Social choice axioms for fuzzy set aggregation
Résumé
The problem of aggregating n fuzzy sets F1, F2,.., Fn on a set ω is viewed as one of merging the opinions of n individuals (e.g. experts) that rate objects belonging to ω. This approach contrasts with the pure set-theoretic point of view, and leads to interpreting already known axioms underlying fuzzy connectives in a way different from that of multiple criteria aggregation. Various natural properties of a voting procedure, including the ones proposed by Arrow are expressed in the fuzzy set setting. A number of conditions limiting the choice of fuzzy set operations are proposed and classified according to whether they are imperative, mainly technical, or facultative. Families of solutions are obtained that include those proposed in earlier works. The case of non-homogeneous groups is briefly examined. Lastly the application of the voting paradigm to the management of antagonistic local decision rules in knowledge-based systems is outlined.