Decision rule approach

Abstract

We present the methodology of Multiple-Criteria Decision Aiding (MCDA) basedon preference modelling in terms of “if.., then ..” decision rules. The basicassumption of the decision rule approach is that the decision maker (DM) acceptsto give preferential information in terms of examples of decisions and looks forsimple rules justifying her decisions. An important advantage of this approachis the possibility of handling inconsistencies in the preferential information, resultingfrom hesitations of the DM. The proposed methodology is based on theelementary, natural and rational principle of dominance. It says that if action x is at least as good as action y on each criterion from a considered family, then x is also comprehensively at least as good as y. The set of decision rules constitutingthe preference model is induced from the preferential information using aknowledge discovery technique properly modified, so as to handle the dominanceprinciple. The mathematical basis of the decision rule approach to MCDA is theDominance-based Rough Set Approach (DRSA) developed by the authors. Wepresent some basic applications of this approach, along with didactic exampleswhose aim is to show in an easy way how DRSA can be used in various contextsof MCDA.