Using surrogate weights for handling preference strength in multi-criteria decisions

Abstract

Various proposals for how to eliminate some of the obstacles in multi-criteria decision making exist and methods for introducing so called surrogate weights have proliferated for some time in the form of ordinal ranking methods for the criteria weights. Considering the decision quality, one main problem is that the input information to ordinal methods is often too restricted. At the same time, decision-makers often possess more background information, for example regarding the relative strengths of the criteria, and might want to use that. Thus, some form of strength relation often exists that can be utilised when transforming orderings into weights. In this article, using a quite extensive simulation approach, we suggest a thorough testing methodology and analyse the relevance of a set of ordering methods including to what extent these improve the efficacy of rank order weights and provide a reasonable base for decision making.