Large-Scale Group Decision-Making Using Analytical Hierarchy Process Based on Bayesian Best Worst Method

Abstract

The most commonly used pairwise comparison method in the literature is the Analytic Hierarchy Process (AHP), with the Best Worst Method (BWM) proposed as an alternative. Both methods can be effectively used in studies with a small number of decision-makers. However, in cases where the number of participants increases, the consistency and applicability of the surveys can sharply decrease. In some instances, it has been observed that the consistency in AHP is considered acceptable up to 10%. In BWM, the Bayesian Best Worst Method (BBWM) is used to address outliers. Therefore, this study aims to address Pairwise Comparison Matrices (PCMs) in multi-criteria decision-making (MCDM) challenges in large group settings by employing an innovative adaptation of BBWM integrated with the AHP. The proposed method starts with the BWM survey, which offers practical comparison opportunities, and then transitions to the AHP matrix. The weights obtained from PCMs show the importance levels of the best and worst criteria in comparison to the other criteria. Normally, these two weights should be consistent, but when they differ, it indicates the indecisiveness of the decision-maker. In both approaches, this indecisiveness is reflected as inconsistency in the matrix, and such evaluations may be disregarded. To prevent this loss of information, in our study, the weights derived from these two PCMs are treated as fuzzy ranges, and fuzzy operations are applied to consider the decision-maker's indecisiveness in producing the final results. The proposed method was applied in the evaluation of mobile shopping applications. The data collected through surveys with 113 users were analyzed using the proposed method, and the rankings of the mobile applications were determined. The obtained rankings were compared with BBWM results to validate the model.