A Prediction Model for Individual Decision Making Behavior under Fuzzy Decision Problems

Most decision problems are inherently fuzzy due to of unable precisely defined or the lack of enough information. Thus these decision problems can properly be formulated by fuzzy sets. Such problems differ from the traditional mathematical programming problem of maximizing a well-defined function over a set of alternatives represented by crisp values. However, when alternatives are represented in terms of fuzzy values, one important question remains: how can one choose one alternative from a given set of alternatives so as to achieve, in some sense, "the best" overall fuzzy value. Current fuzzy decision making methods do not account for subjective values of the decision maker. In this paper the concept of the decision maker's attitude toward fuzziness is defined and a new fuzzy decision making method is derived. The feasibility and validity of the new method is ascertained through experiments. Benchmarking with the indices of preference reversals revealed by the traditional expected utility theory, the new approach leads to better prediction of individual's actual choices.