A novel discrete evidence fusion approach by considering the consistency of belief structures

Fusion of discrete belief structures gets much attention owing to the widespread existence of discrete information in decision analysis, expert system and other fields. Recently, one method that combines discrete evidence was proposed by establishing an optimization model. However, the existing optimization model of discrete belief structure has a problem of high computational complexity since it needs to calculate all the cases of deterministic evidence combination to obtain maximum and minimum values of each focal element in combination result. In this paper, a novel method of discrete evidence fusion is proposed to reduce the computational complexity by considering the consistency of belief structures of evidence groups and finding the most consistent and most inconsistent cases. Example and application are given to illustrate the effectiveness and rationality of the proposed method.