Springer computational intelligence theory and applications (2006) 3540347801

Aroba Department of Electronic Engineering and Computer Science University of Huelva, Spain aroba@diesia.uhu.es Jaakko Astola Tampere University of Technology Institute for Signal Proces

Bernd Reusch (Ed.)Computational Intelligence, Theory and Applications

Advances in Soft Computing

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:

springer.com

First Course on Fuzzy Theory and

Applications, 2004

ISBN 3-540-22988-4

Kwang H Lee (Ed.)

Miguel López-Díaz, Maria A Gil,

Przemyslaw Grzegorzewski, Olgierd

Hryniewicz, Jonathan Lawry (Eds.)

Soft Methodology and Random Information

Marek Kurzynski, Edward Puchala,

Michal Wozniak, Andrzej Zolnierek (Eds.)

Computer Recognition Systems, 2005

ISBN 3-540-25054-9

Abraham Ajith, Yasuhiko Dote,

Takeshi Furuhashi, Mario Köppen,

Azuma Ohuchi, Yukio Ohsawa (Eds.)

Soft Computing as Transdisciplinary

Science and Technology, 2005

ISBN 3-540-25055-7

Marcin Szczuka, (Eds.)

Techniques in Multiagent Systems, 2005

Monitoring, Security, and Rescue

ISBN 3-540-23245-1

Barbara Dunin-Keplicz, Andrzej

Jankowski, Andrzej Skowron,

Frank Hoffmann, Mario Köppen, Frank Klawonn, Rajkumar Roy (Eds.)

Applications, 2005 Soft Computing Methodologies and

ISBN 3-540-25726-8

Mieczyslaw A Klopotek, Slawomir T Wierzchon, Kryzysztof Trojanowski (Eds.)

Intelligent Information Processing and Web Mining, 2005

ISBN 3-540-25056-5

Bernd Reusch, (Ed.)

Computational Intelligence, Theory and Applications, 2005

Web Mining, 2006 Intelligent Information Processing and

ISBN 3-540-33520-X

Ashutosh Tiwari, Joshua Knowles, Eral Auineri, Keshav Dahal, Rajkumar Roy (Eds.)

Applications and Soft Computing, 2006

ISBN 3-540-29123-7

Bernd Reusch

Computational Intelligence, Theory and Applications

ISBN-13

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, recitation, 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 Violations are liable for prosecution under the German Copyright Law.

Springer is a part of Springer Science+Business Media

c

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.

Cover design: Erich Kirchner, Heidelberg

Library of Congress Control Number: 2006930272

ISSN print edition: 1615-3871

3-540-34780-1 Springer Berlin Heidelberg New York

978-3-540-34780-4 Springer Berlin Heidelberg New York

Springer-Verlag Berlin Heidelberg 2006

SPIN: 11757382 89/SPi springer.com

ISSN electronic edition: 1860-0794

Professor Dr Bernd Reusch

For the 9th time since 1991 we invite researchers to participate in theDortmund Fuzzy-Days I am very glad that our conference has establisheditself as an international forum for the discussion of new results in the filed

of Computational Intelligence Again all papers had to undergo a thoroughreview: each one was judged by five referees to guarantee a solid quality ofthe programme

From the beginning of the Fuzzy-Days on Lotfi A Zadeh felt associatedwith the conference I would like to express my gratitude for his encouragementand support and I am particularly glad that he once again delivers a keynotespeech Much to my pleasure Ewa Orlowska, Radko Mesiar together with

have also agreed to present new results of their work as keynote speakers.Many thanks go to my friends Janusz Kacprzyk and Enric Trillas whotogether with Lotfi Zadeh again served as honorary chairmen

Due to my retirement in 2006, these are the last Dortmund Fuzzy Days inthe form we had developed over the years At this point I have to leave open,whether we find another forum or not

I wish to thank all participants of the Dortmund Fuzzy-Days for theircommitment to the conference and the organisers, namely Mrs Ulrike Lippe,for the excellent job they did Last but not least, I am obliged to the Germanresearch council for their valuable financial support

Ewa OrlowskaEndre PapWitold PedryczIrina PerfilievaOlivier PivertSusanne SamingerElie SanchezHideo TanakaEnric TrillasPeter VojtasMichael WagenknechtTakeshi YamakawaAntonio di Nola

Local Organization

Wolfgang HunscherUlrike LippeThomas Wilke

Plenary Talk

From Search Engines to Question-Answering Systems:

The Problems of World Knowledge, Relevance, Deduction,

and Precisiation

Lotfi A Zadeh 1

Invited Session: Fuzzy Multiperson and Multicriteria Decisions Modelling

A Fuzzy Approach to Optimal R&D Project Portfolio

Selection

Christer Carlsson, Robert Full´ er, and P´ eter Majlender 5

Choquet Integration and Correlation Matrices in Fuzzy

Inference Systems

R.A Marques Pereira, P Serra, R.A Ribeiro 15

Linguistic Summarization of Some Static and Dynamic

Features of Consensus Reaching

Janusz Kacprzyk, S
lawomir Zadro˙zny, and Anna Wilbik 19

Consistency for Nonadditive Measures: Analytical and

Algebraic Methods

Antonio Maturo, Massimo Squillante, and Aldo Ventre 29

VIII Contents

Neural Nets

Neuro-Fuzzy Kolmogorov’s Network with a Modified

Perceptron Learning Rule for Classification Problems

Vitaliy Kolodyazhniy, Yevgeniy Bodyanskiy, Valeriya Poyedyntseva,

and Andreas Stephan 41

A Self-Tuning Controller for Teleoperation System using

Evolutionary Learning Algorithms in Neural Networks

Habib Allah Talavatifard, Kamran Razi, and Mohammad Bagher Menhaj 51

A Neural-Based Method for Choosing Embedding Dimension

in Chaotic Time Series Analysis

Sepideh J Rastin and Mohammad Bagher Menhaj 61

On Classification of Some Hopfield-Type Learning Rules

via Stability Measures

Mohammad Reza Rajati, Mohammad Bagher Menhaj 75

Applications I

A New Genetic Based Algorithm for Channel Assignment

Problems

Seyed Alireza Ghasempour Shirazi and Mohammad Bagher Menhaj 85

Max-Product Fuzzy Relational Equations as Inference Engine for Prediction of Textile Yarn Properties

Yordan Kyosev, Ketty Peeva, Ingo Reinbach, and Thomas Gries 93

Automatic Defects Classification and Feature Extraction

Optimization

Bernd Kuhlenk¨ otter, Carsten Krewet, and Xiang Zhang 105

Short-Term Load Forecasting in Power System Using Least

Squares Support Vector Machine

Ganyun LV, Xiaodong Wang and Yuanyuan Jin 117

Plenary Talk

Fifteen Years of Fuzzy Logic in Dortmund

R Mesiar and Vil´ em Nov´ ak 127

Contents IX

Invited Session: Intuitionistic Fuzzy Sets and Generalized Nets I

Intuitionistic Fuzzy Graphs

R Parvathi and M.G Karunambigai 139

On Some Intuitionistic Properties of Intuitionistic Fuzzy

Implications and Negations

Trifon A Trifonov and Krassimir T Atanassov 151

On Intuitionistic Fuzzy Negations

Krassimir T Atanassov 159

Invited Session: Soft Computing Techniques for Reputation and Trust I

A Simulation Model for Trust and Reputation System

Evaluation in a P2P Network

Roberto Aringhieri and Daniele Bonomi 169

A Fuzzy Trust Model Proposal to Ensure the Identity

of a User in Time

Antonia Azzini and Stefania Marrara 181

Quantification of the Effectiveness of the Markov Model for Trustworthiness Prediction

Farookh Khadeer Hussain, Elizabeth Chang, and Tharam S Dillon 191

Applications II

Fuzzy-Genetic Methodology for Web-based Computed-Aided Diagnosis in Medical Applications

F de Toro, J Aroba, J.M Lopez 201

Weight Optimization for Loan Risk Estimation with Genetic Algorithm

Irina Lovtsova 215

A Fuzzy Feature Extractor Neural Network and its

Application in License Plate Recognition

Modjtaba Rouhani 223

X Contents

Invited Session: Intuitionistic Fuzzy Sets and Generalized Nets II

Nearest Interval Approximation of an Intuitionistic Fuzzy

Towards Usage Policies for Fuzzy Inference Methodologies

for Trust and QoS Assessment

Stefan Schmidt, Robert Steele, Tharam Dillon 263

Simulating a Trust-Based Peer-to-Peer Metadata Publication Center

Paolo Ceravolo, Alessio Curcio, Ernesto Damiani and Micol Pinelli 275

The Complex Facets of Reputation and Trust

Karl Aberer, Zoran Despotovic, Wojciech Galuba

and Wolfgang Kellerer 281

Martin Kalina and Alexander ˘ Sostak 301

Lipschitz Continuity of Triangular Norms

Andrea Mesiarov´ a 309

Contents XI

Plenary Talk

Formal Models of Knowledge Operators: Rough-Set-Style

and Fuzzy-Set-Style Approaches

Ewa Or
lowska 323

Invited Session: Looking at Language with Fuzzy Logic

Using a Fuzzy Model for Combining Search Results from

Different Information Sources to Build a Metasearch Engine

Wiratna S Wiguna, Juan J Fern´ andez-´ı´ ebar and Ana Garc´ıa-Serrano 325

Sergio Guadarrama, Eloy Renedo, and Enric Trillas 335

Fuzzy Sets Versus Language

Enric Trillas, Eloy Renedo, and Sergio Guadarrama 353

Theory II

Some Properties of Fuzzy Languages

Claudio Moraga 367

General Form of Lattice Valued Intuitionistic Fuzzy Sets

Andreja Tepavˇ cevi´ c and Marijana Gorjanac Ranitovi´ c 375

A Note on Generated Pseudo-Operations with Two

Parameters as a base for the Generalized Pseudo-Laplace

Optimal Toll Charges in a Fuzzy Flow Problem

Stephan Dempe and Tatiana Starostina 405

Modified Interval Global Weights in AHP

Tomoe Entani and Hideo Tanaka 415

XII Contents

Plenary Talk

Fuzzy Approaches to Trust Management

Elizabeth Chang, Ernesto Damiani, and Tharam Dillon 425

Invited Session: Complex-Valued Neural Networks

Proposal of Holographic 3D-Movie Generation Using

Coherent Neural-Network Interpolation

Akira Hirose and Tomoaki Higo 437

Blur Identification Using Neural Network for Image

Restoration

Igor Aizenberg, Dmitriy Paliy, Claudio Moraga and Jaakko Astola 441

Separable Problems Using a Single Universal Binary Neuron

Szilveszter Kov´ acs 485

Fuzzy Rule Interpolation Based on Polar Cuts

Zsolt Csaba Johany´ ak and Szilveszter Kov´ acs 499

Approximate Reasoning Using Fodor’s Implication

Adrian Giurca and Ion Iancu 513

Plenary Talk

Brain-, Gene-, and Quantum-Inspired Computational

Intelligence: Challenges and Opportunities

Nikola Kasabov 521

Contents XIII

Invited Session: Intelligent Data Mining

Incremental Learning for E-mail Classification

Sigita Misina 545

Reduction of Search Space for Instance-Based Classifier

Combination

Anatoly Sukov and Arkady Borisov 555

Invited Session: Preferences and Decisions

Linguistic Matrix Aggregation Operators: Extensions of the Borda Rule

Jos´ e Luis Garc´ıa-Lapresta, Bonifacio Llamazares and

Miguel Mart´ınez-Panero 561

Evolutionary Algorithms

An Evolutionary Algorithm for the Biobjective QAP

Istvan Borgulya 577

On a Hill-Climbing Algorithm with Adaptive Step Size:

Towards a Control Parameter-Less Black-Box Optimisation

Intragenerational Mutation Shape Adaptation

Stefan Berlik and Bernd Reusch 603

Theory V

The Choquet-Integral as an Aggregation Operator in

Case-Based Learning

Eyke H¨ ullermeier 615

Outlier Resistant Recursive Fuzzy Clustering Algorithms

Yevgeniy Bodyanskiy, Illya Kokshenev, Yevgen Gorshkov, and

Vitaliy Kolodyazhniy 647

Invited Session: Fuzzy Sets – 40 years after

Fuzzy Set Theory – 40 Years of Foundational Discussions

Siegfried Gottwald 653

Fuzzy Control – Expectations, Current State,

and Perspectives

Mirko Navara and Milan Petr´ık 667

Fuzzy Sets in Categories of Sets with Similarity Relations

Jiˇ r´ı Moˇ ckoˇ r 677

Fuzzy Sets as a Special Mathematical Model of Vagueness

Synthesizing Adaptive Navigational Robot Behaviours

Using a Hybrid Fuzzy A* Approach

Antony P Gerdelan and Napoleon H Reyes 699

Fuzzy Impulse Noise Reduction Methods for Color Images

Stefan Schulte, Mike Nachtegael, Val´ erie De Witte, Dietrich Van der Weken, and Etienne E Kerre 711

Use of Variable Fuzzy Sets Methods for Desertification

Evaluation

Wu Li, Guo Yu, Chen Shouyu, and Zhou Huicheng 721

A Genetic Algorithm-Based Fuzzy Inference System

in Prediction of Wave Parameters

M Zanganeh, S.J Mousavi, A Etemad-Shahidi 741

Poster Contributions

Estimation of Degree of Polymerisation and Residual Age

of Transformers Based on Furfural Levels in Insulating Oil

Through Generalized Regression Neural Networks

K.S Madhavan, T.S.R Murthy, and R Sethuraman 751

Fuzzy Shortest Paths in Fuzzy Graphs

Amir Baniamerian and Mohammad Bagher Menhaj 757

Improving Vegas Algorithm Using PID and Fuzzy PID

Controllers

Aria Shahingohar, Mohammad Bagher Menhaj, Mehdi Karrari,

and Mohammad Hossein Yaghamae 765

A Fuzzy-Based Automation Level Analysis in Irrigation

Equipment

Mohsen Davoudi, Mohammad Bagher Menhaj, and Mehdi Davoudi 777

Motorized Skateboard Stabilization Using Fuzzy Controller

Mohsen Davoudi, Mohammad Bagher Menhaj, and Mehdi Davoudi 789

Index 801

Texas A&M University-Texarkana

P.O Box 5518 2600 N Robison Rd

Texarkana, Texas 75505 USA

igor.aizenberg@tamut.edu

Roberto Aringhieri

Dipartimento di Tecnologie

dell’Informazione

Via Bramante 65, Crema I-26013

Italy

roberto.aringhieri@unimi.it

J Aroba

Department of Electronic

Engineering and Computer Science

University of Huelva, Spain

aroba@diesia.uhu.es

Jaakko Astola

Tampere University of Technology

Institute for Signal Processing

CLBME – Bulgarian Academy

of Sciences, P.O Box 12Sofia-1113, Bulgariakrat@bas.bg

Antonia Azzini

di MilanoDipartimento di Tecnologiedell’Informazione, via Bramante 65

26013 Crema (CR), Italyazzini@dti.unimi.it

Adrian I Ban

Department of Mathematicsand Informatics,

University of OradeaUniversitatii 1, 410087 OradeaRomania

aiban@uoradea.ro

Amir Baniamerian

Department of ElectricalEngineering, AmirKabirUniversity of TechnologyBaniamerian@yahoo.com

XVIII List of Contributors

Alan Battersby

School of Computing

and Informatics

Nottingham Trent University

Clifton Lane, Nottingham

NG11 8NS, UK

alan.battersby@ntu.ac.uk

Stefan Berlik

Siegen University

Department of Electrical Engineering

and Computer Science

57068 Siegen, Germany

berlik@informatik.uni-siegen.de

Yevgeniy Bodyanskiy

Control Systems Research

Laboratory, Kharkiv National

Institute for Advanced Management

University

˚Abo, FIN-20520, Finlandchirster.carlesson@abo.fi

Paolo Ceravolo

Dipartimento di Tecnologiedell’Informazione

Milano, Italyceravolo@dti.unimi.it

Elizabeth Chang

Curtin UniversityPerth, AustraliaElizabeth.Chang@cbs.curtin.edu.au

Shouyu Chen

Water Resource Research GroupSchool of Civil Engineeringand Architecture

Dalian University of TechnologyDalian 116024, China

Panagiotis Chountas

University of Westminster-HealthCare Computing Group

HSCS, Northwick ParkLondon, HA1 3TP, UKchountp@wmin.ac.uk

Alessio Curcio

Dipartimento di Tecnologiedell’Informazione

Milano, Italyacurcio@crema.unimi.it

Ermesto Damiani

Dipartimento di Tecnologiedell’Informazione

Milano, Italydamiani@dti.unimi.it

List of Contributors XIX

Francisco de Toro Negro

Signal Theory, Telematics

and Communications Department

University of Granada, Spain

ftoro@ugr.es

Tharam S Dillon

Faculty of Information Technology,

University of Technology, Sydney

Nottingham Trent University

Clifton Lane, Nottingham

NG11 8NS, UK

tarek.elmihoub@ntu.ac.uk

Tomoe Entani

Kochi University2-5-1 Akebono Kochi780-8520, Japanentani@cc.kochi-u.ac.jp

Amir Farshad Etemad-Shahidi

Department of Civil EngineeringIran University of Scienceand Technology, Tehran, Iranetemad@iust.ac.ir

Intelligent Systems Research Group(ISYS-GSI)

jjfernandez@dia.fi.upm.es

Department of Operations Research

H-1117 Budapest, Hungaryrfuller@cs.eltc.hu

Wojciech Galuba

de Lausanne (EPFL)Switzerland

Department of de Econom´ıa

PRESAD Research GroupUniversidad de ValladolidSpain

lapresta@eco.uva.es

XX List of Contributors

Ana Garc´ıa-Serrano

Intelligent Systems

Research Group (ISYS-GSI)

of Radioelectronics, 14Lenin Av., Kharkiv 61166Ukraine

ye.gorshkov@gmail.com

Siegfried Gottwald

Leipzig University, Institute forLogic and Philosophy of ScienceBeethovenstrausse 15

04107 Leipzig, Germanygottwald@uni-leipzig.de

Thomas Gries

Institute for Textile TechnologyRWTH Aachen UniversityEilfschornsteinstr 18

52062 Aachen, GermanyDepartment of TextilesTechnical University of SofiaBul Kliment Ohridski 8Sofia-1000, Bulgaria

Sergio Guadarrama

Departamento de InteligenciaArtificial

de Madrid 28660Boadilla del MonteMadrid, Spainsguada@dia.fi.upm.es

Yu Guo

Water Resource Research GroupSchool of Civil Engineeringand Architecture

Dalian University of TechnologyDalian 116024 China

List of Contributors XXI

Yutaka Hata

Division of Computer Engineering

Graduate School of Engineering

University of Hyogo, 2167 Shosha

Himeji 671-2280, Japan

hata@ieee.org

Tomoaki Higo

Department of Electronic

Engineering, The University

of Tokyo 7-3-1 Hongo, Bunkyo-ku

Nottingham Trent University

Clifton Lane, Nottingham

Farookh Khadeer Hussain

School of Information Systems

Curtin University of Technology

Etienne E Kerre

Fuzziness and UncertaintyModelling Research UnitDepartment of Applied Mathematicsand Computer Science

Ghent University, Krijgslaan 281-S9

9000 Gent, Belgiumhttp://www.fuzzy.ugent.be/

Boyan Kolev

CLBME - Bulgarian Academy ofSciences, Bl 105, Sofia-1113Bulgaria

Yuanyuan Jin

Nanjing Micro One Electronics Inc.11F Huaxin Building, 9 Guanjiaqiao

210005 Nanjin, Chinajinyy@microne.com.cn

XXII List of Contributors

Systems Research Institute

Polish Academy of Sciences

Slovak University of Technology

Auckland University of Technology

Auckland, New Zealand

Division of Computer Engineering

Graduate School of Engineering

University of Hyogo, 2167 Shosha

14, Lenin Av., Kharkiv 61166Ukraine

ikcontact@rambler.ru

Vitaliy Kolodyazhniy

Control Systems ResearchLaboratory, Kharkiv NationalUniversity of Radioelectronics

14, Lenin Av., Kharkiv 61166,Ukraine

kolodyazhniy@ukr.net

Katsuya Kondo

Division of Computer EngineeringGraduate School of EngineeringUniversity of Hyogo, 2167 ShoshaHimeji 671-2280, Japan

kondo@ieee.org

Department of InformationTechnology, University of Miskolc

H-3515, Hungaryszkovacs@iit.uni-miskolc.hu

Faculty of EconomicsMatej Bel University

Bystricapavol.kral@umb.sk

Carsten Krewet

Robotics Research InstituteIndustrial Robotics and HandlingSystems Otto-Hahn-Str.8

Dortmund UniversityD-44221 Dortmund, Germany

List of Contributors XXIII

Robotics Research Institute

Industrial Robotics and Handling

Technical University of Sofia,

Bul Kliment Ohridski 8

Sofia-1000, Bulgaria

info@kyosev.com

Wu Li

Water Resource Research Group

School of Civil Engineering

PRESAD Research Group

Universidad de

Valladolid, Spain

boni@eco.uva.es

J.M Lopez

Signal Theory, Department

of Telematics and Communications

University of Granada, Spain

Irina Lovtsova

Department of Modellingand Simulation

Riga Technical University

1 Kalku Street, Riga, LV - 1658,Latvia

lovcova@inbox.lv

K.S Madhavan

Corporate Researchand DevelopmentBharat Heavy Electricals LimitedHyderabad, India

˚Abo FIN-20520, Finlandpeter.majlender@abo.fi

Ricardo Marques Pereira

26013 Crema (CR), Italymarrara@dti.unimi.it

Miguel Mart´ınez-Panero

Department of de Econom´ıa

PRESAD Research GroupUniversidad de ValladolidSpain

panero@eco.uva.es

XXIV List of Contributors

Department of Social Sciences

Faculty of Social Sciences

Institute for Research

and Application of Fuzzy Modelling

http://www.fuzzy.ugent.be/

Sigita Misina

Department of Modellingand Simulation

Riga Technical University

1 Kalku StreetRiga, LV- 1658, Latviasigita.misina@gmail.com

University of OstravaInstitute for Researchand Applications of Fuzzy Modeling

30, dubna 22, 701 03 Ostrava 1Czech Republic

Jiri.Mockor@osu.cz

Claudio Moraga

Department Computer ScienceUniversity of DortmundD-44221 Dortmund, GermanyEuropean Centre for SoftComputing, E-33600 MieresAsturias, Spain

claudio@moraga.de,claudio.moraga@udo.edu

Seyed Jamshid Mousavi

Department of Civil EngineeringAmirkabir University of TechnologyTehran, Iran

jmosavi@aut.ac.ir

T.S.R Murthy

Corporate Researchand DevelopmentBharat Heavy Electricals LimitedHyderabad, India

Nottingham Trent University

Clifton Lane, Nottingham

Tampere University of Technology

Institute for Signal Processing

TICSP, Tampere University

of Technology, P.O Box 553

Vellalar College for Women

Erode 638052, Tamilnadu, India

paarvathis@rediffmail.com

Ketty Peeva

Technical University of Sofia

Faculty of Applied Mathematics

30, dubna 22, 701 03 Ostrava 1,Czech Republic

Irina.Perfilieva@osu.cz

Milan Petr´ık

Center for Machine PerceptionDepartment of CyberneticsFaculty of Electrical EngineeringCzech Technical University

Milano, Italympinelli@crema.unimi.it

Valeriya Poyedyntseva

Department of Enterprise EconomyKharkiv National Automobileand Highway University

25, Petrovskiy StreetKharkiv 61002, Ukrainepoyedyntseva@gmx.net

Mohammad Reza Rajati

Computational Intelligenceand Control Research CenterDepartment of ElectricalEngineering

XXVI List of Contributors

Technical University of Sofia

Bul Kliment Ohridski 8

28660 Boadilla del Monte

de Novas Tecnologias UNINOVA

Universidade Nova de Lisboa

modjtaba rouhani@yahoo.com

Stefan Schmidt

University of TechnologySydney, P.O Box 123Broadway, NSW 2007Australia

sschmidt@it.uts.edu.au

Stefan Schulte

Fuzziness and UncertaintyModelling Research UnitDepartment of Applied Mathematicsand Computer Science

Ghent UniversityKrijgslaan 281-S9,

9000 Gent, BelgiumStefan.Schulte@Ugent.be

Brno University of TechnologyFaculty of Mechanical EngineeringInstitute of Automation andComputer Science

Czech Republicseda@fme.vutbr.cz

Paulo J.A Serra

Instituto de Desenvolvimento

de Novas Tecnologias UNINOVAUniversidade Nova de LisboaQuinta da Torre 2829-516Caparica, Portugalpja@uninova.pt

Department of Mathematics andInformatics

University of Novi Sad, Trg D

Serbia and Mongenegroseselja@im.ns.ac.yu

List of Contributors XXVII

and Business Sciences

via Nazionale delle Puglie

Robert Steele

University of TechnologySydney, P.O Box 123Broadway, NSW 2007Australia

rsteele@it.uts.edu.au

Andreas Stephan

PSI-Tec GmbHGrenzhammer, 8, D-98693Ilmenau, Germanystephan@psi-tec.de

Anatoly Sukov

Department of Modelling andSimulation

Riga TechnicalUniversityKalku iela 1, Riga LV 1658Latvia

Habib Allah Talavatifard

Department of ElectricalEngineering

Amirkabir University of TechnologyTehran, Iran

Kazuhiko Taniguchi

Kinden Corporation,3-1-1, SaganakadaiKuzu-cho, Souraku-DistrictKyoto 619-0223, Japanktaniguchi@dk.pdx.ne.jp

XXVIII List of Contributors

CLBME - Bulgarian Academy

of Sciences, P.O Box 12

28660 Boadilla del Monte

Madrid, Spain

etrillas@fi.upm.es

Aldo Ventre

University of Napoli,

Department of Culture of the Project

Faculty of Architecture, Abazia di

San Lorenzo ad

Septimum, I-81301 Aversa, Italy

aldoventre@yahoo.it

Institute for Research

and Applications of Fuzzy Modeling

the Czech Republic

Praha 8, Czech Republic

Vilem.Novak@osu.cz

Dietrich Van der Weken

Fuzziness and UncertaintyModelling Research UnitDepartment of Applied Mathematicsand Computer Science

Ghent UniversityKrijgslaan 281-S9

9000 Gent, Belgiumhttp://www.fuzzy.ugent.be/

Wiratna S Wiguna

Intelligent Systems Research Group(ISYS-GSI)

wiratna@dia.fi.upm.es

Fuzziness and UncertaintyModelling Research UnitDepartment of Applied Mathematicsand Computer Science,

Ghent UniversityKrijgslaan 281-S9

9000 Gent, Belgiumhttp://www.fuzzy.ugent.be/

http://www.zjnu.cn

Mohammad Hossein Yaghmai

Department of computer EngineeringFerdowsi University of Mashhadhyaghmae@ferdowsi.um.ac.ir

List of Contributors XXIX

Lotfi A Zadeh

Berkeley Initiative

in Soft Computing (BISC)

Computer Science Division and the

Electronics Research Laboratory

Department of EECS

University of California

Berkeley, CA 94720-1776, USA

zadeh@cs.berkeley.edu

Slawomir Zadro ˙zny

Warsaw Information Technology

(WIT), ul Newelska 6

01-447 Warsaw Poland

zadrozny@ibspan.waw.pl

Morteza Zanganeh

Department of Civil Enginering

Iran University of Science

and Technology, Tehran, Iran

Xiang Zhang

Robotics Research InstituteIndustrial Robotics and HandlingSystems Otto-Hahn-Str.8

Dortmund University,D-44221 DortmundGermany

Huicheng Zhou

Water Resource Research GroupSchool of Civil Engineeringand Architecture

Dalian University of TechnologyDalian 116024, China

From Search Engines to Question-Answering Systems: The Problems of World Knowledge, Relevance, Deduction, and Precisiation

Summary Existing search engines, with Google at the top, have many truly

re-markable capabilities Furthermore, constant progress is being made in improvingtheir performance But what is not widely recognized is that there is a basic capabil-ity which existing search engines do not have: deduction capability – the capability

to synthesize an answer to a query by drawing on bodies of information which reside

in various parts of the knowledge base By definition, a question-answering system,

or a Q/A system for short, is a system which has deduction capability Can a searchengine be upgraded to a question-answering system through the use of existingtools – tools which are based on bivalent logic and probability theory? A view which

is articulated in the following is that the answer is: no

The first obstacle is world knowledge – the knowledge which humans quire through experience, communication, and education Simple examplesare: “Icy roads are slippery,” “Princeton usually means Princeton Univer-sity,” “Paris is the capital of France,” and “There are no honest politicians.”World knowledge plays a central role in search, assessment of relevance anddeduction The problem with world knowledge is that it is, for the most part,perception-based Perceptions – and especially perceptions of probabilities –are intrinsically imprecise, reflecting the fact that human sensory organs, andultimately the brain, have a bounded ability to resolve detail and store in-formation Imprecision of perceptions stands in the way of using conventionaltechniques – techniques which are based on bivalent logic and probabilitytheory – to deal with perception-based information A further complication

ac-is that much of world knowledge ac-is negative knowledge in the sense that itrelates to what is impossible and/or nonexistent For example, “A personcannot have two fathers,” and “Netherlands has no mountains.”

The second obstacle centers on the concept of relevance There is an tensive literature on relevance, and every search engine deals with relevance inits own way, some at a high level of sophistication But what is quite obvious

ex-is that the problem of assessment of relevance ex-is quite complex and far fromsolution

2 L.A Zadeh

There are two kinds of relevance (a) question relevance, and (b) topicrelevance Both are matters of degree For example, on a very basic level, if the

question is q: “Number of cars in California?” and the available information

is p: “Population of California is 37,000,000,” then what is the degree of relevance of p to q? Another example: To what degree is a paper entitled

“A New Approach to Natural Language Understanding” of relevance to thetopic of machine translation

Basically, there are two ways of approaching assessment of relevance (a)semantic, and (b) statistical To illustrate, in the number of cars example,

relevance of p to q is a matter of semantics and world knowledge In existing

search engines, relevance is largely a matter of statistics, involving counts

of links and words, with little if any consideration of semantics Assessment

of semantic relevance presents difficult problems whose solutions lie beyondthe reach of bivalent logic and probability theory What should be noted isthat assessment of topic relevance is more amendable to the use of statisticaltechniques, which explains why existing search engines are much better atassessment of topic relevance than question relevance

The third obstacle is deduction from perception-based information As a

basic example, assume that the question is q: What is the average height

of Swedes?, and the available information is p: Most adult Swedes are tall.

Another example is: Usually Robert returns from work at about 6 p.m What

is the probability that Robert is at home at 6:15 p.m.? Neither bivalent logicnor probability theory provide effective tools for dealing with problems of thistype The difficulty is centered on deduction from premises which are bothuncertain and imprecise

Underlying the problems of world knowledge, relevance, and deduction is

a very basic problem – the problem of natural language understanding Much

of world knowledge and web knowledge is expressed in a natural language

A natural language is basically a system for describing perceptions Sinceperceptions are intrinsically imprecise, so are natural languages

A prerequisite to mechanization of question-answering is mechanization

of natural language understanding, and a prerequisite to mechanization ofnatural language understanding is precisiation of meaning of concepts andproposition drawn from a natural language To deal effectively with worldknowledge, relevance, deduction, and precisiation, new tools are needed Theprincipal new tools are: precisiated natural language (PNL); protoform theory(PFT), and the generalized theory of uncertainty (GTU) These tools aredrawn from fuzzy logic – a logic in which everything is, or is allowed to be, amatter of degree

The centerpiece of the new tools is the concept of a generalized constraint.The importance of the concept of a generalized constraint derives from thefact that in PNL and GTU it serves as a basis for generalizing the univer-sally accepted view that information is statistical in nature More specifically,the point of departure in PNL and GTU is the fundamental premise that, ingeneral, information is representable as a system of generalized constraints,

World Knowledge, Relevance, Deduction, and Precisiation 3with statistical information constituting a special case This, much more gen-eral, view of information is needed to deal effectively with world knowledge,relevance, deduction, precisiation, and related problems.

In summary, the principal objectives of this paper are (a) to make a casefor the view that a quantum jump in search engine IQ cannot be achievedthrough the use of methods based on bivalent logic and probability theory;and (b) to introduce and outline a collection of nonstandard concepts, ideas,and tools which are needed to achieve a quantum jump in search engine IQ

Acknowledgement Research supported in part by ONR N00014-02-1-0294, BT

Grant CT1080028046, Omron Grant, Tekes Grant and the BISC Program of UCBerkeley

A Fuzzy Approach to Optimal R&D Project Portfolio Selection

1 Introduction

A major advance in the development of strategic investment selection toolscame with the application of options reasoning to the fields of Research and

Development (R&D) By real options we understand the opportunity to invest

in and thus support a project opportunity that essentially involves acquisition

or building of real assets In every step of the investment program, whenmaking the appropriate entry (or exit) decisions, we also have to take intoconsideration that the underlying projects can open or close the possibility forfurther options (which might be more profitable) Defining phases and activelyscheduling and managing investment activities, we can collect information todecide whether we are ready to go ahead with the investment or not

Formulating from this point of view, we seek to correct the deficiencies

of traditional investment valuation methods by incorporating the ial flexibility that can (and usually does) bring significant value to projects.From our experience, we found that the main issue in the options approach

manager-to strategic project valuation is the correct characterization of the tical imprecision that we encounter when judging or estimating future cashflows Working out schemes for phasing and scheduling systems of interrelatedprojects, we will develop a basic model for valuing options on R&D invest-ment opportunities, when future revenues and expected costs are estimated bytrapezoidal possibility distributions Furthermore, drawing on our results, weshall present a fuzzy mixed integer programming model for the R&D optimalproject portfolio selection problem

nonstatis-The real options valuation methods were first tried and implemented as tools for working with very large industrial investments, also called as giga-

investments They presented a unique source of income for corporations

through capturing significant market share from their rivals However, thoseopportunities were often left abandoned due to the huge risks and uncer-tainties: there was fear that capital invested in very large projects, with anexpected life time of more than a decade is not very productive and that their

1

The Waeno project; Tekes 40470/00

6 C Carlsson et al.

Investment opportunities of R&D types compete for major portions of therisk-taking capital, and as their outcome is particularly uncertain, compro-mises have to be made on their productivity The short-term productivitymay not be high, although the overall return of the investment program can

be forecasted as very good Another way of motivating an R&D investment

is to point to strategic advantages, which would not be possible without theknowledge that the investment yields Thus, R&D projects do offer some in-direct (intangible) returns as well

Our experience shows that R&D investments made in the paper and pulpindustry face fierce competition and scenarios of slow growth (2–3% p.a.) intheir key market segments However, this environment does not prevent othermore effective competitors to gain footholds in their main markets

There are other issues Global financial markets make sure that capitalcannot be used nonproductively, as its owners are offered other opportunities,and the capital will move (often quite fast) to capture these opportunities Thecapital market has learned “the American way,” i.e., there is a shareholderdominance among the actors, which has often brought short-term shareholderreturn to the forefront as a key indicator of success, profitability, and produc-tivity There are also lessons learned from the Japanese industry, which point

to the importance of immaterial investments They show that investments

in buildings, production, and supporting technologies become enhanced withimmaterial investments, and that these are even more important for furtherinvestments and gradually growing maintenance investments

The core products and services created by R&D investments are enhancedwith life-time services, with gradually more advanced maintenance and finan-cial add-in services These features make it difficult to actually assess theproductivity and profitability of the original R&D project, especially if theproducts and services are repositioned to serve other (e.g., emerging) markets.New technology and enhanced technological innovations have been changingthe life cycle of R&D investments The challenge is to find the right time andthe right innovation to modify the life cycle in an optimal way Technologyproviders are actively involved throughout the life cycle of R&D projects,which actually changes the way we assess the profitability and the productiv-ity of such investments

R&D projects, and in particular, portfolios of R&D projects generate mitments, which possess:

com-1 Long life cycles (taking into account their possible impacts on other vestments)

in-2 Uncertain (i.e., vague), sometimes overly optimistic or pessimistic future

cash flow estimates

3 Uncertain (i.e., biased ), sometimes questionable profitability estimates

4 Imprecise assessments of future effects on productivity, market positions,competitive advantages, and shareholder value

5 The ability to generate series of further investments

A Fuzzy Approach to Optimal R&D Project Portfolio Selection 7Jensen and Warren [14] propose to use options theory to value R&D inthe telecom service sector The reasons are rather similar to those we iden-tified above: research managers are under pressure to explain the value ofR&D programs to the senior management, and at the same time they need

to evaluate individual projects to make management decisions on their ownR&D portfolio The research in real options theory has evolved from generalpresentations of flexibility of investments in industrial cases to more theoret-ical contributions, which resulted in the application of real option valuationmethods to industrial R&D projects The term real option was introduced byKester [15] and Myers [20] in 1984 The option to postpone an investmentopportunity was discussed by McDonald and Siegel [22] Pakes [23] consid-ered patents as options Siegel et al [24] discussed the option valuation ofoffshore oil properties Majd and Pindyck [21] analyzed the optimal time andcomputed the option value of building operations in investment decisions

A fundamental book on managerial flexibility and strategy in resource cation, written by Trigeorgis [25], presented a theory of real options Abel

allo-et al [1] discussed a theory of option valuation of real capital and ments Faulkner [13] discussed the application of real options to the valuation

invest-of R&D projects at Kodak Kulatilaka et al [16] discussed a capability-basedreal options approach to managing information technology investments.The use of fuzzy sets to work with real options is a novel approach, whichhas not been considered and analyzed widely so far One of the first results toapply fuzzy mathematics in finance was presented by Buckley [4], where heworked out how to use fuzzy sets to formulate the concepts of future value,

with fuzzy internal rate of return in the context of investment decisions to

a method for managing capital budgeting problems with fuzzy cash flows.However, there are a growing number of papers in the intersection of the

disciplines of real options and fuzzy sets In one of the first papers on

real option valuation method Muzzioli and Torricelli [19] used fuzzy sets to

optimal timing of investment opportunities with fuzzy real options Carlsson

et al [10,12] developed and tested a method for project selection with optimaltiming and scheduling by using the methodology of fuzzy real options Majlen-der [18] presented a comprehensive overview of the development of investmentvaluation methods in a possibilistic environment

2 Real Options for R&D Portfolios

The options approach to R&D project valuation seeks to correct the cies of traditional methods of valuation that are based on the methodologies

deficien-of net present valuation (NPV) and discounted cash flow (DCF) analyses,

8 C Carlsson et al.

through the recognition of managerial flexibility and interaction with theunderlying investment opportunities This uncertainty can bring significantvalue to a project

Real options in option thinking are based on the same principles asfinancial options In real options, the options involve “real” (i.e., productive)assets as opposed to financial ones, where the options relate to some financialinstruments [2] To have a “real option” means to have the possibility for

a certain period of time to either choose for or against something, withoutbinding ourselves up front The value of a real option is computed by [17]

is the annualized continuously compounded rate on a safe asset, T is the time to maturity of the option in years, and σ stands for the uncertainty of

the probability that a random draw from a standard normal distribution will

Where the maximum deferral time is T , make the investment (i.e., exercise

is positive and attends its maximum value That is,

t=0,1, ,T {V t e −δt N (d1)− Xe −r f t N (d2)} > 0, (1)where

A Fuzzy Approach to Optimal R&D Project Portfolio Selection 9

r is the project-specific risk-adjusted discount rate.

Of course, this decision rule has to be reapplied each time when new mation arrives during the deferral period to see how the optimal investmentstrategy changes in light of the new information From a real option perspec-tive, it can be worthwhile to undertake R&D investments with a negative netpresent value (NPV), when early investment can provide information aboutfuture benefits or losses of the whole investment program

infor-3 A Hybrid Approach to Real Option Valuation

is of the following form:

Usually, the present value of the expected cash flows cannot be terized by a single number However, they can be estimated by a trapezoidalpossibility distribution of the form

charac-˜

That is, the most possible values of the present value of the expected cash

flows lie in the interval [a, b] (which is the core of the trapezoidal fuzzy number

˜

for the present value of the expected cash flows In a similar manner, wecan estimate the nominal value of the expected costs by using a trapezoidalpossibility distribution of the form

˜

X = (a  , b  , α  , β  ).

fuzzy-probabilistic formula for computing fuzzy real option values

˜

C = ˜S e −δT N (d )− ˜ Xe −r f T N (d ), (2)

a similar formula to (1) for the optimal investment strategy in a possibilisticsetting.

4 A Possibilistic Approach to R&D Portfolio Selection

Facing a set of project opportunities of R&D type, the company is usuallyable to estimate the expected investment costs of the projects with a high

a crisp number However, the cash flows received from the projects do involveuncertainty, and they are modeled by trapezoidal possibility distributions Let

us fix a particular project of length L and maximal deferral time T with cash

flows

˜

cfi = (A i , B i , Φ i , Ψ i ), i = 0, 1, , L.

Now, instead of the absolute values of the cash flows, we shall consider

their fuzzy returns on investment (FROI) by computing the return that we

receive on investment X in year i of the project as

2

A Fuzzy Approach to Optimal R&D Project Portfolio Selection 11

We compute the fuzzy net present value of the project by

can be postponed by a maximum of T years, then we will define the value of

its possibilistic deferral flexibility by

F = (1 + σ( ˜ R0))× (1 + σ( ˜ R1))× · · · × (1 + σ( ˜ R T −1))× FNPV,

is the decision variable associated with project i, which takes value one if the project i starts now (i.e., at time zero) and takes value zero if it is postponed

of project i (i.e., the capital expenditure required to keep the associated real

deferral flexibility of project i, respectively, i = 1, , N

In our approach to fuzzy mathematical programming problem (3), we have

ν( F) =E( F) − τ × σ(F) × X,

Since R&D projects are characterized by the long planning horizon andvery large uncertainty, the value of managerial flexibility can be substantial.Therefore, the fuzzy real options model is quite practical and useful The stan-dard work in the field use probability theory to account for the uncertaintiesinvolved in future cash flow estimates This may be defended for financialoptions, for which we can assume the existence of an efficient market with

12 C Carlsson et al.

numerous players and numerous stocks for trading, which in turn justifies theassumption of the validity of the laws of large numbers and thus the use of sta-tistical methods The situation for real options associated with an investmentopportunity of R&D type is quite different The option to postpone an R&Dproject does have consequences, which differs from efficient markets, as thenumber of players producing the consequences is quite small The imprecision

we encounter when judging or estimating future cash flows is nonstochastic

by nature, and the use of probability theory can give us a misleading level ofprecision and a notion that the consequences are somehow repetitive This isnot the case, since in our case the uncertainty is genuine, i.e., we simply donot know the exact level of future cash flows Without introducing fuzzy realoption models, it would not be possible to formulate this genuine uncertainty.The proposed model that incorporates subjective judgments as well as sta-tistical uncertainties can give investors a better understanding of the problemwhen making R&D investment decisions

5 Summary

Multinational enterprises with large R&D departments often face the culty of selecting an appropriate portfolio of research projects The cost ofdeveloping a new product or technology is low as compared to the cost of itsintroduction to the global market The NPV rule and other discounted cashflow techniques for making R&D investment decisions seem to be inappropri-ate for selecting a portfolio of R&D projects, as they favor short-term projects

diffi-in relatively certadiffi-in markets over long-term and relatively uncertadiffi-in markets.Since many new products are identified as failures during the R&D stages,the possibility of refraining from market introduction can add a significantvalue to the NPV of the R&D project Therefore R&D investments can beinterpreted as the price of an option on major follow-on investments

selection problem by a fuzzy 0–1 mathematical programming problem, wherethe optimal solution(s) defined the optimal portfolio(s) of R&D projects withthe biggest (aggregate) possibilistic flexibility value

References

1 A.B Abel, A.K Dixit, J.C Eberly and R.S Pindyck, Options the value of

capital, and investment Quarterly Journal of Economics, 3 (1996) 753–758

2 J Alleman and E Noam (eds.), The New Investment Theory of Real Options

and its Implication for Telecommunications Economics Kluwer, Boston, 1999

3 M Benaroch and R.J Kauffman, Justifying electronic banking network

expan-sion using real options analysis MIS Quarterly, 24 (2000) 197–225

A Fuzzy Approach to Optimal R&D Project Portfolio Selection 13

4 J.J Buckley, The fuzzy mathematics of finance Fuzzy Sets and Systems, 21

(1987) 257–273

Centre for Computer Science Technical Report No 211 (1998)

Mathware and Soft Computing, 6 (1999) 81–89

for Computer Science Technical Report No 367 (2000)

In: Proceedings of the EUROFUSE 2001 Workshop on Preference Modelling and

Applications, pp 235–239, 2001

numbers Fuzzy Sets and Systems, 122 (2001) 315–326

options In: Proceedings of the Second International Symposium of Hungarian

Researchers on Computational Intelligence, pp 81-88, 2001

and Systems, 139 (2003) 297–312

project evaluation In: Y Liu, G Chen and M Ying (eds.), Proceedings of

the Eleventh IFSA World Congress University Press and Springer, Beijing and

Berlin Heidelberg New York, pp 1650–1654, 2005

13 T.W Faulkner, Applying options thinking to R&D valuation Research

Tech-nology Management, (1996) 50–56

14 K Jensen and P Warren, The use of options theory to value research in the

service sector R&D Management, 31 (2001) 173–180

15 W.C Kester, Today’s options for tomorrow’s growth Harvard Business Review,

62 (1984) 153–160

16 N Kulatilaka, P Balasubramanian and J Storck, Managing information nology investments: a capability-based real options approach Boston UniversitySchool of Management, Working Paper No 96-35, 1996

tech-17 K.J Leslie and M.P Michaels, The real power of real options The McKinsey

Quarterly, 3 (1997) 5–22

18 P Majlender, A Normative Approach to Possibility Theory and Soft Decision

Support TUCS – Turku Centre for Computer Science, Dissertation No 54

(2004)

19 S Muzzioli and C Torricelli, Combining the theory of evidence with fuzzy setsfor binomial option pricing, Materiale di discussione n 312 Dipartimento diEconomia Politica, Universita degli Studi di Modena e Reggio Emilia, 2000

20 S.C Myers, Finance theory and financial strategy, lnterfaces 14 (1984) 126–137

21 S Majd and R Pindyck, Time to build, option value and investment decisions

Journal of Financial Economics, 18 (1987) 7–27

22 R.L McDonald and D.R Siegel, The value of waiting to invest Quarterly

Jour-nal of Economica, 101 (1986) 707–27

23 A Pakes, Patents as options: some estimates of the value of holding European

patent stocks Econometrica, 54 (1986) 755–784

24 D.R Siegel, J.L Smith and J.L Paddock, Valuation of offshore oil properties

with option pricing Midland Corporate Finance Journal, (1987), 22–30

25 L Trigeorgis, Real Options: Managerial Flexibility and Strategy in Resource

Allocation MIT, Cambridge, MA, 1996

capital, and investment Quarterly Journal of Economics, (1996) 753–758

2 J Alleman and E Noam (eds.), The New Investment Theory. ..

Researchers on Computational Intelligence, pp 81-88, 2001

and Systems, 139 (2003) 297–312

project evaluation In: Y Liu, G Chen and M Ying (eds.), Proceedings... 2000

20 S.C Myers, Finance theory and financial strategy, lnterfaces 14 (1984) 126–137

21 S Majd and R Pindyck, Time to build, option value and investment decisions

Journal