一种基于相似性测度的证据合成公式

Synthetic Rule of Evidence Based on Similarity Measure

针对经典D-S规则融合冲突证据时会出现结果有悖于常理的问题,提出一种基于相似性测度的证据合成公式。利用相似性测度构造证据源可信度,进而得到焦元可信度。将焦元可信度作为冲突信度分配系数以改进合成公式,利用改进的合成公式对证据源进行数据融合。算例结果表明,该方法能有效解决冲突性证据的融合问题,较之以往方法具有收敛性强和可靠性高的优点。

Aiming at the conflicting evidence fusion with classic D-S rule will appear the abnormal problem,one synthetic rule of evidence based on similarity measure is proposed.Evidence credibility is constructed by using the similarity measure,then the focal element credibility is obtained.The focal element credibility is used as the distribution coefficient of the conflict reliability to improve the synthetic rule,and evidences are fused by employing the improved synthetic rule.Numerical examples illustrate that the proposed method can solve the problem of conflict evidence fusion effectively.Compared with some other methods,it has good characters in strong convergence and high reliability.