一种基于D-S证据理论的Bayes可靠性评定方法  

A Bayesian Reliability Evaluation Method Based on Dempster-Shafer Evidence Theory

作　　者：赵仿泽[1] 

机构地区：[1]西北工业大学航海学院,陕西西安710072

出　　处：《鱼雷技术》2013年第3期175-178,共4页Torpedo Technology

摘　　要：如何合理利用多源先验信息,是小子样Bayes理论方法应用中亟待解决的一个问题。目前,多源信息融合还没有统一的准则,融合理论方法的不同会使得融合结果有所差异。为了解决上述问题,基于证据推理理论,提出了一种多源信息先验分布融合方法,将通过各种融合准则获得的先验分布进行再融合,得到了一种合理的综合先验分布,最后建立了基于证据理论的Bayes可靠性评定模型。仿真结果表明,该融合方法是合理可行的,相比使用单一信息融合方法更科学且合理。Bayesian method is an important and suitable approach for system reliability evaluation in small sample cir- cumstance. In order to reduce the testing data, all kinds of a priori information must be used, so the information fusion of multiple sources becomes a key problem in Bayesian analysis. A new method based on the Dempster-Shafer evidence theory is proposed to realize the information fusion of multiple sources, which can integrate more a priori distributions obtained through different fusion criteria. Simulation shows that the proposed method is effective.

关 键 词：Bayes理论 多源信息 D-S证据推理 信息融合 可靠性评定 

分 类 号：TJ630.1[兵器科学与技术—武器系统与运用工程] TB114.3[理学—概率论与数理统计]