摘要
针对D-S证据理论在处理冲突证据时存在的不足,对不同冲突证据融合算法的性能进行分析。仿真结果表明,Yager算法无法有效处理冲突问题。孙全和Murphy理论主要利用加性策略的可靠性解决冲突证据的融合问题,收敛性差。而邓勇算法采用的加性策略在加性合成同时,引入证据距离函数以得到各证据的可信度作为权值对证据进行加性合成,因此收敛性优于孙全和Murphy算法。关欣算法先对冲突证据进行合理修正,消除证据间冲突,再采用D-S算法进行合成,但该算法证据的冲突由敌方干扰引起,具有一定局限性。
Aiming at limitations of D-S evidence theory in managing conflicting evidences, analyze effects of several conflicting evidence fusion algorithms. Emulational results show that Yager arithmetic is incomplete, while SUN Quan and Murphy theories mainly use add strategy and are bad in astringency; DENG Yong algorithm uses add combination and adopts evidence distance function to get evidence reliability as weighting value and makes add combination for evidence, so its astringency is better than the two above; GUAN Xin algorithm first modifies conflicting evidence, removes conflicts among evidence, and adopts D-S algorithm to synthesize, but enemy disturbance may help conflicts, so has its own restriction.
出处
《兵工自动化》
2008年第2期39-40,48,共3页
Ordnance Industry Automation