Rule-Based Decision Support in Multicriteria Choice and Ranking

摘要

Solving multicriteria decision problems, like choice and ranking, requires the use of DM's preference model. In this paper we advocate for the preference model in terms of "if..., then..." decision rules induced from decision examples provided by the DM. This model has two advantages over the classical models? (i) it is intelligible and speaks the language of the DM, (ii) the preference information comes from observation of DM's decisions. For a finite set A of actions evaluated by a family of criteria, we consider the preference information given in the form of pairwise comparisons of reference actions presented in a pairwise comparison table (PCT). In PCT, pairs of actions from a subset B is contain in A×A are described by preference relations on particular criteria and by a comprehensive outranking relation. Using the dominance-based rough set approach to the analysis of the PCT, we obtain a rough approximation of the outranking relation by a dominance relation. Then, a set of "if..., then..." decision rules is induced from these approximations. The decision rules constitute the preference model which is then applied to a set M is contained in A of (new) actions. As a result, we obtain a four-valued outranking relation on set M. In order to obtain a final recommendation in the problem of choice or ranking, the four-valued outranking relation on set M is exploited using a net flow score procedure.