Pairwise comparison based interval analysis for group decision aiding with multiple criteria

Abstract

Interval AHP (Analytic Hierarchy Process) was proposed to obtain interval weights from a given pairwise comparison matrix showing relative importance between criteria. In this paper, Interval AHP is applied to group decision problems. Interval AHP is first revised suitably for comparing alternatives from the viewpoint that the interval weight vector shows the set of agreeable weight vectors for the decision maker. Under individual interval weight vectors obtained from individual pairwise comparison matrices, three approaches to obtaining a consensus interval weight vector are proposed. One is the perfect incorporation approach that obtains consensus interval weight vectors including all individual interval weight vectors. By this approach, we can count out indubitably inferior alternatives. The second is the common ground approach that obtains consensus interval weight vectors included in all individual interval weight vectors. By this approach we can find agreeable group preference between alternatives when all individual opinions are similar. The third is the partial incorporation approach that obtains consensus interval weight vectors intersecting all individual interval weight vectors. By this approach we can find agreeable group preference between alternatives when individual opinions are not similar. The usefulness of the proposed three approaches is demonstrated by simple numerical examples.