BW-MaxEnt: A Novel MCDM Method for Limited Knowledge

Abstract

With the development of the social economy and an enlarged volume of information, the application of multiple-criteria decision making (MCDM) has become increasingly wide and deep. As a brilliant MCDM technique, the best–worst method (BWM) has attracted many scholars’ attention because it can determine the weights of criteria with less comparison time and higher consistency between judgments than analytic hierarchy process. However, the effectiveness of the BWM is based on complete comparison information among criteria. Considering the fact that the decision makers may have limited time and energy to study all criteria, they cannot construct a complete comparison system. In this paper, we propose a novel MCDM method named BW-MaxEnt that combines BWM and the maximum entropy method (MaxEnt) to identify the weights of unfamiliar criteria with incomplete decision information. The model can be translated into a convex optimization problem that can be solved effectively and has an overall optimal solution. Finally, a practical application concerning the procurement of GPU workstations illustrates the feasibility of the proposed BW-MaxEnt method.