ROOFHUROD WR /OREUHJDW DQG LQ D PLQRU GHJUHH IURP 'LDJRQDO KHDGLQJWRZDUGV/OREUHJDWFRQWULEXWHVWRWKLVH[FHVV Action Planning and Prediction Tasks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tVVLJQDOSODQFRQVWLWXWHVRQHUHFRPPHQGDWLRQRIWKHV\VWHP,QWKHH[DPSOHLW LVVXJJHVWHGWRGLVSOD\FRQJHVWLRQZDUQLQJVDW'LDJRQDOIRUSDQHOV3,93,9 DQG3,9ZKLOHVHWWLQJWKHFRQWHQWLRQOHYHORIUHJXODWRU5WRPHGLXP$VDUHVXOW RIWKLVSURFHVVDVHWRI DOWHUQDWLYH VLJQDO SODQ UHFRPPHQGDWLRQV WRJHWKHU ZLWK WKHLU XWLOLW\LHWKHLUH[SHFWHGUHGXFWLRQRIORFDOWUDIILFH[FHVV LVSURGXFHG Co-ordination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



lo w



 



hig h

m e dium

 



 



























S P EE D (Km /h)



PO S lo w



 

hig h

m e dium



 



 





SA TURA TIO N (Flo w/C ap acity)*1 00

POS 

low

 

hig h

m ed ium



 



 





O C C UP A NC Y  (P ercentage )

Fig. 5.3RVVLELOLW\IXQFWLRQVWRDEVWUDFWWUDIILFYDOXHV

E / OR

D UHJ

W

&DQ&DUDOOHX

'LDJRQDO

&RO OFHU ROD

5 3,9

3,9 3,9

3,9

Congestion warning in Ronda de Dalt at Diagonal 0 Q0M FRQJHVWLRQDW'LDJRQDO

3       

Incident congestion in the central lane at Diagonal 3 75RQGDGH'DOWHQ'LDJRQDO VSHHGORZ RFFXSDQF\KLJK

 

3 75RQGDGH'DOWHQG (VSOXJXHV VSHHGPHGLXPKLJK RFFXSDQF\ORZ

 

2   EHWZHHQ5RQGDGH'DOWHQ'LDJRQDO DQG5RQGDGH'DOWHQG (VSOXJHV H[FHVVYHKK  SDWKV  )URP&ROOFHURODWR/OREUHJDW!>@ )URP'LDJRQDOWR/OREUHJDW!>@

0 L0M FRQJHVWLRQDW'LDJRQDO 0 O0M

FRQJHVWLRQDW'LDJRQDO

8

  8

FRQWHQWLRQOHYHOPHGLXP

= 2   : 

    8 * !>@

0 

= 2   : 

    22  T>@



= 2   : 

       T>@





3     S 

= 2   22   IUHH = 22  *   ZLWKSUREOHPV = *  : 



Fig. 6.$QH[DPSOHVFHQDULR

ZLWKSUREOHPV

(QJLQHHULQJ$JHQW6\VWHPVIRU'HFLVLRQ6XSSRUW

4

Conclusions

,QWKLVSDSHUZHKDYHDUJXHGLQIDYRXURIDQDJHQWEDVHGDSSURDFKWRWKHFRQVWUXFWLRQ RIDGYDQFHG'HFLVLRQ6XSSRUW6\VWHPV:HKDYHLQWURGXFHGWKHFRQFHSWRIDQ,QWHOOL JHQW'HFLVLRQPDNLQJ$VVLVWDQW,'($ DQGRXWOLQHGKRZDJHQWRULHQWHGNQRZOHGJH DQGVRIWZDUHHQJLQHHULQJWHFKQLTXHVFDQEHXVHGWRJXLGHWKHSURFHVVRIDSULQFLSOHG FRQVWUXFWLRQRI,'($VEDVHGRQDPXOWL DJHQWDUFKLWHFWXUH2XUDSSURDFK KDV EHHQ LOOXVWUDWHGLQWKHURDGWUDIILFPDQDJHPHQWGRPDLQ :HDUHDZDUHRIWKHIDFWWKDWDQ,'($EDVHGGHVLJQRI'66SXWVPXFKEXUGHQRQ WKHDJHQWVtNQRZOHGJHPRGHOVZKLFKOHDGVWRUDWKHUu IDWv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

References >@ >@ >@ >@

%UHXNHU - YDQ GH 9HOGH :   2;4*3 :     1A 

J  ,263UHVV %URZQ '& &KDQGUDVHNDUDQ %   *   0   3   0RUJDQ .DXIPDQQ &KDQGUDVHNDUDQ%  7RZDUGV D 7D[RQRP\ RI 3UREOHP 6ROYLQJ 7\SHV 4%%J S #x  &XHQD - 2VVRZVNL 6   'LVWULEXWHG 0RGHOV IRU 'HFLVLRQ 6XSSRUW ,Q :HL HG J 4 3 U 4J 4 *4%0,73UHVV x 

 6DVFKD2VVRZVNLHWDO

>@ >@ >@ >@ >@ >@ >@ >@ >@ >@ >@ >@ >@ >@ >@ >@ >@

&XHQD - +HUQ QGH] - 0ROLQD 0   .QRZOHGJH2ULHQWHG 'HVLJQ RI DQ$SSOLFDWLRQIRU5HDO7LPH7UDIILF0DQDJHPHQW,Q0 %1 %2% 4     1245V:LOH\ 6RQVx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x  0F'HUPRWW -   3UHOLPLQDU\ 6WHSV 7RZDUG D 7D[RQRP\ RI 360V ,Q 0DUFXVHG 4 ;  4 1A 3 .OXZHU 2VVRZVNL 6   &RRUGLQDWLRQ LQ $UWLILFLDO $JHQW 6RFLHWLHV 6SULQJHU 9HUODJ 2VVRZVNL 6 6HUUDQR -0   $JHQWEDVHG $UFKLWHFWXUHV IRU $GYDQFHG 'HFLVLRQ6XSSRUW,Q0 %) .     0  4  )4= 85-& 6LOYHU0 3 3 * J. -RKQ:LOH\ 6RQV 9ODKDYDV , HW DO   $Q ,QWHOOLJHQW 0XOWLDJHQW 6\VWHP IRU :$1 0DQ DJHPHQW111   3 Qx  :RROGULGJH 0 -HQQLQJV 1 .LQQ\ '   7KH *DLD 0HWKRGRORJ\ IRU $JHQWRULHQWHG$QDO\VLVDQG'HVLJQ44 J  3  LL.OXZHU$FDGHPLF3XEOLVKHUVx 

Co-ordinating Heterogeneous Interactions in Systems Composed of Active Human and Agent Societies Konstantinos Prouskas and Jeremy Pitt Intelligent and Interactive Systems Group Department of Electrical and Electronic Engineering Imperial College of Science, Technology and Medicine Exhibition Road, London SW7 2BT, U.K. {k.prouskas,j.pitt}@ic.ac.uk tel. +44 (0)20 7594 6318

Abstract. This paper describes the specification and implementation of the middle layer in a new three-layer time-aware agent architecture. This architecture is designed for applications and environments where societies of humans and agents play equally active roles, but interact and operate in completely different time frames. The middle layer, called the Time-Aware Layer, uses services of the underlying real-time layer to co-ordinate the heterogeneous interactions present in composite humanagent systems. Interactions are unified by abstracting away from their temporal representation, temporal scale and class of parties they involve (be they humans or agents). To achieve this, this paper firstly introduces Availability Functions as the primary mechanism of reasoning about temporal constraints placed on interactions. It subsequently describes their stylised analytic representation and develops a Selective Sampling Algorithm which allows searching through them in bounded time. The resultant implementation allows more effective engineering of the topmost application layer firstly by providing an abstract, unified view of interactions and secondly by predicting and guaranteeing their initiation and completion times.

1

Introduction

Many agent systems operating in the real world affect or depend on humans for their correct and effective operation. In such systems, agents have temporally constrained goals and need to interact with other agents as well as humans in complex patterns in order to achieve them. Their temporal behaviour is determined by the set of temporal constraints placed equally on the computations and the interactions within the system. In this paper we describe the specification and implementation of the middle layer in a new three-layer time-aware agent architecture. Time-aware agents are defined as agents capable of dealing with the hard, fast temporal dimension of agent-to-agent interactions while equally handling the much softer and slower P. Petta et al. (Eds.): ESAW 2002, LNAI 2577, pp. 199–212, 2003. c Springer-Verlag Berlin Heidelberg 2003 

200

Konstantinos Prouskas and Jeremy Pitt

interactions with humans. This architecture is designed for applications and environments where societies of humans and agents play equally active roles, but interact and operate in completely different time frames. We focus on the architecture’s Time-Aware Layer, which uses services of the underlying real-time layer to co-ordinate the heterogeneous interactions present in composite human-agent systems. To facilitate this co-ordination, interactions are unified by abstracting away from their temporal representation, temporal scale and class of parties they involve (be they humans or agents). To achieve this, we firstly introduce Availability Functions as the primary mechanism of reasoning about temporal constraints placed on interactions. We subsequently describe their stylised analytic representation within the layer and develop a Selective Sampling Algorithm which allows searching through them in bounded time. Section 2 provides background information and describes our motivation for this work. Section 3 presents an overview of the overall time-aware architecture and Section 5 focuses on its Time-Aware Layer in which interaction co-ordination takes place. Section 6 performs a brief evaluation before conclusions are presented in Section 7.

2 2.1

Background & Motivation Composite Human-Agent Systems

Practical use of agent systems has seen their application in areas where there is extensive interaction with humans at the most basic level and societal or organisational structures more generally. Interactions between agents and humans play a key role in areas such as agent-enhanced workflow and e-commerce. Such systems comprise two classes of entities: agents and humans, each organised in their respective society (Figure 1). Normally, the agent society is formed by a software multi-agent system, while the human one reflects the roles and relationships of the ‘real’ human societal structure. Each society is not passive and does not function independently from the other, but forms an active and integral element of the environment. As an example of this, consider the following envisaged agent scenario [1]:

Human Society 1

soft (human time) 'at certain times'

4

Agent Society 2

3

hard (agent time) 'at any time'

Fig. 1. Composite systems consisting of agent and human societies

Co-ordinating Heterogeneous Interactions

201

The scenario takes place at a doctor’s office in which a handheld agent is asked to produce a schedule for a physiotherapy treatment. The agent engages in conversation with the doctor’s agent to retrieve the prescribed treatment and subsequently contacts a series of provider agents to determine the best one. After a few minutes the agent returns a proposed schedule of treatment to its human owner. Unfortunately, the schedule is unsatisfactory because it involves the owner driving across town during rush hour. Another agent is then given a stricter set of constraints and asked to come up with an alternative schedule. The second agent returns a better-suited schedule, albeit one that requires its human owner to cancel some of their less important appointments. The system is initially unsuccessful in its attempt to provide a solution because it has not taken into account that the affected party is a human and therefore bound by a wholly different set of temporal constraints (in this case, the concept of rush hour) than a software agent. Furthermore, there is unnecessary iteration because humans are never treated as active parts of the system. Instead, they are external, passive elements: the system is a black box in which humans feed some input and receive some output. This results in a trial-and-error approach: problems are not detected until the results are produced, at which point a human has to make modifications and retry. To avoid these limitations it is therefore desirable to treat both humans and agents as active parts of these systems. 2.2

Co-ordinating Heterogeneous Interactions

The challenge of making agents and humans equally active participants lies in the fundamentally different temporal nature of interactions that it would encompass: agent-agent interactions (within the agent society), human-human interactions (within the human society), and also human-agent interactions in both directions (originating at either society). From a temporal viewpoint, agent-agent interactions are well defined, fast and predictable, while human-agent interactions are very loosely defined and constrained and operate on an entirely different temporal scale. Human-human interactions are even more loosely defined as they introduce human uncertainty at both ends. Agent-agent interactions normally assume availability of both parties (they can occur ‘at any time’) and response times in the order of seconds at most. Additionally, they are commonly iterative in nature and follow to a greater or lesser degree some protocol convention. Human-agent interactions are constrained either by physical location limitations, as when a human is ‘out of reach’ of an agent, or by higher-level concepts such as attention, interest, mood etc. (that is, they only occur ‘at certain times’). Further factors that may affect human-agent interactions include day/night-time, time of day, working hours, bank holidays, illness, social commitments, biological reasons etc. Human-human interactions are a generalisation of human-agent interactions where both parties are human, and are the least predictable of the three.

202

Konstantinos Prouskas and Jeremy Pitt

This discrepancy stems partially from the difference in which agents and humans understand and deal with time. Agents, being principally software computing entities, have an implementation-level representation of time such as vector clocks commonly found in distributed systems, whereas humans are used to dealing with abstract relative concepts such as ‘today’ and ‘tomorrow’. Agents also have a much finer-grain representation of time like ‘seconds elapsed since January 1st, 1970’, whereas humans normally only need a much coarser granularity like ‘3 o’clock’ or ‘this afternoon’. Finally, agents have well-defined means of measuring temporal intervals (by measuring the difference of two UTC-synchronised clocks, for example), whereas humans rely on a variety of calendar systems and intervals of variable duration such as ‘evening’, ‘month’, ‘year’ and so on. 2.3

Related Work

There are several examples of real-time agent architectures, both ‘hard’ [2, 3] and ‘soft’ [4]. Real-time agent systems are systems in which agents are temporally constrained in achieving their goals. They combine real-time (RT) and artificial intelligence (AI) characteristics by either embedding AI in RT, embedding RT in AI, or by adopting a co-operative RT-AI approach [5]. However, applying these architectures to this particular problem presents two key limitations. The first is that they mainly consider systems comprised of a single class of entities, namely software agents. Adding a human class, however, is problematic as humans ‘are not real-time’, that is, they cannot be considered to be the well-defined, predictable, principally computational entities their agent counterparts are. The second limitation is that these architectures assume the internal deliberative process of an agent is exposed and focus on temporally constraining it in real time. In extending this to the type of composite human-agent systems we are considering, however, expressing temporal constraints on the deliberative process of a human actor is not appropriate. This is similar to deontic specifications in software engineering, where placing obligations on human actors is equally inappropriate. On the other hand, work on collaborative systems between agents and humans either does not address temporal constraints [6] or, when it does [7], it is not from a real-time, predictability perspective. There are also examples of “dramatic failures” in such systems [8] because the temporal distinction between humans and agents has not been made explicit in the design. There is therefore a need to identify the different classes of parties present; secondly to reason about how their interactions (rather than their deliberative processes) affect the system’s temporal behaviour; and finally be able to handle all interactions in a uniform manner, irrespective of the parties they involve.

3

A Time-Aware Agent Architecture

We have developed a real-time architecture, comprising time-aware agents, that better meets the requirements described above. Time-aware agents are defined

Co-ordinating Heterogeneous Interactions

203

as agents capable of dealing with the hard, fast temporal dimension of agentto-agent interactions while equally handling the much softer and slower interactions with humans. Time-aware agent systems can deal with an amalgam of hard, soft, human and non-real-time interactions, reason about the temporal constraints placed on the system by each type of interaction, make transformations between them and co-ordinate (schedule) activities seamlessly irrespective of their constituent constraints. In this architecture, we consider systems comprised of a number of parties, agents or humans, logically or physically distributed over an interconnected network environment such as the Internet. Agents are lightweight computational entities that cannot share any data except by means of interaction through message passing. They are trustworthy and work co-operatively toward achieving the system’s goals. We primarily consider the co-ordination of agent-agent and human-agent interactions. Human-human interactions are considered only as far as they are agent-mediated, that is, emergent from a specific pattern of interaction (namely 1 → 2 → 3 → 4 in Figure 1). We treat deliberation as internal to both agents and humans, treating both opaquely. We instead focus on providing the environmental and architectural mechanisms by which agents can be aware of the different types of interactions they are engaged in, as well as their temporal repercussions in achieving their goals. We have realised this architecture using a prototype implementation in the Agent Process Interaction Language (April) [9], consisting of three layers: the April Real-Time Run-Time layer, the Time-Aware Layer (TAL) and the Application Agent Layer (AAL). The ART layer forms the underlying real-time agent platform. Similar to a real-time operating system, the ART provides temporal guarantees to the rest of the architecture. More information on the ART can be found in [10]. The AAL is the overlying application layer; in it, software agents exist and work toward their goals. The discussion in this paper focuses on the middle of the three layers: the Time-Aware Layer.

4

The Time-Aware Layer

The Time-Aware Layer (TAL) is the layer that provides a unified temporal view of the entire system. The TAL abstracts away from different time representations, disparate temporal scales and class of parties (human or agent) engaged in interactions. The main components of the TAL are (Figure 2): Human/Agent Handles: Both humans and agents are uniquely identified by their handle. Handles are an abstraction mechanism by which parties in the system can be uniformly addressed. Task Representation: Tasks are hierarchically decomposed sequences of either processing or interacting (communicating). Unified Time Representation: The Unified Time Representation (UTR) is an April macro layer responsible for transforming different human and agent

204

Konstantinos Prouskas and Jeremy Pitt

Task representation Unified time representation Human/agent handles

Interaction scheduling

ART layer

Availability functions

Fig. 2. The Time-Aware Layer

time representations, units, and calendars into a common representation (namely, Universal Coordinated Time). Human concepts like time zones, working hours, next working day, etc. are handled by the UTR. Availability Functions: Availability Functions are associated with each party to express that party’s degree of presence within the system. They are fully described in Section 5.1. Interaction Scheduling: Interaction scheduling involves co-ordinating and guaranteeing initiation and completion times of interaction sequences (generated from tasks). Interaction scheduling is detailed in Section 5.2. Schedules generated by this process are passed on to the underlying ART layer.

5

Realisation

Interaction scheduling is central to the TAL and Availability Functions play a key role in this process. In order to make interaction scheduling fast and predictable enough for real-time application, we constrain the Availability Function representation and discuss a Selective Sampling Algorithm. 5.1

Availability Functions

The TAL uses the concept of availability and Availability Functions (AFs) to reason about the temporal nature of interactions. The availability of a party i at time t, given by Vi (t), is that party’s degree of ‘presence’ in the system, as far as interactions with other parties are concerned. For agents, availability is determined by the intervals during which they are reachable via the network. Some network connections (such as local area network links) may be persistent, others (such as satellite links) may be intermittently connected for fixed periods of time, while yet others (such as dial-up links) may have connection times and intervals which can only be statistically characterised. For humans, availability is principally affected by the time spent sitting in front of their terminal (for fixed contact points) or the times they are in the vicinity of the relevant device (for mobile contact points such as cellular phones). Human AFs closely follow the natural daily cycle. AFs are mathematically expressed as probabilities. However, AFs are not probability distributions. They rather function as look-up tables that provide a statistical measure of availability at any point in time.

Co-ordinating Heterogeneous Interactions

205

AFs are assumed to be externally specified by either the relevant parties themselves or a third party such as the application programmer. Realistic AF generation is important as it directly affects the quality of the predictions, but is outside the scope of this research. Nevertheless, we have used screensaver data to generate experimental human AFs for validation purposes. Simplified versions of this data have been used to produce the figures presented in this paper. 5.2

Interaction Scheduling

Interaction scheduling is the process of co-ordinating the system’s interactions such that the time of their completion is known or can be statistically predicted and guaranteed in advance. Interaction scheduling (in the TAL) is the counterpart of computation time scheduling (in the ART). A one-way interaction (a notification) completes when that notification is received and the recipient becomes aware of it. Similarly, a two-way interaction (a request-reply) completes when the reply reaches the initiator and the initiator becomes aware of it. We therefore distinguish between the time a message is delivered from the time it is handled. AFs drive not only the completion time of an interaction, but also the time of its initiation. For a guarantee level g (expressed as a probability) and a party i, an interaction can be initiated or completed only when Vi (t) ≥ g. For scheduling interactions interspersed with processing, we also distinguish between off-line and on-line processing. Off-line processing at a party can proceed independently of its availability, while on-line processing can proceed only when availability allows it, that is, when Vi (t) ≥ g. For example, a request for a human to review a paper can proceed off-line, while requests for e-mail composition are normally done on-line. Figure 3 shows an example of interaction scheduling for a typical requestreply scenario for g = 0.5. At t = t1 , some off-line processing has completed at j’s end and a request is ready to be sent to k. This interaction will not be initiated until t = t2 , when Vj (t2 ) ≥ 0.5. Party k is not aware of this until t = t3 , when Vk (t3 ) ≥ 0.5. As a result of the request, some on-line processing needs to be performed, proceeding only when Vk (t) ≥ 0.5. At t = t4 processing is complete. The reply can be initiated immediately, however it will not be handled by j until t = t5 , when Vj (t5 ) ≥ 0.5. Interaction scheduling can be conceptually reduced to repeated applications of variations on a primitive operation mathematically formulated as: Given a function f (t), a reference point t = tref and a condition C on f (t), find:        (1) tx : (tx ≥ tref ∧ C (f (tx ))) ∧ tx : tx > tx ≥ tref ∧ C f tx Practically, arbitrarily-shaped functions f (t) can either be stored symbolically in their analytic form, or numerically as discrete data samples. In our context, sampling has two undesirable properties:

1.00 0.75

Konstantinos Prouskas and Jeremy Pitt availability

206

0.50

j 0.25

k

0.00 0.00

time(hrs) t1

t2

24.00

t3

48.00

72.00

96.00

120.00

t4

144.00

t5

168.0

Fig. 3. Interaction scheduling for a request-reply scenario between parties j and k, and a guarantee level g = 0.5

– The success of sampling is dependent on a good choice of sampling frequency (step size). A large step may result in solutions landing between sampled points. A small step causes unnecessary over-sampling leading to unfeasibly large computation times. Unfortunately, in composite human-agent systems, we are dealing with interactions with greatly disparate temporal scales. On one hand, it is desirable to detect sub-second availability periods for agents, while on the other, it may be several hours or even days before human availability rises to the desired level. There is therefore no single best step size. – Sampling is an iterative process which cannot be temporally bound without restricting the search range. This has adverse implications for any real-time system attempting to provide deterministic guarantees. A match may be found immediately or it may be, for example, that a human is on vacation and several weeks must be searched through until one is found. To address this we have developed a stylised analytic representation for AFs (Section 5.3) which, when used in conjunction with a Selective Sampling Algorithm (Section 5.4), allows searches to be performed in bounded time. 5.3

Availability Function Representation

AF representation forms the first half of making interaction scheduling fast and predictable enough for real-time application. We constrain AF representation in the following ways: – AFs are analytically stored and represented as hierarchical operator trees named environments. Primitive (leaf) nodes of environments are named spans. For example, a ‘working’ span for a human might set the availability to 0.9 whereas a ‘sleeping’ span might set the availability to zero. – Spans have a constant availability value v. – The periods of composed environments must be harmonics (integral multiples of one another). This is a requirement for the Selective Sampling Algorithm to work correctly. Non-harmonic environments must be broken down

modulation V1×V2

availability

subtraction V1-V2

availability

addition V1+V2

availability

original functions

availability

availability

1.00

availability

Co-ordinating Heterogeneous Interactions

demodulation V1+(1-V1)×V2

207

selection V1

0.75

V1

0.50

V2

0.25

time(s)

0.00

time(s)

time(s)

time(s)

time(s)

time(s)

Fig. 4. Results of Availability Function composition for common composition functions f (addition, subtraction, modulation, demodulation, selection) into a series of sub-environments such that they have harmonic repetitions. Fortunately, most human environments follow the natural daily cycle and therefore do not require any transformation. Environments go through a series of operations (dictated by the nodes in the tree) to arrive at the final composite AF. Nodes in the AF trees can be one of the following types: Span (v): Sets the availability value to v ∈ IR.1 Bound (V, i): Bounds the value of environment V to the interval i = [t1 , t2 ). V remains unchanged within [t1 , t2 ) but becomes undefined outside it. Undefined environments do not affect the availability value. Repeat (V, p): Repeats environment V with a period of p seconds. Offset (V, o): Translates (offsets) environment V by o seconds to the right. Compose (f, V1 , V2 , . . .): Composes two or more environments together using the function f as an infix operator. Composition only has an effect at points where at least two environments are defined. Function f can be any binary function, but it is normally one of addition (= V1 +V2 ), subtraction (= V1 − V2 ), modulation (= V1 · V2 ), demodulation (= V1 + (1 − V1 ) · V2 ) or selection (= V1 ). Environment composition for common functions f is illustrated in Figure 4. Repeated compositions result in more complex AFs, such as the ones shown in Figure 5 (the notation in this figure has been simplified for clarity). Given the AF’s tree representation, an evaluation function is provided to return V (t) at any point t. Simplified pseudocode for evaluating the AF node at point t is given in Algorithm 1. 5.4

Selective Sampling Algorithm

The task of the Selective Sampling Algorithm (SSA) is to generate a minimal set of sampled points while guaranteeing that, if the solution of (1) exists, it will be among those points. 1

Note that the value of v is not constrained at this stage. This allows for a simpler tree representation of AFs. However, since AFs represent probabilities, final results are clipped to lie within [0, 1].

Konstantinos Prouskas and Jeremy Pitt

Compose *

Repeat 24

Compose *

Span 0.40

1.00

Bound [0,6)

Span 0.10

Bound [9,17)

Span 2.00

Bound [12,13)

Span 0.25

Bound [15,16)

Span 0.35

Bound [0,24)

Span 0.10

Offset 144

Bound [0,24)

availability

208

0.75 0.50 0.25

time(hrs) Span 0.10

0.00 0.0

24.0

48.0

72.0

96.0

120.0

144.0

168.0

Fig. 5. Tree representation and resultant Availability Function Informally, the SSA operates on the premise that, since spans are constant, only their (bounded) extremities are relevant. Furthermore, repeated environments generate relevant points for only their first full period – any solution found in subsequent periods will have a corresponding solution in an earlier period. Lastly, environment compositions are periodic with a period that is the maximum of the periods of their constituent environments. The SSA can be specified as follows: Input: An AF Vi (t) and a reference point t = tref at which the search starts. Output: A set of time points where spans may cause a change in the availability value. It can be shown that this is a superset of the set of points S that satisfy (1) for each discrete value that Vi (t) can take, and that if a solution tx exists, then it is always contained in S. Steps: There are two stages. In the first stage, a set of relevant points is generated from tref . A point is relevant if it is the earliest point where a change in spans may cause a change in the availability value. The reference point is always relevant. The theorems which drive relevant point generation for a reference point tref are summarised below (proofs are omitted): 1. Span (v) generates a single relevant point at t = tref . 2. Bound (V, I) generates a relevant point at each edge of I. 3. Repeat (V, p) generates relevant points only in its relevant period. The relevant period is the first full period nearest to tref in the direction of the search. Algorithm 1 Availability Function evaluation algorithm evaluate(node, t) { case node of type Span(v): v Bound(V,I): evaluate(V,t) if t lies in I, undefined otherwise Repeat(V,p): evaluate (V,remainder(t/p)) Offset(V,o): evaluate (V,t-o) Compose(f,V1,V2,...): evaluate(V1,t) f evaluate(V2,t) ... }

Co-ordinating Heterogeneous Interactions

209

4. Offset (V, o) has a relevant point at tx if and only if the environment V has a relevant point at tx − o. 5. Compose (f, V1 , V2 , . . .) generates relevant points which always lie in the relevant periods of its constituent environments (V1 , V2 , . . .). In the second stage, each of the relevant points of the first stage functions as a reference point to add its own relevant points to the set. This stage accounts for possible availability values resulting from the composition of two or more spans.

1.00

availability

After the second stage, the list of relevant points can be passed to the evaluation function to test (sample) each point. Figure 6 illustrates relevant point generation in the first and second stages of the SSA. Based on this specification, correctness, completeness and complexity properties of the SSA can be proven. We omit the full format proof of these results here, however we present a sketch of the computational complexity result as it is the limiting factor for time-aware agents. The first stage is O (n) with the number of nodes in the AF representation, generating a linearly proportional number of relevant points (a maximum of two per node). The second stage is identical to the first, only tref takes values from results of the first stage. The total algorithmic complexity is therefore  the O n2 . Theoretically, up to 4n2 points will need to be sampled in the worst case, although practically not every node will generate relevant points and many points generated from the first and second stages will overlap. The main advantage of the SSA is therefore that it completes in bounded time. Another advantage is that, as it operates on analytic representations of AFs, it will find a solution (if one exists) irrespective how slim or far into the future it is. Agent functions that change in a temporal scale of milliseconds can be handled in the same way as human functions that may be specified in terms of hours. Searching in ‘forward’ time can be used to predict the completion time of current interactions, while searching in ‘backward’ time can be used to determine the latest start time for an interaction sequence such that it completes by a specified deadline.

first stage second stage

0.75

0.50

0.25 time(hrs) 0.00 0.00

72.00

144.00

216.00

288.00

360.00

432.00

504.00

Fig. 6. Relevant point generation in the first and second stages of the Selective Sampling Algorithm for a reference point at t = 0

210

Konstantinos Prouskas and Jeremy Pitt

The SSA can be used as the basis for a wide range of interaction scheduling operations, including (i) finding the first point at which availability exceeds or falls below a threshold (ii) finding the maximum and minimum value of availability within an interval (iii) summing the length of time for which the availability lies above or below a threshold within an interval and (iv) calculating the value

t of functions of the form h = t12 (Vi (t) − g) dt, which give a measure of how much ‘better’ or ‘worse’ availability is during [t1 , t2 ] with respect to g.

6

Evaluation

The SSA and the encompassing TAL layer will be evaluated from two distinct perspectives: raw real-time performance will be addressed first, followed by a brief discussion of experimental results in a practical application. There are two versions of the SSA: one implemented in April and one in C (still accessible through April). The April version executes much slower than its C counterpart. However, its main advantage is that it is interruptible and resumable at any point by the April virtual machine. Other advantages are that Compose nodes are much more flexible, allowing any composition function (standard or user-defined) and the composition of an arbitrary number of environments in the same node. The C version is faster than the April one by approximately 350 times. This improves the minimum temporal scale that can be handled and allows much more complex AFs without adversely affecting the system’s performance. However, once the C algorithm is started, it must execute to completion. Another limitation is that Compose nodes are restricted to pre-defined composition functions and only two environments can be composed in a single node. Experimentally, the number of relevant points generated on average by the SSA was much lower than the worst case, lying between 1% and 2% of the theoretical 4n2 limit. In absolute terms, relevant point generation is very fast, with the C version generating relevant points for a 50-node AF within 160μs on a 1GHz processor. Evaluating the same function at each point takes approximately 1.2μs per point. We are also in the process of applying this architecture in a workflow management system to schedule typical processes in Institutions of Higher Education, where some interactions are scheduled with millisecond accuracy, while the overall temporal scale can be of the order of years. Figure 7 shows interaction scheduling for filling in a six-month progress report. Participating parties are shown on the vertical axis. Solid bars denote that a local party process is executing, dashed bars denote that the process is blocked waiting for its dependencies to be satisfied. Arrows indicate messages (interactions) between parties. Left-pointing triangles indicate the delay in interaction completion (as determined by availability), while right-pointing triangles indicate the delay in its initiation. Availability functions are plotted immediately below their corresponding party. It can be seen how certain interactions occurring over weekends will not be initiated or completed until the next working day

Co-ordinating Heterogeneous Interactions

211

root ITsystem meeting scheduler B

DPGS PGT mentor supervisor A

A

A

student

time

Monday, June 3rd

Monday, June 10th

Monday, June 17th

Monday, June 24th

Fig. 7. Interaction schedule for a “progress report” workflow process, showing (A) interactions scheduled around weekends and (B) processing time affected by periods of unavailability (labelled as A in the figure), as well as how on-line processing time is affected by periods of unavailability (labelled as B). This application has also demonstrated that prediction quality is directly affected by how accurately AFs reflect ‘real’ availability. The TAL and the encompassing time-aware architecture represent ongoing research work and the current iteration has several limitations: – The architecture is unsuited to hard real-time environments. Guarantees associated with schedules generated by the TAL are very soft and should be treated as guidelines only. This is a consequence of the presence of interactions which wholly or partly depend on humans for completion. – In all calculations, message delivery latency is assumed to be negligible. This is a valid assumption for local controlled network environments but is not necessarily true for large-scale distributed systems over the Internet. – Commitments to tasks may translate to changes in the AFs. For example, availability may be decreased for the duration of a task so that other tasks’ interactions are scheduled around it. These changes are currently done on a FIFO basis. This may cause a lower-priority task to be scheduled in preference to a higher-priority one. The selective re-scheduling of a minimal set of affected tasks to avoid these priority inversions has not yet been addressed.

7

Conclusions

We have presented the specification and realisation of the Time-Aware Layer in a time-aware architecture which co-ordinates agent-to-agent and human-toagent interactions. To facilitate this co-ordination, interactions are unified by abstracting away from their temporal representation, temporal scale and class of parties they involve. This is achieved by a combination of a Unified Time Representation, common human/agent identifiers and Availability Functions. We have used Availability Functions to model and predict the initiation and completion time of any kind of interaction. By restricting the Availability Func-

212

Konstantinos Prouskas and Jeremy Pitt

tion form with a stylised analytic representation and by taking advantage of their piecewise constant and repetitive nature, we have developed a Selective Sampling Algorithm which allows searching through them in bounded time. The research work described in this paper is ongoing. There are still several outstanding issues, such as the effects of wide-scale distribution and the interplay between computational and interactional scheduling, which will need to be addressed in the near future. Nevertheless, initial evaluation results have been encouraging and we are applying the described time-aware architecture to implement a workflow management system.

Acknowledgements This work has been undertaken with the financial support of the EPSRC-funded project TARA2 (GR/N24940).

References [1] Berners-Lee, T., Hendler, J., Lassila, O.: The Semantic Web. Scientific American (2001) 35–43 200 [2] Musliner, D., Durfee, E., Shin, K.: CIRCA: A Co-operative Intelligent Real-time Control Architecture. IEEE Transactions on Systems, Man, and Cybernetics 23 (1993) 1561–1574 202 [3] Botti, V., Carrascosa, C., Julian, V., Soler, J.: The ARTIS Agent Architecture: Modelling Agents in Hard Real-Time Environments. In: Proceedings of the 9th European Workshop on Modelling Autonomous Agents in a Multi-Agent World, Springer-Verlag (1999) 63–76 202 [4] Wagner, T., Garvey, A., Lesser, V.: Criteria-Directed Heuristic Task Scheduling. International Journal of Approximate Reasoning, Special Issue on Scheduling 19 (1998) 91–118 202 [5] Musliner, D., Hendler, J., Agrawala, A., Durfee, E., Strosnider, J., Paul, C.: The Challenge of Real-Time Artificial Intelligence. Computer 28 (1995) 58–66 202 [6] Babaian, T., Grosz, B., Shieber, S.: A Writer’s Collaborative Assistant. In: Proceedings of the International Conference on Intelligent User Interfaces. (2002) 7–14 202 [7] Payne, T., Lenox, T. L., Hahn, S. K., Sycara, K., Lewis, M.: Agent-Based Support for Human/Agent Teams. In: CHI 2000 Conference on Human Factors in Computing Systems, ACM Press (2000) 202 [8] Chalupsky, H., Gil, Y., Knoblock, C. A., Lerman, K., Oh, J., Pynadath, D. V., Russ, T. A., Tambe, M.: Electric Elves: Agent Technology for Supporting Human Organizations. Artificial Intelligence 23 (2002) 11–24 202 [9] McCabe, F., Clark, K.: April: Agent Process Interaction Language. In Jennings, N., Wooldridge, M., eds.: Intelligent Agents. Volume 890. Springer-Verlag (1995) 203 [10] Prouskas, K., Pitt, J.: Towards a Real-Time Architecture for Time-Aware Agents. In: Proceedings of the First International Joint Conference on Autonomous Agents and Multi-Agent Systems. Volume 1. (2002) 92–93 203

SABPO: A Standards Based and Pattern Oriented Multi-agent Development Methodology 2÷X]'LNHQHOOLDQG5L]D&HQN(UGXU (JH8QLYHUVLW\'HSDUWPHQWRI&RPSXWHU(QJLQHHULQJ %RUQRYD,]PLU7XUNH\ {oguzd,erdur}@staff.ege.edu.tr

Abstract. 2UJDQL]DWLRQDO PHWDSKRU LQ ZKLFK HDFK DJHQW SOD\V D VSHFLILFUROHWRDFKLHYHRUJDQL]DWLRQt VJOREDOJRDOV VHHPVDVWKHPRVW VXLWDEOH DSSURDFK WR PRGHO PXOWLDJHQW V\VWHPV ,Q WKLV SDSHU D QHZ PHWKRGRORJ\ ZKLFK XVHV RUJDQL]DWLRQDO PHWDSKRU DV LWV EDVLV DQG LQWHJUDWHV WKLV PHWDSKRU WR WKH ),3$ VWDQGDUGV DQG ZHOONQRZQ LQWHUDFWLRQSURWRFROVLQDV\VWHPDWLF ZD\LVLQWURGXFHG,Q DGGLWLRQ WR WKHSURSRVHGPHWKRGRORJ\DQHZLQWHUDFWLRQSDWWHUQZKLFKLVFDOOHGDV u RQWRORJ\ LQFHSWLRQv  LV LQWURGXFHG 7KLV SDWWHUQ LOOXVWUDWHV KRZ WKH PHWKRGRORJLVWV FDQ PLQH QHZ DJHQW LQWHUDFWLRQ SDWWHUQV IURP WKH UHTXLUHPHQWV DQG KRZ WKHVH SDWWHUQV FDQ EH LQVHUWHG LQWR WKH ),3$ VWDQGDUGV WR FUHDWH D SDWWHUQ FDWDORJXH 1DWXUDOO\ LQWHUDFWLRQV RI HDFK LQWHUDFWLRQSDWWHUQDGGQHZWDVNVWRWKHSDUWLFLSDQWDJHQWV:HEHOLHYH WKDWWKHUHTXLUHPHQWVRIWKHVHQHZWDVNVDQGKRZWKH\DUHLQVHUWHGLQWR WKH DJHQWt V LQWHUQDO DUFKLWHFWXUH PXVW EH GRFXPHQWHG DV SDUW RI WKH SDWWHUQ NQRZOHGJH 7KXV LQWHUQDO UHTXLUHPHQWV RI WKH u RQWRORJ\ LQFHSWLRQv  SDWWHUQ DUH GLVFXVVHG IURP WKH SHUVSHFWLYH RI WKH SDUWLFLSDWLQJDJHQWV %\ WKLV ZD\ XQGHUVWDQGDELOLW\ RI WKH SDWWHUQ ZLOO EHLPSURYHG

1

Introduction

&KDUDFWHULVWLFVRIVRIWZDUHV\VWHPVDUHJRLQJWRFKDQJHGUDPDWLFDOO\7RGD\VRIWZDUH GHYHORSHUVKDYHWRGHVLJQV\VWHPVWRH[HFXWHLQDZRUOGZKHUHDXWRQRPRXVVRIWZDUH FRPSRQHQWVDUHDOUHDG\H[HFXWLQJDQGORFDOO\LQWHUDFWLQJZLWKHDFKRWKHUDQGZLWKWKH G\QDPLFDOO\ FKDQJLQJ HQYLURQPHQW >@ ,Q VXFK DQ HQYLURQPHQW FRPSRQHQW DEVWUDFWLRQLVQRWVXIILFLHQWDQGGHYHORSHUVQHHGDQDEVWUDFWLRQZLWKFDSDELOLWLHVVXFK DVJRDOGHOHJDWLRQVHPDQWLFVEDVHGLQWHUDFWLRQDQGUHDVRQLQJWRJHQHUDWHNQRZOHGJH IURPHYHQWVDQGPHVVDJHV>@7KHVHFDSDELOLWLHVDUHSURYLGHGE\DQDJHQWDEVWUDFWLRQ DQG V\VWHPV H[HFXWLQJ RQ VXFK HQYLURQPHQWV KDYH WR EH GHYHORSHG DV PXOWLDJHQW V\VWHPV ,Q WKH ODVW IHZ \HDUV VHYHUDO PXOWLDJHQW V\VWHP GHYHORSPHQW PHWKRGRORJLHV ZKLFK WDNH DJHQWV DV KLJKOHYHO DEVWUDFWLRQV DQG GHYHORS WKH PHWKRGRORJ\DURXQGWKHDJHQWFRQFHSWKDYHEHHQSURSRVHG0RVWRIWKHVHSURSRVDOV HLWKHUH[WHQGREMHFWRULHQWHG 22  PHWKRGRORJLHV XVLQJ DJHQW DEVWUDFWLRQ LQVWHDG RI 33HWWDHWDO(GV (6$:/1$,SS ©6SULQJHU9HUODJ%HUOLQ+HLGHOEHUJ

2÷X]'LNHQHOOLDQG5L]D&HQN(UGXU

REMHFW RU IRFXV RQ WKH LQWHUQDO GHVLJQ RI WKH LQGLYLGXDO DJHQW DUFKLWHFWXUH >@>@>@>@ +RZHYHU 0$6V EHKDYH OLNH VRFLDO RUJDQL]DWLRQV ZKHUH HDFK DJHQWSOD\VDVSHFLILFUROHZLWKLQWKHRUJDQL]DWLRQWRVDWLVI\WKHRUJDQL]DWLRQt VJOREDO JRDOV+HQFHZHQHHGDGLIIHUHQWDSSURDFKWKDWWDNHVRUJDQL]DWLRQDODEVWUDFWLRQDVD NH\FRQFHSWWRPRGHOWKHPXOWLDJHQWV\VWHPVZRUNLQJRQG\QDPLFHQYLURQPHQWV *DLD >@ LV RQH RI WKH ZHOONQRZQ PHWKRGRORJLHV WKDW XVHV WKH RUJDQL]DWLRQDO YLHZWRFRQVWUXFWWKHGHYHORSPHQW PHWKRGRORJ\,Q*DLD DQ DJHQW RUJDQL]DWLRQ LV GHILQHG E\ FROOHFWLRQ RI UROHV HDFK UROHt V UHVSRQVLELOLWLHV DQG LQWHUDFWLRQV EHWZHHQ WKRVH UROHV $OO RI WKHVH HQWLWLHV WU\ WR VDWLVI\ WKH RUJDQL]DWLRQ V JOREDO JRDOV DOO WRJHWKHU +HQFH *DLD DVVLVWV WKH GHYHORSHUV WR V\VWHPDWLFDOO\ GLVFRYHU WKH DEVWUDFW HQWLWLHV RI WKH RUJDQL]DWLRQ VXFK DV UROHV UHVSRQVLELOLWLHV LQWHUDFWLRQ SURWRFROV DQG DFWLYLWLHV DQGWRPDSWKHVHHQWLWLHVWRWKHFRQFUHWHRQHVVXFKDVDJHQWVVHUYLFHVDQG DFTXDLQWDQFHV %XW*DLD KDVEHHQLQLWLDOO\ GHYHORSHG IRU PRGHOLQJ RQO\ WKH FORVHG V\VWHPV ZKHUHRUJDQL]DWLRQDO VWUXFWXUH RI WKH V\VWHP LV VWDWLF DQG GR QRW FKDQJH DW UXQWLPH ,W LV D YHU\ FULWLFDO VKRUWFRPLQJ VLQFH 0$6 GHYHORSHUV DUH PDLQO\ LQWHUHVWHGLQRSHQDQGG\QDPLFHQYLURQPHQWV 7KLV VKRUWFRPLQJ RI WKH *DLD PHWKRGRORJ\ KDV EHHQ DGGUHVVHG E\ =DPERQHOOL -HQQLQJV:RROGULGJH>@2PLFLQL>@DQG%HUQRQHWDO>@=DPERQHOOL-HQQLQJV DQG :RROGULGJH LQWURGXFHG WKUHH DGGLWLRQDO RUJDQL]DWLRQDO FRQFHSWV RYHU WKH *DLD RUJDQL]DWLRQDO UXOHV RUJDQL]DWLRQDO VWUXFWXUHV DQG RUJDQL]DWLRQDO SDWWHUQV 2UJDQL]DWLRQDO UXOHV DUH XVHG WR GHILQH JOREDO UHTXLUHPHQWV RI WKH RUJDQL]DWLRQ +HQFH QHZ DJHQWV FDQ OHDUQ RUJDQL]DWLRQt V JOREDO UXOHV ZKHQ WKH\ MRLQ WKH RUJDQL]DWLRQLQDQRSHQHQYLURQPHQW2UJDQL]DWLRQDOVWUXFWXUHGHILQHVWKHDUFKLWHFWXUH RIWKHRUJDQL]DWLRQZKLFKLVVHOHFWHGDPRQJPDQ\SRVVLEOHDOWHUQDWLYHVDFFRUGLQJWR WKH VSHFLILHG UHTXLUHPHQWV 2UJDQL]DWLRQ SDWWHUQV H[SUHVV SUHGHILQHG RUJDQL]DWLRQDO VWUXFWXUHVWKDWFDQEHUHXVHGIURPV\VWHPWRV\VWHP 2PLFLQLt V PHWKRGRORJ\ ZKLFK LV FDOOHG 62'$ H[SOLFLWO\ WDNHV WKH DJHQW HQYLURQPHQWLQWRDFFRXQW(QYLURQPHQWLV PRGHOHGDVVHUYLFHV DQG DJHQWV XVH WKHVH VHUYLFHV WR ZRUN SURSHUO\ ZLWKLQ WKH RSHQ HQYLURQPHQW DFFRUGLQJ WR WKH V\VWHPt V UHTXLUHPHQWV 62'$ DOVR VHSDUDWHV WKH UROHt V WDVNV LQWR LQGLYLGXDO DQG VRFLDO RQHV 6RFLDOWDVNVDUHXVHGWRPRGHOWKHRUJDQL]DWLRQDODVSHFWVRID0$6 $'(/)(>@LVDQRWKHUPHWKRGRORJ\WKDWGLUHFWO\IRFXVHVRQWKHDGDSWLYHQDWXUHRI WKH HQYLURQPHQW ,W LQWURGXFHV WKH FRQFHSW RI  u 1RQ &RRSHUDWLYH 6LWXDWLRQVv  1&6 WKDW GHILQHV WKH EHKDYLRXU RI ORFDO DJHQWV ZKHQ WKH\ HQFRXQWHU DQ XQSUHGLFWDEOH VLWXDWLRQ LQ D G\QDPLF HQYLURQPHQW 'XULQJ WKH GHYHORSPHQW SURFHVV WKH\ WU\ WR LGHQWLI\VXFK1&6WRPRGHOORFDODJHQWVt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

6$%32$0XOWLDJHQW'HYHORSPHQW0HWKRGRORJ\

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u RQWRORJ\LQFHSWLRQv 7KLVSDWWHUQ FDQ EH XVHG WR FRQVWUXFW DGDSWDEOH PXOWLDJHQW RUJDQL]DWLRQV ZRUNLQJ LQ WKH HQYLURQPHQWZKHUHRQWRORJLHVFKDQJHDWUXQWLPHDQGQHZRQWRORJLHVPD\EHDGGHG WRWKHV\VWHPDWDQ\WLPH 7KHUHVWRIWKHSDSHULVRUJDQL]HGDVIROORZV6HFWLRQJLYHVDQRYHUYLHZRIWKH 6$%32PHWKRGRORJ\HVSHFLDOO\GLVFXVVLQJKRZWKH),3$VWDQGDUGVDUHXVHGGXULQJ WKHPRGHOLQJSURFHVV7KHXVHRIWKHPHWKRGRORJ\LVLOOXVWUDWHGE\PHDQVRIDFDVH VWXG\LQVHFWLRQ7KHu RQWRORJ\LQFHSWLRQv SDWWHUQLVLQWURGXFHGLQVHFWLRQ7KLV LQWURGXFWLRQ LQFOXGHV WKH PRWLYDWLRQ EHKLQG WKH u RQWRORJ\ LQFHSWLRQv  SDWWHUQ LWV LQWHUDFWLRQPHFKDQLVPDQGLWVDIIHFWWRWKHSDUWLFLSDWLQJDJHQWt VLQWHUQDODUFKLWHFWXUH /DVW VHFWLRQ FRQFOXGHV WKH VWXG\ DQG RXWOLQHV PDLQ FRQWULEXWLRQV RYHU WKH SUHYLRXV VWXGLHV

2

Overview of the SABPO Methodology

2XUDSSURDFKSXWVWKH),3$t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t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

2÷X]'LNHQHOOLDQG5L]D&HQN(UGXU

LQWHUDFWLRQ SURWRFRO FDQ EH FRQVLGHUHG DV D JHQHULF SDWWHUQ ZKHUH HDFK SDUWLFLSDWLQJ DJHQWSOD\VDVSHFLILFUROHZLWKLQWKHGHILQHGVHPDQWLFWRVROYHDSDUWLFXODUUHFXUULQJ SUREOHP 6LQFH ZH DFFHSW LQWHUDFWLRQ SURWRFROV DV JHQHULF SDWWHUQV ZH XVH WKH FRQFHSWVRIWKHLQWHUDFWLRQSURWRFRODQGWKHLQWHUDFWLRQSDWWHUQLQWHUFKDQJHDEO\ZLWKLQ WKHSDSHU  Q

UH J LV WH UV

A gent

X VHV

 Q

 Q VHQ GHU UH F H LY H U

 Q

 Q

VHDUFK HV 





 Q

D ir e c to r y S e r v ic e

T ra n sp o r t-m e ssa g e

O n to lo g y S e r v ic e

  Q

LV V H Q WUH F H LY H G E \

F R Q WD LQ V

 Q

 Q  Q

M e s s a g e -tr a n s p o r t-s e r v ic e

D ir e c to r y E n tr y

Fig. 1.%DVLFDJHQWUHODWLRQVKLSV

,Q LWLD WR U

3 D UWLF LS D Q W T X H U\ LI

[

T X H U\ UH I Q R WX Q G H UV WR R G UH IX VH

[

ID LOX UH LQ IR UP

Fig. 2.),3$TXHU\LQWHUDFWLRQSURWRFRO

'LUHFWRU\VHUYLFHZKLFKLVDQPDQGDWRU\FRPSRQHQWRI),3$DEVWUDFWDUFKLWHFWXUH PXVWEHDQDO\]HGFDUHIXOO\GXULQJWKHPRGHOLQJSURFHVVVLQFHWKLVVHUYLFHLVWKHKHDUW RIDQ\PXOWLDJHQWRUJDQL]DWLRQUXQQLQJLQDQRSHQHQYLURQPHQW$GLUHFWRU\VHUYLFH SURYLGHV D VKDUHG LQIRUPDWLRQ UHSRVLWRU\ LQ ZKLFK DJHQWV PD\ SXEOLVK WKHLU

6$%32$0XOWLDJHQW'HYHORSPHQW0HWKRGRORJ\

FDSDELOLWLHV DQG LQ ZKLFK WKH\ PD\ VHDUFK IRU RWKHU DJHQWV RI LQWHUHVW $ GLUHFWRU\ VHUYLFHSURYLGHVWKHIROORZLQJVWDQGDUGVHUYLFHV    

$QDJHQWPD\UHJLVWHUDGLUHFWRU\HQWU\ZLWKWKHGLUHFWRU\VHUYLFH $Q DJHQW PD\ PRGLI\ D GLUHFWRU\HQWU\ WKDW KDV EHHQ UHJLVWHUHG ZLWK D GLUHFWRU\VHUYLFH $QDJHQWPD\GHOHWHDGLUHFWRU\HQWU\IURPDGLUHFWRU\VHUYLFH $QDJHQWPD\TXHU\DGLUHFWRU\VHUYLFHWRORFDWHDQDJHQWRILQWHUHVW

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



+HOSLQJD),3$DJHQWLQVHOHFWLQJDVKDUHGRQWRORJ\IRUFRPPXQLFDWLRQ &UHDWLQJDQGXSGDWLQJDVSHFLILFRQWRORJ\RURQO\VRPHWHUPVRIWKHRQWRORJ\ 5HVSRQGLQJWRTXHULHV IRU UHODWLRQVKLSV EHWZHHQ WHUPV RU EHWZHHQ RQWRORJLHV 6L[UHODWLRQVKLSVDUHGHILQHGLQWKHVSHFLILFDWLRQVRQWRORJLHVFDQEHLGHQWLILHG HTXLYDOHQWH[WHQVLRQRIWKHRWKHUZHDNO\WUDQVODWDEOHVWURQJO\x WUDQVODWDEOHRU DSSUR[LPDWHO\WUDQVODWDEOH 2$ VWRUHV WKH GHILQHG UHODWLRQVKLSV EHWZHHQ WKH RQWRORJLHVDQGDJHQWVLQWKHRUJDQL]DWLRQFDQTXHU\WKLVNQRZOHGJH 7UDQVODWLQJ H[SUHVVLRQV EHWZHHQ GLIIHUHQW RQWRORJLHV RI FRXUVH RQWRORJLHV PXVWEHWUDQVODWDEOHDQGWUDQVODWLRQUXOHVPXVWEHGHILQHGE\DQDXWKRULW\

$JHQWV TXHU\ WKH 2$ WR JHW WKH GHVLUHG VHUYLFH XVLQJ WKH ),3$ LQWHUDFWLRQ SURWRFROVZKHUHFRQWHQWRIWKHHDFK$&/ZLWKLQWKHSURWRFROLVZULWWHQLQ),3$0HWD 2QWRORJ\ $VDFRQFOXVLRQ6$%32LGHQWLILHVWDVNVRIDOODEVWUDFWHOHPHQWVRIWKHDUFKLWHFWXUH DQG ),3$ EDVHG LQWHUDFWLRQ SDWWHUQV EHWZHHQ WKHVH HOHPHQWV WR UHDOL]H D FRQFUHWH DUFKLWHFWXUH ZKLFK VDWLVILHV WKH V\VWHPt V UHTXLUHPHQWV 6$%32 RQO\ GHILQHV WKH DQDO\VLVDQGGHVLJQSKDVHVRIWKHGHYHORSPHQWSURFHVV0RGHOVFUHDWHGDQGDFWLYLWLHV LQYROYHGLQWKHVHSKDVHVDUHGHVFULEHGEHORZ

2÷X]'LNHQHOOLDQG5L]D&HQN(UGXU

2.1

Analysis Phase

6$%32 DQDO\VLV SKDVH H[SORLW WZR GLIIHUHQW PRGHOV WKH UROH PRGHO DQG WKH LQWHUDFWLRQ PRGHO DV LQ WKH *DLD PHWKRGRORJ\ %XW ),3$ DEVWUDFW DUFKLWHFWXUH DQG ),3$LQWHUDFWLRQSDWWHUQVDUHWDNHQLQWRFRQVLGHUDWLRQGXULQJWKHPRGHOLQJSURFHVV Role Model. 7KLV PRGHO PDLQO\ GHILQHV WKH UROHV RI WKH RUJDQL]DWLRQ DQG UHVSRQVLELOLWLHV RI WKHVH UROHV WR VDWLVI\ WKH RUJDQL]DWLRQt V JOREDO JRDOV 7KHVH UROHV FDQ EH FRQVLGHUHG DV DEVWUDFW DJHQW HOHPHQWV RI WKH ),3$t V DEVWUDFW DUFKLWHFWXUH VSHFLILFDWLRQ 7KHQ 6$%32 LQWURGXFHV WZR QHZ UROHV EDVHG RQ WKH DEVWUDFW DUFKLWHFWXUH VSHFLILFDWLRQ 7KHVH UROHV DUH u 'LUHFWRU\ 6HUYLFH 3URYLGHUv  DQG u 2QWRORJ\6HUYLFH3URYLGHUv 6HUYLFHVJLYHQE\WKHGLUHFWRU\VHUYLFHDQG2$RI WKH DEVWUDFW DUFKLWHFWXUH DUH DFFHSWHG DV WKRVH UROHV  UHVSRQVLELOLWLHV ,Q 6$%32 UROH PRGHOLVGRFXPHQWHGXVLQJWKH*DLDt VPRGHOLQJVW\OH Interaction Model.7KLVPRGHOGHILQHVWKHLQWHUDFWLRQSURWRFROVEHWZHHQWKHDJHQWV 7R FUHDWH D ),3$FRPSOLDQW DUFKLWHFWXUH u 'LUHFWRU\ 6HUYLFH 3URYLGHU v  DQG u 2QWRORJ\ 6HUYLFH 3URYLGHUv  UROHV SDUWLFLSDWH LQ LQWHUDFWLRQV XVLQJ WKH LQWHUDFWLRQ SDWWHUQVGHILQHGLQWKH),3$VSHFLILFDWLRQV7KHVHLQWHUDFWLRQVDUHGRFXPHQWHGXVLQJ $JHQW80/>@QRWDWLRQDQG),3$t VFRPPXQLFDWLRQDFWVDUHH[SOLFLWO\LQGLFDWHGRQ WKH LQWHUDFWLRQ GLDJUDPV ,QFOXVLRQ RI D GLUHFWRU\ VHUYLFH SURYLGHU DQG DQ RQWRORJ\ VHUYLFHSURYLGHUIRUFHVWKHGHYHORSHUVWRPDNHVDPHGHFLVLRQVLQWKHDQDO\VLVSKDVH 7KHVHGHFLVLRQVDUH    

2.2

1DPHDQGJHQHUDOREMHFWLYHVRIWKHJOREDORQWRORJ\PXVWEHLGHQWLILHG 5HODWLRQVKLSV EHWZHHQ WKH JOREDO RQWRORJ\ DQG LI WKHUH DUH DQ\ RWKHU NQRZQ RQWRORJLHVLQWKHGRPDLQWKHQWKH\PXVWEHLGHQWLILHG ,I PRUH WKDQ RQH RQWRORJ\ LV LGHQWLILHG IRU WKH RUJDQL]DWLRQ WKHQ WKH VSHFLILF RQWRORJ\XVHGIRUHDFKUROHPXVWEHLGHQWLILHG 6HUYLFHV JLYHQ E\ WKH GLUHFWRU\ VHUYLFH SURYLGHU DQG WKH RQWRORJ\ VHUYLFH SURYLGHUUROHVGXULQJWKHLQWHUDFWLRQVPXVWEHVSHFLILHGDFFRUGLQJWRWKHUHODWHG ),3$VWDQGDUGV Design Phase

,Q 6$%32 GHVLJQ SKDVH D ),3$ EDVHG FRQFUHWH DUFKLWHFWXUH LV FRQVWUXFWHG E\ WUDQVIRUPLQJWKHDEVWUDFWHQWLWLHVLGHQWLILHGLQWKHDQDO\VLVSKDVHWRWKHFRQFUHWHRQHV XVLQJ),3$VSHFLILFDWLRQV'HVLJQSKDVHJHQHUDWHVWKUHHPRGHOVZKLFKDUHQDPHGDV WKHRQWRORJ\PRGHOWKHDJHQWPRGHODQGWKHGHWDLOHGLQWHUDFWLRQPRGHO Ontology Model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

6$%32$0XOWLDJHQW'HYHORSPHQW0HWKRGRORJ\

=DPERQHOOL-HQQLQJVDQG:RROGULGJH VVWXG\t VDSSURDFKDQGVWRUHGDVDSDUWRIWKH RQWRORJ\NQRZOHGJH Agent Model.7KLVPRGHOGHILQHVWKHDJHQWW\SHVDQGDVVLJQVWKHUROHVGHILQHGLQWKH DQDO\VLVSKDVHWRWKHDJHQWW\SHV7KLVPRGHODOVRGHILQHVWKHVHUYLFHVRIHDFKDJHQW DFFRUGLQJWRWKHUHVSRQVLELOLWLHVRIWKHUROHVDVVLJQHGWRWKHDJHQWV+HQFHWKLVPRGHO FRPELQHV WKH $JHQW 0RGHO DQG 6HUYLFH 0RGHO RI WKH *DLD PHWKRGRORJ\ u 2QWRORJ\ 6HUYLFH 3URYLGHUv  UROH DQG u 'LUHFWRU\ 6HUYLFH 3URYLGHUv  UROH DQG WKHLU GHILQHGVHUYLFHVDUHDVVLJQHGWRERWKWKHRQWRORJ\DJHQWDQGWKHGLUHFWRU\ IDFLOLWDWRU DJHQWZKLFKDUHWKHDJHQWVWKDWDUHXVHGIRUWKHVHUROHVLQ),3$VSHFLILFDWLRQV6LQFH HDFK DJHQW PXVW DGDSW LWVHOI WR WKH RQWRORJLHV XVHG LQ WKH RUJDQL]DWLRQ PDSSLQJ EHWZHHQ DJHQWt V LQWHUQDO NQRZOHGJH DQG V\VWHPt V RQWRORJLHV DUH GHILQHG IRU HDFK DJHQW7KLVPDSSLQJGLIIHUVIRUGLIIHUHQWW\SHVRIDJHQWVZLWKLQWKHRUJDQL]DWLRQ)RU H[DPSOHDJHQWVSOD\LQJWKHUROHRILQIRUPDWLRQSURYLGHUVQHHGDPDSSLQJEHWZHHQLWV ORFDONQRZOHGJHDQGWKHRQWRORJLHVWKH\VXSSRUW2Q WKH RWKHU KDQG DJHQWV SOD\LQJ WKHUROHRILQIRUPDWLRQUHTXHVWHUVMXVWQHHGSURSHUXVHULQWHUIDFHGHILQLWLRQVEDVHGRQ WKH RQWRORJLHV WKH\ VXSSRUW 7KHVH PDSSLQJV DUH GRFXPHQWHG XVLQJ WKH H[WHQVLEOH VW\OH VKHHW ODQJXDJH IRU WUDQVIRUPDWLRQV ;6/7  >@ VWDQGDUG DQG VWRUHG IRU HDFK DJHQW Detailed Interaction Model. 7KLV PRGHO PDSV WKH LQWHUDFWLRQ SDWWHUQV LGHQWLILHG LQ WKHDQDO\VLVSKDVHWRWKH),3$VSHFLILFDWLRQV)RUWKLVSXUSRVHHDFKPHVVDJHEHWZHHQ WKH DJHQWV LV UHZULWWHQ XVLQJ WKH ),3$ $&/ VSHFLILFDWLRQ 7KH FRQWHQW RI HDFK PHVVDJH LV DOVR ZULWWHQ LQ RQH RI WKH FRQWHQW ODQJXDJHV VXSSRUWHG E\ ),3$ VXFK DV ),3$5') $WWKHHQGRIWKHGHVLJQSKDVHDOORIWKHFRQFUHWHHQWLWLHVRIWKHRUJDQL]DWLRQDUH UHDG\ WR EH LPSOHPHQWHG XVLQJ DQ\ RI WKH ),3$FRPSOLDQW DJHQW IUDPHZRUNV VLQFH WKH\WRWDOO\FRQIRUPWKHVSHFLILFDWLRQVRIWKH),3$

3

Applying the SABPO Methodology

7RLOOXVWUDWHKRZ6$%32LVDSSOLHGWRHQJLQHHUDUHDOV\VWHPZHZLOOXVHDSDUWRI WKHUHDOLVWLF%%DSSOLFDWLRQDVDQH[DPSOH7KLVDSSOLFDWLRQLVXVHGE\DGHYHORSHU FRPSDQ\ ZKLFK JDWKHUV UHTXLUHG SDUWV IURP GLIIHUHQW VXSSOLHUV +HQFH WKH DSSOLFDWLRQVKRXOGVXSSRUWWKHIROORZLQJIXQFWLRQDOLWLHV   

,W VKRXOG VXSSRUW WKH GHYHORSHU FRPSDQ\ XVHUV WR GHILQH WKH GHVLUHG SDUWt V NQRZOHGJH ,WVKRXOGILQGWKHSRWHQWLDOVXSSOLHUVRIWKHGHVLUHGSDUWV ,WVKRXOGVHQGDUHTXHVWWRFROOHFWWKHLQIRUPDWLRQDERXWWKHGHVLUHGSDUWVWRWKH VXSSOLHUVDQGVKRZVWKHUHVXOWVWRWKHGHYHORSHUFRPSDQ\t VLQWHUHVWHGXVHUV %HFDXVHRIWKHVSDFHOLPLWDWLRQRQO\DQDO\VLVSKDVHRIWKH6$3%2LVLOOXVWUDWHG IRUWKHUHTXLUHPHQWVOLVWHGDERYH

Role Model.u 'HYHORSHU&RPSDQ\v UROHDQGu 6XSSOLHUv UROHFDQEHLGHQWLILHGHDVLO\ IURPWKHUHTXLUHPHQWV)ROORZLQJWKH*DLDPHWKRGRORJ\u 'HYHORSHU&RPSDQ\v UROH PXVW KDYH D u SDUW GHILQLWLRQv  DFWLYLW\ ZKLFK KHOSV WKH XVHUV WR GHILQH WKH GHVLUHG

2÷X]'LNHQHOOLDQG5L]D&HQN(UGXU

SDUWV$u UHTXHVWv SURWRFROPXVWDOVREHDGGHGWRWKHu 'HYHORSHU&RPSDQ\v UROHWR VHQGWKHSDUWUHTXHVWWRWKHSURSHUVXSSOLHUV2QWKHRWKHUKDQGu 6XSSOLHUv UROHPXVW KDYHDu JHWSDUWNQRZOHGJHv DFWLYLW\DQGDQu LQIRUPv SURWRFROWRUHWXUQGHVLUHGSDUW NQRZOHGJHWRWKHu 'HYHORSHU&RPSDQ\v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nteraction Model. ,QWHUDFWLRQ 0RGHO ,Q WKH LQWHUDFWLRQ PRGHO DOO SURWRFROV LGHQWLILHGLQWKHUROHPRGHODUHGHILQHGXVLQJWKH$JHQW80/>@$OVRWKHJHQHUDO LQWHUDFWLRQ PHFKDQLVP LV YLVXDOL]HG XVLQJ WKH FROODERUDWLRQ GLDJUDP RI WKH $JHQW 80/)LJVKRZVWKHFROODERUDWLRQGLDJUDPRIWKHUROH VLQWHUDFWLRQ 7KHLQIRUPDWLRQIORZRIWKHLQWHUDFWLRQLQILJLVJLYHQLQGHWDLOEHORZ    

'HYHORSHU &RPSDQ\ UHTXHVWV WKH SRVVLEOH VXSSOLHUV VXSSRUWLQJ WKH SDUW RQWRORJ\IURPWKH'LUHFWRU\6HUYLFH3URYLGHU 'LUHFWRU\ 6HUYLFH 3URYLGHU LQIRUPV WKH 'HYHORSHU &RPSDQ\ E\ UHWXUQLQJ WKHSRVVLEOHVXSSOLHUV 'HYHORSHU&RPSDQ\UHTXHVWVIURPWKHVXSSOLHUIRUWKHGHVLUHGSDUWXVLQJWKH SDUWRQWRORJ\ 6XSSOLHUVUHWXUQWKHUHTXHVWHGSDUWLQIRUPDWLRQWRWKH'HYHORSHU&RPSDQ\ ' LUHFWR U\ 6 HUY LFH 3 UR Y LG HU

 5 HT X HVW6 X S S OLHU VX S S R UWLQ J S DUWR Q WR OR J \

 ,Q IR UP 6 X S S OLHU 5 HT X HVWS DUWN QRZ OHGJH

' HY HOR S HU & R P S DQ\

6 X S S OLHU  ,Q IR UP S DUWN Q R Z OHG J H

Fig. 3.&ROODERUDWLRQGLDJUDPRIWKHUROH VLQWHUDFWLRQ

6$%32$0XOWLDJHQW'HYHORSPHQW0HWKRGRORJ\

' LUH F WR U\ 6 H UY LF H 3 UR Y LG H U  5 H T X H V W $ OO 6 X S S OLH UV

 5 H T X H V W 2 Q WR OR J \ 6 H U Y LF H 3 UR Y LG H UV

 ,Q IR UP  6 X S S OLH U V

' H Y H OR S H U & RP SDQ\

 ,Q IR UP 6 X S S OLH UV

 7 U D Q V OD WH

 ,Q IR UP   5 H T X H VW3 D UW . Q R Z OH G J H

2 Q WR OR J \ 6 H UY LF H 3 UR Y LG H U

 ,Q IR UP 3 D UW. Q R Z OH G J H

6 X S S OLH U

Fig. 4. &ROODERUDWLRQGLDJUDPRIWKHUROH VLQWHUDFWLRQIRUDGLIIHUHQWVLWXDWLRQ

1RZ ZH ZLOO FKDQJH WKH VLWXDWLRQ DQG LOOXVWUDWH KRZ 6$%32 KDQGOHV GLIIHUHQW FRQGLWLRQVXVLQJWKH),3$VSHFLILFDWLRQV/HWXVDVVXPHWKDWWKHUHLVDQRWKHURQWRORJ\ LQWKLVGRPDLQVXSSRUWHGE\VRPHVXSSOLHUVDQGFDOOWKLVRQWRORJ\DVSDUWRQWRORJ\ 'HYHORSHU&RPSDQ\VXSSRUWVRQO\WKHSDUWRQWRORJ\EXWLWDOVRNQRZVWKDWWKHUH DUH VRPH VXSSOLHUV DURXQG VXSSRUWLQJ GLIIHUHQW RQWRORJLHV 7R KDQGOH WKLV VLWXDWLRQ WKHUH PXVW EH D UHODWLRQVKLS EHWZHHQ WKHVH RQWRORJLHV DQG WKH 2QWRORJ\ 6HUYLFH 3URYLGHU VKRXOG EH DEOH WR SHUIRUP WKH WUDQVODWLRQ VHUYLFH EHWZHHQ DQ\ WZR RQWRORJLHV DV GHILQHG LQ WKH ),3$ 2QWRORJ\ 6HUYLFH 6SHFLILFDWLRQ >@ 7KH FROODERUDWLRQGLDJUDPWRKDQGOHWKHGHILQHGVLWXDWLRQLVVKRZQLQILJ ([SODQDWLRQRIWKHLQIRUPDWLRQIORZLVJLYHQEHORZ   

    

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

7KLV QHZ VLWXDWLRQ RI FRXUVH FUHDWHV QHZ DFWLYLWLHV IRU HDFK DJHQW DQG QHZ LQWHUDFWLRQSURWRFROVDVVKRZQLQILJ

2÷X]'LNHQHOOLDQG5L]D&HQN(UGXU

$V D VXPPDU\ 6$%32 ILUVW GHILQHV WKH UROHV RI WKH RUJDQL]DWLRQ DQG WKHLU UHVSRQVLELOLWLHV LQ WKH DQDO\VLV SKDVH 7KHVH LQGLYLGXDO UROHV DUH DFFHSWHG DV WKH DEVWUDFWDJHQWHOHPHQWRIWKH),3$ VDEVWUDFWDUFKLWHFWXUH7KHQLW IXOO\H[SORLWVWKH DEVWUDFW DUFKLWHFWXUH E\ WDNLQJ WKH GLUHFWRU\ DQG WKH RQWRORJ\ VHUYLFHV DV UROHV DQG GHILQLQJWKHLQWHUDFWLRQSURWRFROVEHWZHHQWKHUROHVWRVDWLVI\V\VWHP VUHTXLUHPHQWV

4

Ontology Inception Pattern

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

Motivation and Interaction Mechanism

7RGRFXPHQWWKH0$6LQWHUDFWLRQSDWWHUQVZHGHILQHWKHSDWWHUQIURPWKUHHGLIIHUHQW SHUVSHFWLYHV )LUVW PRWLYDWLRQ EHKLQG WKH SDWWHUQ LGHQWLILFDWLRQ LV LQWURGXFHG 7KHQ LQWHUDFWLRQ PHFKDQLVP RI WKH SDUWLFLSDWLQJ DJHQWV LV GHILQHG DFFRUGLQJ WR WKH ),3$ VWDQGDUGV)LQDOO\JXLGHOLQHVDUHGHILQHGWRPDNHWKHSDWWHUQ),3$FRPSOLDQW Motivation.7R XQGHUVWDQG WKH PRWLYDWLRQ EHKLQG WKH 2QWRORJ\ ,QFHSWLRQ SDWWHUQ OHWXVORRNDWWKHRSHQHQYLURQPHQWIURPWKHRQWRORJLFDOSHUVSHFWLYHDQGDVVXPHWKH IROORZLQJFKDUDFWHULVWLFV  

7KHUHDUHDODUJHQXPEHURIGRPDLQGHSHQGHQWRQWRORJLHVZKLFKPD\FKDQJH DWUXQWLPH $WDQ\WLPHQHZRQWRORJLHVPD\EHDGGHGWRWKH0$6$OVRVRPHH[LVWLQJ RQWRORJLHVPD\GHFLGHGQRWWREHXVHGDQ\PRUH

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

6$%32$0XOWLDJHQW'HYHORSPHQW0HWKRGRORJ\

' LU H F WR U\ 6 H U Y LF H 3 UR Y LG H U  5 H T X H V W

 , Q I R U P

2 Q WR OR J \ $ JHQW

 , Q I R U P  1 H Z 2 Q WR OR J \  7 UD Q V IH U

$ JHQW  5 H T X H V W2 Q WR OR J \ 7 U D Q V I H U

Fig. 5.,QWHUDFWLRQPHFKDQLVPRIRQWRORJ\LQFHSWLRQSDWWHUQ

Interaction Mechanism: ,QWHUDFWLRQ PHFKDQLVP GHILQHV WKH LQWHUDFWLRQ SURWRFRO EHWZHHQWKHDJHQWVWKDWFROODERUDWHWR VROYH WKH SUREOHP DW KDQG $OVR DFWLYLWLHV RI WKHSDUWLFLSDWLQJDJHQWVDUHLGHQWLILHGDQGPRGHOHGDVVHUYLFHVXVLQJ*DLDVW\OH7KH LQWHUDFWLRQPHFKDQLVPLVLOOXVWUDWHGLQILJXVLQJDFROODERUDWLRQGLDJUDP ,QIRUPDWLRQIORZEHWZHHQWKHSDUWLFLSDWLQJDJHQWLVGHVFULEHGEHORZ     

2QWRORJ\DJHQWUHTXHVWVWKHGLUHFWRU\IDFLOLWDWRUDJHQWWRILQGWKHDJHQWZLWKLQ WKHRUJDQL]DWLRQ 'LUHFWRU\IDFLOLWDWRUVHQGVWKHOLVWRIDFWLYHDJHQWVWRWKHRQWRORJ\DJHQW 2QWRORJ\DJHQWLQIRUPVHDFKDJHQWWKDWWKHUHLVDQHZRQWRORJ\LQWKHV\VWHP ,QWHUHVWHGDJHQWVVHQGDUHTXHVWWRWKHRQWRORJ\DJHQWIRURQWRORJ\WUDQVIHU 2QWRORJ\DJHQWWUDQVIHUVWKHRQWRORJ\WRWKHLQWHUHVWHGDJHQWV

,WFDQEHVHHQIURPWKHLQWHUDFWLRQVWKDWWKHDJHQWUROHQHHGVDQDFWLYLW\WRGHILQH WKHUHTXLUHGPDSSLQJVEHWZHHQWKHWUDQVIHUUHGRQWRORJ\DQGLWVORFDONQRZOHGJH7KLV DFWLYLW\ LV FDOOHG DV ORFDOL]H DQG GHILQHG DV WKH VHUYLFH RI WKH DJHQW UROH $OVR RQWRORJ\DJHQWQHHGVDQHZFRPPXQLFDWLRQDFWWRWUDQVIHUDVSHFLILFRQWRORJ\WRWKH LQWHUHVWHGDJHQW7KLV QHZFRPPXQLFDWLRQDFWLVQDPHGDV WUDQVIHU DQG GHILQHG LQ $JHQW80/ Guidelines for FIPA Comformance. *XLGHOLQHV IRU ),3$FRPSOLDQFH 7KH IROORZLQJ JXLGHOLQHVDUHGHILQHGWRFRQIRUPWKH2QWRORJ\,QFHSWLRQSDWWHUQ WR WKH ),3$VSHFLILFDWLRQ   

$ QHZ VHUYLFH VKRXOG EH GHILQHG IRU WKH RQWRORJ\ DJHQW WR WUDQVIHU WKH UHTXLUHGRQWRORJ\ ),3$0HWD2QWRORJ\VSHFLILFDWLRQVKRXOGEHH[WHQGHGZLWKUHTXLUHGVHPDQWLF WRVXSSRUWWKHWUDQVIHUVHUYLFH ),3$ VWDQGDUGV FRPPLWWHH VKRXOG GHFLGH HLWKHU WR GHILQH D QHZ FRPPXQLFDWLYH DFW IRU WUDQVIHU SURWRFRO RU XVLQJ LQIRUP DFW LQVHUWLQJ WKH UHTXLUHGNQRZOHGJHWRWKHFRQWHQWRIWKHPHVVDJH

2÷X]'LNHQHOOLDQG5L]D&HQN(UGXU

4.2

Run-Time Behaviour of the Ontology Inception Pattern

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

5

Conclusion

,Q WKLV SDSHU D QHZ PHWKRGRORJ\ ZKLFK XVHV RUJDQL]DWLRQDO PHWDSKRU DV WKH EDVLV DQGLQWHJUDWHVWKLVPHWDSKRUWRWKH),3$VWDQGDUGVLVLQWURGXFHG,QDGGLWLRQDQHZ LQWHUDFWLRQSDWWHUQZKLFKLVFDOOHGDVu RQWRORJ\LQFHSWLRQv LVGHILQHG7KHSURSRVHG PHWKRGRORJ\DQGWKHRQWRORJ\LQFHSWLRQSDWWHUQKDYHEHHQXVHGLQWKHGHYHORSPHQW RI D PXOWLDJHQW V\VWHP LQIUDVWUXFWXUH IRU VRIWZDUH FRPSRQHQW PDUNHWSODFH >@ ZKLFKKDVDOOWKHEDVLFWKHFKDUDFWHULVWLFVRIRSHQHQYLURQPHQWV

6$%32$0XOWLDJHQW'HYHORSPHQW0HWKRGRORJ\

References >@

>@ >@ >@

>@

>@

>@ >@

>@ >@ >@ >@ >@

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t  0DGULG6SDLQ  %HUQRQ&*OHL]HV03H\UXTXHRX6DQG3LFDUG*$'(/)(DPHWKRGRO RJ\IRUDGDSWLYHPXOWLDJHQWV\VWHPV,Q3HWWD37RONVGRUI5=DPERQHOOL )DQG2VVRZVNL6HGV :RUNVRS1RWHVRIWKH7KLUG,QWHUQDWLRQDO:RUNVKRS (QJLQHHULQJ6RFLHWLHVLQWKH$JHQWV:RUOG(6$:t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

2÷X]'LNHQHOOLDQG5L]D&HQN(UGXU

>@ >@ >@

>@

>@ >@ >@ >@

>@

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t  0DGULG6SDLQ  =DPERQHOOL ) -HQQLQJV 1 DQG :RROGULGJH 0 2UJDQLVDWLRQDO UXOHV DV DQ DEVWUDFWLRQIRUWKHDQDO\VLVDQGGHVLJQRIPXOWLDJHQWV\VWHPV,QW-RXUQDORQ 6RIWZDUH(QJLQHHULQJDQG.QRZOHGJH(QJLQHHULQJ  

Modelling a Multi-agent System Environment 6HKO0HOORXOL*X\:0LQHDXDQG'DQLHO3DVFRW /DYDO8QLYHUVLW\&RPSXWHU6FLHQFH'HSDUWPHQW *934XHEHF&DQDGD {sehl.mellouli,guy.mineau}@ift.ulaval.ca [email protected] http://www.ift.ulaval.ca/~lic

Abstract. $ PXOWLDJHQW 6\VWHP 0$6  FDQ EH YLHZHG DV D VRIWZDUH V\VWHP HYROYLQJ LQ VRPH HQYLURQPHQW ZLWK ZKLFK LW KDV WR LQWHUDFW 'XULQJ WKH 0$6 OLIH F\FOH PDQ\ VLWXDWLRQV PD\ RFFXU VLWXDWLRQV WKDW DUHFRQVLGHUHGWREHFULWLFDODQGWKDWZRXOGKDYHDQHIIHFWRQWKH0$6 SURPSWLQJLWWRTXLFNO\DGMXVWWRWKLVQHZVLWXDWLRQ$GHVLJQPHWKRGRO RJ\RID0$6VKRXOGKHOSWKHGHVLJQHUWRUHSUHVHQWLQIRUPDWLRQDERXWD FKDQJLQJ HQYLURQPHQW DQG LWV HIIHFWV RQ WKH 0$6 DQ DVSHFW RI WKH PRGHOOLQJWDVNZKLFKLVFXUUHQWO\ODFNLQJIURPDJHQWGHVLJQPHWKRGROR JLHV :H SURSRVH WR DGG  WZR QHZ GLDJUDPV DQ HQYLURQPHQW GLDJUDP DQG DQ DJHQW GLDJUDP WR 0$6 PRGHOOLQJ PHWKRGRORJLHV 7KHVH GLD JUDPVZLOOFRQFHSWXDOLVHWKHLPSDFWRIWKHHQYLURQPHQWRQWKHVWUXFWXUH RIWKH0$6DQGWKHUHIRUHVKRXOGJXLGHWKHGHYHORSPHQWRIWKHDFWXDO LPSOHPHQWDWLRQRIWKH0$6

1

Introduction

$PXOWLDJHQWV\VWHP0$6 FRQVLVWVRIDQXPEHURIDJHQWVWKDWRSHUDWHWRJHWKHULQ RUGHUWRDFKLHYHVRPHJRDOV ,W FDQ EH YLHZHG DV DQ agent RUJDQL]DWLRQ E\ DQDORJ\ ZLWKhuman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machineorganismbrainDQGpoliticalPHWDSKRUV%DVHGRQWKHVHPHWD SKRUVZHKDYHLGHQWLILHGILYHNH\GLPHQVLRQVWRGHVLJQDKXPDQRUJDQL]DWLRQWKDWZH DGDSWWRWKHGHVLJQRIDQDJHQWRUJDQL]DWLRQ>@7KHVHGLPHQVLRQVDUHWKHQDWXUHRI WKHHQYLURQPHQWWKHWDVNVWREHSHUIRUPHGWKHFRQWUROFRPPXQLFDWLRQDQGFROODER UDWLRQUHODWLRQVKLSVEHWZHHQDJHQWV 33HWWDHWDO(GV (6$:/1$,SS ©6SULQJHU9HUODJ%HUOLQ+HLGHOEHUJ

6HKO0HOORXOLHWDO

/RRNLQJDWDJHQWRULHQWHGPHWKRGRORJLHVVXFKDV*DLD>@$80/>@DQG0$6 &RPPRQ.$'6 >@ ZH ILQG WKDW WKH\ GR QRW LQWHJUDWH DQ\ PHFKDQLVP WR H[SOLFLWO\ PRGHOWKHHQYLURQPHQWLWVHYROXWLRQDQGLWVLPSDFWRQWKHVWUXFWXUHRIWKH0$6>@ 7KH\KDYHQRIRUPDOUHSUHVHQWDWLRQRIWKHVSHFLILFDWLRQVRIWKH0$6DQGWKHUHIRUH WKHUHLVQRZD\WRLPSOHPHQWVRPHHDUO\YDOLGDWLRQRIWKHEHKDYLRXURIWKH0$6XQGHU GHVLJQ :HGHILQHWKHHQYLURQPHQWDVDVHWRIVLWXDWLRQVWKDWFDQRFFXUGXULQJWKHRUJDQL ]DWLRQ OLIH F\FOH DQG ZKHUH HDFK VLWXDWLRQ KDV DQ LPSDFW RQ WKH VWUXFWXUH RI WKH RU JDQL]DWLRQ DQG RQ LWV PHDQLQJIXO EHKDYLRXU :H IHHO WKDW WKH PRGHOOLQJ RI WKH HQYL URQPHQWDQGLWVLPSDFWRQWKHRUJDQL]DWLRQVWUXFWXUHRIWKH0$6LVYLWDOWRWKHXVHIXO QHVVRIWKHV\VWHPVLQFHLWFKDOOHQJHVWKHDGDSWDELOLW\RIWKHV\VWHP :H SURSRVH WR DGG WZR GLDJUDPV WR 0$6 PRGHOOLQJ PHWKRGRORJLHV LQ RUGHU WR FRQFHSWXDOLVH DQG UHSUHVHQW WKH HQYLURQPHQW DQG LWV LPSDFW RQ WKH RUJDQL]DWLRQDO VWUXFWXUH RI WKH 0$6 7KHVH GLDJUDPV DUH WKH HQYLURQPHQW GLDJUDP DQG WKH DJHQW GLDJUDP 7KHHQYLURQPHQWGLDJUDP,WUHSUHVHQWVDVWDWHWUDQVLWLRQGLDJUDPEHWZHHQWKHGLIIHU HQW VLWXDWLRQV WKDW WKH HQYLURQPHQW FDQ XQGHUJR $ FKDQJH LQ D VLWXDWLRQ ZLOO DIIHFW PRVW SUREDEO\ WKH UROHV WKDW WKH DJHQWV SOD\ DQG WKHLU UHODWLRQVKLSV 7KLV GLDJUDP LV QHHGHG LQ WKH PRGHOOLQJ RI WKLV GHSHQGHQF\ EHWZHHQ HQYLURQPHQWDO VLWXDWLRQV DQG D 0$6:HGHILQHORRVHO\WKHHQYLURQPHQW DV EHLQJ WKH VHW RI WKH GLIIHUHQW VLWXDWLRQV WKDWFRQGLWLRQWKHEHKDYLRXURIWKHV\VWHPDQGWKHSRVVLEOHVHTXHQFHRIRFFXUUHQFHV RI WKHVH VLWXDWLRQV 7KURXJK GRPDLQ DQDO\VLV DQG EDFNJURXQG NQRZOHGJH ZH GHWHU PLQHZKDWVLWXDWLRQVDUHOLNHO\WRKDYHDWUHPHQGRXVLPSDFWRQWKHRUJDQL]DWLRQRIWKH 0$6DQGIRUREYLRXVFRPSOH[LW\UHDVRQVZHUHVWULFWRXUVHOYHVWRWKHP 7KHDJHQWGLDJUDP'HILQHGLQUHODWLRQWRWKHHQYLURQPHQWGLDJUDPLWUHSUHVHQWVWKH VWUXFWXUHRIWKHDJHQWVUROHV DQGWKHLUUHODWLRQVKLSVFRQWUROFROODERUDWLRQDQGFRP PXQLFDWLRQUHODWLRQVKLSV DQGKRZWKHHQYLURQPHQWDOVLWXDWLRQVDIIHFWWKHP 7KLVSDSHULVRUJDQL]HGDVIROORZV,QVHFWLRQZHSUHVHQWWKHOLPLWVRIWKH*DLD >@ $80/ >@ DQG 0$6&RPPRQ.$'6 >@ PHWKRGRORJLHV ZLWK UHJDUG WR WKH PRGHOOLQJRIWKHHQYLURQPHQW6HFWLRQSUHVHQWVWKHHQYLURQPHQWGLDJUDP DQG 6HF WLRQ  SUHVHQWV WKH DJHQW GLDJUDP EDVHG RQ VRPH VRFFHU JDPH H[DPSOH 6HFWLRQ  FRQFOXGHV

2

Limits of Agent Oriented Methodologies

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

0RGHOOLQJD0XOWL$JHQW6\VWHP(QYLURQPHQW

WLRQ DVEHLQJLPSRUWDQWWRWKHPRGHOOLQJRIDQ\RUJDQL]DWLRQDUHWDNHQLQWRDFFRXQWE\ WKHVHPHWKRGRORJLHV)URPWKDWSUHVHQWDWLRQZHDLPDWVKRZLQJWKDWQRSURYLVLRQZDV PDGHWRH[SOLFLWO\PRGHOWKHHQYLURQPHQWXQGHUWKHVHWKUHHPHWKRGRORJLHV 2.1

The Gaia Methodology

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t V QHW 7KLV FRQVWLWXWHV WKH VDIHW\ UH VSRQVLELOLWLHV

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

6HKO0HOORXOLHWDO Table 1.6FKHPDIRUUROH'HIHQGHU

5ROHVFKHPD 'HVFULSWLRQ

'HIHQGHU 6WRSVDGYHUVHIRUZDUGSOD\HUV

3URWRFROVDQG$FWLYLW\

VWRS3OD\HUVJXLGH'HIHQVH

3HUPLVVLRQV

NLFN%DOO

5HVSRQVLELOLWLHV /LYHQHVV 6DIHW\ 2.2

VWRS3OD\HUV JXLGH'HIHQVH 1RDGYHUVHSOD\HUVFRUHV

AUML

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efender

Stops adverse striker

stopPlayers guideDefence

Fig. 1.$SDUWRIWKHDJHQWFODVVDVGHILQHGLQ$80/

0RGHOOLQJD0XOWL$JHQW6\VWHP(QYLURQPHQW

Defender isDefender adverseStrikerPos.

Fig. 2.0$6&RPPRQ.$'6FHQWUDOGHIHQGHUDJHQWFODVV

2.3

MAS-CommonKADS

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

6HKO0HOORXOLHWDO

DJHQWV DUH DOVR UHSUHVHQWHG :H ZLOO DOVR SURSRVH WR DGG D FRQWURO UHODWLRQ EHWZHHQ DJHQWVLQWKHDJHQWGLDJUDPVLQFHWKHHQYLURQPHQWGLDJUDPPD\LPSDFWRQLW

3

The Environment Diagram

$0XOWL$JHQW6\VWHP0$6 LVGHVLJQHGWRDFKLHYHVSHFLILFJRDOV$0$6LVSODFHG LQDQHQYLURQPHQWZLWKZKLFKLWLQWHUDFWV7KH0$6PXVWGHFLGHZKDWWRGRDQGGH YHORSDVWUDWHJ\LQRUGHUWRDFKLHYHLWVDVVLJQHGJRDOV)RUWKLVWKH0$6PXVWKDYHD UHSUHVHQWDWLRQ RU D PRGHO RI WKH HQYLURQPHQW LQ ZKLFK LW HYROYHV :H QRZ DVN WKH TXHVWLRQhow can the environment be modelled? $ SDUW RI WKLV DQVZHU LV JLYHQ E\ 0F&DUWK\ LQ LWV VLWXDWLRQ FDOFXOXV >@ )RU 0F&DUWK\ WKH HQYLURQPHQW RU ZRUOG  LV FRPSRVHG RI VLWXDWLRQV $ VLWXDWLRQ LV WKH FRPSOHWHVWDWHRIWKHZRUOGDWDQLQVWDQWRIWLPHDVQDSVKRWRIWKHZRUOGDWVRPHLQ VWDQW (DFK VLWXDWLRQ FDQ EH IROORZHG E\ DQ HYHQW WKDW OHDGV WR D QHZ VLWXDWLRQ 7KH VLWXDWLRQ FDOFXOXV DOORZV WR LQIHU VRPHWKLQJ DERXW IXWXUH VLWXDWLRQV /HW XV GHILQH Vt  UHVXOWHV ZKHUHVt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critical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⊆3M,IDOVRWKHYDOXH YMRIDQHOHPHQWSM LVHTXDOWKHYDOXHYLRIWKHHOHPHQWSMLQ3L3L⊆3M ZHZLOOVD\ WKDWVLLVDJHQHUDOL]DWLRQRIVMDQG VM LV D VSHFLDOL]DWLRQ RI VL :H FRQFOXGH WKDW LI D VLWXDWLRQVNOHDGVWRVLWXDWLRQVMVRLWOHDGVWRVLWXDWLRQVLDQGLIWKHVLWXDWLRQVLOHDGVWR DVLWXDWLRQVNWKHQWKHVLWXDWLRQVMOHDGVWRWKHVLWXDWLRQVN )RUPDOO\WKHHQYLURQPHQWVWUXFWXUH6LVGHILQHGDV6 ,:(5)!ZKHUH,LV WKHVHW RI LQLWLDO VLWXDWLRQV : LV WKH VHW RI DOO VLWXDWLRQV , ⊆ :  ( LV WKH VHW RI DOO SRVVLEOHHYHQWV5LVDUHODWLRQWKDWUHSUHVHQWVWUDQVLWLRQVEHWZHHQVHWRIVLWXDWLRQVZLWK  :HFDQH[WHQGWKHIRUPDOLVPIRUPXOWLYDOXHGSDUDPHWHU

0RGHOOLQJD0XOWL$JHQW6\VWHP(QYLURQPHQW

5⊆℘: [([℘: ZKHUH℘: LVWKHSRZHUVHWRI:5LVDWULSOHW6LH6M ZKHUH6LDQG6MDUHLQ℘: DQGHLVLQ(7KHWULSOHW6LH6M PHDQVWKDWIURPDQ\ HOHPHQWRI6LLIWKHHYHQWHRFFXUVWKHV\VWHPZLOOVXEVHTXHQWO\EHLQVLWXDWLRQ6M$OO WKHHOHPHQWVRI6LUHVSHFWLYHO\6MDUHUHODWHGWRHDFKRWKHUE\WKHUHODWLRQRIJHQHUDOL ]DWLRQVSHFLDOL]DWLRQDVGHILQHGDERYH7KHVHW)LVWKHVHWRIILQDOVLWXDWLRQVWKDWFDQ EHUHDFKHG)⊆: 7KHUHDUHQRSRVVLEOHVLWXDWLRQVDIWHUWKHHOHPHQWVRI)LHQR WULSOHWRI5KDVWKHVWUXFWXUHIHIt IRUZKDWHYHUHLQ(DQGIt LQ:7KHVLWXDWLRQVDQG WKHLUWUDQVLWLRQVGHILQHWKHHQYLURQPHQWGLDJUDP>@ZKLFKLVDVWDWHVLWXDWLRQ WUDQVL WLRQGLDJUDP,WLVDJUDSKLFDOUHSUHVHQWDWLRQRI5VXFKWKDWLWLVDJUDSKZKHUHQRGHV UHSUHVHQW HOHPHQWV LQ : DQG ZKHUH DQ DUF H EHWZHHQ WZR VLWXDWLRQV VL DQG VM LQ : H[LVWVLIWKHUHH[LVWVWZRVHWV6∈℘: DQG6∈℘: ZKHUHVL∈6VM∈6DQG 6H6!∈5 ,QWKHVRFFHUH[DPSOHZHKDYHLGHQWLILHGWKUHHSDUDPHWHUVIRUVLPSOLFLW\UHDVRQV ZKLFKWRJHWKHUGHILQHVLWXDWLRQVWKDWPD\LPSDFWWKHUROHVRIDJHQWV7KHVHSDUDPH WHUV DUH WKH 6FRUH 'LIIHUHQFH 6'  WKH 5HPDLQLQJ 7LPH 57  DQG WKH 'HSOR\PHQW 7DFWL&7& ,I6' vIRUv∈1 LWPHDQVWKDWWKHWHDPLVOHDGLQJE\vJRDOV,I 6' vWKHQLWLVORRVLQJE\vJRDOV7KH57SDUDPHWHUWDNHVYDOXHVRIWKHIRUPRSW ZKHUHW∈1 DQGRS∈^ ”•`)LQDOO\7&LVGHILQHGDV QQQVXFKWKDW QQQ QXPEHURISOD\HUVRQWKHILHOGIRUWKHVDPHWHDP ZKHUHQLVWKHQXPEHU RIGHIHQGHUVQLVQXPEHURIPLGILHOGHUVDQGQLVWKHQXPEHURIVWULNHUV $Q HQYLURQPHQW GLDJUDP VXFK DV ZKDW ZH SURSRVH LH D VLWXDWLRQ WUDQVLWLRQ GLDJUDP ZRXOG KHOS WR IRUPDOO\ UHSUHVHQW WKH FKDUDFWHULVWLFV RI WKH HQYLURQPHQW ,Q HDFKVLWXDWLRQZHJLYHDQDPHWKDWLGHQWLILHVWKHVLWXDWLRQDQGYDOXHVWRWKHGLIIHUHQW SDUDPHWHUV FKDUDFWHUL]LQJ LW )RU WKH VRFFHU H[DPSOH ZH UHVWULFW RXU VWXG\ WR RQO\ HLJKWVLWXDWLRQVVHH)LJXUH $PRQJWKHVHVLWXDWLRQVWKHUHDUHVSHFLDORQHVVVV VDQGVVLVWKHLQLWLDOVLWXDWLRQLGHQWLILHGE\DEROGUHFWDQJOH,WLVWKHILUVWVLWXDWLRQ WKDWFRXOGRFFXUDQGUHSUHVHQWVWKHHQYLURQPHQWZKHQWKHJDPHVWDUWVVVVDQGV DUHILQDOVLWXDWLRQVWKHUHDUHQRSRVVLEOHVLWXDWLRQVDIWHUWKHP (DFKVLWXDWLRQLVGHVFULEHGE\WKHWKUHHSDUDPHWHUV6'57DQG7&)RUH[DPSOH WKHVLWXDWLRQVUHSUHVHQWVDVLWXDWLRQZKHUHWKHVFRUHGLIIHUHQFH6'  LVHTXDOWR ZKHUHWKHUHUHPDLQPLQXWHVRISOD\WLPH57  DQGZKHUHWKHDGRSWHGWDFWLFLV 7&   $FFRUGLQJWRWKHGHILQLWLRQRIWKHHQYLURQPHQW,:(5)!WKHVHW, ^V` :  ^V V V V V V V V` (  ^ER[BJRDO VFRUHVBJRDO WLPHBB LQ MXUHGBSOD\HU UHFRYHUHGBSOD\HU  5  ^^V` ER[BJRDO ^V`  ^V` VFRUHVBJRDO ^V` ^V`WLPHBB^V` ^V`ER[BJRDO^V` ^V`VFRUHVBJRDO^V` ^V` LQMXUHGBSOD\HU^V` ^V`UHFRYHUHGBSOD\HU^V` `DQG) ^VVVV` 2QFHWKHVLWXDWLRQVDUHLGHQWLILHGLWLVLQWHUHVWLQJWRVHHWKHLULPSDFWRQWKHVWUXF WXUHRIWKH0$6PRUH VSHFLILFDOO\ RQ WKH UROHV SOD\HG E\ WKH LQGLYLGXDO DJHQWV DQG WKHLUUHODWLRQVKLSV)RUWKLVSXUSRVHZHQHHGWRXVHDQDJHQWGLDJUDP>@WRUHSUHVHQW DJHQWVDQGWRGHVFULEHWKHLUUROHVDQGUHODWLRQVKLSV7KLVDJHQWGLDJUDPLVGUDZQHDUO\ RQ ZKHQ GHVLJQLQJ D 0$6 7KLV NLQG RI DJHQW GLDJUDP LV QRW GHILQHG LQ DQ\ RI WKH PHQWLRQHGPHWKRGRORJLHV*DLD$80/ DQG 0$6&RPPRQ.$'6 7KLV LV ZKDW ZH SUHVHQWLQWKHQH[WVHFWLRQ

6HKO0HOORXOLHWDO S1

S0 SD = 0 RT = 90 TC = 4-4-2

SD = -1 RT < 25 TC = 4-3-3

box_goal

time_= 1 RT < 25 TC = 4-3-3 box_goal

scores_goal

SD >= 1 RT < 25 TC = 4-2-4 injured_player

S5

SD = 0 RT < 25 TC = 4-4-2

S7

S6 SD >= 1 RT < 25 TC = 3-5-2

SD >= 1 recovered_player RT < 25 TC = 3-5-1

Fig. 3.7KHHQYLURQPHQWGLDJUDPRIDVRFFHUJDPH

4

The Agent Diagram

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ituation name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

)RUQRZZHZLOOFRQVLGHUFROODERUDWLRQDVDNLQGRIFRPPXQLFDWLRQ

0RGHOOLQJD0XOWL$JHQW6\VWHP(QYLURQPHQW

Agent name = A1

Situation s0

Role = Goal keeper Control Communication

Agent name = A3 Role = Defender

Agent name = A2

Agent name = A4

Agent name = A5

Role = Defender

Role = Defender

Role = Defender

Agent name = A8 Role = Midfielder

Agent name = A9

Agent name = A7 Role = Midfielder

Agent name = A6

Role = Midfielder

Role = Midfielder

Agent name = A10

Agent name = A11

Role = Striker

Role = Striker

Fig. 4.7KHDJHQWGLDJUDP'DVVRFLDWHGZLWKVLWXDWLRQV

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`ZKHUH$LVWKH VHWRIDJHQWLQVWDQFHV&LVWKHVHWRIUHODWLRQVKLSVWKDWFDQEHGHILQHGEHWZHHQDJHQWV DQG5DLVDUHODWLRQWKDWGHWHUPLQHVWKHUHODWLRQVKLSVEHWZHHQDJHQWV5D⊆$[&[$ (DFKDJHQWGLDJUDPLVUHODWHGWRDSDUWLFXODUVLWXDWLRQLQWKHHQYLURQPHQWGLDJUDP:H GHILQH'DVDVHWRIYDULRXVDJHQWGLDJUDPV'DVR' ^'D`:HGHILQHDUHODWLRQ5VD⊆ :['WKDWUHODWHVDVLWXDWLRQ:LQWKHHQYLURQPHQWGLDJUDPWRVRPHDJHQWGLDJUDP ,QWKHVRFFHUJDPHH[DPSOHDFFRUGLQJWR)LJXUHV' EHORQJVWR5VD 7KHDJHQWGLDJUDP'LVGHILQHGDV' ^$&5`ZKHUH$ ^$$$$ $ $ $` &  ^FRQWURO FRPPXQLFDWLRQ` DQG 5D  ^$ FRQWURO $  $ FRPPXQLFDWLRQ$ $FRPPXQLFDWLRQ$ $FRQWURO$ $FRPPXQLFD WLRQ $  $ FRPPXQLFDWLRQ $  $ FRQWURO $  $ FRPPXQLFDWLRQ $  $ FRPPXQLFDWLRQ $  $ FRQWURO $  $ FRPPXQLFDWLRQ $  $ FRPPXQLFDWLRQ$ $FRQWURO$ $FRPPXQLFDWLRQ$ $FRPPXQLFD

6HKO0HOORXOLHWDO

WLRQ$ $FRQWURO$ $FRPPXQLFDWLRQ$ $FRPPXQLFDWLRQ$  $ FRPPXQLFDWLRQ $  $ FRPPXQLFDWLRQ $  $ FRPPXQLFDWLRQ $  $ FRPPXQLFDWLRQ $  $ FRPPXQLFDWLRQ $  $ FRPPXQLFDWLRQ $ `

5

Applications

7KHHQYLURQPHQWDQGDJHQWGLDJUDPVDUHXVHGWRGHVLJQPXOWLDJHQWV\VWHPVZKHUHWKH HQYLURQPHQWLV G\QDPLF HYROYHV RYHU WLPH  DQG ZKHUH WKH RUJDQL]DWLRQ VWUXFWXUH RI WKH0$6 HYROYHV RYHU WLPH ZKHUH WKH UROHV RI DJHQWV DQG WKHLU UHODWLRQV FKDQJH DF FRUGLQJWRWKHHQYLURQPHQWHYROXWLRQ7\SLFDOGRPDLQDSSOLFDWLRQVFRXOGEHFRPSXWHU JDPHV PLOLWDU\ DSSOLFDWLRQV RU ,QWHUQHW DSSOLFDWLRQV ,Q ZKDW IROORZV ZH JLYH DQ H[DPSOHRIDPLOLWDU\DSSOLFDWLRQ 5.1

The Military Application

:HSUHVHQWWKHH[DPSOHRIDJURXSRIILYH)ILJKWHUVSLORWHGE\DXWRPDWHGVRIWZDUH DJHQWVWKDWKDYHWRDWWDFNDQGGHVWUR\DQXFOHDUSODQW7KLVLVDKLJKO\XQSUHGLFWDEOH HQYLURQPHQW VLQFH PDQ\ WURXEOHVRPH VLWXDWLRQV PD\ RFFXU VXFK DV DQ LQWHUFHSWLRQ DWWHPSWIURPHQHP\ILJKWHUVRUIURPWKHHQHP\t VDLUGHIHQVHV\VWHP)RUHDFKVLWXD WLRQWKHILJKWHUVPXVWUHRUJDQL]HWKHPVHOYHVDVWKH\PD\KDYHWRUHJURXSLQVPDOOHU VTXDGURQVWRHQJDJHWKHHQHP\7KH0$6JXLGLQJWKHDWWDFNPXVWTXLFNO\UHDFWWRDQ\ QHZVLWXDWLRQDQGWU\WRUHDFKLWVWDUJHWDVPXFKDVSRVVLEOHRUUHWUHDW7KHUHLVPXFK LQIRUPDWLRQWKDWPXVWEHVKDUHGDWDOOWLPHVEHWZHHQWKHILJKWHUVRIWKHVTXDGURQ,WLV LPSRUWDQWWRLGHQWLI\WKLVLQIRUPDWLRQDQGGHVFULEHYDULRXVVWDWHVRIWKHHQYLURQPHQW WKDWVKRXOGKDYHDQLPSDFWRQWKH0$6EHKDYLRU7KHXVHIXOLQIRUPDWLRQWRWKHILJKW HUVLQWKHLUYDULRXVWDVNVDUHDPRQJRWKHUV   

7KH SRVLWLRQ RI WKH WDUJHW HDFK DJHQW PXVW NQRZ KRZ IDU WKH SODQW LV DQG ZKHUHLWLVORFDWHGDFFRUGLQJWRWKHSRVLWLRQRIWKHJURXS 3RVLWLRQZLWKLQWKHJURXSHDFKILJKWHUKDVDSRVLWLRQZLWKLQWKHVTXDGURQWKDW GHSHQGVRQWKHDFWXDOIRUPDWLRQ 7KH SRVLWLRQ RI HQHP\ ILJKWHUV NQRZLQJ WKLV SRVLWLRQ WKH 0$6 FDQ UH RUJDQL]H LWVHOI WR DYRLG DQ DWWDFN DQG WR GHIHQG LWV XQLW\ &RXQWHUDWWDFN FRXOGDOVREHRUJDQL]HG

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

0RGHOOLQJD0XOWL$JHQW6\VWHP(QYLURQPHQW

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gent name = A1

Situation s2

Role = Goal keeper Control Agent name = A3

Communication

Role = Defender

Agent name = A2

Agent name = A4

Agent name = A5

Role = Defender

Role = Defender

Role = Defender

Agent name = A8

Agent name = A9

Role = Midfielder

Role = Midfielder

Agent name = A7 Role = Midfielder

Agent name = A10

Agent name = A6

Agent name = A11

Role = Striker

Role = Midfielder

Role = Striker

Fig. 5.7KHDJHQWGLDJUDP'DVVRFLDWHGZLWKVLWXDWLRQVWKHUROHRI$KDVFKDQJHGDQGVR KDYHWKHUHODWLRQVKLSVZLWKWKHQHLJKERXULQJDJHQWV

S0 Position of the target plant : 25∞north Position within the group: as planned Position of enemy fighters: nothing

(YHQW %RPEDUGPHQW

S1 Position of the target plant : 25∞north Position within the group: the head fighter is missing Position of enemy fighters: nothing S2

Position of the target plant : 25∞north Position within the group: head fighter under attack Position of enemy fighters: 2 enemy fighters attacking from behind

(YHQW (QHP\ ILJKWHUVDWWDFNLQJ

Fig. 6. (QYLURQPHQWGLDJUDPRIWKHDLUSODQHDWWDFN

6HKO0HOORXOLHWDO

)RUH[DPSOHDSDUWRIWKHDJHQWGLDJUDPIRUWKHILUVWHQYLURQPHQWVWDWHRIRXUH[ DPSOHLVJLYHQLQ)LJXUH Situation s0 $JHQWQDPHILJKWHU 5ROHKHDGSLORWILJKWHU

$JHQWQDPHILJKWHU 5ROH$X[LODU\KHDGSLORWILJKWHU 7DVNV &ROOHFWLQIRUPDWLRQ $VVLVWLQSODQQLQJ $WWDFN

7DVNV 'HFLGH 3ODQ $WWDFN

Fig. 7.3DUWRIWKHDJHQWGLDJUDPDWVWDWHV

$SDUWRIWKHDJHQWGLDJUDPIRUWKHVHFRQGVWDWHLVVKRZQLQ)LJXUH Situation s1 $JHQWQDPHILJKWHU 5ROHKHDGSLORWILJKWHU 7DVNV 'HFLGH 3ODQ $WWDFN

$JHQWQDPHILJKWHU 5ROH$X[LODU\KHDGSLORWILJKWHU 7DVNV &ROOHFWLQIRUPDWLRQ $VVLVWLQSODQQLQJ $WWDFN

Fig. 8.3DUWRIWKHDJHQWGLDJUDPDIWHUKHDGSLORWLVVKRWGRZQ

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

6

Conclusion and Future Work

%DVHGRQRUJDQL]DWLRQWKHRULHVZHKDYHLGHQWLILHGILYHNH\GLPHQVLRQVWKDWPXVWEH FRQVLGHUHG ZKHQ PRGHOOLQJ VRPH RUJDQL]DWLRQ HLWKHU KXPDQ RU VRIWZDUHEDVHG WKH HQYLURQPHQWWKHQDWXUHRIWKHWDVNVWREHSHUIRUPHGWKHFRPPXQLFDWLRQFRQWURODQG FROODERUDWLRQ OLQNV EHWZHHQ WKH GLIIHUHQW HQWLWLHV DJHQWV  RI WKH V\VWHP 7KURXJK DQ DQDO\VLVRIWKHWKUHHPRVWSRSXODU0$6PRGHOOLQJPHWKRGRORJLHV*DLD$80/DQG 0$6&RPPRQ.$'6ZHFRXOGSLQSRLQWWKDWDOORIWKHPGRQRWSURYLGHGLDJUDPVIRU

0RGHOOLQJD0XOWL$JHQW6\VWHP(QYLURQPHQW

WKHH[SOLFLWPRGHOOLQJRIWKHHQYLURQPHQWLQZKLFKWKH0$6ZLOOHYROYH,IWKLVHQYL URQPHQWLVG\QDPLFWKH0$6PD\KDYHVRPHLQVWDEOHEHKDYLRXU 7KHUHIRUH ZH SURSRVH WZR GLDJUDPV WR DGG WR DJHQW RULHQWHG PRGHOOLQJ PHWKRG RORJLHVWKHenvironment diagramDQGWKHagent diagram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

References >@ >@

>@

>@

>@

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t :RUNVKRSRQDJHQW7KHRULHV$UFKLWHFWXUHVDQG/DQ JXDJHV3URYLGHQFH86$-XO\ 0F&DUWK\ - 6LWXDWLRQ &DOFXOXV :LWK &RQFXUUHQW (YHQWV DQG 1DUUDWLYH KWWSZZZIRUPDOVWDQIRUGHGXMPFQDUUDWLYHQDUUDWLYHKWPO

6HKO0HOORXOLHWDO

>@

>@ >@

>@ >@

>@

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

Organizational model Legend: role agent

Social model

Interaction model

structural interaction actual interaction (contract) Fig. 1. 2UJDQL]DWLRQDOIUDPHZRUNIRUDJHQWVRFLHWLHV

+DQV:HLJDQGHWDO

$FWXDO LQWHUDFWLRQV EHWZHHQ DJHQWV DUH VSHFLILHG LQ WKH LQWHUDFWLRQ PRGHO 7KHVH LQWHUDFWLRQVFDQIROORZIL[HGSURWRFROVHJWKHFRQWUDFWQHWSURWRFRO EXWFDQEHWKH UHVXOW RI FRPPXQLFDWLRQ EHWZHHQ DJHQWV LQ ZKLFK WKH LQWHUDFWLRQ IRUPV UHVXOWV DQG GXWLHVIRUDOOSDUWLHVKDYHEHHQQHJRWLDWHG7KHW\SHRILQWHUDFWLRQDQGWKHIUHHGRPIRU DJHQWV WR QHJRWLDWH VSHFLILF LQWHUDFWLRQ IRUPV GHSHQGV RI WKH VRFLHW\ PRGHO ,Q QHWZRUN PRGHOVLWLVSRVVLEOHWRFUHDWHQHZ LQWHUDFWLRQV EXW PDUNHW DQG KLHUDUFKLFDO VRFLHWLHVSURYLGHDWOHDVWVRPHIL[HGIRUPV ,Q RUGHU WR GHYHORS D FRPPRQ XQGHUVWDQGLQJ DQG D IHHOLQJ RI WKH SRVVLEOH DSSOLFDWLRQV RI WKH DJHQW VRFLHW\ IUDPHZRUN LW LV QHFHVVDU\ WR GHYHORS VFHQDULRV LQ GLIIHUHQW DUHDV VXFK DV HFRPPHUFH FRRSHUDWLYH ZRUN DQG PHGLDWLRQ 7KH DJHQW VRFLHW\IUDPHZRUNLVHVSHFLDOO\VXLWDEOHLQVLWXDWLRQVZKHUHFRQWURODQGLQLWLDWLYHDUH GLVWULEXWHGDQGVRFLDORUGHULVDQLVVXH,QKXPDQRUJDQL]DWLRQVDQGVRFLHWLHVQRUPV DQGFRQYHQWLRQVDUHXVHGWRVSHFLI\WKHEHKDYLRUWKDWVRFLHW\PHPEHUVDUHH[SHFWHGWR FRQIRUP WR 6XFK QRUPV DUH XVXDOO\ EDFNHG E\ D YDULHW\ RI VRFLDO LQVWLWXWLRQV WKDW HQIRUFH ODZ DQG RUGHU PRQLWRU IRU DQG UHVSRQG WR HPHUJHQFLHV DQG SUHYHQW DQG UHFRYHU IURP XQDQWLFLSDWHG GLVDVWHUV ,Q WKLV ZD\ FLWL]HQV RI FLYLOL]HG VRFLHWLHV FDQ XWLOL]HUHODWLYHO\VLPSOHDQGHIILFLHQWUXOHVRIEHKDYLRURIIORDGLQJWKHSUHYHQWLRQDQG UHFRYHU\RIPDQ\SUREOHPW\SHVWRVRFLDOLQVWLWXWLRQVWKDWFDQKDQGOHWKHPHIILFLHQWO\ DQGHIIHFWLYHO\E\YLUWXHRIWKHLUHFRQRPLHVRIVFDOHDQGZLGHO\DFFHSWHGOHJLWLPDF\ 7KHYLUWXHRIDQLQVWLWXWLRQUHVLGHVLQLWVSRWHQWLDOWROHQGOHJLWLPDF\WRLWV PHPEHUV DQGVHFXULW\WRLWVWUDGHUVE\HVWDEOLVKLQJQRUPV6HYHUDOUHVHDUFKHUVKDYHUHFRJQL]HG WKDW WKH GHVLJQ RI DJHQW VRFLHWLHV FDQ EHQHILW IURP DEVWUDFWLRQV DQDORJRXV WR WKRVH HPSOR\HG E\ RXU UREXVW DQG UHODWLYHO\ VXFFHVVIXO VRFLHWLHV DQG RUJDQL]DWLRQV >  @(OHFWURQLFLQVWLWXWLRQVFDQHQJHQGHUWUXVWWKURXJKFHUWLILFDWLRQRIDQDJHQWDQGE\ WKH JXDUDQWHHV WKDW LW SURYLGHV WR EDFN FROODERUDWLRQ )XUWKHUPRUH WKH HOHFWURQLF LQVWLWXWLRQ FDQ DOVR IXQFWLRQ DV WKH LQGHSHQGHQW SODFH LQ ZKLFK DO W\SHV RI DJHQW LQGHSHQGHQWLQIRUPDWLRQDERXWWKHLQWHUDFWLRQEHWZHHQWKHDJHQWVZLWKLQWKHVRFLHW\LV VWRUHG>@

3

Requirements for Interaction Specification

,Q RUGHU WR UHSUHVHQW LQWHUDFWLRQV EHWZHHQ DJHQWV LQ VXFK D FRQWH[W D IUDPHZRUN LV QHHGHGWKDWLVDEOHWRVSHFLI\VHTXHQFHVRILQWHUDFWLRQVFHQHV • •

LQGHSHQGHQWO\ IURP WKH LQWHUQDO GHVLJQ RI WKH DJHQW LQWHUQDO DXWRQRP\ UHTXLUHPHQW DQG ZLWKRXW IL[LQJ WKH LQWHUDFWLRQ VWUXFWXUHV FRPSOHWHO\ LQ DGYDQFH FROODERUDWLRQ DXWRQRP\UHTXLUHPHQW 

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

6SHFLILFDWLRQE\5HILQHPHQWDQG$JUHHPHQW

s UXQWLPHt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• •

•

Extensibility QRW HYHU\WKLQJ QHHGV WR EH DQG RIWHQ FDQQRW EH NQRZQ EHIRUHKDQG7KHVRFLHW\FDQHYROYHVPRRWKO\ZKHQQHZLQWHUDFWLRQVRUZD\VWR SHUIRUPWKHVHLQWHUDFWLRQVDUHEHLQJGHYHORSHG Flexibility VSHFLI\LQJ D FHUWDLQ LQWHUDFWLRQ VWUXFWXUH EHIRUHKDQG RIWHQ PHDQV WKDW GHVLJQ FKRLFHV DUH PDGH WKDW DUH QRW UHDOO\ QHFHVVDU\ )RUHVWDOOLQJ WKHVH FKRLFHV  JLYHV WKH DJHQWV WKH SRVVLELOLW\ WR PDNH WKHVH FKRLFHV WKHPVHOYHV JHDUHGWRWKHLUSDUWLFXODUVLWXDWLRQ Perusability QRW RQO\ LV LW SRVVLEOH WR GHYHORS QHZ LQWHUDFWLRQ VWUXFWXUHV ZLWKLQ WKH VRFLHW\ EXW LW LV DOVR HDVLHU WR UHXVH LQWHUDFWLRQ VWUXFWXUHV DFURVV VRFLHWLHV

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x  WKH\ FDQ XVH D

+DQV:HLJDQGHWDO

PRQLWRULQJDJHQWRUQRWDU\VHUYLFHVSURYLGHGE\WKHVRFLHW\IRUWKDW +RZHYHULWLV WKHUHVSRQVLELOLW\RIWKH VRFLHW\ WR FKHFN WKDW WKH DJHQWV IXOILOO WKHVH UHVSRQVLELOLWLHV WKLVLVUHDOL]HGE\PHDQVRIVRFLDOFRQWUDFWVFIVHFWLRQ  u 5HILQHPHQWv LVGHILQHGE\WKH20*DVu DPRUHGHWDLOHGGHVFULSWLRQWKDWFRQIRUPV WR DQRWKHU LWV DEVWUDFWLRQ v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

4

Organizational Model

7KH RUJDQL]DWLRQDO PRGHO VSHFLILHV WKH VWUXFWXUH RI DQ DJHQW VRFLHW\ LQ WHUPV RI H[WHUQDOO\REVHUYHGEHKDYLRUDQGFRPSRQHQWV$QRUJDQL]DWLRQDOPRGHOLVDWXSOH20  5 &) , 1  ZKHUH 5 LV WKH VHW RI UROH GHVFULSWLRQV &) LV WKH FRPPXQLFDWLYH IUDPHZRUN,LVWKHLQWHUDFWLRQVWUXFWXUHDQG1LVWKHVHWRIVRFLHW\QRUPVRUUXOHV7KH HOHPHQWVRI20FDQEHUHIHUUHGWRE\ roles20  5 comm-framework20  &) interactions20  , norms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

6SHFLILFDWLRQE\5HILQHPHQWDQG$JUHHPHQW

RUJDQL]DWLRQRIWKHFRQIHUHQFHDQGDQRSHUDWLRQDOSDUWZKHUHWKHFRQIHUHQFHSURJUDP LV JHQHUDWHG DQG FRQVXPHG Operational VWDNHKROGHUV LQ WKLV VRFLHW\ VSHFLILHG DV UROHV LQ WKH RSHUDWLRQDO FRPSRQHQW RI WKH PRGHO DUH DXWKRUV 3& PHPEHUV DQG SDUWLFLSDQWV 7KH JRDO RI DXWKRU LV WR JHW SDSHUV DFFHSWHG WKH 3& PHPEHU DLPV DW DVVXULQJ WKH TXDOLW\ RI WKH SURJUDP DQG WKH SDUWLFLSDQW KRSHV IRU D KLJK TXDOLW\ WR FRQVXPH Facilitation DFWLYLWLHV FDQ EH GHVFULEHG LQ WHUPV RI DQ RUJDQL]HU UROH WKDW DGPLQLVWUDWHVWKHFRQIHUHQFHDQGDFKDLUSHUVRQUROHUHVSRQVLEOHWRUHJXODWHFRQIHUHQFH VHVVLRQV 4.1

Interaction Structures

7KH SRVVLEOH DFWLRQV RI D UROH GHWHUPLQH LQWHUDFWLRQV ZLWK RWKHU DJHQWV 7KHVH LQWHUDFWLRQVDUHDUWLFXODWHGLQWRLQVWLWXWLRQDOL]HG LQWHUDFWLRQSDWWHUQVEHWZHHQDJHQWV WKDWHVWDEOLVKWKHSRVVLEOHGLDORJXHVDQGLQWHUFKDQJHVEHWZHHQDJHQWV6XFKSDWWHUQV DUH GHVLJQHG LQ VXFK D ZD\ WKDW WKH VRFLHW\ DLPV DUH DFKLHYHG ([DPSOHV DUH WKH QHJRWLDWLRQ GLDORJXH QHHGHG WR FORVH D EXVLQHVV WUDQVDFWLRQ RU WKH ZRUNIORZ VXSSRUWLQJDGHOLYHU\SURFHVV:HGHILQHDQinteraction structureDVDSDLU, 67 ZKHUH 6 LV D VHW RI VFHQHV UHSUHVHQWHG E\ LQWHUDFWLRQ VFULSWV  DQG UHODWLRQ 7 6×6 JLYHVWKHVFHQHWUDQVIRUPDWLRQPDWUL[)RUVVt ∈6VVt ∈7PHDQVWKDWVt FDQEH UHDFKHGIURPV 7KHVFHQHWUDQVIRUPDWLRQLQGXFHVDSDUWLDORUGHULQJ7 RQWKHVHWRIVFHQHV6FHQH WUDQVIRUPDWLRQ7KDVFRQVHTXHQFHVIRUroles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interaction script GHVFULEHV D VFHQDULR RI DFWLYLW\ WKDW LV KRZ UROHV LQWHUDFW DQG HYROYH LQ WKH FRQWH[W RI D VFHQH ,QWHUDFWLRQ VFULSWV VHUYH DV D EOXHSULQW IRU WKH DFWXDO LQWHUDFWLRQV EHWZHHQ DFWRUV /DQGPDUNV DUH ORJLFDO H[SUHVVLRQV WKDW GHVFULEH WKH FKDUDFWHULVWLFV IRU LQVWDQFH JRDOV DQG DFWLRQ SODQV  RI WKH VFHQH 7KH OHYHO RI VSHFLILFDWLRQ RI ODQGPDUNV GHWHUPLQHV WKH GHJUHH RI IUHHGRP WKH DFWRUV KDYH DERXW WKHLUSHUIRUPDQFH1RUPVDUHGHRQWLFH[SUHVVLRQVWKDWUHJXODWHWKHDFWLYLW\RIDFWRUV org, ⊥ participant

5HJLVWHU

or g

pa rti cip a

org, ⊥ pcmember

VWDUW

)RUP3&

org, pcmember

org

6HQG&DOO org, ⊥ author 3DSHU IRU3DSHUV VXEPLVVLRQ

pcmember  chair

g or

org

5HYLHZ SURFHVV

nt

author

3DSHU $FFHSWDQFH

&RQIHUHQFH auth o pres r  ente r

6HQG&DOOIRU 3DUWLFLSDWLRQ

Fig. 2.,QWHUDFWLRQVWUXFWXUHIRUs FRQIHUHQFHt VRFLHW\

HQG

+DQV:HLJDQGHWDO

Review process Roles: Organizer (1) PCmember (N) Landmarks: ∀ p ∈Papers, ∀ p ∈Papers,

get_reviews(p, review1, review2). paper_decision(p, decision, review1, review2).

DONE(Organizer, get_reviews(P, Rev1, Rev2)) ≡ DONE(Organizer, assign_paper(P, PCmemberA, DeadlineR) BEFORE DeadlineA) AND DONE(Organizer, assign_paper(P, PCmemberB, DeadlineR), BEFORE DeadlineA) AND DeadlineA BEFORE DeadlineR AND DONE(PCmemberA, receive_review(Org, P, Rev1) BEFORE DeadlineR) AND DONE(PCmemberB, receive_review(Org, P, Rev2) BEFORE DeadlineR) Norms: PERMITTED(Organizer, assign_paper(P, Reviewer, Review_deadline) ). OBLIGED(Reviewer, receive_review(Org, P, Rev) BEFORE Review_deadline). OBLIGED (Organizer, paper_decision(P, Decision, Rev1, Rev2) BEFORE Decision_deadline).

Fig. 3.,QWHUDFWLRQVFHQHs 5HYLHZSURFHVVt

$QLQWHUDFWLRQVFULSW,6∈6LVGHILQHGE\,6 3&)t / ZKHUH   

3LVWKHVHWRISDUWLFLSDWLQJUROHV3⊆roles20  &)t  LV WKH FRPPXQLFDWLYH IUDPHZRUN XVHG LQ WKH VFHQH SRVVLEO\ D VSHFLDOL]DWLRQRIcomm-framework20  / DUH WKH ODQGPDUNV GHVFULELQJ WKH LQWHUDFWLRQ SURSRVLWLRQV WKDW KROG IRU WKH VFHQHLQFOXGLQJWKHQRUPV 

)RU H[DPSOH LQ WKH FRQIHUHQFH VRFLHW\ WKH LQWHUDFWLRQ VFULSW IRU VFHQH s 5HYLHZ SURFHVVt  LV SDUWLDOO\ GHVFULEHG LQ ILJXUH  7KH DULW\  RU 1  RI UROHV LQGLFDWHV WKH DOORZHG QXPEHU RI GLIIHUHQW DFWRUV WKDW FDQ SOD\ WKH VDPH UROH LQ WKH VFHQH 7UDQVIRUPDWLRQV LQGLFDWHV WKH SRVVLEOH FRQVHTXHQFHV IRU DFWRUV RI D UROH LQ IXWXUH LQWHUDFWLRQVZKHUH⊥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

5

Matching OM and Agent Autonomy: Contracts

7KH2UJDQL]DWLRQDO0RGHODVGHVFULEHGDERYHLVQRWHQRXJKWRGHVFULEHWKHDFWLYLW\RI DQDJHQWVRFLHW\7KH20SURYLGHVWKHVSHFLILFDWLRQRIWKHVRFLHW\EHKDYLRUDVGHVLUHG E\ WKH VRFLHW\ GHVLJQHUV +RZHYHU WKH DFWXDO EHKDYLRU RI WKH VRFLHW\ HPHUJHV IURP WKHJRDOSXUVXLQJEHKDYLRUIURPWKHLQGLYLGXDODJHQWVZLWKLQWKHFRQVWUDLQWVVHWE\WKH 2UJDQL]DWLRQDO 0RGHO )URP DQ RUJDQL]DWLRQDO SRLQW RI YLHZ WKLV FUHDWHV D QHHG WR

6SHFLILFDWLRQE\5HILQHPHQWDQG$JUHHPHQW

FKHFN FRQIRUPDQFH RI WKH DFWXDO EHKDYLRU WR WKH GHVLUHG EHKDYLRU ,Q WKH Social Model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actor  )XUWKHUPRUH DFWXDO LQWHUDFWLRQV EHWZHHQ DFWRUV PXVW DOVR EH PDGH H[SOLFLW :H FDOO WKLV WKH Interaction Model ,0  ,Q WKH DJHQW PRGHO LQWHUDFWLRQ VFHQHV DUH LQVWDQWLDWHGRUFUHDWHGDVDUHVXOWRISUHYLRXVVFHQHVIRULQGLYLGXDODJHQWV,QWHUDFWLRQ VFHQHV DUH GHVFULEHG IRU JHQHULF DJHQW UROHV DQG FDQ EH LQVWDQWLDWHG IRU HQDFWLQJ DJHQWV $ VFHQH LQVWDQWLDWLRQ LV GHVFULEHG LQ D contract EHWZHHQ WKH LQWHUDFWLQJ DJHQWV ,Q WKH IROORZLQJ ZH FROOHFW WKH PRVW LPSRUWDQW UHTXLUHPHQWV IRU D FRQWUDFW ODQJXDJHx DIRUPDOORJLFWKDWPHHWVWKHVHUHTXLUHPHQWVLVGHVFULEHGLQ>@ 5.1

Contract Requirements

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able 1.&RRUGLQDWLRQLQDJHQWVRFLHWLHV

7\SHRIVRFLHW\ $JHQWs YDOXHVt

Market 2SHQ 6HOILQWHUHVW

)DFLOLWDWLRQ UROHV

0DWFKPDNLQJ %DQNLQJ

Network 7UXVW 0XWXDOLQWHUHVW &ROODERUDWLRQ *DWHNHHSLQJ 5HJLVWU\ 0DWFKPDNLQJ

Hierarchy &ORVHG 'HSHQGHQF\ ,QWHUIDFH &RQWURO

+DQV:HLJDQGHWDO

)XUWKHUPRUHYHULILFDWLRQRIWKHEHKDYLRURIDQRSHQVRFLHW\ZKHUHWKHGHVLJQRI SDUWLFLSDWLQJ DJHQWV FDQQRW EH FRQWUROODEOH PXVW EH EDVHG RQ WKH H[WHUQDOO\ REVHUYDEOH HIIHFWV RI DJHQW DFWLRQV 7KDW LV IURP WKH VRFLHW\ SHUVSHFWLYH GLIIHUHQW DFWLRQVWKDWEULQJDERXWWKHVDPHVWDWHRIDIIDLUVLQWKHZRUOGFDQQRWEHGLVWLQJXLVKHG

6

Society Types

7KH ZD\ RUJDQL]DWLRQDO FRRUGLQDWLRQ PRGHOV DFKLHYH FRRUGLQDWLRQ LV GHWHUPLQDQW WR WKHPRWLYDWLRQRIFRRUGLQDWLRQLQDJHQWVRFLHWLHV7DEOHGHULYHGIURP>@JLYHVDQ RYHUYLHZRIWKHFKDUDFWHULVWLFVRIDJHQWVRFLHWLHVZLWKGLIIHUHQWFRRUGLQDWLRQPRGHOV ,QmarketsDJHQWVDUHVHOILQWHUHVWHGGHWHUPLQHDQGIROORZWKHLURZQJRDOV DQG YDOXHWKHLUIUHHGRPRIDVVRFLDWLRQDQGRZQMXGJHPHQWDERYHVHFXULW\DQGWUXVWLVVXHV 2SHQQHVV LV WKXV SHU GHILQLWLRQ D IHDWXUH RI PDUNHWV )DFLOLWDWLRQ LV LQ WKH PRVW H[WUHPH FDVH OLPLWHG WR LGHQWLILFDWLRQ DQG PDWFKPDNLQJ DFWLYLWLHV ,QWHUDFWLRQ LQ PDUNHWVRFFXUVWKURXJKFRPPXQLFDWLRQDQGQHJRWLDWLRQ Network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hierarchy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market WKH 20 LV UHODWLYHO\ VLPSOH ,W FRQWDLQV WKH UROHV RI %X\HU DQG 6HOOHU DV ZHOO DV VRPH IDFLOLWDWLQJ UROHV VXFK DV 0DUNHW 2ZQHU 0DWFKPDNHU DQG %DQN 7KH ,QWHUDFWLRQ 6FHQHV IL[ WKH LQWHUDFWLRQ EDVHG RQ VWDQGDUG SURWRFROV VXFK DV DXFWLRQLQJ SURWRFROV ,Q D W\SLFDO PDUNHW WKH DJHQWV FDQQRW UHILQH WKHVH SURWRFROV LQ RWKHU ZRUGV WKH FRQWUDFWV DW ,0 OHYHO DUH LGHQWLFDOWRWKHVFHQHVRIFRXUVHWKHSDUDPHWHUVDUHGLIIHUHQW IRUHDFKLQVWDQFHEXW WKHFRQWUDFWVWUXFWXUHLVIL[HG +RZHYHUDPDUNHWPD\DOVRDOORZIRUUHILQHPHQWLQ ZKLFKFDVHWKHVRFLHW\ ZLOOVWDUWWRUHVHPEOHWKH QHWZRUNW\SH 7KH6RFLDO&RQWUDFW 60OHYHO VSHFLILHVIRUDQDJHQWLILWFDQSOD\WKHUROHRI%X\HU6HOOHURUERWK7KH VRFLDOFRQWUDFWDOVRVSHFLILHVZKDWFDQKDSSHQLQWKHFDVHRIPLVEHKDYLRU+RZHYHU WKH SDUWLHV WKHPVHOYHV DUH UHVSRQVLEOH IRU PRQLWRULQJ WKH DFWXDO SHUIRUPDQFH ,I WKH FRXQWHUSDUW\ GRHV QRW EHKDYH DFFRUGLQJ WR WKH VFHQHFRQWUDFW DQG WKH LVVXH LV QRW VROYHGELODWHUDOO\RQHSDUW\FDQUDLVHDFODLPZLWKWKH0DUNHW2ZQHU'HSHQGLQJRQ

6SHFLILFDWLRQE\5HILQHPHQWDQG$JUHHPHQW

WKH6RFLDO&RQWUDFWWKH0DUNHW2ZQHUPD\WKHQGHFLGHWRVDQFWLRQWKHQRQEHKDYLQJ SDUW\ ,Q D W\SLFDO hierarchy WKH 20 KDV D ULFK GLIIHUHQWLDWLRQ ,W FRQWDLQV YDULRXV IXQFWLRQDOUROHGHVFULSWLRQVDQGWKH,QWHUDFWLRQ6FHQHELQGVWKHVHUROHVWRJHWKHULQWR SURFHVVPRGHOV2QHRUPRUH&RQWUROOHUUROHVPXVWEHGHILQHGWKDWKDYHWKHWDVNRI PRQLWRULQJWKHSHUIRUPDQFHRIDIXQFWLRQDOUROHFRPSDULQJLWZLWKWKHRUJDQL]DWLRQDO JRDOV RU VRPH WDUJHWV GHULYHG IURP WKHP DQG SRVVLEO\ LQLWLDWLQJ VRPH UHPHGLDO DFWLRQ7KHH[WUHPHFDVHLVWKDWWKH&RQWUROOHUUROHLVSHUIRUPHGE\RQHDJHQWDQGWKH 20 GRHV QRW DOORZ IRU UHILQHPHQW ,Q VXFK D VLWXDWLRQ WKH DGYDQWDJHV RI DQ DJHQW DUFKLWHFWXUH VXFK DV ORFDO DXWRQRP\ DQG IOH[LELOLW\ EHFRPH SUREOHPDWLF 7KH RWKHU H[WUHPHLVWKDWWKH&RQWUROOHUUROHLVGLVWULEXWHGRYHUPDQ\DJHQWVDQGWKDWUHILQHPHQW RIFRQWUDFWVLVSRVVLEOH,QWKLVVLWXDWLRQWKH KLHUDUFK\ VWDUWVWRUHVHPEOHD QHWZRUN 1RUPDOO\WKHVWUXFWXUHRIWKHKLHUDUFK\ZLOOEHVRPHZKHUHEHWZHHQWKHVHH[WUHPHV GHSHQGLQJRQWKHOHYHORIVWDQGDUGL]DWLRQWKDWLVGHVLUDEOHLQWKHJLYHQHQYLURQPHQW 7KHDVVXPSWLRQWKDWDJHQWVDUHFRRSHUDWLYHWKDWLVIROORZXSUHTXHVWVLVSUHIHUDEO\ QRWWDNHQIRUJUDQWHGEXWSXWH[SOLFLWO\LQWKH6RFLDO&RQWUDFW$JHQWVDUHVXSSRVHGWR IROORZXSDOOUHTXHVWVDXWKRUL]HGGLUHFWO\RULQGLUHFWO\E\WKH6RFLDO&RQWUDFW2QWKLV SRLQWWKHKLHUDUFK\LVIXQGDPHQWDOO\GLIIHUHQWIURPERWKWKHPDUNHWDQGWKHQHWZRUN 7KH network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standardization DQG collaboration autonomy &ROODERUDWLRQ DXWRQRP\ FRUUHVSRQGV ZLWK ZKDW 0LQW]EHUJ >@ KDV FDOOHG mutual adjustmentDQGZLWKUHVSHFWWRVWDQGDUGL]DWLRQKHKDVGLVWLQJXLVKHGVHYHUDOW\SHVRI VWDQGDUGL]DWLRQ RI ZRUN SURFHVVHV RXWSXWV VNLOOV RU QRUPV  7KH VL[WK RUJDQL]DWLRQDO FRRUGLQDWLRQ PHFKDQLVP GLVWLQJXLVKHG E\ 0LQW]EHUJ LV GLUHFW VXSHUYLVLRQ ZKLFK ZH ZRXOG OLNH WR JHQHUDOL]H WR authorization DQG ZKLFK LV WKH EDVLFVWUXFWXUHRIKLHUDUFKLHV

7

Conclusion

,QWKLVSDSHUZHKDYHGHVFULEHGDQDSSURDFKWRWKHVSHFLILFDWLRQRILQWHUDFWLRQVWKDW UHVSHFWVERWKWKHLQWHUQDODXWRQRP\RIDJHQWVDVZHOODVWKHFROODERUDWLRQDXWRQRP\ ,Q RUGHU WR SURILW IURP WKH IOH[LELOLW\ RI DXWRQRPRXV DJHQWV ZKLOH DW WKH VDPH WLPH HQVXULQJWKHUREXVWQHVVRIWKHFRPSOHWHV\VWHPZHKDYHDUJXHGIRUWKHQHHGWRGHILQH DQ RUJDQL]DWLRQDO VWUXFWXUH IRU WKH PXOWLDJHQW V\VWHP VLPLODU WR UHDOOLIH

+DQV:HLJDQGHWDO

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u UHILQHPHQWv DQGu DJUHHPHQWv LQRXUSURSRVDOLVUDWKHUXQLTXH 5HILQHPHQW LQ LWVHOI LV QRW QHZ IRU H[DPSOH LW LV DOVR XVHG XQGHU WKH QDPH RI u OHYHOLQJv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

References >@

>@ >@

$QGUDGH / )LDGHLUR / *RXYHLD - .RXWVRXNRV * /RSHV $ DQG :HUPHOLQJHU 0 &RRUGLQDWLRQ 3DWWHUQV IRU &RPSRQHQW%DVHG 6\VWHPV 0XVLFDQWH 0 +DHXVOHU + (GV  3URF RI %UD]LOLDQ 6\PSRVLXP RQ 3URJUDPPLQJ/DQJXDJHV8)35 % $UWLNLV $ 3LWW - $ )RUPDO 0RGHO RI 2SHQ $JHQW 6RFLHWLHV 3URF $XWRQRPRXV$JHQWV  %RQV 5 /HH 5 :DJHQDDU 5 :ULJOH\ & 0RGHOLQJ ,QWHURUJDQL]DWLRQDO 7UDGH 3URFHGXUHV 8VLQJ 'RFXPHQWDU\ 3HWULQHWV 3URF +DZDLL ,QWHUQDWLRQDO &RQIHUHQFH 

6SHFLILFDWLRQE\5HILQHPHQWDQG$JUHHPHQW

>@

>@ >@

>@ >@ >@

>@ >@ >@ >@ >@

>@ >@ >@

&DVWHOIUDQFKL & (QJLQHHULQJ 6RFLDO 2UGHU 2PLFLQL $ 7RONVGRUI 5 =DPERQHOOL ) (GV  (QJLQHHULQJ 6RFLHWLHV LQ WKH $JHQWV :RUOG )LUVW ,QWHUQDWLRQDO:RUNVKRS(6$:%HUOLQ*HUPDQ\/1$,6SULQJHU 9HUODJ x  'DVWDQL 0 'LJQXP 9 'LJQXP ) 2UJDQL]DWLRQV DQG 1RUPDWLYH $JHQWV 6XEPLWWHG WR WKH )LUVW (XUDVLDQ &RQIHUHQFH RQ ,&7 7HKHUDQ ,UDQ 2FWREHU   'HOODURFDV & &RQWUDFWXDO $JHQW 6RFLHWLHV 1HJRWLDWHG VKDUHG FRQWH[W DQG VRFLDO FRQWURO LQ RSHQ PXOWLDJHQW V\VWHPV 3URF :6 RQ 1RUPV DQG ,QVWLWXWLRQV LQ 0XOWL$JHQW 6\VWHPV $XWRQRPRXV $JHQWV %DUFHORQD   'LJQXP)$XWRQRPRXV$JHQWVZLWK1RUPV,Q$,DQG/DZ7  x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utonomous Agents and Multi-Agent Systems x   .XPDU 6 +XEHU 0 &RKHQ 3 0F*HH ' 7RZDUGV D )RUPDOLVP IRU &RQYHUVDWLRQ 3URWRFROV 8VLQJ -RLQW ,QWHQWLRQ 7KHRU\ &KDLEGUDD % 'LJQXP) (GV  &RPSXWDWLRQDO ,QWHOOLJHQFH -RXUQDO 6SHFLDO ,VVXH $JHQW &RPPXQLFDWLRQ/DQJXDJH To appear. 0LQW]EHUJ+0LQW]EHUJRQ0DQDJHPHQW7KH)UHH3UHVV1HZ@$QRWKHULVWKH6WUDWHJLF5DWLRQDOH PRGHOZKLFKGHVFULEHVDQGVXSSRUWVWKHUHDVRQLQJWKDWHDFKDFWRUKDVDERXWLWVUHOD WLRQVKLSV ZLWK RWKHU DFWRUV LWV DOWHUQDWLYH PHDQV WR DFKLHYH LWV JRDOV DQG KRZ WKH TXDOLWDWLYH H[SHFWDWLRQV RI DFWRUV DUH VDWLVILHG E\ WKHVH DOWHUQDWLYHV $FWRUV LQ L IUDPHZRUNDUHVWUDWHJLFEHFDXVHWKH\FDQHYDOXDWHWKHLUVRFLDOUHODWLRQVKLSVLQ WHUPV RIRSSRUWXQLWLHVWKDWWKH\RIIHUDQGYXOQHUDELOLWLHVWKDWWKH\PD\EULQJ

$Q$JHQWDQG*RDO2ULHQWHG$SSURDFKIRU9LUWXDO(QWHUSULVH0RGHOOLQJ

2.1

Strategic Dependency Model

$ 6WUDWHJLF 'HSHQGHQF\ 6'  PRGHO FRQVLVWV RI PDQ\ QRGHV DQG OLQNV (DFK QRGH UHSUHVHQWVDQDFWRUDQGHDFKOLQNEHWZHHQWZRDFWRUVLQGLFDWHVWKDWRQHDFWRUGHSHQGV RQ WKH RWKHU 7KH GHSHQGLQJ DFWRU LV FDOOHG WKH GHSHQGHU DQG WKH DFWRU ZKR LV GH SHQGHGRQLVFDOOHGGHSHQGHH7KDWPHDQVWKHGHSHQGHUQHHGVVRPHKHOSWRDWWDLQWKH JRDOIURPWKHGHSHQGHH7KHGHSHQGHUFDQDFKLHYHJRDOVWKDWRWKHUZLVHFDQQRWEHUH DOL]HGZLWKRXWWKHGHSHQGHQF\RUQRWDVHDVLO\RUQRWDVHIIHFWLYHO\$WWKHVDPHWLPH WKH GHSHQGHU EHFRPHV YXOQHUDEOH ,I WKH GHSHQGHH IDLOV WKH GHSHQGHU ZRXOG EH DG YHUVHO\ DIIHFWHG LQ LWV DELOLW\ WR DFKLHYH LWV JRDOV 7KHUH DUH IRXU GHSHQGHQFLHV LQ 6WUDWHJLF 'HSHQGHQF\ PRGHO JRDO GHSHQGHQF\ WDVN GHSHQGHQF\ UHVRXUFH GHSHQG HQF\DQGVRIWJRDOGHSHQGHQF\,QDJRDOGHSHQGHQF\DQDFWRUGHSHQGVRQDQRWKHUWR PDNH D FRQGLWLRQ FRPH WUXH ,Q D WDVN GHSHQGHQF\ DQ DFWRU GHSHQGV RQ DQRWKHU WR SHUIRUP DQ DFWLYLW\ ,Q D UHVRXUFH GHSHQGHQF\ DQ DFWRU GHSHQGV RQ DQRWKHU IRU WKH DYDLODELOLW\RIDQHQWLW\7KHVRIWJRDOGHSHQGHQF\LVDYDULDQWRIWKHJRDOGHSHQGHQF\ DQG LW LV QR a priori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t VRUGHULVWKHUHTXLUHPHQWRI WKHPDUNHW:KHQWKHFXVWRPHUKDVDGHVLUHWRSXUFKDVHDQDLUVHSDUDWRULWLVYHU\GLI ILFXOWWRJHWDSURGXFWLQVWRFNEHFDXVHWKLVNLQGRISURGXFWLVFXVWRPL]HGDQGFDQt W EHSURGXFHGLQEDWFK6RWKHFXVWRPHU PXVW KDQGRYHUWKHRUGHUWRWKH3URGXFW'H VLJQLQJDQG4XRWLQJ*URXSLQRUGHUWRFXVWRPL]HWKHSURGXFW7KH3URGXFW'HVLJQLQJ DQG 4XRWLQJ *URXS ZLOO GHVLJQ WKH SURGXFW LQ DFFRUGDQFH ZLWK WKH SHUIRUPDQFH RI SURGXFW DQG JLYH WKH TXRWHG SULFH DQG WKH GHOLYHULQJ GDWH )RU WKH FXVWRPHU LW FDQ HYDOXDWHWKHGHVLJQSULFHDQGGHOLYHULQJGDWHDQGGHFLGHWRDFFHSWRUQRW,WFDQJHW WKHUHODWHGLQIRUPDWLRQVXFKDVSULFHDQGSURGXFLQJWLPHIURPWKH'HYLFHVDQG0D FKLQHV 0DQXIDFWXULQJ *URXS 3DUWV 6XSSO\LQJ *URXS DQG 3URGXFW $VVHPEOLQJ *URXS,IQRWLWFDQQHJRWLDWHZLWKWKH3URGXFW'HVLJQLQJDQG4XRWLQJ*URXS,IWKH QHJRWLDWLRQ IDLOVWKHYLUWXDO HQWHUSULVH ZLOO QRW EH EXLOW 7KH 3URGXFW 'HVLJQLQJ DQG 4XRWLQJ *URXS ZLOO GHVLJQ WKH SURGXFW DFFRUGLQJ WR WKH SHUIRUPDQFH RI SURGXFW ,Q RUGHU WR HQKDQFH FRPSHWLWLRQ RI WKH SURGXFW WKH 3URGXFW 'HVLJQLQJ DQG 4XRWLQJ *URXS VKRXOG JLYH WKH &XVWRPHU D JRRG GHVLJQ LQFOXGLQJ JRRG SHUIRUPDQFH JRRG SULFHDQGVKRUWGHOLYHULQJWLPH*RRG SHUIRUPDQFH PHDQV WKH SURGXFW VKRXOG VXIIL FLHQWO\PHHWWKHQHHGRIWKHSURGXFW,WGHPDQGVWKH3URGXFW'HVLJQLQJDQG4XRWLQJ *URXSWRKDYHVXIILFLHQWH[SHULHQFHVLQSURGXFWGHVLJQ,QWKH PHDQWLPHWKH*URXS VKRXOGFRQVLGHUWKHUHODWHGFRQVWUDLQWVLQWKHGRZQVWUHDPRISURGXFLQJSURFHVVLQDG YDQFH LQFOXGLQJ WKH FRQVWUDLQWV LQ WKH PDQXIDFWXULQJ SURFHVV DVVHPEOLQJ SURFHVV DQGVRRQ,WFDQDYRLGGRLQJRYHUDJDLQLQRUGHUWRVKRUWWKHGHOLYHULQJWLPHDQGLW DOVR FDQ HIIHFWLYHO\ UHGXFH WKH GHVLJQLQJ VXUSOXV LQ RUGHU WR FXW GRZQ WKH SURGXFW

=KL/LXDQG/LQ/LX

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

Fig. 1.7KHVWUDWHJLFGHSHQGHQF\PRGHORIWKHYLUWXDOHQWHUSULVH

$Q$JHQWDQG*RDO2ULHQWHG$SSURDFKIRU9LUWXDO(QWHUSULVH0RGHOOLQJ

%HVLGHVWKHFRRSHUDWLRQDERYHWKH3URGXFW'HVLJQLQJDQG4XRWLQJ*URXSZLOOFR RSHUDWH ZLWK WKH 3DUWV 6XSSO\LQJ *URXS DQG 3URGXFW $VVHPEOLQJ *URXS 7KH 3DUWV 6XSSO\LQJ*URXSZLOOSXUFKDVHDQGSURYLGHDOONLQGVRIVWDQGDUGSDUWVRULQVWUXPHQWV IRUWKHSURGXFW7KH3URGXFW$VVHPEOLQJ*URXSZLOOJHWDOONLQGVRIGHVLJQLQIRUPD WLRQGHYLFHVDQGPDFKLQHVPDQXIDFWXUHGDQGVWDQGDUGSDUWVDQGWKHQDVVHPEOHWKHP WRJHWKHU7KHSURGXFWZLOOEHJLYHQDWHVWUXQDQGDFKHFNDQGZLOOEHDFFHSWHGE\WKH 3URGXFW'HVLJQLQJDQG4XRWLQJ*URXS,WZLOOEHVHQWWRWKH&XVWRPHUE\WKH*URXS DQGHQWLUHYLUWXDOHQWHUSULVHZLOOEHGLVPLVVHG 2.2

Strategic Rationale Model

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u 5HTXLUHPHQWV RI 3URGXFW %H 'HFLGHGv  7KLV WDVN FDQ EH GHFRP SRVHGLQWRWKUHHVXEWDVNVVXFKDVu 3HUIRUPDQFHV%H'HFLGHGv u 3ULFH%H'HFLGHGv  DQGu 'HDGOLQH%H'HFLGHGv )RUWKH3URGXFW'HVLJQLQJDQG4XRWLQJ*URXSLWVJRDOLV u 3URGXFW 'HVLJQLQJ DQG 4XRWLQJv  7KHUH DUH WKUHH SRVVLEOH ZD\V WR DFFRPSOLVK LW VXFKDVu 1HZB3URGXFW'HVLJQLQJDQG4XRWLQJv RUu 7UDQVIRUPHGB3URGXFW'HVLJQLQJ DQG 4XRWLQJv  RU  u 0RGLILHGB3URGXFW 'HVLJQLQJ DQG 4XRWLQJv  7KHVH WDVNV UHVSHF WLYHO\UHVSRQVHWRGLIIHUHQWSURGXFWVDQGHDFKWDVNLVFRPSRVHGRIWZRVXEWDVNVUH ODWHG SURGXFW GHVLJQLQJ DQG TXRWLQJ ,Q DQ\ GHVLJQLQJ ZD\ WKHUH DUH ILYH PDLQ GH VLJQLQJ SURFHVVHV LQFOXGLQJ u *HQHUDO 'HVLJQLQJv  u 'HYLFH 'HVLJQLQJv  u 0DFKLQH 'HVLJQLQJv u 3RZHU6XSSO\6\VWHP'HVLJQLQJv DQG u 'HWHFW6\VWHP 'HVLJQLQJv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

=KL/LXDQG/LQ/LX

FRPSRVH WKH GHVLJQLQJ ZRUN DQG ILQG SDUWQHUV WR DFFRPSOLVK LW 7KHVH SDUWQHUV FDQ FRRSHUDWHZLWKHDFKRWKHULQVLGHRURXWVLGHWKHLURZQFRPSDQLHV ,Q ILJXUH  VKRZV WKH 6WUDWHJLF 5DWLRQDOH PRGHO WR VXSSRUW UHDVRQLQJ DERXW WKH SURGXFWGHVLJQLQJDQGTXRWLQJSURFHVV(DFKSRVVLEOHZD\KDVGLIIHUHQWLPSOLFDWLRQV IRUWKHTXDOLWLHVRIJRDOVRUVRIWJRDOV$VRIWJRDOLVXVXDOO\QRWHDV\WRTXDQWLILHG7KH FRQWULEXWLRQVWRVRIWJRDOVFDQEHSRVLWLYHRUQHJDWLYHDGHTXDWHRULQDGHTXDWH7KH

Fig. 2.7KHVWUDWHJLFUDWLRQDOHPRGHORIWKHYLUWXDOHQWHUSULVH

$Q$JHQWDQG*RDO2ULHQWHG$SSURDFKIRU9LUWXDO(QWHUSULVH0RGHOOLQJ

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

Roles, Positions and Agents

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

=KL/LXDQG/LQ/LX

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

Fig. 3.7KHVWUDWHJLFGHSHQGHQF\PRGHOZLWKUROHVSRVLWLRQVDQGDJHQWV

$Q$JHQWDQG*RDO2ULHQWHG$SSURDFKIRU9LUWXDO(QWHUSULVH0RGHOOLQJ

Fig. 4.7KHFRRSHUDWLRQUHODWLRQVKLSVLQWKHYLUWXDOHQWHUSULVH

3

Study of Cooperation in the Virtual Enterprise

3.1

Cooperation in the Virtual Enterprise

:KHQ WKH YLUWXDO HQWHUSULVH LV EXLOW WKH SDUWQHUV ZLOO FRRSHUDWH ZLWK HDFK RWKHU DF FRUGLQJ WR WKHLU GHSHQGHQF\ UHODWLRQVKLSV 7KH GHSHQGHQFLHV SURSDJDWLQJ RYHU WKH QHWZRUNRIGHSHQGHQF\UHODWLRQVKLSVPDNHWKHGHSHQGHUVDQGGHSHQGHHVFRRSHUDWHHI IHFWLYHO\ ,Q ILJXUH  WKH &XVWRPHU GHSHQGV RQ WKH 3URGXFW 'HVLJQLQJ DQG 4XRWLQJ *URXS JRDOGHSHQGHQF\ WRFXVWRPL]HWKHSURGXFWDQGGRHVQt WFDUHKRZWKH*URXSDFKLHYHV WKHJRDO,QWKHYLUWXDOHQWHUSULVHWKH3URGXFW'HVLJQLQJDQG4XRWLQJ*URXSWDNHVRQ DOHDGHUUROHWRHVWDEOLVKDQGRUJDQL]HWKHYLUWXDOHQWHUSULVHLQRUGHUWRDFFRPSOLVKWKH WDVN7KHVXEWDVNVLQWKH3URGXFW'HVLJQLQJDQG4XRWLQJ*URXSVXFKDV*HQHUDO'H VLJQLQJ'HYLFH'HVLJQLQJ0DFKLQH'HVLJQLQJZLOOILQLVKWKHJHQHUDOSURGXFWGHVLJQ DQG DOO NLQGV RI GHVLJQV RI GHYLFHV DQG PDFKLQHV LQ WKLV SURGXFW 7KH\ UHVSHFWLYHO\ GHSHQG RQ 'HYLFHV DQG 0DFKLQHV 0DQXIDFWXULQJ WDVNGHSHQGHQF\  WR PDQXIDFWXUH WKH GHYLFHV DQG PDFKLQHV GHSHQGV RQ 3DUWV 6XSSO\LQJ WDVNGHSHQGHQF\  WR VXSSO\ WKHSDUWVGHSHQGVRQ3URGXFW$VVHPEOLQJWDVNGHSHQGHQF\ WRDVVHPEOHWKHSURGXFW

=KL/LXDQG/LQ/LX

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

Cooperative Architecture for the Virtual Enterprise

%DVHGRQWKHDQDO\VLVZLWKWKHL IUDPHZRUNDQDJHQWDQGJRDORULHQWHGPRGHOIRUWKH YLUWXDO HQWHUSULVH FDQ EH EXLOW %HFDXVH WKH YLUWXDO HQWHUSULVH LV D G\QDPLF DOOLDQFH ZKHQWKH JRDOVDQGWDVNVKDYHEHHQDOORFDWHGWR SDUWQHUV WKH DOO\LQJ SURFHVV LV ILQ LVKHG,QRUGHUWRVXSSRUWWKHDOO\LQJSURFHVVDQGFRRSHUDWLRQLQYLUWXDOHQWHUSULVHZH SURSRVHDKLHUDUFKLFDODQGIHGHUDWHGDUFKLWHFWXUHLQILJXUH,QWKLVDUFKLWHFWXUHWKH SDUWQHUVFDQFRPPXQLFDWHZLWKHDFKRWKHUZLWKWKHZHOONQRZQ&RQWUDFW1HW3URWRFRO >@ ZKHQ WKH\ DOO\ ZLWK HDFK RWKHU 7KRVH DJHQWV QRW DW WKH ORZHVW OHYHO  FDQ G\ QDPLFDOO\ WDNH WZR UROHV PDQDJHU RU FRQWUDFWRU >@ $V D PDQDJHU WKH DJHQW ZLOO GHFRPSRVHWKHWDVNRUVXEWDVNVDQGDOORFDWHWKHPWRLWVSDUWQHUV,WFDQDQQRXQFHWKH WDVNRUVXEWDVNVWRUHOHYDQWSDUWQHUVDQGFROOHFWWKHELGVIURPSDUWQHUVDQGDZDUGWKH WDVNRUVXEWDVNVWRWKHEHVWELGGHUV:KHQWKHWDVNRUVXEWDVNVKDYHEHHQDOORFDWHG WRWKHSDUWQHUVWKHPDQDJHUDQGWKHSDUWQHUVIRUPDIHGHUDWLRQ7KHPDQDJHUZLOOFR RSHUDWH DQG FRRUGLQDWH ZLWK WKH SDUWQHUV DQG VXSSRUW FRPPXQLFDWLRQ DPRQJ WKH DJHQWVLQLWVIHGHUDWLRQ$JHQWVLQDIHGHUDWLRQFDQFRPPXQLFDWHZLWKRWKHUVLQGLIIHU HQWIHGHUDWLRQVWKURXJKWKHLUUHVSHFWLYHXSSHUPDQDJHU7KLVPDNHVWKHUHDOL]DWLRQRI FRPPXQLFDWLRQPRUHFRQYHQLHQW ,QILJXUHWKHKLHUDUFKLFDODQGIHGHUDWHGDUFKLWHFWXUHQRWRQO\VXSSRUWVWKHWDVNWR EHGHFRPSRVHGDQGDOORFDWHGEXWDOVRWRIDFLOLWDWHWKHWDVNVWREHPDQDJHG,QDIHG HUDWLRQWKHVHSDUWQHUVFRRSHUDWHZLWKHDFKRWKHULQRUGHUWRDFFRPSOLVKWKHWDVNVRU VXEWDVNV7KHVHSDUWQHUVPD\FRPHIURPGLIIHUHQWFRPSDQLHVRUFRPHIURPGLIIHUHQW ZRUN XQLWV LQ WKH VDPH FRPSDQ\ 7KH\ FDQ RYHUODS WR LPSOHPHQW WKHLU UHVSHFWLYH WDVNVRUVXEWDVNVDOORFDWHGWRWKHPEDVHGRQYDULRXVGHSHQGHQF\UHODWLRQVKLSV7KDW PHDQVWKRVHFRPSDQLHVFRRSHUDWLQJZLWKHDFKRWKHULQWKHYLUWXDOHQWHUSULVHDVGLIIHU HQWYLUWXDO JURXSVVXFKDV**RU*UHVSRQGGLIIHUHQWWDVNV RU VXEWDVNV LQ WKH KLJKOHYHOFRRSHUDWLRQ,QWKHORZOHYHOFRRSHUDWLRQWKHGHYHORSHUVDQGZRUNXQLWVLQ VDPHFRPSDQ\RULQGLIIHUHQWFRPSDQLHVFDQ IRUPDOO NLQGV RI ZRUN WHDPV VXFK DV 7 WR WDNH RQ WKH GHFRPSRVHG WDVNV RU VXEWDVNV 7KLV PDNHV WKH YLUWXDO HQWHUSULVH PRUHIOH[LEOHDQGPRUHFRPSHWLWLYH,QWKLVSDSHUDSDUWQHUFDQEHDFRPSDQ\RUD ZRUNXQLWRUDGHYHORSHULQDFRPSDQ\

$Q$JHQWDQG*RDO2ULHQWHG$SSURDFKIRU9LUWXDO(QWHUSULVH0RGHOOLQJ

&RPPHQWV

3'4*3URGXFW'HVLJQLQJDQG4XRWLQJ*URXS36*3DUWV6XSSO\LQJ*URXS '00*'HYLFHDQG0DFKLQH0DQXIDFWXULQJ*URXS3$*3URGXFW$VVHPEOLQJ*URXS 34*3URGXFW4XRWLQJ*URXS3'*3URGXFW'HVLQJLQJ*URXS *'*HQHUDO'HVLJQHU'''HYLFH'HVLJQHU 366'3RZHU6XSSO\6\VWHP'HVLJQHU0'0DFKLQH'HVLJQHU '6''HWHFW6\VWHP'HVLJQHU'0'HYLFH0DQXIDFWXULQJ 000DFKLQH0DQXIDFWXULQJ3,63DUWVLQ6WRFNV 333DUWV3XUFKDVLQJ Fig. 5.$KLHUDUFKLFDODQGIHGHUDWHGDUFKLWHFWXUHIRUWKHYLUWXDOHQWHUSULVH

4

Discussion

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

=KL/LXDQG/LQ/LX

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

Acknowledgements 7KHZRUNRIWKLVSDSHULVJUHDWO\PRWLYDWHGDQGLQVSLUHGE\P\FRRSHUDWLRQZLWK3UR IHVVRU-RKQ0\ORSRXORVt VUHVHDUFK+LVNLQGVXSSRUWDQGHQFRXUDJHPHQWLVJUDWHIXOO\ DFNQRZOHGJHG

References >@ >@

>@ >@

>@

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

>@ >@

(ULF