Study on the method and application of revising the weight coefficient of three times throughout the process of evaluation

Abstract

A crucial step in the analysis of multi-attribute group decision-making is to determine the weights of experts and attributes, which will pose great impacts on the result of multi-attribute group decision-making. This paper proposes a dynamic and full process weight evaluation method with the attribute weight and the expert weight being taken into account. We employ three times of weight revising with the improvement of the information conditions in the evaluation process: first is the `ex ante evaluation', which defines a general weight coefficient based on the importance of multiple attribute; then, `ex post evaluation' is performed to revise weight coefficient by the entropy method based on the difference of attribute value; finally, clustering analysis is used to revise weight coefficient according to the discrete degree of score assigned by the experts. The paper describes the process of the calculation in detail and illustrates its feasibility with a practical example.