基于证据特征的区间证据合成方法研究

Research on combination of interval-valued belief structures based on features of evidence

基于区间值的证据合成理论可以有效地描述和处理不确定性问题,针对目前区间证据合成方法存在融合失效的问题,分析了其产生的原因并提出新的区间证据组合方法。对证据源本身的特征进行提取,依此对证据进行修改,并提出两种优化方法用于区间信度结构下的证据合成。数值实验结果表明,提出的方法收敛性、鲁棒性较好,在面对冲突区间证据融合时得到了合理的结果。

Combination of interval-valued belief structures can describe and deal with uncertainty problems effectively. In order to solve the problems of fusion failure existed in current methods, its reasons are analyzed. The new method extracts the characteristics of evidence source and modifies the evidence according to the char- acteristics. The evidence is modified according to the characteristics extracted from the evidence, and two kinds of optimization models are proposed to combine the interval-valued belief structure. The experiment results show that the proposed method is convergent and robustness. And reasonable result can be obtained in the face of combination of interval-valued belief structure with eonfliction.