D-S证据理论中一种新的冲突证据融合方法

摘要

针对D-S证据理论在融合高度冲突的证据时可能导致与直观结果相悖的问题,通过分析现有的几种冲突证据融合方法,提出了一种有效处理冲突证据的融合方法.该方法综合了模型修正和规则修正这两类方法的优点.首先基于冲突系数矩阵计算证据的可信度,并以此作为折扣因子来修正证据结构,然后基于局部冲突、局部分配原则确定证据合成时的冲突分配空间和分配权重.实验结果表明,该方法对合理的合成方向具有较快的收敛速度,同时降低了决策风险.%A new method was developed for dealing with seriously conflicting evidence when the Dempster combination result can not identify the actual conditions. The method combines the advantages of justification of evidence information and combination rules. The reliability of evidence, computed by a conflict matrix, is taken as discount factor to correct the belief functions. And then the problem of space and weight for conflict redistribution is addressed based on the idea of distributing local conflict to local focal elements. Compared with the existing methods, the new method can applied to both eases of conflicting and coincidence. And the numerical simulation shows that the method improves the reliability and rationality of combination results,and also accelerates convergence and reduces the decision risk.