Thèse de doctorat
Résumé : PROMETHEE belongs to the family of outranking multi-criteria methods. As most outranking methods, they may suffer from the so-called rank reversal phenomenon. The legitimacy of methods suffering from this phenomenon has been largely debated in the literature. The aim of this work is twofold. First, it provides a further investigation of this phenomenon in the specific framework of the PROMETHEE II method which is aimed at building complete rankings. This is performed by both further analysing in which situations rank reversal can happen and by performing an axiomatic characterization of PROMETHEE II ’s net flow scores. This axiomatic characterization is focused on the highlighting of the importance of third alternatives in the building of the final ranking and therefore presents rank reversals as an inherent characteristic of the method. Then, the PROMETHEE method aimed at producing partial rankings, PROMETHEE I , is studied. It is shown that the incomparability relationdefined with this method is to our point of view not compatible with the general signification of incomparability. Further shortcomings of PROMETHEE I will also be highlighted. To tackle these shortcomings, a new version of PROMETHEE , denoted as PROMETHEE γ, is proposed.