基于置信熵差的冲突证据融合方法

Conflict evidence combination method based on belief entropy difference

在证据理论中,加权平均是解决冲突证据融合的有效方法。然而,如何合理地确定证据权重是一个值得研究的问题,现有的一些方法利用置信熵自身来确定证据权重,在一定程度上解决了冲突问题,但当冲突证据置信熵值大于或等于正常证据时,系统将赋予冲突证据同等甚至更大的权重值,导致冲突得不到有效处理。本文提出基于置信熵差的冲突证据融合方法,利用证据内对应焦元的邓熵值差度量证据之间的差异,从而合理地确定证据权重值。算例结果表明,该方法能够克服现有方法的不足,赋予冲突证据更小的权重值,且具有更高的正确目标可信度。

In evidence theory,the weighted average is an effective method to resolve conflict evidence combination.However,how to reasonably determine the weight of evidence is a problem worth studying.Some existing methods propose to use confidence entropy itself to determine the weight of evidence,which solves the conflict problem to a certain extent.When the confidence entropy value of conflict evidence is greater than or equal to normal evidence,the system will assign the same or even greater weight value to the conflict evidence and resulting in conflicts that are not handled effectively.To solve this problem,this article put forward the conflict evidence combination method based on belief entropy difference.In this method,the difference between evidence is measured by the difference in Deng entropy value of corresponding focal element,so as to reasonably determine the weight value of each evidence.Numerical example results show that the proposed method can overcome the shortcomings of existing methods,and give conflict evidence even smaller weight value,and has a higher reliability target correctly.