Inferring criteria's relative importance coefficients in PROMETHEE II

Abstract

Most of decision aid models involve parameters that allow modelling the decision-maker's preferences. These parameters' values must be determined with the decision-maker's collaboration. Notably, this is the case for the criteria's relative importance coefficients (RICs). The determination of criteria's RICs represents an important task for the decision-maker, which is generally difficult to accomplish. The information provided at this level is unavoidably subjective and often partial, and then it should be well exploited. This is precisely the target of this paper, for the criteria's RICs of PROMETHEE method. We will use mathematical programming in order to develop a preference disaggregation approach. The programme built will contain multiple solutions, but will allow taking out intervals for the RICs' values. Hurwicz's criterion will permit us to explicit a value for each RIC.