Ranking alternatives on the basis of generic constraints and examples - A possibilistic approach

Romain Gérard (1) , Souhila Kaci (2) , Henri Prade (1)

1 IRIT-ADRIA - Argumentation, Décision, Raisonnement, Incertitude et Apprentissage
2 CRIL - Centre de Recherche en Informatique de Lens

The paper presents and discusses a method for rank-ordering alternatives on the basis of constraints induced by generic principles (expressing for instance the relative importance of criteria), or by examples of orderings between particular alternatives, without resorting to the use of an aggregation operation for evaluating the alternatives. The approach, which remains qualitative, is based on the minimal specificity principle of possibility theory in order to complete the constraints. It is compared on an illustrative example to an aggregation-based approach using Choquet integral. The way constraints expressed in the Choquet integral setting translate into constraints in the proposed approach is discussed.

Domaines: Intelligence artificielle [cs.AI]

Romain Gérard, Souhila Kaci, Henri Prade. Ranking alternatives on the basis of generic constraints and examples - A possibilistic approach. 12th International Joint Conference on Artificial Intelligence (IJCAI 2007), Jan 2007, Hyderabad, India. pp.393-398. ⟨hal-00438313⟩