模糊先验信息情形下Bayes可靠性评估的新思路  

A New Bayesian Method for Estimation of Reliability with Fuzzy Prior Information

作　　者：周涛[1] 胡昌华[1] 张伟[1] 

机构地区：[1]第二炮兵工程学院,西安710025

出　　处：《弹箭与制导学报》2008年第1期216-218,226,共4页Journal of Projectiles,Rockets,Missiles and Guidance

摘　　要：针对模糊先验信息情形下的Bayes可靠性评估问题,提出了一种综合利用模糊先验信息和试验数据进行可靠性评估的新方法:首先基于试验数据诱导出一个置信分布,并将此置信分布视为先验分布,而将模糊先验信息视为一个模糊观测,最后,通过一个扩展意义上的Bayes公式将上述两类信息进行有效综合,从而得出更加可靠的评估结果。该方法回避了从模糊先验信息构造先验分布的困难,为模糊先验信息在可靠性评估中的应用提供了一条新的途径。A new Bayesian method is presented for evaluation of reliability, which can avoid difficulties of modeling prior distribution with fuzzy prior information. The main idea of the method is as follows： the confidence distribution of reliability induced by test data is considered as prior distribution, and the fuzzy observation data as some of test data, then posterior distribution is computed by a general Bayes analysis program. The result of this new method is more credible.

关 键 词：可靠性评估 BAYES分析 模糊先验信息 等效数据 

分 类 号：TB114.3[理学—概率论与数理统计]