An Approach to Robust Multiattribute Concept Selection

In this paper, we investigate and extend a method of selecting among a set of concepts or alternatives using multiple, potentially conflicting criteria. This method, called the Hypothetical Equivalents and Inequivalents Method (HEIM), has been shown to avoid the many pitfalls of already existing methods for such problems, such as pair-wise comparison, ranking methods, rating methods, and weighted sum approaches. The existence of multiple optimal sets of attribute weights based on a set of stated preferences is investigated. Using simple visualization techniques, we show that there is a range of weights that satisfy the constraints of HEIM. Depending on the attribute weights used, multiple possible alternative winners could exist. The visualization techniques, coupled with an indifference point analysis, are then used to understand the robustness of the solution obtained and determine the appropriate additional constraints necessary to identify a single robust optimal alternative.