成败型产品可靠性的Bayes评估

本文讨论了成败型产品可靠性的Bayes评估问题，通过引入继承因子，合理地考虑了产品在设计和改进过程中的各种信息，进而把继承因子看作随机变量，得出了产品可靠性的验后密度函数。仿真结果表明，在小子样前提下，经典统计方法得到的可靠性置信下限过于保守，传统Bayes方法虽考虑了相关或相似产品的试验信息，但却忽略了它们之间的差异，而文中的方法不仅考虑了相关产品的试验信息，同时还考虑了产品之间的差异，评估结论是合理的。

BAYESIAN ASSESSMENT FOR PRODUCT RELIABILITY USING PASS-FAIL DATA

Bayesian assessment for product reliability using pass fail data is discussed Various information during product design and improvement is appropriately considered by introducing factors of inheritance Further, the inheritance factor is thought as a random variable, and posterior density function of product reliability is obtained Results from simulation show that confidence lower limits of product reliability are too conservative according to classical methods using small binomial samples Although traditional Bayesian analysis can consider test information from related or similar products, it ignores differences between such products The methods proposed in this paper consider not only the testing information from related products but also differences between such products, so the conclusion following them is reasonable