Séminaire Donnees et APprentissage Artificiel
Fuzzy sets and multicriteria decision making
Intervenant(s) : Radko MESIAR (STU Bratislava, Slovakia)
Some fuzzy set theory-based methods of building preference structures for alternatives described by score vectors are given. In the first part, fuzzy measures based methods are discussed and a limit based approach yielding two different extensions of leximax/leximin for score vectors of different size is introduced. Second part is devoted to the fuzzy logic-based construction of preference relations, including several examples. Finally, we deal with fuzzification of score vectors into fuzzy quantities and then the orderings of fuzzy quantities can be applied. As an example, MOM-defuzzification-based ordering gives a method generalizing the original Yager penalty-based approach to aggregation. Moreover, this method allows to introduce the weights (importances) of single criteria even into originally non-symmetric aggregation, as well as to generalize OWA operators.
Plus d'informations ici
Javier.Diaz (at) null