Rectification of Preferences in a Fuzzy Environment

Abstract

In this paper we consider a complete fuzzy preference structure, defined by means of a constructive approach associated to the necessary learning process in a decision making problem. Preference relations are successively assigned over a given set of alternatives with the possibility of revision and rectification, subject to a certain set of beliefs which sustain the construction of different viewpoints. Measures of information are then considered in this process for a formal identification of uncertainty due to ignorance, a concept that is closely related to ambiguity as defined by Fishburn (1993). Finally, a coherence criterion is proposed in order to study the importance of preferences, in such a way that an indirect and comparative analysis between them is needed so outcomes may be coherently organized.