Distance-based aggregation in group AHP

Abstract

ABSTRACT The aggregation of evaluators’ preferences is a key problem in group decision making. We examine the recently proposed distance-based techniques and compare their efficiency to the traditional aggregation of individual preferences (AIP) methods in simulated Analytic Hierarchy Process (AHP) cases. We use the Kendall W statistic to measure the rank correlation among the individual priority vectors of the group and the common priority vector for the different aggregation approaches. Extensive simulations (altogether 88000 cases) show that both the Euclidean Distance-Based Aggregation Method (EDBAM) and the Aitchison Distance-Based Aggregation Method significantly outperform the traditional techniques in case of smaller and mid-sized priority vectors (at most six items to be compared). However, EDBAM outperform the AIP methods for all dimensions that is conventionally used in AHP, and its computation time is also low.