Fuzzy multiple objective decision making. Methods and applications (Q1323497)

This work is the continuation of a monograph by the same authors [`Fuzzy mathematical programming. Methods and applications', Lect. Notes Econ. Math. Syst. 394 (1992; Zbl 0793.90094)]. It is aimed to ``provide readers with a capsule look into all existing fuzzy programming methods for solving multiple objective decision making, fuzzy multiple objective decision making and possibilistic multiple objective decision problems'' (cf. page 1). We are not sure that this ambitious goal was really achieved. What we can say is that this book contains a worth studying collection of real-life problems reflecting difficulties encountered in the practice of decision making by using ``crisp'' optimization methods when given objectives and restrictions are ``fuzzy''. In spite of the fact that the presentation and the classification of various types of problems are rather confuse and in spite of the fact that, similarly to what happens in the book cited above, the authors sometimes make licentious use of mathematical terms, notations and results. Reading this book through mathematical glasses leads to some interesting conclusions. It seems to us that the most important of them is that in many real-life decision making problems ``softening'' the objectives and the restrictions by requiring not ``the best'' decision but one which is ``good enough'' is not only acceptable but even desirable from a practical point of view. What ``good enough'' means is now a matter of debate. The authors suggest that a proposed solution is ``good enough'' when its membership degrees to the sets of optimal and feasible solutions is sufficiently high. How membership degrees to various sets are determined is a matter of experience. Readers who disagree with some artificial methods of computing membership degrees involved in the book under review, may try some other methods themselves. Of course, changing the way in which membership degrees are computed may completely change the ``good enough solution'' the computations produce. The question is which of the good enough solutions is acceptable from a practical point of view. In many of their examples, the authors have defined membership degrees in such a way that the good enough solutions they obtain are practically acceptable. However, this is no proof that employing similar methods for solving other decision making problems will lead to practically acceptable conclusions. Therefore, students and practitioners trying to use this book as a source of rules of solving decision making problems should be careful. Nevertheless, critical use of the material summarized in this work can help them improve their personal experience. This also applies to the recepees of solving decision making problems the author call ``algorithms''. For mathematicians having some experience with linear programming techniques it may be easy to produce numerical examples to which application of one ``algorithm'' or another will fail for lack of feasible and/or optimal solutions at various computational stages. Fact is that building mathematically valid algorithms for solving decision making problems involving fuzziness of objectives and restrictions is not so easy and this is clearly revealed by a careful reading of the book under review.