可靠性评估中利用先验信息的一种新思路

Bayes分析的一般程序是先验分布 数据=后验分布.提出一种新的Bayes分析程序:先基于数据诱导出参数的置信分布,而将参数的原先验分布等效成某种类型数据;然后将置信分布作为参数的"先验分布",而将先验信息的等效数据作为"试验数据";最后采用通常的Bayes方法进行可靠性评估.该新思路主要用于多参数寿命分布的可靠性评估.它克服了多参数模型下确定先验分布的困难及计算上的困难.仿真结果表明该方法具有较高的精度,且计算简单,可以应用于实际.

A new method for reliability estimation by using prior information

The general process for Bayes analysis is:prior distribution data=posterior distribution.A new Bayes analysis routine is put forward,which can get rid of the difficulties in prior distribution determination for multi-parameters model.The main idea of the method is as follows:First,determining the confidence distribution of the parameter based on test data,making the prior distribution of the parameter equivalent to a certain type data;then taking the confidence distribution as the "prior distribution"of the parameter,and the equivalent data of prior information as"test data"finally,using general Bayes analysis for reliability estimation.The new method is mainly used for reliability estimation of multi-parameter life distribution.Simulation result showed that the method has high precision and simple calculation.