模糊证据理论的深入的研究.pdfVIP

Proceedings of the 26th Chinese Control Conference July 26-31, 2007, Zhangji aj ie, Hunan, China 糊证据理论的深入研究* 马小平，汪永东，樊 阳 国矿业大学信息与电气工程学院，徐州 221008 E-mail: xpma@cumt.edu.cn 要：把证据理论向模糊集推广，利用模糊集的隶属函数提出一种构造证据理论 的基本概率赋值函数的方法， 使得证据理论应用于实际更加方便和有效；针对 D-S 证据组合规则和其他一些改进的组合规则对证据冲突处理 存在的不足，提出改进的证据合成规则。该规则兼顾了证据之间的冲突性和一致性，对冲突性证据以证据源的可 信度加权合成，对一致性证据采用 映聚焦程度的与运算。理论分析和数值实验表明改进后的规则对高冲突性证 据和一致性证据都非常有效，特别对高冲突性证据，该规则比其他一些规则能更快地收敛到合理的目标。 关键词：模糊集，证据理论，模糊证据理论，合成规则 A Further Study on Fuzzy Evidence Theory Ma Xiaoping, Wang Yongdong, Fan Yang School of Information Electrical Engineering, China University of Mining and Technology, Xuzhou, 22 1008, P. R. China E-mail: xpma@cumt.edu.cn Abstract: This paper generalizes the evidence theory and makes it can be used in fuzzy sets in order to make it can be used more effectively. With the fuzzy membership function, this paper presents a method to set up the basic probability assign- ment function (mass function) in evidence theory, and integrates fuzzy theory and evidence theory effectively. This method brings the evidence theory nearer to practice. This paper presents an improved combination rule to deal with consistent or inconsistent evidences obtained from multiple sources. The improved rule reflects the intersection of focus elements and adapts AND-operation to combine consistent evidences, and allocates the conflicts to vary focus elements according to relevant evidences creditability. From the theoretic analysis and the results of the numerical examples, the given rule is very rational and effective for both high conflictive evidences and consistent evidences. Especially, the given rule can pro- vide reasonable results with better