Fuzzy cooperative games. Cooperation with vague expectations (Q5942975): Difference between revisions

Traditional game theory describes conflict situations in which, once each side has picked up its strategy, we know the set of possible outcomes, the relative importance of each outcome to each player, and the probability of different outcomes. In reality, often, we can only make vague expert predictions about the probabilities and/or utility of different outcomes, expert predictions that are formulated only in terms of words from natural language, like ``It is somewhat possible that the negotiations will fail''. Such statements can be formalized by using fuzzy logic and related techniques that were specifically invented to handle such ``linguistic'' uncertainty. It is therefore important to extend traditional game theory techniques to games in which for each combination of strategies and for each player, the expected outcome is characterized not by a single number, but rather by a fuzzy number -- i.e., by a function \(\mu(x)\) that maps, to every possible real number \(x\), a degree to which this \(x\) can be the actual expected outcome. There has been a lot of research in fuzzy games; the author concentrates on fuzzy cooperative games. He starts with overviewing the known results and going into the detailed analysis of how different solution concepts can be extended from non-fuzzy (``crisp'') games to general fuzzy games and whether the known properties of these solution concepts still hold for fuzzy games. For fuzzy games with side payments, the main emphasis is on the core and Shapley value. In particular, it turns out that a natural (``verbatim'') fuzzy analog of the Shapley value does not satisfy the usual properties of the non-fuzzy Shapley value; the author shows, however, that it is possible to construct a more sophisticated analog that satisfies these properties. For fuzzy games without side payments, the emphasis is on the core and effective coalitions. The author also expands his results to a more general case of partial side payments that includes games with side payments and games without side payments as particular cases. The book starts with a brief introduction to fuzzy techniques and to traditional (non-fuzzy) cooperative game theory. As a result, this book can be used not only by specialists in fuzzy games, but also by fuzzy researchers who are not very familiar with game theory, and by students and researchers in game theory who are not familiar with the basics of fuzzy techniques.