AN IMPROVED MULTIPLE CRITERIA VISUAL INTERACTIVE METHOD WITH STEPWISE FOCUSED CONTOURING OF EFFICIENT CRITERION VECTORS

Abstract

In spite of their advantages over other methods in MCDM, existing VIMs have some limitations: that is, the convergency of solutions depends much on how good the DM is in specifying search directions, and thus there have been few studies which successfully showed their effectiveness by simulated tests. This paper introduces an improved multiple criteria visual interactive method ASEOV-VIM with a procedure ASEOV. The procedure ASEOV (approximation of the set of efficient objective vectors) presents the stepwise focused representative subset of efficient criterion vectors (N) and proposes candidate goals for the new search direction of ASEOV-VIM. To contract the criterion space where more detailed contouring or candidate goal searching is recommended, the polyhedral cone constructed from the DM's prior-expressed preference is utilized. With these decision aids, ASEOV-VIM reduces the DM's burden in specifying every new search direction. An illustrative example and simulated results, which show fast convergency to the optimum for various problems, are presented.