Incremental elicitation of the criteria weights of SRMP using a regret-based query selection strategy

International audienceThe Simple Ranking Method using Reference Profiles (or SRMP) is a Multi-Criteria Decision Aiding (MCDA) technique based on the outranking paradigm, which allows to rank decision alternatives according to the preferences of a decision maker (DM). Inferring the preference parameters of such a model can lead to a cognitive fatigue of the DM, who is often asked to express several preferential statements about pairs of alternatives during the elicitation process. To limit the DM's effort, an incremental elicitation process can be used to select informative pairs of alternatives to be presented to the DM sequentially with the aim of refining the SRMP model until a robust recommendation can be made. In this paper, we propose a heuristic for selecting the pairs of alternatives submitted sequentially to the DM for evaluation, using a regret-based elicitation approach. We restrict our proposal to the elicitation of a subset of parameters of the SRMP model, namely the criteria weights, and show that in this context, the proposed heuristic outperforms previously studied query selection criteria