Multicriteria choice and ranking using decision rules induced from rough approximation of graded preference relations

Abstract

The approach described in this paper can be applied to support multicriteria choice and ranking of actions when the input preferential information acquired from the decision maker is a graded pairwise comparison (or ranking) of reference actions. It is based on decision-rule preference model induced from a rough approximation of the graded comprehensive preference relation among the reference actions. The set of decision rules applied to a new set of actions provides a fuzzy preference graph, which can be exploited by an extended fuzzy net flow score, to build a final ranking.