An approach to evaluate the methods of determining experts' objective weights based on evolutionary game theory

摘要

In group decision making problems, it is usually necessary to measure the relative importance of experts objectively based on their provided information, according to specific techniques such as entropies, distance measures, and optimization models. However, it is hard for decision organizers to select a proper method for specific applications. Considering the fact that, in practice, an expert might be invited to do the same or similar evaluations repetitively in a long term, it is natural to anticipate that a weighting method should drive the expert to complete the evaluations more and more actively. Thus, we investigate the repetitive games between the active experts and the passive experts, develop an approach to figure out how a weighting method can encourage the active work and restrain the passive work. Based on which, we present an experimental analysis to evaluate some famous objective weighting methods. Some interesting conclusions regarding the characteristics of methods and suggestions for applications are exploited. The main contribution of this study is to develop a quantitative approach for evaluating the objective weighting methods. (C) 2019 Elsevier B.V. All rights reserved.