Article révisé par les pairs
| Résumé : | In multi-criteria decision aid, one usually distinguishes three main families of methods: aggregating, outranking and interactive approaches. Among the outranking methods, PROMETHEE II allows decision makers to obtain a complete ranking of the alternatives regarding multiple – usually conflicting – criteria. Despite its popularity and the various theoretical contributions that have been made to it, few work has been focused on the influence of the alternatives’ evaluations on the final ranking. In this contribution, this issue is addressed by proposing a solution to the following question: given a ranking of the alternatives, what are the minimal modifications required on the evaluations of a given alternative in order to improve its position? A novel exact algorithmic procedure is proposed to answer this question even when multiple criteria are involved. In such cases the solution takes the form of a Pareto optimal front. A proof of the correctness of the algorithm and a complexity analysis are provided. Finally, experimental tests of the method on a real life dataset is discussed along with some indications on how it could be presented to decision makers. |
