Uma Abordagem para Jogos Bayesianos Fuzzy com base na Probabilidade Fuzzy de Buckley

This paper presents an approach for fuzzy Bayesian games, based on the notion of fuzzy probabilities to estimate unknown types of agents in strategic situations. For that, we consider the notion of fuzzy probability introduced by Buckley and Eslami, who use fuzzy numbers to represent uncertain and imprecise probabilities, without, however, applying the fuzzy probability theory, but considering an arithmetic restriction to the interval [0; 1]. The paper formalizes this notion of fuzzy probability, defining fuzzy Bayesian games of 2 agents with n possible types. We also discuss, by analyzing some examples, the dependence of the model solution on the order relation adopted in the fuzzy number space. Then, we suggest the use of the total order relation of Asmus-Dimuro in order to find the Nash equilibrium solution.