Which Prioritization Method Is Better for Deriving Priority from Best-Worst Preferences? A Theoretical and Experimental Analysis

The Best-Worst Method (BWM) is a popular multi-criteria decision-making tool to prioritize alternatives or criteria via a set of subjective pairwise judgments. Deriving the priority weights from best-to-others and others-to-worst preferences is one of the key issues, and several prioritization methods have been proposed to address it. However, their behavior and performances in different situations are yet to investigate. In this study, we analyze the performance of four prioritization methods from theoretical and experimental perspectives. For this purpose, we first show that when the given preference is fully multiplicative consistent, the prioritization methods produce the same weight priority, and it can directly obtain through the analytic formulae without solving the optimization model. For inconsistent preferences, the prioritization methods are compared in terms of deviation from the original preferences and total order violation measures. Simulation experiments suggest that Euclidean distance and order violations metric based measures could lead to different choices of prioritization methods.