231 118 86MB
English Pages 428 Year 1989
Systems Analysis and Simulation 1988 I
Mathematical Research
Mathematische Forschung
Wissenschaftliche Beiträge herausgegeben von der Akademie der Wissenschaften der DDR Karl-Weierstraß-Institut für Mathematik
Band 46 Systems A n a l y s i s a n d S i m u l a t i o n
1988
Systems Analysis and Simulation 1988 I: Theory and Foundations Proceedings of the International Symposium held in Berlin (GDR), September 12-16,1988
edited by Achim Sydow Spyros G. Tzafestas Robert Vichnevetsky
Alcademie-Verlag Berlin 1988
Herausgeber: Prof. Dr. Achim Sydow, Zentralinsti tut für Kybernetik und prozesse der Akademie der Wissenschaften der DDR, Berlin Prof. Dr. Spyros G. Tzafestas, Computer Engineering Technical University of Athens, Athens P r o f . Dr. Robert Vichnevetsky, Universi ty, New Brunswick
Die T i t e l dieser Schriftenreihe Autoren reproduziert.
ISBN ISSN
Dept.
Division,
of Computer Science,
werden
vom Originalmanuskript
Informations-
National
Rutgers
der
3-05-500Ö54-2 0138-3019
Erschienen irn Akademie-Verlag Berlin,DDR-1086 Berlin,Leipziger Str. (c) Akademie-Verlag B e r l i n 1p88 Lizenznummer! 202.100/500/88 Printed in the German Democratic Republic Gesamtherstellung; VEB Kongreß- und Werbedruck, 9273 Oberlungwitz LSV 1095 Bestellnummer1 763 958 1 (2182/46) 05600
3-4
PREFACE The
present volume contains the papers which were accepted for
tation
at
the
3rd International Symposium for
Systems
presen-
Analysis
and
Simulation held in Berlin (GDR), September 12-16, 1986. It
is
already a tradition to meet a broad international
experts sium.
in systems analysis,
community
modelling and simulation at this
of
sympo-
This fact shows the requirements for a forum of presentation
and
discussion of new developments and applications of modelling and simulation in systems analysis. To
the great interest in this field one has to take into
con-
sideration the developed role of computer simulation as a powerful
realize
tool
of problem solving. been
More and more areas in sciences and production have
investigated by mathematical models and computer simulation.
Bio-
logical sciences and social sciences are even by now influenced by
this
trend. The
model
use on the computer has been very much improved in
decision
suppprt systems.
Parallel simulation will provide drastic shortening of
computing
Parallel simulation and model based
time.
decision
support
systems are brought in the focus of international activities. Numerical mathematics, algorithms or
systems theory and control sciences provide with
supporting the modelling process itself based on
analytic
modelling
methods.
and graphics for representing results are real model
systems. A
simulation
Such simulation systems equipped with tools
for
support
*
new important impact comes from artificial intelligence by
processing.
knowledge
Expert systems may help decision making in case of
mathematical models.
missing
Expert systems may also support teaching and using
simulation systems. New teria
application areas are investigated. control
problems
of
qualitative
Applications sciences
Complex systems with multicri-
problems are in the scope of the symposium analysis of small
in engineering sciences,
scale
as
nonlinear
economy and management,
well
as
systems. natural
and social sciences are examined but also mixed problems
from
different areas. The
state
efficient
of
computer technique and programming
conditions for simulations.
environment^
set
Personal computer are even
up used
for simulation more and more. The
symposium
reflects
the state of the art
and
trends
in
systems
analysis, modelling and simulation.
5
The 3rd International Symposium is organized by the Central Institute of Cybernetics and Information Processes of the Academy of Sciences of the GDR (ZKI) with cosponsorship of the -International Association for Mathematics and Computers in Simulation (IMACS), -International Federation of Automatic Control (IFAC), -International Institute for Applied Systems Analysis Laxenburg (IIASA), Scientific Society of Measurement and Automation (WGMA) in the Chamber of Technology (KdT) of the GDR, -Mathematical Society (MG) of the GDR. The papers included in these proceedings were not formally refereed. The authors themselves are fully responsible. The international Program Committee consisted of: W. Ameling (FRG), P. Borne (France), L. Dekker (The Netherlands), S. Deng (PRC), A. A. Dorodnicyn (USSR), K. H. Fasol (FRG), W. Findeisen (Poland), 0. I. Frankseri (Denmark), V. Hamata (Czechoslovakia), C. Hu (PRC), A. Javor (Hungary), K. Kabes (Czechoslovakia), V. V. Kalashnikov (USSR),- V. Kempe (GDR), E. J. H. Kerckhoffs (The Netherlands), R. Klotzler (GDR), R. Kulikovski (Poland), A. Kurzhansky (Austria, USSR), N. Levan (USA), A. H. Levis (USA), T. I. Oren (Canada), M. Peschel (GDR), P. D. Dieu (Vietnam), F. Pichler (Austria), K. Reinisch (GDR), W. Schirmer (GDR), B. Schmidt (FRG), V. V. Solodovnikov (USSR), F. Stanciulescu (Romania), J. M. Svirezhev (USSR), M. Thoma (FRG), I. Troch (Austria), S. G. Tzafestas (Greece), G. C. Vansteenkiste (Belgium), R. Vichnevetsky (USA) , A. Sydow (GDR). Many thanks should be given to the members of this committee for the very helpful cooperation. Special thanks are said to Prof. Dr. V. Kempe, Director of the ZKI, for his great support in preparing and performing the symposium. Furthermore, great gratitude is to express to Prof. Dr. R. Vichnevetsky (USA), IMACS-President, Prof. Dr. B. Tamm (USSR), IFAC-President, Prof. Dr # R. H. Pry (USA), IIASA-Director, Prof. Dr. H. Richter (GDR), Chairman of WGMA, Prof. Dr. R. Klotzler (GDR), Chairman of MG for help and encouragement. A lot of the hard preparation work was done by the Department for Systems Analysis and Simulation of the Central Institute of Cybernetics and Information Processes. The editor expresses his thanks to all col-
6
leagues and friends who were very much engaged in the research work and in the preparation. First of all I would like to name Dr. P. Rudolph and Dr. A. Wittmilß who helped to prepare the proceedings. Furthermore I thank these colleagues and Dr. K.Bellmann, Dr. W.Jansen, Dr. E.Matthäus, Dr. R. St.raubel and all the other colleagues for engaged cooperation for years in developing this research area. Mrs. Ch. Fröhlich and Mrs. J. Obretenov should be named for speedy service in preparing the manuscript. Mrs. S. Böttcher made an excellent job as organizer. Last not least I would give my thanks to the publishers, especially Mrs. R. Helle and Mrs. G. Reiher, for their assistance and cooperation. Finally I would like to express my expectation also on behalf of the coeditors Prof. Dr. S. G. Tzafestas and Prof. Dr. R. Vichnevetsky that also the third symposium will be a contribution to the further development in systems analysis, modelling and simulation as well as a place for cooperation and communication like the first both.
April 1988
Achim Sydow On behalf of the editors
7
TABLE OF CONTENTS 1. Methods and Fundamentals 1.1. Theory of Discrete Systems D.Garte; J.Haufe; St.Ruelke: A Language to Describe and to Simulate Digital Systems
15
J.Voeroes: The State Space Approach to Discrete Event Dynamic Systems
22
the
Analysis
of
1.2. Theory of Continuous Systems - Qualitative Theory G.-Jumarie /INVITED PAPER/: Catastrophe,Chaos,Synergetics and Thermodynamics. A Unified Approach via Information of Deterministic Maps
28
W.Metzler: A Route t'o Chaos
34
V.S. Anishchenko; T.E. Vadivasova; M.A. Safonova: Bifurcations of Two-Dimensional Tori and Chaos in Dissipative Systems
40
H.G.Bothe: riant Sets
42
Shape and Dimension of Certain Hyperbolic
Inva-
W.Jansen; U.Feudel: CANDYS/QA - A Software System for Qualitative Analysis of the Behaviour of the Solutions of NonLinear Dynamical Systems
45
W.Metzler; H.Krieger: Qualitative Behaviour of Ordinary Differential Equation Models Describing Forest Growth Under Air Pollution
48
1.3. Decision Support Systems L.Cserny: The Analysis of Decision Making Systems
53
P.Bronisz; L.fCrus: Interactive Procedures for Multi-Criteria Decision Support in Bargaining Problem
59
Z.Strezova: A Procedure for Decision Support Systems Design: Modelling and Simulation Environment
63
G.Kreiselmeier; R.Seidl: Textile Technology
67
Knowledge-Based
Decision Aid
in
R.Schmidt; B.Koch: A Multiobjective Decision Support System for the Top Management
71
1.4. Modelling K.-H.Schmelovsky /INVITED PAPER/: Modelling and Simulation
75
A.G.Ivakhnenko /INVITED PAPER/: Self-Organizing Methods in Modelling and Clustering: GMDH Type Algorithms
86
H. Schwarz; H.T. Dorissen; I.. Guo: Bi1inearization of Nonlinear Systems
89
Weijian Zhang: Analytical Analysis of a Stochastic Differential Equation
97
Partial
G. Bohlender: Is Floating-Point. Arithmetic Still Adequate?
105
R.H. Adams: Using Systems of Incomplete, Often Inconsistent, Models 109 W. Borutsky: Top-Down Modelling of Complex Systems by means of Word Bond Graphs 113 9
Yi Yunwen; Zhang Lu: An Approach of the Grey System Modelling and Simulation for Complex Systems 117 V. Wenzel; SONCHES
E. Matthaeus; M. Flechsig: Generic Modelling in
121
G. Hertel: Statement and Tendencies of Models for Complicated Technical Systems 125 A.A. Lebedev: A Mathematical Model for Description of Random and Indefinite Factors from Unified Positions 130 G. Dzemyda: The Algorithms of Extremal Parameter Grouping
133
T. Lange: A New Approach for Structural Modelling
137
E. Apelt; D. Apelt: Problems of Qualitative Change of Parameters in Different Hierarchical Levels 141 L. Fortuna; A. Gallo; G. Nunnari: Studying the Interactions Among Model Reduction Algorithms via CAD Technique 146 J. Halawa; A. Trzmielak-Stanislawska: Determination of Simplified Models.by means of Chebyshev Polynomials 151 J. Halawa: A Note on Simplification of Large Dynamic Systems Using a Moment Technique 155 R. Boettner: Model Reduction and Stability Analysis of Nonlinear Dynamical Systems by means of Centre Manifold Theory 159 M. Kejak; P. .Javarsky: Method of Automated Construction of System Dynamics Models 163 S. Krueger; W. Mylius: A Modular Computer-Aided Modelling and Simulation System in Chemical Engineering 165 R. Funke: CANDYS/CM - A Dialogue System for Modelling Continuous Dynamical Systems with Chain Structure by Differentia] Equations 169 M.N. Krasilshchikov; V.I. Karlov: Control of the Observation Process by Probability Criterion 172 J. Markowski; M. Popkiewicz: Simulations Analysis of a Nonparametric Algorithm for Identification of Discrete-Time Hammerstein System 175 E. Jesierski: Remarks on Pole Assignment by Constant Output Feedback 179 P.S. Szczepaniak; A. Maiolepszy: On the Computational Solution ,of Differential Equations with Delay 183 Ch. Dahme: A Theory of Elementary Social Systems as a Basis for the Analysis and Modelling of Decision Situations 189 Th. Hager; Ch. Dahme: An Approach to the Development of Supporting Systems for Analysis and Construction/Influencing Social Systems 193 Z. Mital: Distributed Message Exchange System Modelling
196
E.-G. Woschni: Analysis and Optimization by means of Estimations in Measurement 200 M. Kejak: Simulation of Fuzzy System Dynamics Models 10
205
1.5. Multiobjective Optimization J. Ester /INVITED PAPER/: Multicriteria Fuzzy Decisions Anlan Song; Wei-Min Cheng: A Method for Multi-Criteria Decision Making
Multihuman
209 and
213
1.6. Methods of Optimisation and Control H.P. Schwefel /INVITED PAPER/: Evolutionary Learning Optimum Seeking on Parallel Computer Architectures 217 I. Troch /INVITED PAPER/: Control Design
Optimisation and Simulation
in
M. Peschel; H.-M. Voigt ; W. Mende; F. Breitenecker /INVITED PAPER/: System-Engineering Methodology for Simulation and Control of Dynamical Networks
226
232
J. Alder; K.J. Reinschke: Modelling of Large Processes Containing Continuously and Binarly Controlled Parts 237 Z. Emirsajlow: Integral Riccati Equations for a Feedback Solution of LQCP with a Terminal Inequality Constraint 243 J. Gondzio: Stable Variant of the Simplex Method for Solving Supersparse Linear Programs 247 •J. Fischer: Some Remarks on Optimizing Simulated Systems
251
Xu Kekang; Wang Zhenquan: D-Controllability and Strong D-Controllabi1ity and Control of Multiparameter and Multiple Time-Scale Singularly Perturbed Systems 255 J. Doleial: On-Line Optimal Control of Nonlinear Systems by Singular Perturbation Techniques 259 M. Schwaar: Design of Optimal Feedback Controllers for Some Classes of Nonlinear Systems 263 P. Javorsky: Systems
An Algorithm for Optimal Control of Nonlinear
J. Cretnik; S. Strrncnik: Design of a Combustion Controller
267 270
H.-M. Voigt; I. Santibanez-Koref: Solving Assignment Problems by Selection Pressure Controlled Replicator Networks 274 A. Grzech: Structures. viour
Local Area Networks with Different Topological Analysis of Qualitative and Quantitative Beha-
278
E. Szlachcic: Bicriterial Optimization of Structure of Complex Network 282 V. Sima: OPTPACK - An Interactive Package for Personal Computers 2.
Optimization
Software
286
Simulation Techniques
2.1. Simulation of Discrete Systems M. Aicardi; F. Davoli; R. Minciardi: Approximate Performance and Sensitivity Analysis of Closed Queueing Networks 290 S. Vincze: Computer-Aided Asynchronous Synthesis Procedure
294
K. Irmscher: Performance vices in CIM Environments
298
Evaluation of Communication Ser-
11
2.2. Simulation of Continuous Systems G.C. Vansteenkiste /INVITED PAPER/: ches to Ill-Defined Systems B. Schmidt /INVITED PAPER/: Simulation System SIMPLEX II
New Simulation Approa-
Methodological Basis
of
the
302 309
A. Dzieliriski: Real-Time Simulation of Nonlinear Quadratic Gaussian Adaptive Control Systems 319 V. Ceric: Simulation of Complex Real Systems: Practice 2.3. Parallel Simulation D.J. Evans; G.M. Megson: A Systolic Extrapolation
Theory
and 323
Design
32V
S. Szejko: SLA - A Language for Simulational Evaluation of Concurrent Systems Performance 331 2.4. Software Support K. Wang; 0. BjjSrke: The Off-Line Motion Planning via the Computer Graphics Simulation System
335
2.5. Simulation Environment E.J. Kerckhoffs /INVITED PAPER/: Mini-Supercomputers: Perspectives in Scientific Computation and Simulation
New
350
Duan Ping: The Simulated Performance of a Real-Time InterEvent-Driven processor Synchronization Algorithm Based on Method 356 M. Flechsig; E. Matthaeus; V. Wenzel: Simulation Environment in SONCHES 360 M. Marx; R. Czerner: A Program Generator for a Model-Based Simulation System 365 N.E. Madjarov; St.B. Maleshkov: Software Package for Linear Nonstationary Systems Analysis and Simulation 369 Th. Schulze: SIMPC - An Implementation of GPSS for Personal Computer 373 3.
Methods of Knowledge Processing for Systems Analysis S. Tzafestas /INVITED PAPER/: Expert Systems in CIM Operations: Key to Productivity and Quality 379 F. StSnciulescu /INVITED PAPER/: Construction of a Knowledge Base for Simulation and Control of Large Scale and Complex Systems. Applications 387 A. Javor /INVITED PAPER/: Knowledge Based Inference Controlled Logic Simulation 397 Dj.B. Petkovski /INVITED PAPER/: Knowledge-Based Systems for Distributed Decision-Making 406 A. Lehmann /INVITED PAPER/: Knowledge-Based Modelling and Simulation: Restrictions, Alternatives and Applications 412 B. Boehme; Control
R.
Wieland; U. Starke: Knowledge Based Process
A. Ligeza; M. Szymkat: Computer-Aided Modelling 12
A Symbolic-Numerical Support
for
419 423
LIST OF AUTHORS Adams, R. H. Aicardi, M. Alder, J. Anishchenko,V.S. Anlan Song Apelt.D. ApeIt,E.
109 290 237 40 213 141 141
Bjarke,0. Boehme, B. Boettner, R. Bohlender,G. Borutzky, W. Bothe, H. G. Breitenecker, F. Bronisz,P.
335 419 159 105 113 42 232 59
Öeri/ t h e r e is moreover the architectural level r e p r e s e n t e d by t h e P M S l a n g u a g e . T h i s l a n g u a g e is n o t y e t i n c l u d e d in the HB language For the purpose of structural and functional simulation of d i g i t a l s y s t e m s , a p a r t i c u l a r s i m u l a t o r of e a c h D D L ( B O O L E A N e q u a t i o n s or F 7 7 ) d e s i g n e r d e f i n e d f u n c t i o n i s g e n e r a t e d a n d l i n k e d t o s i m u l a t o r l^OSIM T h e c o m p l e t e s y s t e m is i l l u s t r a t e d i n f i g 1. The verification of every DDL unit is performed separately secause it is inefficient within the structural and functional simulator and because the internal f a c i l i t i e s of DDL unit are not directly visible For that reason, the generated particular DDL s i m u l a t o r is l i n k e d t o a simulator frame developed for it. The designer view for this simulator w a s d e v e l o p e d a n a l o g o u s l y to t h e d e s i g n e r v i e w of K O S I M s i m u l a t o r . The DDL simulator includes time controlled data input like the k O S I M simulator. T h e H B D - s y s t e m r u n s on the computer K1B40 from ROBOTRON and applies operating system SVP1800.
?..
HB
language
description
T h e H B d e s c r i p t i o n l a n g u a g e is based on the NBS34 and KOSIM network d e s c r i p t i o n l a n g u a g e e x t e n d e d for d e s i g n e r d e f i n e d functions. It c o n s i s t o f a s t r u c t u r a l p a r t f o r d e s c r i b i n g n e t w o r k o f e l e m e n t s a n d several designer defined u n i t s and a f u n c t i o n a l part for d e s c r i b i n g these e l e m e n t s and these designer defined units included the pin functions of the circuit. W i t h i n the n e t w o r k e a c h e l e m e n t or u n i t is f e a t u r e d by i t s n a m e , i t s t y p e a n d i t s t e r m i n a l s . Optionally you can indicate parameters for instance time delay for the outputs. The s t a n d a r d d e l a y is o n e t i m e u n i t . All these i n f o r m a t i o n s have to meet the type declaration in the functional part T h e t e r m i n a l s can be busses, too. 1) Central Institute K u r s t r a s s e 33, B e r l i n .
of Cybernetics 1080, GDR
and
Information
Processes,
15
Here is an e x a m p l e (Fig. 2): a c o n t r o l l e d 8 bit circle counter. It c o n s i s t s of 2 elements: AND1 of type A N D and 0R1 of type OR and 2 shifter r e g i s t e r units SRI and S R 2 of type SR. First, there is a c o n n e c t i o n d e s c r i p t i o n for AND1. OR1, SRI and SR2 with the name of e l e m e n t or unit, c o n n e c t i o n list and type of element or unit Moreover, the structural description of the functions for the pins INP, TOR and RING by a table c o n t r o l l e d g e n e r a t o r INPUT of type T A B S D and for the pin TAKT by the clock generator LPER is following The types T A B S D and LPF.R are s t a n d a r d types of the K O S I M - l i b r a r y . Second, the e l e m e n t d e c l a r a t i o n s are following beginning with 'E: ' for the types AND, OR, T A B S D and LPER and the unit d e c l a r a t i o n for the type SR beginning with 'F ', latter are continued with the language type the functional description 'DDL'. In the case of B O O L E A N e q u a t i o n s units is c o n t i n u e d with 'BOOLE', of F77 units with 'F77' The syntax of all unit t e r m i n a l s is equal to that of the elements Input, o u t p u t and b i d i r e c t i o n a l t e r m i n a l s are following in turn A n o t h e r m o d i f i e d D D L d e s c r i p t i o n - a 4 bit counter d e s c r i p t i o n - is shown in fig 3. B O O L E A N e q u a t i o n s and r e g i s t e r t r a n s f e r s can be declared outside automata (global declaration) or within the automata (local declaration) Moreover, they can be declared within the automaton o u t s i d e the state d e c l a r a t i o n or w i t h i n the state declaration. A c c o r d i n g l y , the header example is e q u i v a l e n t to SR: < T I > TKT < T E > EIN- TOR, OUT. in D D L language. 3. D D L model and
of SR d e c l a r a t i o n
in HB language
of
this
simulation
The r e g i s t e r t r a n s f e r language DDL is based on the m u 1 t i a u t o m a t o n model (see fig 4). For the s i m u l a t i o n this model is t r a n s f o r m e d in to a single a u t o m a t o n model (see fig 5) which c o n s i s t s of a set of B O O L E A N e q u a t i o n s and a set of r e g i s t e r transfers. The set of B O O L E A N e q u a t i o n s is solved by m e a n s of static simulation. Incorrect cycles in the BOOLEAN equations are indicated. F e a t u r e s the s e p a r a t e D D L s i mu1 a tor are. - The s e p a r a t e D D L simulator frame is based on the developed DDL debugger Its commands are formally i d e n t i c a l l y with the d e b u g g e r c o m m a n d s of the used operating s y s t e m SVP 1800 Its r e s u l t s shown a c o m p r o m i s e b e t w e e n the s y s t e m d e b u g g e r r e s u l t s and the K O S I M s i m u l a t o r results. - R e f e r i n g to its d e f i n e d t e r m i n a l s there are only inputs, o u t p u t s and internal terminals but no b i d i r e c t i o n a l t e r m i n a l s The latters are n e c e s s a r y to link D D L units in a circuit network Therefore, they have to be considered It d e p e n d s on a u t o m a t o n state w h e t h e r the b i d i r e c t i o n a l terminal o p e r a t e s like an input or like an output. As an input the terminal is on the r i g h t side of a r e g i s t e r t r a n s f e r s or c o n n e c t i o n s o t h e r w i s e as o u t p u t on the left hand side. In dependence on g i v e n internal c o n d i t i o n s the b i d i r e c t i o n a 1 terminal in the o u t p u t mode is set to lower i m p e d a n c e o t h e r w i s e to h i g h e r impedance. - The i m p l i c a t i o n of the D D L s y s t e m clock ( c l o c k . ) is shown in fig. 6 (see a l s o part 4. ). E s p e c i a l l y if no s y s t e m clock is d e f i n e d the D D L unit is started by an event of its inputs and works autonomously as long as no internal state changes occur This i n t e r p r e t a t i o n is based on the p r o c e s s model from ISPS In order to reduce the description predefined operators are developed as usual in K O S I M and ISPS. In the HB s y s t e m there is a library of these p r e d e f i n e d operators. T h e s e o p e r a t o r s can be called as macros.
16
4.
HB
model
and
simulation
T h e H B s i m u l a t o r is b a s e d o n t h e e v e n t o r i e n t e d simulator KOSIM also. T h e e v e n t o r i e n t e d c a l l of t h e D D L u n i t w i t h i n K O S I M s i m u l a t o r s e p a r a t e d in a B O O L E A N e q u a t i o n call a n d in a r e g i s t e r transfer call accordingly fig 5 is s h o w n in f i g . 6. H e r e f o r t h e e v a l u a t i o n of the c o m b i n a t o r i c part f r o m fig 5 t h e B O O L E A N c a l l is p r e s e n t a n d f o r t h e e v a l u a t i o n of t h e s e q u e n c e p a r t t h e r e g i s t e r t r a n s f e r call. T h e v i s i b i l i t y of developed DDL debugger
the internal DDL f a c i l i t i e s ( s e e a l s o p a r t 3. ).
5
simulation
Design
strategy
and
is
obtained
by
the
dialogue
T h e H B s y s t e m c a n b e u s e d by t h e t o p douin d e s i g n (see fig 7) started for e x a m p l e on the r e g i s t e r t r a n s f e r l e v e l and s t o p p e d on the logic g a t e level. On this o c c a s i o n the D D L units are t r a n s f o r m e d into a n e t w o r k of l o g i c a l s t a n d a r d g a t e f u n c t i o n s . The HB s u p p o r t it.
6
dialogue
commands
are
deduced
from
this
strategy
and
Examples
T h e f o l l o w i n g e x a m p l e s c h a r a c t e r i z e the a p p l i c a t i o n v o l u m e of HB language: - counters, shifters - asynchronous automata. - a u t o m a t a net, coupled automata. - processors, multi processor systems and cellular automata.
7 /l/
/2/
/3/
/4/
/5/
/6/
References B a r b a c c i , M. R ; S i e w i o r e c k , DP.: "The design and analysis of instruction set processes" McGraui Hill, 1982 D i e t m e y e r , D o n a l d L. 470 " L o g i c D e s i g n of D i g i t a l Systems" Allyn and Bacon. Inc. . A t l a n t i c A v e n u e . B o s t o n 1971 D o n a t h , U. ; S c h w a r z . P. ; T r a p p e , P. "Dynamische Logiksimu1 ation auf Bitund Wortniveau" 19.Fachko11oquium Informationstechnik Dresden 1986 Issel, U et al "NBS-84: A structural description language for VLSI design", Circuit theory and design 8 5 , P r o c e e d i n g s of t h e 1 9 8 5 E u r o p e a n Conference, p 62-5 Sc hwar z , F "A p r o g r a m for the m i x e d - l e v e l s i m u l a t i o n of digital integrated circuits" Proc E C C T D ' 8 5 , P r a g 1 9 8 5 . p. 133-136 Siewioreck, D P et al . " P r o p o s a l for r e s a e r c h on DEMETER — a design methodology and e n v i r o n m e n t " R e s e a r c h R e p o r t No. C M U C A D - P 3 - 5 , Jan. 1983, CMU
8.
Figures
¡HB d e s c r i p t i o n : i s t r u c t u r a l and ! functional part V H B - c omp i 1 e r !HB f u n c t i o n a l p a r t ! with DDL units, ! BOOLEAN equations ! and F77 units ! ! > 1 ! >
!HB s t r u c t u r a l p a r t 1 with NBS84 included ! single BOOLEAN equations ! I ¡initialization i i of r e g i s t e r s a n d ! i memories» ! !test pattern, ! ¡simulator ! ! commands
units
V
V simulator
with
debugger
V results Fig. 1:
18
View
of
the
of
system
simulation
V
C I R C L E COUNTER: 2 DDL S H I F T R E G I S T E R S INCLUDED S T A N D A R D AND INPUT E L E M E N T S W I T H K O S I M INPUT L A N G U A G E L I N K E D TKT TOR OR_L + + INP
OR
RING
A N D — ' -I— ' VI I + + I AND T
! •»
EIN1 +
! SR 1 SR 'DDL' (RI. . R4)
SR I SR 'DDL' ! 'a ,* '1,2..,n ,
with the notation
n 2
f10"1)
which characterizes how uniformly distributed is H ^ ^ j (f(. ,z) ;iî) around its mean value. Entropie Distance. Given a map 1(.), we shall identify it with the map f z (-) whene ver the following condition is satisfied, that is < H b,c(z) ( f z ( - ) ; "' D ) - HcU(.);fl>2
i
ko2f
where k denotes a positive constant the value of which is chosen via practical considerations. • 10.3 Application to Potential Functions These results apply directly to the potential functions V(x,u) of the
catastrophe
theory, and one can then envisage the exhaustive study of the latter in terms of
entropy
of deterministic maps. 11.
CONCLUDING
REMARK
The present theory of entropy of deterministic maps provides new basic
approaches
to fuzzy sets, probabilistic sets, pattern recognition, logical inference, combination of evidences, approximate reasoning, and so on.
REFERENCES [1] [2] [3] [4] [5] [6] [7] [8] [9] 110] [11]
ACZEL, J; DAROCZY, Z; On Measures of Information and Their Characterizations, Acade mie Press, New York, 1975 GUY, A. G.; The scientific revolution of 1987. Abdication of Boltzmann's Entropy, Applied Physics Communication, Vol 7, No 3, 217-235, 1987 HAKEN, H.; Synergetics, Springer Verlag, New York, Berlin, 1978 HAKEN, H.; Advanced Synergetics, Springer Verlag, New York, Berlin, 1983 JUMARIE, G.; Subjectivity, Information, Systems. Introduction to a Theory of Relati vistic Cybernetics, Gordon and Breach, New York, London, 1986 JUMARIE, G.; New results on the information theory of patterns and forms,J. Systems Analysis, Modelling and Simulation, Vol 4, No 6, 483-520, 1987 JUMARIE, G.; A Minkowskian theory of observation. Application to uncertainty and fu zziness, Fuzzy Sets and Systems, Vol 24, No 2, 231-254, 1987 JUMARIE, G.; Relative Information. Theories and Applications, Springer Verlag, New York, London (to appear) ROSEN, R. ; On information and complexity, in Complexity, Language and Life: Mathematical Approaches; Biomathematics, Vol 16, 174-196, 1986 SCHUSTER, H.G.; Deterministic Chaos, Physik-Verlag, Weinhein 1984 THOM, R.; Structural Stability and Morphogenesis, translated by D.H. Fowler, Benjamin, New York, 1975
33
A ROUTE TO CHAOS Wolfgang Metzler Abstract. The route to chaos of bifurcation from a stable fixed Hopf bifurcation, quasiperiodic the iteration scheme tends to a Eiffel tower.
the coupled logistic map is studied. We observe a flip point to a stable period-2 orbit which is followed by a behaviour and periodic orbits. At the end of the route fascinating strange attractor looking like the
1. INTRODUCTION Chaotic behaviour of simple dynamical systems is today widely believed to model temporarily irregular phenomena in many fields of science [4, 8 ] , Driven by graphic computers, chaotic dynamical systems also deliver to us a large variety of new fantastic insights into mathematical structures (Julia sets, strange attractors; cf.[2,12,14,15,17]). This paper provides an example of a plane dynamical system where computer graphics have opened a wide field of interesting mathematical questions (stability, bifurcations and chaos) and helped to formulate hypotheses and to prove theorems. 2. THE MAP It is well known that the iteration scheme x
k+l
=
x
k
+
h x
k(1 " xk) .
h
> 0 .
(1)
which is Euler's method for solving the logistic equation x = x ( l - x ) , may be transformed into
V i = ruk0
-
Department of Mathematics, University of Kassel, P.O. Box 101380, D-3500 Kassel, F.R. Germany.
Relative to (4), M = {(x,y)e]R
| x = y} is an invariant manifold. Reduction of (4) onto
M results in the one-dimensional iteration scheme k+1
= x. + h (2x. -x.f)
(5)
which may be transformed into (2) by u^ = (h/ (1 + 2h) Jx^ and r = l + 2h. The dynamics of the coupled system (4) will be discussed in this paper. It is based on an unpublished note [1]
which has been the reason for a computergraphics film [16].
Related couplings of nonlinear oscillators have been studied by Hogg and Hubermann [9] as well as Waller and Kapral [18]. Kaneko [10] investigated the transition to chaos of a coupled logistic map with a linear coupling term. 3. PERIOD-DOUBLING Considering eq. (4) we find, as the parameter h is varied, that there are two fixed points given by E 1 = (0,0) ,
(6)
E 2 = (2.2).
Linearizing (4) in the neighbourhood of E p we obtain the eigenvalue equation DF h (0,0) - A I
1 + h-A
h
h
1+h-A
(7)
= 0 ,
where DF^ denotes the Jacobian matrix of the first partial derivatives of the'function F h (x,y) = [x + h ( x - x 2 + y),y + h ( y - y 2 + x)j defined by (4). Because of its roots
= 1 + 2h and
(8) = 1 the fixed point E^ is un-
stable for h>0. For E 2 we obtain the eigenvalue equation 1 - 3h - A
DF h (2,2)-Al
h
h 1 -3h - A
(2)
= 0
(9)
(2)
with the eigenvalues AJ '= l - 2 h and A£ '= l - 4 h . Obviously E^ is a sink only if 0 < h < 0 . 5 . For h>0.5 the period-1 cycle E^ is unstable. There is a critical situatic at h = 0.5. If h >0.5, then °h
:
= {
(
M
^
'
H
^
(10)
'
is a hyperbolic period-2 orbit. It is born in the equilibrium point E^ = (2,2) and moves with increasing h to E^ = (0,0) on a circle, which is given by (x - l) 2 + (y - l) 2 = 2. In particular we have 0 Q
5
= {(2,2)} and 0^ = {(0,0)}.
The first part of this statement on 0, follows by inserting x, h 3 1 1 \/ k y we get k = h~h ™ t o eq" k+i - M
and X
k+2 = y k+l = x k
aS Wel1
aS y
^
k+2 = X k+1 = y k
h
h (11) (12)
35
Furthermore, using (12), eq. (4) can e a s i l y be shown to be equivalent to 2 h y k + 2 h x k - hy k 2 - h x k 2 = 0
(13)
and t h i s proves the second part of the statement. Moreover, for 0 . 5 < h < 0 . 6 proved
i s a stable o r b i t . I f h > 0 . 6 , 0^ i s unstable. This can be
straightforward looking at F ^ x . y ) = x + h ( x - x 2 + y)
(14a)
+
(14b)
F2(x,y) = y
h(y-y
2
+ x)
and ^(x.y)
:= F 1 ( F 1 ( x , y ) , F 2 ( x , y ) )
F 2 ( x , y ) := F 2 ( F 1 ( x , y ) , F 2 ( x , y ) ) respectively, for x =
(15a) ,
(15b)
V2h-1 , y = £ - ^ V 2 h - l
.
A lengthy c a l c u l a t i o n y i e l d s F j i i . y ) = 5 - 1 0 h + 2h2 = ^ F 2 ( i , y )
^
(16)
and ^FjU.i)
= 4 h V 2 h - l - 2h + 2 h 2 ,
(17a)
^F2(x,y)
= - 4 h V 2 h - 1 - 2h + 2 h 2 .
(17b)
This leads to the c h a r a c t e r i s t i c equation 5 - lOh + 2 h 2 - A -4h V 2 h - 1 - 2h + 2 h
4h V 2 h - 1 - 2h + 2 h 2 2
5 - lOh + 2 h 2 - A
0,
(18)
equivalent to A 2 + (-10 + 20h - 4 h 2 ) A + (25 - lOOh + 100h 2 ) = 0. The eigenvalues of (18) are Ax
2
= (5 - lOh + 2 h 2 ) ± V 2 0 h 2 - 4 0 h 3 + 4h 4 2
3
(19) 4
Analyzing the roots of the radicant 20h - 4 0 h + 4 h , one can e a s i l y show that
and
|A.| < 1
if
0.5 < h < 0.6,
(20a)
IA.| = 1
if
h = 0.6,
(20b)
|A.| > 1
if
h > 0.6.
(21)
The considerations concerning ( y , x ) are analogous. Thus (20) and (21) prove the above s t a b i l i t y statements about the period-2 o r b i t 0^. 4. THE ROUTE TO CHAOS Now we i n v e s t i g a t e the system's behaviour in the range 0 . 6 s h s 0 . 6 8 4 . As shown in the section before, there e x i s t s a stable period-2 o r b i t f o r 0 . 5 < h < 0 . 6 . F i r s t we consider the b i f u r c a t i o n at h = 0.6. We know that the one parameter family F ^ : 3 F^ • F^ s a t i s f i e s
36
F h (x,y) = (x,y) for h > 0 . 5 with x = ^ + ^ V 2 h - l and y = ¿ - £ V 2 h - l . Its Jacobian h h trix D F ^ x . y ) has two non-real eigenvalues A^ and A^ satisfying |A^| < 1 for h < 0.6 and |A^| > 1 for or h > O.i 0.6. Also we have d |A^(h) dh
10 > 0 .
(22)
h = 0.6
These three properties of the map (4) and further numerical studies (cf.[14]) indicate a Hopf bifurcation at (see [7], p. 162) h = 0.6. Accordingly for (y,x).
Fig. 1. Route to cnaos (after [15]): (a) Stable loops (h = 0.614), (c) Period-26 attractor (h = 0.678),
(b) Overlapping loops (h = 0.66), (d) "Eiffel tower" (h = 0.684). 2.1
V V^
. .
\ ••
- •• v
V •'
3
1.9
2.1
1.9
Fig. 2. The attractor of the map (4) at h = 0.684 with a blow up of the squared region [1.9,2.1] x [1.9,2.1],
37
Next we outline some computational results about the map (4) in the range 0 . 6 s h s O . 6 8 4 . These calculations have been performed on a graphic computer with double precision arithmetic. First we iterate the map 10.000 times to avoid transients. For h > 0 . 6 the periocH2 orbit blossoms out into two stable loops (Fig. la). The loops grow by turn with various periods, overlap themselves (Fig. lb) by turn with other periods (Fig. lc) and finally reach the structure of the Eiffel tower (Fig. Id). A blow up of the Eiffel tower's topstar (Fig. 2b) delivers a fractal structure which indicates sensitive dependence on the initial conditions. Indeed, one computes a positive characteristic exponent (cf. [6])
for h > 0.651, as shown in Fig. 3. Increasing
the values of h, the chaotic attractor of the map (4) continuously changes its shape, interrupted by different periodic orbits neglected in Fig. 3. The attractor's final shape at h = 0.684 is shown in Fig. 2. It is followed by escape which can be observed for h > 0.686.
Fig. 3. Spectrum of the characteristic exponents lim 1/n log
( X q , ) | for
0.6165 < h S 0.685 and (x 0 ,y Q ) = (0.4,0.5). REFERENCES [ 1] Beau, W., W. Metzler and A. Überla: The Route to Chaos of Two Coupled Logistic Maps. Preprint (1986). [ 2] Beau, W., W.H. Hehl, W. Metzler: Computerbilder zur Analyse chaoserzeugender Abbildungen. Informatik Forsch. Entw. 2 (1987), 122-130. [ 3] Collet, P. and J.-P. Eckmann: Iterated Maps on the Interval as Dynamical A. Jaffe and D. Ruelle (eds.). Birkhäuser, Basel/Boston/Stuttgart 1980. [ 4] Cvitanovic, P. (ed.):.Universality in Chaos. Adam Hilger Ltd., Bristol
Systems.
1983.
[ 5] Feigenbaum, M.: The Universal Metric Properties of Nonlinear Transformations. J. Stat. Phys. 21 (1979), 669-706. [ 6] Feit, S.D.: Characteristic Exponents and Strange Attractors, Commun. math. Phys. 61 (1978), 249. [ 7] Guckenheimer, J., P. Holmes: Nonlinear Oscillations, Dynamical Systems and Birfurcations of Vector Fields. Springer, New York/Berlin/Heidelberg/Tokyo 1986 2 . [ 8] Haken, H. (ed.): Evolution of Order and Chaos in Physics, Chemistry, and Biology. Springer, Berlin 1982. [ 9] Hogg, T. and B.A. Huberman: Generic Behavior of Coupled Oscillators, Phys. Rev. A 29 (1984), 275.
38
[10] Kaneko, K.: Transition from Torus to Chaos Accompanied by Frequency Lockings with Symmetry Breaking. Prog. Theor. Phys. 69 (1983), 1427. [11] Li, T.Y. and J.A. Yorke: Period Three Implies Chaos. Amer. Math. Monthly (1975), 958-992. [12] Mandelbrot, M.S.: The Fractal Geometry of Nature. Freeman, San Francisco 1982. [13] May, R.B.: Simple Mathematical Models with very Complicated Dynamics. Nature 261 (1976), 459-467. [14] Metzler, W., W. Beau, W. Frees, A. Überla: Symmetry and Self-similarity with Coupled Logistic Maps. Z. Naturforsch. 42a (1987), 310-318. [15] Metzler, W.: Chaos und Fraktale bei zwei gekoppelten nichtlinearen Modelloszi11atoren. PdN Physik 7/36 (1987), 23-29. [16] Metzler, W., W. Beau, A. Überla: A Route to Chaos. Computergraphies Film. Inst. 'f.d. Wiss. Film, C 1641, Göttingen 1987. [17] Peitgen, H.O., P.H. Richter: The Beauty of Fractals. Springer, Berlin/Heidelberg/ New York/Tokyo 1986. [18] Waller, H. and Kaprai, R.: Spatial and Temporal Structure in Systems of Coupled Nonlinear Oscillators. Phys. Rev. A 30 (1984), 2047.
39
BIFURCATIONS OF TWO-DIMENSIONAL TORI AND CHAOS IN DISSIPATIVE SYSTEMS V.S.Anishchenko, T.E.Vadivasova, M.A.Safonova Transition to dynamical chaos in different distributed and multidimensional systems is often preceded by a quasiperiodic motion bifurcations. In simplest case, chaos arises via distinction of two-dimensional torus (Tg) . This communication represents the results of computer and physical experiments on the investigation of torus distinction regularities, mechanisms of appearance of quasiattractors (CA.j) and their characteristics in different flow and diskrete systems. The following systems realising regim of quasiperiodic oscillations were investigated: driven generator with inertial nonlineority, two coupled generators and discrete system of coupled Feigenbaum maps. Numerical simulation was carried out with help of computer programs permitting to calculate the lines of limit cycle bifurcations on the parameter plane, various dynamical and statistical characteristics of oscillation regimes: Poincare section, probability distribution density, power spectra, Lapunov characteristic oxponents. Random noise generator was added in numerical scheme for fluctuation excitation simulation. Universal character of following bifurcation mechanisms of transition T CA^ predicted by torus distinction theorem was confirmed £2,3] 1. loss of smoothness and break-down of ergodic torus with soft appearance of torus-attractor. Smoothness loss effect can be preceded by the ergodic torus period-doubling, 2. Loss of smoothness of torus with resonance structure on it, its distinction on the resonance cycle stability loss line and appearance of chaos within looking region through Feigenbaum sequence of perioddoubling or through emergence of new torus and its following breakdown. 3. Hard appearance of torus-attractor via the saddle-node bifurcation on unsmooth torus or after its distinction. 4. Torus with resonance structure breaks also on tho line of homoclinic tangency of resonance saddle cycle manifolds. However, appearing hyperbolic subset of trajectories is nonattracting. Numerical experiments permitted to establish the role of natural fluctuations in situation under consideration. Noise influence on system dynamics in the regime of stable limit cycle with homoclinic structure in its neighbourhood leads to chaos arising. For example let us regard the results of numerical analysis of discrete system [3]
*~5 ' Saratov State University, Departaraent of Physics, SU-410601.Saratov, USSR 40
which simulates dynamics of two coupled generators [ 2 ]
.
Bifurcational diagram on parameter plane in neighbourhood of phase locking with winding number 0 = 2:5 is represented in Fig. 1.
Pig. 1. 1 0 - the Hopfe bifurcation line; li - the saddle-node bifurcation line; l* - the invariant manyfold homoclinic tangency line; lcr - the boundary of C A t - region; O - multipliJ cator of limit cycle. The digits on Fig. 1. indicate the motion directions on parameter plane where the bifurcational mechanisms mentioned above are realised. Regularities of transition "torus-chaos" predicted by one-dimensional circle mapping theory are regarded for flow systems. Quantitative relations in oscillation power spectrum at the T 2 -distruction moment are investigated in physical and computer experiments. Universal regularity in frequency distribution of spectral lines proved for model circle mapping are confirmed. Good agreement with one-dimensional theory results is obtained for fractal dimension of the set of irrational values of winding number near the torus break-down line. References [1] Anishchenko V.S.: Dynamical Chaos-Basic Concepts. Teubner-Verlag, Leipzig 1987. (2J Afraimovich V.S», Shilnikov L.P.: Invariant two-dimensional tori, their distruction and stochastisity. Methods of qualitative theory of differential equations. University Gorky, 1983, 3-25. [3] Anishchenko V.S.: Distruction of quasiperiodic oscillations and chaos in dissipative systems. Journal of Technical Physics, ¿6, (1986) 2, 225-237.
41
SHAPE AND DIMENSION OF CERTAIN HYPERBOLIC INVARIANT SETS H.G. Botho In many casas the evolution of a system whose state in each moment is determined by a point in a phase space P can be described by a dynamical system on P, i.e. by a family ^ l t f T of mappings vf" : P — > P depending on a time parameter t belonging to a set T which consists either of all real numbers, all non-negative real numbers, all integers or all non-nogative integers. These mappings satisfy . . -f S+t = Kf oSy t f O - id, where o denotes the composition of mappings. If at the beginning (i.e. at time 0) the system has the state corresponding to the point pQ in P, then its state at timo t corresponds to the point ^ (P 0 )' and the whole evolution of the system is described by the orbit W^Po^teT
startin
9
at
P0-
If a dynamical system •W^tj.T- o n a space P is given then it suggests itself to look for invariant sets A in P, i.e. for sets is an attractor, i.o. if for each satisfying (A) = A (t£T). If A point p in P sufficiently close to A we have
increases. Before analyzing this matter of fact, for the sake of completeness we compile all system parameters, constants and coefficients of (2.5) in Table 1. Via two pollution functions ^(ij)) and e[0,1.5] (df. [4], p.168), i.e., l=l(t,i|>) and a=a(t,t). The first quadrant D=IR*x]R* is an invariant set for the flow of (2.5). In D, (2.5) has two fixed points: (0,0) is asymptotically stable for all values of \|ie[0,1.5], and the nontrivial
49
c1 = 7[a]: average lifespan of needles Cj = 16[kg ASSl/(kg LEAF-a)]: optimal net photosynthesis rate c, = 241[kg H,0/(kg ASS1)]: transpiration coefficient c, - 2000[kg H,0/(kg ROOT-a)]: water supply coefficient cs = 6[t ASSI/(a • ha)]: growth of wooden biomass pollution functions: constants:
*!(*) = O.S^O.ô-O.A*): efficiency of needles = »(l+i»)" 1 : proportion of damaged needles • *c[0,1.5]: pollution parameter coefficients: c^jc,,"1*^ • (c^'CsO.Oô + c,'1! + yu = (1 - CjC^Oc^! - C!/6, Xi = 0.06y3c1"1c5, ». = -yiy,.
x, = -y^s,
*5 = yiy..
) + (1/6+ cs) =: y ^ y j .
y5 = », = -y,y5 .
*« = ys(y.-y,).
= y:y.-
Table 1. System parameters of (2.5), y 3 free for choice (cf. (2.3)) equilibrium ( l 1 , a 1 ) = (11(«|i),ai_(iti))1n the interior of D is an unstable saddle point for all parameter values with (1 x(4i),a1(4>))ED. After [5] there exists i|>0, 0> variations of the free parameter y 3 (as well as c2
and c 5 ) only change the position of
(l^ax).
1=0
Fig. 2. Decomposition of the phase space Fig. 3. Limiter functions Li (3.1) and L 2 of (2.5) for fixed E is attract((3.2), a = 0.5, b = 0.3) for ed by the origin (after [5]). xe[-0.5,1.5].
3. THE LIMITER-FUNCTION CONCEPT We will focuss now on another basic principle of forest ecosystems modelling turning out to be responsible for the distinct behavioural modes mentioned above. It is often useful to compare the resource demand of a specific internal process (e.g. assimilate demand for growth) with the appropriate supply (photosynthetic production). If the demand can be matched, the process will work unaffected. In case of insufficient supply the activity of the process must be diminished proportional to the supply demand ratio, while it doesn't make sense to increase the activity if the demand is oversupplied. Therefore, the supply demand ratio is restricted artificially to values lower than 1 before using 50
it to control the process. Additionally, the ratio is often restricted to positive values. The simplest approach to realize this kind of restriction is the LbniteJi-¿unatton M x )
:=
0 if x< 0 x if x e [ 0 , 1 ] 1 if x> 1
.
(3.1)
As shown in Fig 3, it is also possible to apply the more smooth function L 2 (x) := 1/(1 + exp(-2*(x-a)/b))
(3.2)
with real parameters a,b. KRIEGER [2] shows that the complex forest ecosystem model described in [1] can be reduced to a two-dimensional, parameter-dependent model (3.3) containing only one simple-type limiter-function and showing qualitatively the same behavioural modes as the complex version (cf. Fig. 1(a),(b)) dl dr
=
- c 1 - » I - ( c 2 + » ; ) - l 2 + c 3 -(c u -t 1 -L 1 (r-(c 1 -art of the problems the decision situation is stochastic-like, i.e. some probability distribution can be assigned to each of the alternatives presuming a /subjective/ probability field (fi,B,P). For a given E the risk-event a €Q is the following 55
where a^j
is the a l t e r n a t i v e is the c o n d i t i o n a l
to be r e a l i z e d . T h u s
p. = P ( a . . , a ) = P ( l s . - s. .I>6 e' -j -jt J
P: A x Si c.)The
decision,
After having
> V = [ 0 , 1 ] £ R
the
solved
the a l t e r n a t i v e s ,
decision
should
would be his
decision.
If A is an o r d e r i n g X: Z x
(17)
make a
relation
/subjective/
T h e DM's
— — >
>
t h e set of
s t r a t e g i e s of t h e i - t h D M w h i c h m a y be u s e d in the c a s e of
o o 1 li o a e C !-• h- n> p B 1
w s t-1 M s ö g
•t) CÖ > "Li
l
65
As for the g e n e r a l technology, it concerns a class of r e a l - w o r l d problems i n the m a n a g e m e n t and p o l i c y analysis area. The specific DSS applications are c o n c e r n e d w i t h information and computer resource allocation p r o b l e m s . A n u m b e r of s u c h problems, r e l a t e d to analysis of r e g i o n a l (local) and sectoral p o l i c y of c o m p u t e r and i n f o r m a t i o n resource a l l o c a t i o n has b e e n performed. T o r s u p p o r t i n g all m o d e l l i n g , s i m u l a t i o n and v a l i d a t i o n functions of the
entity
of specific DSS a m o d u l a r software system has b e e n developed. SIMOS I consists of f o u r s t a n d a r d p r o g r a m packages
(discriminant analysis, ILP, G P S S , linear r e g r e s -
s i o n analysis), a centralized i n f o r m a t i o n b a s e , and a module of special p u r p o s e p r o g r a m s . T h e i n f o r m a t i o n b a s e includes d a t a and i n f o r m a t i o n c o n c e r n i n g the stated r e a l - w o r l d problem, model d e s c r i p t i o n d a t a , s i m u l a t i o n results
, validation data,
as w e l l as all m o d e l l e d and simulated alternatives, p r e s e n t e d explicitly. Designed i n such a way SIMOS I can b e considered as a step toward a n integrated m o d e l l i n g environment. A I - S U P P O R T IN MBDSS D E S I G N AND
IMPLEMENTATION
A r t i f i c i a l Intelligence (Al) is the subfield of the c o m p u t e r s c i e n c e w h i c h has p e r h a p s the greatest potential w i t h respect to d e c i s i o n m a k i n g . Expert i n c o r p o r a t e d i n MBDSS can assist a number of tasks. The so-called
systems,
"semi-expert"
systems /2/ combine the p o w e r of the expert system technology and the k n o w l e d g e a l r e a d y available - about problems w h o s e solution requires e x t e n d e d support, as this t e r m was described above. T h e p r a c t i c a l r e a l i z a t i o n of s u c h systems w i l l a l low to switch from the q u e s t i o n "what...if" to " w h a t ought to be" in areas w h i c h are challenge for technologies such as D S S , modelling, s i m u l a t i o n ajid A I . I n F i g . 1 are p o i n t e d out the stages of the procedure w h i c h c a n be supported b y a semi-experted s y s t e m (SES). S o m e SES components are included in SIMOS I. CONCLUSION The p r o c e d u r e for MBDSS, c o n s i d e r e d b r i e f l y above, is consistent w i t h the requirements of the concept of f l e x i b i l i t y of d e c i s i o n options and those of the v e r s a t i lity a p p r o a c h /1 / . If applied adequately this procedure c a n provide extended s u p port to ill-defined problems solving. REFERENCES 979)1. B o n d e r , S.: C h a n g i n g the future of operations r e s e a r c h . Opns Res 2J_ 2 . Keen, P.G.W.: D e c i s i o n support systems: the next d e c a d e . D e c i s i o n S u p p o r t S y s t e m s , ¿ (1987), 33. Keen, P.G.tf. and M.S. S c o t t Morton: D e c i s i o n S u p p o r t S y s t e m s : A n O r g a n i z a t i o nal P e r s p e c t i v e . A d d i s o n - W e s l e y , R e a d i n g , MA, 1978. 4 . S p r a g u e , R.H. and E.D. Carlson: B u i l d i n g Effective D e c i s i o n S u p p o r t S y s t e m s . P r e n t i c e Hall, Englewood Cliffs, N.J., 1982. 5. S t r e z o v a , Z.: T e c h n i c a l Reports IR 109, 111, 115. SC MIS, 1981-83. 'in B u l g . ) 6. S t r e z o v a , Z.: A n approximate a p p r o a c h to the d e s i g n of d e c e n t r a l i z e d m a n a g e m e n t systems structures. P r o c e e d i n g s of the 8th IFAC C o n g r e s s , v o l . 7, P e r g a m o n Press, 1982. 7. S t r e z o v a , Z.: V a l i d a t i o n in ill-defined problem m o d e l l i n g : resource a l l o c a t i o n p o l i c y analysis. Proceedings of the 5th IFAC/lFORS Dynamic M o d e l l i n g Conference. P e r g a m o n P r e s s , 1;)86.
66
KNOWLEDGE-BASED DECISION AID IN TEXTILE TECHNOLOGY *. ) Dr.rer.nat Gert Kreiselaeier
*. ) , Dr.-Ing.
Roland
Seidl
1. DECISION SITUATIONS IN TEXTILES Textile
technology
Technicians a
huaan
i n t e g r a t e s the skill of
» a n y special
disciplines.
and the o p e r a t o r s of t e x t i l e p l a n t s fit the d e f i n i t i o n
expert
co»pletely and provide
the
potential
for
textile
t e c h n o l o g y to take a d v a n t a g e of k n o u l e d g e - b a s e d c o » p u t e r p r o g r a m s . capture
a n d w i d e l y a p p l y the e x p e r i e n c e a n d e x p e r t i c e of t h i s
is the a i l of d e v e l o p i n g a u t o m a t i c d e c i s i o n To u n d e r s t a n d in
the
To
experts
aid.
in which r e s p e c t t h e s e c o a p u t e r p r o g r a a a s c a n be
field
of
of t e x t i l e s suppose uhat the f o l l o w i n g t a s k s
useful have
in
coaaon? d i a g n o s i n g fabric or a a c h i n e
faults,
c h o o s i n g an a p p r o p r i a t a a c h i n e to p r o d u c e a d e s i r e d
fable,
c o n f i g u r i n g the aany c o a p o n e n t s t h a t Bake u p a c l o t h i n g In
all
cases
singulare experts
decision
data do
is
nade
and i n f o r a a t i o n .
well.
in a c o a p l e x This
As
experts
s p r e a d out tions
This
Although
involved
the n e e d for soae d e c i s i o n s u p p o r t
is due to the fact that is rare,
is r e q u i r e d
coaputers
process
t h u s c o a p u t e r p r o g r a m s which do this c o n t i n u e
huaan expertise
huaen expertise
feu
people
thuab.
in the field of t e x t i l e s face a more c o a p l e x a n d
urgent and
aany a
Their reasoning
includes use of judgaent a n d r u l e s of
w o r l d w i t h i n which to o p e r a t e , becoaing
froa only
E x p e r t s are people who can do t h i n g s o t h e r
c a n n o t do b e c a u s e of t r a i n i n g a n d e x p e r i e n c e . in e a c h c a s e
situation
is a t a s k s that
for the c o n s i d e r e d
in h i g h d e a a n d a n d
in reaote
tools
in » a n y t e x t i l e p r o c e s s e s , e x p e r t s a b i l i t y to
there
become
often
indespensible
is c o n t i n u e d r e l i a n c e on
identify a n d s y n t h e s i z e d i v e r s e
fora j u d g e a e n t s , e v a l u a t e a l t e r n a t i v e s a n d to a i d
*. ) F o r s c h u n g s i n s t i t u t
applica-
expensive,
areas.
a n d 3 i a u l a t i o n a o d e l s have
huaan
is to
für T e x t i 1 t e c h n o l o g l e
factors,
the to
derisions.
Kar 1 - H a r x - S t a d t ,
DDR
67
the s t r a t e g y
it f o l l o w s
Is c a l l e d the control, s t r a t e g y .
is to scan t h r o u g h the r u l e s until one facts and
in the d a t a base, the
scanning
different
strategy
This
strategy
is f o u n d w h o s e c o n d i t i o n s
the r u l e is a p p l i e d ,
resuaes.
One
is
known
as
is to s e l e c t a goal to be a c h i e v e d a n d
t h e r e are no r u l e s to e s t a b l i s h the new s u b g o a l , for the n e c e s s a r y
strategy We
is known as
scan
that
over an e x c h a n g a b l e
the rule base at
a
hypothesis.
This
situation
general
characteristics.
If the
rule
This
on
interpreter
in
r u l e - b a s e a n d the b a s i c C13.
base.
This
that
is
PROLOG
facts s u p p l i e d by the to
establish
the
basis
of
infor»ation
i n f o r m a t i o n will be a c q u i r e d
is is,
user
truth
p r o b l e m can be u s e d to c l a s s i f y an the
which
idea of w h i c h
The e s s e n c e of the p r o g r a m
is u s e d to e x a m i n e
conclusion,
event
or
goal.
the p r o g r a m a s k s
facts a n d e n t e r s t h e » in the d a t a
a c c o r d i n g to S t e r l i n g a n d S h a p i r o arrive
A the
backward-chaining.
constructed a backward-chaining
operates
base
forward-chaining.
rules to find t h o s e w h o s e c o n s e q u e n t a c t i o n c a n a c h i e v e the user
Batch
u p d a t i n g the d a t a
to
of
a
objekt,
about
its
sequentially
in
d i a l o g with the user. The
s y s t e m a s k s q u e s t i o n s of the user
which
cannot
information.
be found
about
in the k n o w l e d g e
"primitive"
information,
base or d e r i v e d
The user can r e s p o n d to such q u e s t i o n s
in two
from
by s u p p l y i n g the r e l e v a n t
i n f o r m a t i o n as an y e s / n o - a n s w e r
q u e r y , c h o o s i n g an a n s w e r
from a menue or t y p i n g a n u m e r i c a l
- ask the system, why this The into
latter o p t i o n
is useful
the s y s t e m s c u r r e n t
information
that the s y s t e m s q u e r y a p p e a r s the From
query
The user can ask
irrelevant,
would r e q u i r e a d d i t i o n a l
the system
is a s k i n g
that
information.
Suppose,
than 5 k n i t t i n g - s y s t e m s
for
kind
or
"why"
value
the
is worth the e x t r a e f f o r t of
r e q u i r e d ?".
may type
user.
informaobtaining
is a s k i n g
Then the user,
cases
answering
judge wether the
for e x a m p l e , the s y s t e m
insight
in
in c a s e s that
e f f o r t on the part of
the c o n s e q u e n c e s of a yes or no r e p l y , uhat
the
is needed.
the s y s t e m s e x p l a n a t i o n s the user will
tion
to
in order to e n a b l e the user to get
intentions.
other
ways:
"are
not yet
more
knowing
"why" r e p e a t l y to
of k n i t t i n g m a c h i n e the s y s t e m has c h o o s e n as its
see
actuall
h y p o s e s i s. O n c e the s y s t e m has come up with an a n s w e r to the users q u e s t i o n , user may answering
like to see h o w t h i s c o n c l u s i o n was r e a c h e d . 3uch
a how-question
supplied primitive
68
information
the
A p r o p e r w a y of
is to d i s p l a y the s u b g o a l s from w h i c h the c o n c l u s i o n was
and
user-
reached.
2. E X P E R T S Y S T E M S P U T E X P E R I E N C E IN T H E P R O D U C T I O N In recent years, research •ade many been
An
in the field of artificial
important successes.
the
"Expert
developement
"Expert
require
of powerful
system"
new computer
is a computer programm
procedures
to
significant
human expertise
the knowledge base of the
inference engine
the explanatory
that uses
for
important respects. pertinent
general general
solution.
rules
In a conventional
computer p r o g r a m s
computer program,
to the problem and means and methods Intermixed,
so that
for
modify the
knowledge about the problem and the methods for a p p l y i n g
the
In an expert system there
The systeat can be changed by adding
3. O N E POPULAR APPROACH FOR KNOWLEDGE
REPRESENTATION
researchers have worked out a variety of m e t h o d s
and U3ing knowledge research
expert knowledge
in computers. This
in AI circles.
for
representing
is still a matter of d e b a t e and
One popular approach
is to use situation-action
can be connected to each other to form rule
to
which
networks.
such networks provide a s u l t i b l e leans for
ting knowledge,
required by the textile expert to decide
situation.
set of rules
The
represent
or production r u l e s ,
Once assembled,
is often referred to as the
represen-
In a certain "rule-base"
part of the program that d e c i d e s which rule to apply as
interpreter" or task
in
know-
utilizing
it is difficult to
base.
The
to main
component.
knowledge to the problem.
the
The
reasoning
or subtracting rules in the knowledge
active
and
enough
is a clear separation of
program.
"rule
as
inter face(why/how-component)
this knowledge are all
and
knoun
knowledge
complex
their
expert system differs from more conventional
several
AI
has has
are:
facts and for
the knowlede acquisiton
the
software
solve poblems that are
components of an expert system
ledge
intelligence
Among the most significant of these
systems".
inference
An
PROCESS
of the
inference
the
engine.
inference engine
Is to decide which rules to
apply,
69
The
diagnosis-shell
WIDIHO is in practical use on 16-bit PC for
the
following problems: - selection of Suitable knitting machines on the basis of desired properties of # the textile fabric; - »achine fault diagosis on multi shed weaving looms on the basis of information about elemantary defects; - selection
of suitable standards for special textile testing
problems; - interpretation of faults in textile fabrics during finishing processes; and - predetermination of properties in yarn production. These
problems are complicated enough to justify developement and use
of expert systems.
4. DECISION AID FOR DIAGNOSING FAULTS ON WEAVING LOOMS One major problem In de3iging an expert system is to find a method for the
acquisition and modelling of expert knowledge and for effectively
overcoming
the
psychological
"knowledge engineering
and analysis of expertise are important. to
bottleneck".
aspects of knowledge ellcitatlon,
extract
knowledge
from
textile
In
this
field
interview techniques
We used interview techniques experts
and
formulated
und
structured this knowledge in knowledge rules about weaving looms. knowledge space,
has
been represented as a hlerachleally
organised
search
as an AND/OK tree. The body of this tree is formed from inter-
related rules with the facts occupying terminal positions. ledge
base
fault
occurs.
150
rules.
This know-
is consulted by WIDIMO so that advice can be given
if
At present the knowledge base contains 90 objects
This
fault
diagnostic system beats
the
human
tenance
can
be put into effect.
less skilled staff is available or if the plan,t Is working
KEFEKENCES £1] Sterling, L., E. Shapiro: The Art of Prolog. The HIT Press, Cambridge, Mass., London, 1987.
70
main-
It al3o can be effectively used
reduced personnel.
a and
textile
experts in speed and availability so that a new area of machine only
The
if with
A MULTI-OBJECTIV DECISION SUPPORT SYSTEM FOR TOP MANAGEMENT (DSS-CAPS)
THE
Dipl.-Ing. Rolf Schmidt,VEB Trafowerk "K.Liebknecht", M a r k e t i n g , 1 1 6 0 B e r l i n , W i l h e l m i n e n h o f s t r . 83-86 D r . - I n g . B e r n d K o c h . A d W d e r D D R , Z e n t r a l i n s t i t u t für K y b e r n . u n d I n f o r m a t i o n s p r o z e s s e , 1 0 8 6 B e r l i n , K u r s t r a P e 33 Introduction: The c o n t i n u o u s s a t i s f a c t i o n of d e m a n d s of the national economy and international market requires a higher quality of m a r k e t i n g . In c a s e o f s u p p l y m i s s i n g (material, cooperation) o f t e n the e x a c t l y o b s e r v a n c e of the production and sales p l a n e is n o t p o s s i b l e in a l l d e t a i l s . By m e a n s of the programsystem CAPS, an i n t e r a c t i v e c o m p u t e r a i d e d production and sales decision support system (DSS-CAPS) using the a l g o r i t h m of R E H / S t r a u b e l - 8 5 - 1 / , / S t r a u b e l - 8 6 - 1 / , it is p o s s i b l e to c o m p e n s a t e or b e t t e r to r e d u c e d e v i a t i o n s of the p r o d u c t i o n p l a n a n d its c o n s e q u e n c e s o n the f u l f i l m e n t of the o v e r a l l t a r g e t s of the e n t e r p r i s e . Description: The f u n d a m e n t a l a i m w a s to c r e a t e a s i m p l e and robust model of p r o d u c t i o n a n d sale. The computer based system ha6 to be e a s y by u n d e r s t o o d a n d a p p l i e d by the user and particularly by the (top) m a n a g e m e n t . Paet experience hase s h o w n us t h a t e v e n the b e s t m o d e l / s y s t e m w i l l n e v e r be used if it is n o t c o m p r e h e n s i b l e a n d if the user ( top management ) c a n n o t g r a s p its f u n c t i o n . The above mentioned system is u s e d for d e c i s i o n m a k i n g for the direction of p r o d u c t i o n a n d d e l i v e r y w i t h i n the c u r r e n t y e a r of planning. The s t a r t i n g p o i n t for the c u r r e n t y e a r is t h a t the production a n d d e l i v e r y p l a n s are f i x e d f o r the c o n s i d e r e d period of time. The r e l a t i o n c a n be i l l u s t r a t t e d thus: Balance = Plan = Contracts
Pic. Rl.l.
(delivery).
Modelpart I- Data bank system of
RABS1
71
The production plan as well as the stocks which result from that are fixed regarding time and the stocks are largly linked with contracts. The thus chosen solutions for production and delivery of goods present efficient alternatives from computer- aided and determined compromise solutions ( illustration Rl.l and R1.2). The respective complexes and the existing connections and their interdependences will not be described in this article. As for the modelpart RABS1 we refer to /Schmidt-871/. The applied basic model is elucidated in pictures Rl.l and R1.2.
Pic. R.1.2.
Modelpart II of DSS-CAPS
Picture Rl.l illustrates the flow of information within operative working of customers'orders. The results of this model form the bases (a priori information) for the decision support system CAPS. The production managemant. is faced with the task of making optimum use of the existing production capacity.That means to realise the plannend production targets , which are assigned by top management, with a minimal expenditure of material, working time and man-power. A necessery consequence of this is , among others, the reduction standstills of machineries and its most efficient employment especially with regard to optimal production quantity. The 6ales management is confronted with the task of satisfying the needs of the market in such a way that the position in the market can be guaranteed in future and its own share in the market can be increased lucratively and profitably. The application of the DSS-CAPS realises for the first time the coherence or better interdependence of targets of pro-
72
duction w i t h t h o s e of t h e s a l e s m a n a g e m e n t . The production realisation according to the g i v e n p r o d u c t i o n p l a n depends mainly on the r e a l p r o d u c t i o n c a p a c i t y at the respective period of time . Unintentional production deficits can be caused for e x a m p l e b y s u p p l y w h i c h h a s gone missing, been d e l a y e d o r w h i c h is of p o o r q u a l i t y b o t h f r o m t h e h o m e m a r k e t and abroad and by unintentional demage6. This i n f l u e n c e s o n the p r o d u c t i o n p r o c e s s c a n a f f e c t t h e temporal c o n d i t i o n s as w e l l as s h i f t s in the a s s o r t m e n t of t h e p r o d u c e d g o o d s . As a r e s u l t n e g a t i v e e f f e c t s c a n c o m e into existence as to the r e a l i s a t i o n of c o n t r a c t s and the t o t a l t u r n o v e r - p o l i c y o f the p r o f i t - s e e k i n g e n t e r p r i s e . The f u n d a m e n t a l r e q u i e m e n t of t h e e n t e r p r i s e h o w e v e r reads : f u l f i l l m e n t of the m o n t h l y a n d c u m u l a t i v e e c o n o m i c t a r g e t s b y a p p l i c a t i o n of the m i n i - m a x - m e t h o d . By m e a n s of C A P S it is p o s s i b l e to c o m p e n s a t e or better to r e d u c e d e v i a t i o n s of the p r o d u c t i o n p l a n a n d its c o n s e q u e n c e s on the f u l f i l m e n t of the o v e r a l l t a r g e t s o f the enterprise within the current p e r i o d of p l a n n i n g ( c u r r e n t year of planning). We proceed from the idea t h a t the production management is a b l e to o f f e r v a r i o u s a l t e r n a t i v e s of product i o n for the r e s p e c t i v e types o f p r o d u c t s ( e i t h e r - or ). The c h o i c e of an e f f i c i e n t p r o d u c t i o n a l t e r n a t i v e c a n n o t a n d must n o t d e c i d e d by the p r o d u c t i o n roanagemant. It h a s not got either the n e c e s s a r y s u r v e y n o r the insight into the complexity of marketing processes. From this follows the necessity for the p r o d u c t i o n m a n a g e m e n t as w e l l as the sale management to suggest v a r i a n t s of decisions to the top m a n a g e m e n t of the e n t e r p r i s e . The v a r i a n t s s h o u l d be p a r e t o optimal solutions of a set of c o m p r o m i s e p o i n t s which are calculated by m e a n s of CAPS o n the b a s e s o f the following targets : b a c k l o g of c o m m o d i t y p r o d u c t i o n c u m u l a t i v e a n d m o n t h l y (ad v a l o r e m ) backlog of p r o d u c e d g o o d s a c c o r d i n g to the production p l a n - b a c k l o g of s o l d g o o d 6 (ad valorem) b a c k l o g s of f u l f i l m e n t of c o n t r a c t s backlog of the sold commodity production c u m u l a t i v e u n t i l the e n d of the c u r r e n t l y y e a r difference between the d a t e of realisation fixed in the c o n t r a c t a n d the a c t u e l d a t e of delivery d i f f e r e n c e b e t w e e n the f i n a l s t o c k according to the p r o d u c t i o n p l a n a n d the a c t u a l stock availabel deviations from t h e a s s o r t m e n t a c c o r d i n g to the p r o d u c t i o n p l a n w h i c h c a n n o t bo s o l d or linked with contracts within the current p e r i o d of p l a n n i n g
> Min > Min > Min > Min > Min > Min > Min
What alternative is g i v e n p r i o r i t y by t h e top management depends on the subjective preference structure of the d e c i s i o n s m a k e r s a n d thu6 it is the r e s p o n s i b i l i t y of the t o p management himself. The d e c i s i o n m a k e r h a s t h e possibility for assigning to the c u s t o m e r s different priorities. The customers and following from that the contracts are
73
prioritised taking into consideration economical and commercial aspects. The procedure CAPS offers the possibility of involving s t a t e o r d e r s as restrictions too. But this a p p r o a c h is left to t o p m a n a g e m e n t ' s d i s c r e t i o n . A d e s c r i p t i o n in d e t a i l o f t h e a l g o r i t h m o f REH y o u f i n d in /Straubel-83-1/, /Straubel-85-1/, /Straubel-86-1/ and /Wittmü(3-83-l/. Experience/Conclusions: The DSS-CAPS realises the connection between the t a r g e t s o f p r o d u c t i o n a n d s a l e s m a n a g e m e n t with the a i m for r e a l i z i n g the h i g h e s t p o s s i b l e p r o f i t . Consequently CAPS is a multi-criterical DSS using the algorithm of R E H . B y m e a n s of C A P S t h e d e c i s i o n m a k e r and t h u s the t o p m a n a g e m e n t h a s the p o s s i b i l i t y to d e c i d e f o r o n e of the p a r e t o - o p t i m a l s o l u t i o n of a s e t o f c o m p r o m i s e p o i n t s r e l a t i v e to p r o d u c e d a n d s o l d g o o d s . In d e p e n d e n c e from the choosen alternative all target results are shown to the decision maker (record). The data b a n k s y s t e m R A B S 1 is u s e d in m o r e t h a n 50 enterprices. T h e D S S - C A P S w i l l b e i n t r o d u c e into t h i s e n t e r p r i c e s in 1989. A m o n g o t h e r t h i n g s one r e s u l t a n d s u c c e s s is the improvement in the field of management in q u a l i t y and quantity. References /Schmidt.-87-1/ S c h m i d t , R .
RABS1- rechnergestützter A b s a t z 1. S t u f e , N e u e T e c h n i k im B ü r o 4/87 VEB Verlag Technik
/ S c h m i d t - 8 7 - 2 / S c h m i d t , R . D a t a B a n k S y s t e m s in t h e M a n a g e m e n t N e u s e T e c h n i k im B ü r o 6/87 VEB Verlag Technik lecture.no publeshed manuscript / S t r a u b e l - 8 5 - 1 / R. S t r a u b e l , A. W i t t m ü ß , R. R o s e n m ü l l e r in Proceedings of the international symposium held in B e r l i n (GDR) 1985, e d i t e d by Achim Sydow, Manfred Thoma and Robert Vichnevetsky A k a d e m i e V e r l a g B e r l i n 1 9 8 5 , M a t h , r e s e a r c h 27 / S t r a u b e l - 8 6 - 1 / R. Straubel, A. Wittmüß, Das Programmsystem REH z u r r e c h n e r g e s t ü t z t e n E n t s c h e i d u n g s h i l f e ZKI I n f o r m a t i o n 2 / 1 9 8 6 Akademie der Wissenschaften der DDR / S t r a u b e l - 8 3 - 1 / R. S t r a u b e l , A. W i t t m ü ß , R. R o s e n m ü l l e r eine rechnergestützte Entscheidungshilfe makroökonomischen Planansatzrechnungen ZKI I n f o r m a t i o n 4 / 1 9 8 3 p. 106- 156 /Wittmüß-83-1/
74
bei
A. W i t t m ü ß , Ein sequentielles Entscheidungsverfahren zur Bestimmung einer mehrkriterialen "optimalen" Steuerung ZKI I n f o r m a t i o n 2 / 1 9 8 3 , p. 94- 113
MODELING AND OPTIMIZATION 1
Schmelovsky, Karl-Heinz
This
paper
is an
'
attempt to get
physical modeling
optimization somewhat closer together, state
using
the
and
mathematical
concept of
enlarged
space. Here, every influence, assumed to be neither exactly known
nor totaly random, is described by state variables. This leads automatically to markoff processes in that state space. Furthermore state models can
always
be
formulated so that observables
depend
only
on actual
state and - possible - on white time discrete noise.
,
?j-y(e)
*
• can
e
'
( & ) u j
y ,
-
~
mostly by neglected.
strategy its only necessary to find respective next
order
to get it,
steps. sequence
80
U
are defined similar to before and
_ A Dependance of 0Cf g on For
'
This
meens
step.
its necessary to regard its influence on that
not only the reaction of object
But
in
consecutive on
a
given
of control steps but the sequence itself must be extrapolated.
Dynamical
programming
solves this problem by a s s u m i n g ,
d e c i s i o n s are optimal
also.
The problem is,
"as good as possible" with than existing Aim of optimization
is here to minimal
consecutive
that "optimal" m e a n s realy
k n o w l e d g e and processing
total
(more exact the integral) of immediate
that
loss, expressed
means.
by the sum
losses
J*-' 17 JA Immediate
loss can mostly
be expressed
control
costs
(e.g.energy)
10
ffcjtl
i Uj)Sf(*j+1~
The
basic
use
introducing
>
the value of a state
gain obtainable starting this
problem, by
+
error
loss and
wanted
terminius
also,
G(z)
i.e.
the
a
sum
of immidiate loss and value
of
maximal
steps.
minimum
i. e. real value has opposite sign. V a l u e function
minimizing
influence
from this s t a t e in remaining
in spite of that we treat
direct
trajectory.
"deterministic" approximation n e g l e c t s statistical
in this task, total
by control
We loss
is obtained
consecutive
state
recursively [*?]•
G
j-1 (zj-i) = Min
{ 0. Then { ri„; n = 0, 1, 2, • • - } is an ONB in H_„. which is a Hilbert scale space [2] with generating operator A and inner product