- Email: [email protected]
Fuzzy decision procedures with binary relations. Towards a unified theory
384
Bulletin
that these contributions have been prepared with greater care than the theoretical chapters provided by the editors.) Let me finally relativize the above criticisms as follows: I am very happy to inform you that I know other books from Prentice Hall, which are on a much higher level with respect to both content and form. Besides, from the middle of 1992 until the end of October 1993 there have appeared (especially in Germany and in German) such a lot of books and reports (I suppose, most of them under very strong time restrictions) that, realistically, one simply has to expect some of them to be of low quality. I could give examples from at least three otherwise respectable publishing companies other than Prentice Hall. It's a pity that during the current 'fuzzy rush hours' cash sometimes seems to be the strongest argument. Dr. Herbert Toth
RiedstraBe 17a A-1140 Wien, Austria
Fuzzy Logic. State of the
Art
'The present book finds its root in the IFSA'91 World Conference held in Brussels in July 1991. A number of contributions to that conference were selected via a refereeing process and were consequently chosen as a basis for extended contributions in the present volume. Since the final contributions were prepared approximately one year after the conference, all authors were requested to include also their more recent findings in their contributions, even if this meant deviating considerably from the original contribution for the conference. The contents of this volume are therefore more recent than its link with IFSA'91 leads to believe. For this reason the title also does not refer to IFSA'91. Nevertheless, as was the case for the conference itself, the contributions are divided over four topical chapters. Each chapter dealing with a major area of research of fuzzy logic or fuzzy set theory in general' (p. ix, Preface). This part from the preface by the editors (R. Lowen and M. Roubens) gives a very good and informative characterization of the book. And as one can easily conclude, the book (Kluwer Academic Publishers, Dordrecht, 1993, xi + 587 pp) presents a considerable number of papers concerning with various aspects of fuzzy set theory. So the title of the book follows the questionable new trend to use the specialized notion of fuzzy logic instead of the general and correct notion of fuzzy set theory. This is, however, only a minor objection. Let me now give a very short overview about the contents and structure of the book: (1) Mathematics (pp. 1-177; 17 papers). Besides six papers on e.g. MV-algebras, relations etc. we can also find interesting contributions of a more fundamental nature dealing with e.g. possibility/probability transfor-
mations (Dubois et al.) and empirical possibility, minimal information distortion (Joslyn) and reasoning with conditional information (Nguyen). (2) Artificial intelligence (pp. 179-261; 7 papers). This part presents contribution on e.g. theorem provers, rules of inference and approximate reasoning, knowledge bases and pattern classification. (3) Computer, management and systems science (pp. 263-462; 18 papers). Papers on fuzzy (linear, non-linear, 0-1) programming and optimization are accompanied by contributions on classifier systems, fuzzy data analysis as well as on the context model (Kurse et al.), on the meaning of fuzziness in random experiments (Gil) and on information-preserving probability-possibility transformations (Klir and Geer). (4) Engineering (pp. 463-587; 11 papers). The last section of the book deals with e.g. self-tuning methods of inference rules, stability analysis of fuzzy control systems, adaptive fuzzy control and various other aspects of this research area. As you do perhaps already suppose from the above, I do not hesitate to recommend the book as a valuable source of interesting and high-quality information on various topics of fuzzy set theory. However, I also do not hesitate to close with my main objection which concerns the price of the book (Dfl 320.00/US$186.50/UK£131.50) which makes it a must for libraries rather than for single persons.
Dr. Herbert Toth
RiedstraBe 17a A- 1140 Wien, Austria
Fuzzy Decision Procedures with Binary Relations. Towards a Unified Theory The book (Kluwer Academic Publishers, Dordrecht, 1993, ISSN 0-7923-2367-X) is a well-thought out presentation of a self-contained approach designed by the author Leonid Kitainik, possessing both theoretical a n d though less directly-practical value. It is interesting to note that the approach starts with the classical 'crisp' notions connected with the binary preference relations and rationality, to then go on to respective fuzzy theoretical considerations, linked already more directly with the choice procedures and ranking of alternatives. The notions and properties of fuzzy inclusion, fuzzy implication, contensiveness and efficiency form the basis for establishment of effective fuzzy decision procedures. The chapter devoted to fuzzy inclusion seems to be of special interest, with a number of original and important theoretical results concerning properties of fuzzy inclusions. One of the points raised by the author is the two-way traffic of results and ideas between the crisp and the
Bulletin fuzzy formalisms and approaches. Thus, for instance, there is a special chapter devoted to application of some of the results obtained for the fuzzy environment to the crisp choice rules, although, in fact, many of the results presented on the way apply either equally to fuzzy and crisp elements of decision theories or to both of them in a specific manner. The last chapter describes the applications to Decision Support System construction and operation, based upon the results presented earlier in the book. The main body of the book is devoted to discussion of the foundations for the integration of descriptive and prescriptive decision approaches to the preference domain, and for the comparative study of the choice rules available, and in fact special attention is paid to the latter. Thus, with respect to the choice rules it is stated that the traditional way of assessment derived from the choice theory, founded on the combinations of rational choice axioms, turns out to be too restrictive, while relaxation of these rationality requirements yields results which are dissatisfying, since there is little to be said of these relaxed, and sometimes extravagant rules. Thus, the author proposes a new approach, the approach of the Fuzzy Decision Procedures with Binary Relations, founded upon the analysis of rationality precepts of more general nature. The author suggests thereafter that application of what he calls the 'extended choice logic', related, in particular, to adoption of multifold choice, allows absorption of uncertainty by the decision rules developed. Furthermore, the properties obtained make it possible to narrow down the class of acceptable decision rules and to question some of the well established ones. The literature quoted, though not very extensive, is very much to the point, without unnecessary references to a vast body of publications on the subject, of which only a part would be in fact useful for the discourse presented. This feature is also related to the nature of the book, which is, as said, to a large extent selfcontained. If at this point the book leaves something to be desired it is a discussion of some classes of real cases, which would make the use of certain properties better justified, since decision making procedures, whether fuzzy or crisp, whether allowing multifold choices or leading to point-wise solutions, may not always be applied via automatic devices but will often be confronted with human analysts and experts, who might have definite views of particular aspects of the decision process itself, and not only of the subject matter. The book is certainly a valuable theoretical contribution to the domain, and can also serve as the source of information for those informed developers of decision support systems who would like to extend their scope of justifications and instruments. Jan W. Owsifiski Systems Research Institute Polish Academy of Sciences Newelska 6, 01-447 Warszawa Poland
385
Conference Announcements and Call for Papers EUFIT "94 The ELITE Foundation (European Laboratory for Intelligent Techniques Engineering) is pleased to announce EUFIT ' 9 4 - T h e Second European Congress on Intelligent Techniques and Soft Computing. EUFIT '94 aims to bring together scientists and practitioners from academic, governmental, and industrial institutions to discuss new developments and results in the field of intelligent technologies. The congress will take place in Aachen (Aix la Chapelle), Germany, on September 20-23, 1994. Charlemagne, the then Emperor of Europe, resided in the medieval Aachen in the 8th century. Here, the charm of 1 200 years of history merges with the most advanced technological research, since the 'RW'rH', the largest European Institute of Technology is located in Aachen. Structure of the Congress Tutorials: September 20, 1994 Conference: September 21-23, 1994 Exhibition: September 21-23, 1994 Working Groups: September 20-23, 1994 Tutorials The tutorials (1 day and ~ day) are offered as an introduction for newcomers to the areas. The lectures will be presented by well known researchers and recognized practitioners. The tutorials will facilitate the understanding of the presentations of the conference. They include the basics and applications, e.g. on Fuzzy Technology, Neural Nets, Fuzzy Control, Genetic Algorithms, Data Analysis, and other topics. Conference The conference will be composed of plenary presentations, semi-plenary sessions, panel sessions, invited and contributed sessions, and other events. There will be parallel and interactive sessions. They include scientific and applied presentations in Fuzzy Technology, Neural Nets, and Genetic Algorithms. In interactive sessions speakers can communicate freely with the interested audience up to 60 minutes. Topics to be covered will be: • Basic concepts and mathematical aspects. • Methodologies such as Artificial Intelligence, Expert Systems, Decision Support Systems, Pattern Recognition, Computer Vision, Clustering, Control, Supervision, Adaptation, Data Analysis, Software and Hardware Support. • Applications, for example, in Engineering, Process Control, Quality Control, Measurement, Signal Analysis, Production, Medicine, Finance. Exhibition There will be an international software, hardware and book exhibition (like EUFIT '93 with at least 20 ex-
Bulletin
that these contributions have been prepared with greater care than the theoretical chapters provided by the editors.) Let me finally relativize the above criticisms as follows: I am very happy to inform you that I know other books from Prentice Hall, which are on a much higher level with respect to both content and form. Besides, from the middle of 1992 until the end of October 1993 there have appeared (especially in Germany and in German) such a lot of books and reports (I suppose, most of them under very strong time restrictions) that, realistically, one simply has to expect some of them to be of low quality. I could give examples from at least three otherwise respectable publishing companies other than Prentice Hall. It's a pity that during the current 'fuzzy rush hours' cash sometimes seems to be the strongest argument. Dr. Herbert Toth
RiedstraBe 17a A-1140 Wien, Austria
Fuzzy Logic. State of the
Art
'The present book finds its root in the IFSA'91 World Conference held in Brussels in July 1991. A number of contributions to that conference were selected via a refereeing process and were consequently chosen as a basis for extended contributions in the present volume. Since the final contributions were prepared approximately one year after the conference, all authors were requested to include also their more recent findings in their contributions, even if this meant deviating considerably from the original contribution for the conference. The contents of this volume are therefore more recent than its link with IFSA'91 leads to believe. For this reason the title also does not refer to IFSA'91. Nevertheless, as was the case for the conference itself, the contributions are divided over four topical chapters. Each chapter dealing with a major area of research of fuzzy logic or fuzzy set theory in general' (p. ix, Preface). This part from the preface by the editors (R. Lowen and M. Roubens) gives a very good and informative characterization of the book. And as one can easily conclude, the book (Kluwer Academic Publishers, Dordrecht, 1993, xi + 587 pp) presents a considerable number of papers concerning with various aspects of fuzzy set theory. So the title of the book follows the questionable new trend to use the specialized notion of fuzzy logic instead of the general and correct notion of fuzzy set theory. This is, however, only a minor objection. Let me now give a very short overview about the contents and structure of the book: (1) Mathematics (pp. 1-177; 17 papers). Besides six papers on e.g. MV-algebras, relations etc. we can also find interesting contributions of a more fundamental nature dealing with e.g. possibility/probability transfor-
mations (Dubois et al.) and empirical possibility, minimal information distortion (Joslyn) and reasoning with conditional information (Nguyen). (2) Artificial intelligence (pp. 179-261; 7 papers). This part presents contribution on e.g. theorem provers, rules of inference and approximate reasoning, knowledge bases and pattern classification. (3) Computer, management and systems science (pp. 263-462; 18 papers). Papers on fuzzy (linear, non-linear, 0-1) programming and optimization are accompanied by contributions on classifier systems, fuzzy data analysis as well as on the context model (Kurse et al.), on the meaning of fuzziness in random experiments (Gil) and on information-preserving probability-possibility transformations (Klir and Geer). (4) Engineering (pp. 463-587; 11 papers). The last section of the book deals with e.g. self-tuning methods of inference rules, stability analysis of fuzzy control systems, adaptive fuzzy control and various other aspects of this research area. As you do perhaps already suppose from the above, I do not hesitate to recommend the book as a valuable source of interesting and high-quality information on various topics of fuzzy set theory. However, I also do not hesitate to close with my main objection which concerns the price of the book (Dfl 320.00/US$186.50/UK£131.50) which makes it a must for libraries rather than for single persons.
Dr. Herbert Toth
RiedstraBe 17a A- 1140 Wien, Austria
Fuzzy Decision Procedures with Binary Relations. Towards a Unified Theory The book (Kluwer Academic Publishers, Dordrecht, 1993, ISSN 0-7923-2367-X) is a well-thought out presentation of a self-contained approach designed by the author Leonid Kitainik, possessing both theoretical a n d though less directly-practical value. It is interesting to note that the approach starts with the classical 'crisp' notions connected with the binary preference relations and rationality, to then go on to respective fuzzy theoretical considerations, linked already more directly with the choice procedures and ranking of alternatives. The notions and properties of fuzzy inclusion, fuzzy implication, contensiveness and efficiency form the basis for establishment of effective fuzzy decision procedures. The chapter devoted to fuzzy inclusion seems to be of special interest, with a number of original and important theoretical results concerning properties of fuzzy inclusions. One of the points raised by the author is the two-way traffic of results and ideas between the crisp and the
Bulletin fuzzy formalisms and approaches. Thus, for instance, there is a special chapter devoted to application of some of the results obtained for the fuzzy environment to the crisp choice rules, although, in fact, many of the results presented on the way apply either equally to fuzzy and crisp elements of decision theories or to both of them in a specific manner. The last chapter describes the applications to Decision Support System construction and operation, based upon the results presented earlier in the book. The main body of the book is devoted to discussion of the foundations for the integration of descriptive and prescriptive decision approaches to the preference domain, and for the comparative study of the choice rules available, and in fact special attention is paid to the latter. Thus, with respect to the choice rules it is stated that the traditional way of assessment derived from the choice theory, founded on the combinations of rational choice axioms, turns out to be too restrictive, while relaxation of these rationality requirements yields results which are dissatisfying, since there is little to be said of these relaxed, and sometimes extravagant rules. Thus, the author proposes a new approach, the approach of the Fuzzy Decision Procedures with Binary Relations, founded upon the analysis of rationality precepts of more general nature. The author suggests thereafter that application of what he calls the 'extended choice logic', related, in particular, to adoption of multifold choice, allows absorption of uncertainty by the decision rules developed. Furthermore, the properties obtained make it possible to narrow down the class of acceptable decision rules and to question some of the well established ones. The literature quoted, though not very extensive, is very much to the point, without unnecessary references to a vast body of publications on the subject, of which only a part would be in fact useful for the discourse presented. This feature is also related to the nature of the book, which is, as said, to a large extent selfcontained. If at this point the book leaves something to be desired it is a discussion of some classes of real cases, which would make the use of certain properties better justified, since decision making procedures, whether fuzzy or crisp, whether allowing multifold choices or leading to point-wise solutions, may not always be applied via automatic devices but will often be confronted with human analysts and experts, who might have definite views of particular aspects of the decision process itself, and not only of the subject matter. The book is certainly a valuable theoretical contribution to the domain, and can also serve as the source of information for those informed developers of decision support systems who would like to extend their scope of justifications and instruments. Jan W. Owsifiski Systems Research Institute Polish Academy of Sciences Newelska 6, 01-447 Warszawa Poland
385
Conference Announcements and Call for Papers EUFIT "94 The ELITE Foundation (European Laboratory for Intelligent Techniques Engineering) is pleased to announce EUFIT ' 9 4 - T h e Second European Congress on Intelligent Techniques and Soft Computing. EUFIT '94 aims to bring together scientists and practitioners from academic, governmental, and industrial institutions to discuss new developments and results in the field of intelligent technologies. The congress will take place in Aachen (Aix la Chapelle), Germany, on September 20-23, 1994. Charlemagne, the then Emperor of Europe, resided in the medieval Aachen in the 8th century. Here, the charm of 1 200 years of history merges with the most advanced technological research, since the 'RW'rH', the largest European Institute of Technology is located in Aachen. Structure of the Congress Tutorials: September 20, 1994 Conference: September 21-23, 1994 Exhibition: September 21-23, 1994 Working Groups: September 20-23, 1994 Tutorials The tutorials (1 day and ~ day) are offered as an introduction for newcomers to the areas. The lectures will be presented by well known researchers and recognized practitioners. The tutorials will facilitate the understanding of the presentations of the conference. They include the basics and applications, e.g. on Fuzzy Technology, Neural Nets, Fuzzy Control, Genetic Algorithms, Data Analysis, and other topics. Conference The conference will be composed of plenary presentations, semi-plenary sessions, panel sessions, invited and contributed sessions, and other events. There will be parallel and interactive sessions. They include scientific and applied presentations in Fuzzy Technology, Neural Nets, and Genetic Algorithms. In interactive sessions speakers can communicate freely with the interested audience up to 60 minutes. Topics to be covered will be: • Basic concepts and mathematical aspects. • Methodologies such as Artificial Intelligence, Expert Systems, Decision Support Systems, Pattern Recognition, Computer Vision, Clustering, Control, Supervision, Adaptation, Data Analysis, Software and Hardware Support. • Applications, for example, in Engineering, Process Control, Quality Control, Measurement, Signal Analysis, Production, Medicine, Finance. Exhibition There will be an international software, hardware and book exhibition (like EUFIT '93 with at least 20 ex-