R. A. Aliev, B. Fazlollahi, R. R. Aliev

Soft Computing and its Applications in Business and Economics

Studies in Fuzziness and Soft Computing, Volume 157

Editor-in-chief Prof. Janusz Kacprzyk Systems Research Institute Polish Academy of Sciences ul. Newelska 6 01-447 Warsaw Poland E-mail: kacprzyk@ibspan.waw.pl

Further volumes of this series can be found on our homepage: springeronline.com

Vol. 141. G.C. Onwubolu, B.V. Babu New Optimzation Techniques in Engineering, 2004 ISBN 3-540-20167-X

Vol. 142. M. Nikravesh, 1.A. Zadeh, V. Korotkikh (Eds.) Fuzzy Partial Differential Equations and Relational Equations, 2004 ISBN 3-540-20322-2

Vol. 143. 1. Rutkowski New Soft Computing Techniques for System Modelling, Pattern Classification and Image Processing, 2004 ISBN 3-540-20584-5

Vol. 144. Z. Sun, G.R. Finnie Intelligent Techniques in E-Commerce, 2004 ISBN 3-540-20518-7

Vol. 145. J. Gil-Aluja Fuzzy Sets in the Management of Uncertainty, 2004 ISBN 3-540-20341-9

Vol. 146. J.A. Gamez, S. Moral, A. Salmer6n (Eds.) Advances in Bayesian Networks, 2004 ISBN 3-540-20876-3

Vol. 147. K. Watanabe, M.M.A. Hashem New Algorithms and their Applications to Evolutionary Robots, 2004 ISBN 3-540-20901-8

Vol. 148. C. Martin-Vide, V. Mitrana, G. Päun (Eds.) Formal Languages and Applications, 2004 ISBN 3-540-20907-7

Vol. 149. J.J. Buckley Fuzzy Statistics, 2004 ISBN 3-540-21084-9

Vol. 150.1. BuH (Ed.) Applications of Learning Classifier Systems, 2004 ISBN 3-540-21109-8

Vol. 151. T. Kowalczyk, E. PleszczyTIska, F. Ruland (Eds.) Grade Models and Methods for Data Analysis, 2004 ISBN 3-540-21120-9

Vol. 152. J. Rajapakse, 1. Wang (Eds.) Neural Information Processing: Research and Development, 2004 ISBN 3-540-2ll23-3

Vol. 153. J. Fulcher, 1.C. Jain (Eds.) Applied Intelligent Systems, 2004 ISBN 3-540-21153-5

Vol. 154. B. Liu Uncertainty Theory, 2004 ISBN 3-540-21333-3

Vol. 155. G. Resconi, J.1. Jain Intelligent Agents, 2004 ISBN 3-540-22003-8

Vol. 156. R. Tadeusiewicz, M.R. Ogiela Medical Image Understanding Technology, 2004 ISBN 3-540-21985-4

Rafik A. Aliev Bijan Fazlollahi Rashad R. Aliev

Soft Computing and its Applications in Business and Economics

~ Springer

Professor Rafik Aziz Aliev

Department of Control Systems

Azerbaijan State Oil Academy

Joint MBA Program (USA, Azerbaijan)

20 Azadlyg Avenue

Baku azl0l0

Azerbaijan

E-mail: raliev@iatp.az

Professor Bijan Fazlollahi

Department of Computer Information Systems &

Institute ofInternational Business

Robinson College ofBusiness

Georgia State University

Atlanta, GA 30303

USA

E-mail: DSCBBF@langate.gsu.edu

ISSN 1434-9922

Professor Rashad Rafik Aliev

Department ofMathematics

Eastern Mediterranean University

Gazimagusa

Turkish Republic ofNorth Cyprus

Mersin 10

Turkey

E-mail: rashad.aliyeV@emu.edu.tr

ISBN 978-3-642-53588-8 ISBN 978-3-540-44429-9 (eBook) DOI 10.1007/978-3-540-44429-9

Library of Congress Control Number: 2004106462

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, reuse of illustrations, recitations, broadcasting, reproduction on microfilm 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 to prosecution under the German Copyright Law.

Springer is apart of Springer Science+Business Media springeronline.com

© Springer-Verlag Berlin Heidelberg 2004 Softcover reprint of the hardcover 1st edition 2004

The use of general descriptive names, registered names trademarks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use.

Typesetting: camera-ready by authors Cover design: E. Kirchner, Springer-Verlag, Heidelberg Printed on acid free paper 62/3020/M - 5 4 3 2 1 0

Foreword

"Soft Computing and its Applications in Business and Economics," or SC-BE for short, is a work whose importance is hard to exaggerate. Authored by leading contributors to soft computing and its applications, SC-BE is a sequel to an earlier book by Professors R.A. Aliev and R.R. Aliev, "Soft Computing and Its Applications," World Scientific, 200l. SC-BE is a self-contained exposition of the foundations of soft computing, and presents a vast compendium of its applications to business, finance, decision analysis and economics. One cannot but be greatly impressed by the wide variety of applications - applications ranging from use of fuzzy logic in transportation and health case systems, to use of a neuro-fuzzy approach to modeling of credit risk in trading, and application of soft computing to e-commerce.

To view the contents of SC-BE in a clearer perspective, a bit of history is in order. In science, as in other realms of human activity, there is a tendency to be nationalistic - to commit oneself to a particular methodology and relegate to a position of inferiority or irrelevance all alternative methodologies.

As we move further into the age of machine intelligence and automated reasoning, we run into more and more problems which do not lend themselves to solution through the use of our favorite methodology. In this event, it may be expedient to explore the possibility of using an alternative methodology, or a combination of our favorite methodology with an alternative methodology, or a combination of alternative methodologies. This is the key idea that was the genesis of soft computing.

The idea became a reality in 1991, when the Berkeley Initiative in Soft Computing (BISC) was launched at UC Berkeley. Today, the so-called BISC Group is a worldwide internet-linked community with over 5000 members.

Basically, soft computing is a coalition of methodologies whose principal constituents are: fuzzy logic, neurocomputing, evolutionary computing, proba-bilistic computing, chaotic computing, rough set theory and parts of machine leaming. The central thesis of soft computing is that, in general, best results are obtained when the constituent methodologies of soft computing are used in combination rather than in a stand-alone mode.

SC-BE is an outstanding treatise on soft computing and its applications, especially in the realms of business and economics. One of the many outstanding features of SC­BE is its excellence of organization. More concretely, the first two chapters are devoted to exposition of the basics of the principal constituents of soft computing: fuzzy set theory and fuzzy logic; neurocomputing; probabilistic computing; evolutionary computing; and chaotic computing. Following Chapter 2, Chapter 3 describes how the constituent methodologies may be used in combination to achieve superior performance. The principal modes of combination are neurofuzzy, neuro-genetic, fuzzy-genetic and neuro-fuzzy-genetic. These modes of combination underlie the recently developed (a) fuzzy-logic-based methodology of computing with words and perceptions; and (b)

multi-agent distributed intelligence systems, which are -described at the end of the chapter. The material in this chapter attests to the authors mastery of the theory of soft computing and extensive experience in the realm of its applications.

At this juncture, the most widely used combination is that of neuro-fuzzy systems. A goal example relates to what may be called the Arabshahi idea - a basic idea which was described by Arabshahi in the context of the backpropagation algorithm, but, in fact, has a much broader applicability. Specifically, most algorithms contain some parameters which can be adjusted to fit a particular problem. In many cases, the choice of values of such parameters is a matter of judgment based on experience. What fuzzy logic - and only fuzzy logic provides - is a language for representing judgment and perceptions derived from experience in the form of fuzzy if-then rules. Addition of such rules to the main algorithm leads to a combined algorithm - an algorithm which reduces the need for experience and im-proves performance.

The remaining six chapters provide an exceptionally complete, insightful and up­to-date coverage of a wide variety of applications of soft computing to business, economics, finance, forecasting, decision analysis and related fields. Among the numerous applications which are discussed are those relating to fuzzy linear prograrnming, probabilistic scheduling for oil refinery, decision support, medicine, credit assessment, inventory control, stock market prediction, portfolio selection, risk management and e-commerce. These chapters place on the table a vast compendiurn of hard-to-get information about the ways in which soft computing can be employed to formulate and solve important real-world problems in business, economics and re1ated fields.

What is obvious and yet not widely recognized, is that soft computing, taken as a whole, is much more powerful than any of its constituents taken in isolation. What this implies is that a strong case can be made for the thesis that a course on soft computing should be a standard course in every engineering curriculum. This is not the case at present, but is likely to become a reality in the not distant future. In this context, SC-BE could play an important role both as a reference and a textbook.

In summary, SC-BE is an outstanding work - a source of many new ideas and techniques, and a vast resource of illurninating information about soft computing and its numerous applications. It is a must reading for anyone who is interested in the conception, design and utilization of systems which are called upon to perform complex tasks requiring a high level ofmachine intelligence. The authors, Professors R.A. Aliev, B. Fazlollahi and R.R. Aliev; the publisher, Springer-Verlag; the series editor, Professor J. Kacprzyk; and the book editor, T. Ditzinger, deserve our thanks and congratulations for producing an outstanding book that is certain to contribute to a wider acceptance and use of soft computing as a powerful tool for constructing systems with superior MIQ (Machine IQ).

April 15, 2004 Lotfi A. Zadeh

Dedication

Dedicated to Professor Lotfi Zadeh, the founder of Soft Computing

Preface

Soft Computing is of central importance among postmodern infonnation tech­nologies for the creation of hybrid intelligent systems with high MIQ (Machine IQ). It was born about a decade ago, when there was intense competition between various emerging technologies. The founder of Soft Computing theory, Professor Lotfi Zadeh had perceived that more could be gained by cooperation than by claims and counter-claims of superiority of these new technologies. Soft Comput­ing concept is based on the idea that in general, better results can be achieved by employing the constituent methodologies in combination than in a stand-alone mode. Soft Computing is consortium of such intelligent paradigms as Fuzzy Logic (FL), Neurocomputing (NC), Evolutionary Computing (EC), Probabilistic Com­puting (PC), and Chaotic Computing (CC) that enable one to solve many impor­tant real-world problems, which could not have been solved by mentioned and other existing technologies.

"Soft Computing and Its Applications" by R.A.Aliev and R.R.Aliev, World Scientific, 2001, a best-selling textbook, presents a unified framework and was the first to provide a systematic account of the major concepts and methodologies of Soft Computing. It introduces the general theory, foundations, and design princi­pIes of Soft Computing based systems and covers mainly applications in engineer­ing problems.

Significant progress in Soft Computing technology and the need for results that rely on more realistic assumptions inspired new researchers to revisit business problems, i.e. the problems that have been traditionally tackled by introducing simplifying assumptions in the past. Today businesses are undergoing a major paradigm shift, moving from traditional management into a worlu of agile and smart organizations and processes. An agile and smart corporation is able to rap­idly and intelligently respond to market changes. For this reason, corporations have been seeking to develop and adopt new technologies, such as soft computing technology, in management of business processes that can assist in developing more realistic solutions quickly and intelligently.

In recent years, a great number of papers and some books have explored the use of fuzzy logic ("Fuzzy Logic for Business, Finance, and Management" by G.Bojadziev and M.Bojadziev, World Scientific, 1997), Neural Networks ("Busi­ness Applications of Neural Networks" by P.Lisboa, A.Vellido, and B.Edinbury, World Scientific, 2000), Genetic Aigorithms ("Evolutionary Computation in Eco­nomics and Finance" by Shu-Heng Chen, Physica-Verlag, 2002) and Soft Com­puting ("Soft Computing for Risk Evaluation and Management: Applications in

X Preface

Technology, Environment and Finance", D. Ruan, M. Fedrizzi, J. Kacprzyk (eds.), Springer-Verlag, 2001) as a tool for designing intelligent systems in business, fi­nance, management and economics.

Books on applications of emerging technologies in business and economics, in­cluding the books mentioned, present recent progress in the application of con­stituent methodologies, in particular neural networks, fuzzy logic, chaos etc. This book highlights some of the recent developments in practical applications of soft computing in business and eccnomics. It is the first book on application of SC based hybrid methods combining fuzzy logic, neuro-computing, evolutionary computing, probabilistic computing and chaotic computing in functional areas of business and economics.

This book is organized into 9 chapters. The first chapter introduces the general concept of Soft Computing technology and answers to the question ''what is Soft Computing?" It also introduces the comparative features of the constituents of SC and their intelligent synergy. In order to make the book self-contained, a review of the theories of constituents of SC: fuzzy logic, neuro-computing, evolutionary computing, probabilistic computing and chaotic computing is given in chapter 2. The reader is expected to have a basic knowledge of these emerging technologies. The integrations ofthose constituents methodologies formed the core ofSC is sub­ject of chapter 3. A distinguishing feature ofthis book is that it deals with all main synergy of different constituent paradigms. Much attention is given to very widely-used in practice, neuro-fuzzy technology. Fuzzy logic and genetic algo­rithms, together with neurocomputing technologies are recognized as one of ma­jor parts of SC, consequently a special place in this chapter is given to hybridiza­tion between these constituents. This chapter also introduces the reader to neuro­genetic, fuzzy-genetic and other SC technologies. It also includes a new aspect of Distributed Intelligence, namely, Soft Computing based multi-agent systems.

Soft Computing in this book is presented not only with the theoretical devel­opment but also with a large variety of realistic applications to business and eco­nomics problems, which are considered in six consequent chapters - chapters 4-9.

Chapter 4 deals with Soft Computing based qualitative and quantitative fore­casting methods.

It represents neuro and fuzzy computing based time series forecasting, fuzzy Delphi method, Soft Computing based prediction in chaotic time series. Chapter 5 covers a wide range of methods and procedures for creation of new generation of Decision making systems and DSS based on SC technology. It introduces the evolutionary and fuzzy chaos approaches to fuzzy linear programming, fuzzy de­cision making, and multi-agent distributed intelligent systems. The emphasis is on development of Soft Computing based multi-agent DSS, hybrid DSS based on ge­netic algorithms and simulation. Application of Soft Computing technology in marketing is subject of chapter 6. This chapter includes soft computing based marketing analysis ofa customer's behavior, credit evaluation, mud detection and service quality evaluation.

Chapter 7 highlights some of the recent practical applications of soft computing in operations management. It covers scheduling fuzzy-probabilistic expert system for oil refineries, fuzzy regression based quality evaluation and neuro-fuzzy pat-

Preface XI

tern recognition in manufacturing, soft computing based inventory control and project scheduling and other related problems.

Soft computing provides effective tools for dealing with complex problems in finance characterized almost always with uncertainty, vagueness and imprecision. Such tools are considered in chapter 8 for creation of intelligent stock market pre­dicting, loan assessment, and risk management systems. This chapter also covers Soft Computing approaches to portfolio selection, trading DSS and solution of other important problems in finance.

A new way of conducting business over the Internet, Electronic Commerce (EC) is growing exponentially. Because EC is relatively new and available infor­mation regarding this kind of business strategy and environment is often inexact and vague, the application of Soft Computing technology to solve EC problems seems very appropriate. Chapter 9 inc1udes Soft Computing based methods for creation of a multi-agent system for EC decisions, personalization of EC, and so­lution of a very important problem in EC, namely, risk analysis.

This book will be valuable aid to anyone considering the application of Soft Computing theory and technology to real problems of business and economics, because it contains a number of detailed accounts of such applications. It brings in a systematic way Soft Computing into the university and college educational sys­tems and may be basic text for introducing business managers, teachers, and scien­tists from various fields of business and economics to the Soft Computing tech­nology, enabling them to initiate projects and make applications.

We would like to express our thanks to professor L. Zadeh, founder of Soft Computing theory for his constant and invaluable support of our research and for his help in the publication of this book. Special thanks are due to Professors Mo. Jamshidi, 1. Turksen, J. Kacprzyk, T. Whalen, S.Ulyanov and V. Loia for helpful discussions on various topics of Soft Computing and its applications.

We are grateful to our colleagues dr. R. Vahidov and B. Guirimov for many en­joyable and productive conversations and collaborations.

R. A. Aliev B. Fazollahi R. R. Aliev

Contents

I Introduction to Soft Computing .......................................................................... I 1.1 Basic Concepts of Soft Computing ............................................................. 1 2.2 Combination of Constituents of Soft Computing ........................................ 4 References ......................................................................................................... 8

2. Constituent Methodologies of Soft Computing ............................................... 11 2.1 Elements of Fuzzy Sets Theory ................................................................. 11

2.1.1 Fuzzy Sets and Operations Over Them ............................................... 11 2.2.2 Mathematics ofFuzzy Computing ...................................................... 31 2.1.3 Fuzzy Logic and Approximate Reasoning .......................................... 54 2.1.4 Probability and Fuzziness ................................................................... 80 2.1.5 Fuzzy Sets and Possibility Theory ...................................................... 81

2.2 Foundations ofNeurocomputing ............................................................... 82 2.2.1 Basic Types and Architeetures ofNeural Networks ........................... 82 2.2.2 Learning Algorithms ofNeural Networks .......................................... 88

2.3 Probabilistic COmputing....................................... .......................... ......... 111 2.3.1 Bayesian Approach ........................................................................... 112 2.3.2 Dempster-Shafer Theory ofBelief.. .................................................. 114

2.4 Evolutionary Computing ......................................................................... 119 2.4.1 Evolution Programming and Genetic Algorithms ............................. 119 2.4.2 Computation with Genetic Aigorithms ............................................. 125

2.5 Chaotic Computing ................................................................................. 145 2.5.1 Elements ofChaotic COmputing ....................................................... 145 2.5.2 Non-Linear Dynamics and Chaotic Analysis .................................... 146 2.5.3 Empirical Chaotic Analysis .............................................................. 151

References ..................................................................................................... 152 3. Emerging COmbined Soft Computing Technologies .................................... 159

3.1 Neuro-Fuzzy Technology ........................................................................ 159 3.2 Neuro-Genetic Approach ........................................................................ 170 3.3 Fuzzy Genetic Paradigm ......................................................................... 175 3.4 Genetic Algorithms with Fuzzy Logic .................................................... 185 3.5 Neuro-Fuzzy-Genetic Paradigm .............................................................. 186 3.6 Multi-Agent Distributed Intelligent Systems Paradigm .......................... 193 3.7 Computing with Words Technology ....................................................... 205 References ..................................................................................................... 208

4. Soft Computing Technologies in Business and Economic ............................ 219 Forecasting 4.1 Neuro-Computing and Forecasting ......................................................... 219

XIV Contents

4.2 Fuzzy Time Series Based Forecasting ...................................................... 220 4.3 Fuzzy Delphi Method ............................................................................... 228 4.4 Soft Computing Based Forecasting Complex Time Series ...................... 229 4.5 Soft Computing Based Prediction Ensemble for Forecasting in ............. 235

Chaotic Time Series References ...................................................................................................... 240

5 Soft Computing Based Decision Making and DSS ......................................... 243 5.1 Fuzzy Linear Programming ...................................................................... 243 5.2 Evolutionary Algorithm Based Fuzzy Linear Programming .................... 256 5.3 Fuzzy Chaos Approach to Fuzzy Linear Programming Problem ............. 258 5.4 Fuzzy-Probabilistic Scheduling for Oil Refinery ..................................... 259 5.5 Fuzzy Decision Making ........................................................................... 271 5.6 Multi-Agent Distributed Intelligent System Based on Fuzzy .................. 287

Decision Making 5.7 Soft Computing and Data Mining ............................................................ 294 5.8 Soft Computing Based Multi-Agent Marketing DSS ............................... 297 5.9 Hybrid DSS Based on Simulation and Genetic Aigorithms ..................... 299 5.10 Soft Computing Based Alternatives Generations by Decision ............... 312

Support Systems References ..................................................................................................... .326

6 Soft Computing in Marketing .......................................................................... 333 6.1 Marketing Analysis of a Customer's Purchasing Behavior ...................... 333 6.2 Customer Credit Evaluation ..................................................................... 335 6.3 Soft Computing Based Fraud Detection ................................................... 338 6.4 Fuzzy Evaluation ofService Quality ........................................................ 341 6.5 Application ofFuzzy Programming to Hospital's Service ...................... 343

Performance Evaluating References ..................................................................................................... .349

7 Soft Computing Applications in Operations Management .............................. 351 7.1 Application ofFuzzy Logic in Transportation Logistics .......................... 351 7.2 Scheduling Fuzzy Expert Systems with Probabilistic Reasoning ............ 354

for Oil Refineries 7.3 Detection and Withdrawal ofDefect Parts in the Computer ................... 360

Aided Manufacturing of Evaporators 7.4 Genetic Aigorithms Based Fuzzy Regression Analysis and Its ................ 366

Applications for Quality Evaluation 7.5 An Intelligent System for Diagnosis ofthe Oil-Refinery Plant.. .............. 375 7.6 Neuro-Fuzzy Pattern Recognition in Manufacturing ............................... 381 7.7 Soft Computing Based Inventory Control ................................................ 389 7.8 Fuzzy Project Scheduling ......................................................................... 392 7.9 CW Based Decision Analysis on Risk Assessment of an ....................... 397

Engineering Project References ..................................................................................................... .400

8 Soft Computing in Finance ............................................................................. .403 8.1 Soft Computing Based Stock Market Predicting System ......................... 403 8.2 Fuzzy Nonlinear Programming Approach to Portfolio Selection ............. 406

Contents XV

8.3 Neuro-Fuzzy Approach to Modeling of Credit Risk in Trading ............. 409 Portfolios

8.4 A Fuzzy Approach to the Credit Portfolio Constructing ......................... 412 8.5 Soft Computing Based TDSS Multi-Agent Systems in Finance ............. 415 8.6 Neural Nonlinear Modeling for Risk Management in Banking .............. 421 8.7 Neuro-Fuzzy Loan Assessment System .................................................. 422 References ..................................................................................................... 428

9 Soft Computing in Electronic Business .......................................................... 431 9.1 A Multi-Agent System for E-Commerce Decisions ................................ 431 9.2 Soft Computing and Personalization of Electronic Commerce ............... 440 9.3 Risk Analysis in Electronic Commerce Using Fuzzy Weighted ............. 442

Average References ..................................................................................................... 445