双参数指数分布在步降应力模型下的Bayes估计  

The Bayes Estimation of Two-parameter Exponential Distribution under the Step-down-stress

作　　者：刘平清 Liu Pingqing(School of Mathematics and Statistics,Liupanshui Normal University,Liupanshui 553001,China)

机构地区：[1]六盘水师范学院数学与统计学院,贵州六盘水553001

出　　处：《六盘水师范学院学报》2022年第6期105-112,共8页Journal of Liupanshui Normal University

基　　金：贵州省教育厅自然科学研究项目“高维情形下个性化特征选择研究”(黔教合KY字〔2020〕111);贵州省教育厅高等学校教学内容和课程体系改革项目项目“大数据驱动背景下地方高校经济统计学专业人才培养模式的改革研究”(2021287);六盘水师范学院科学研究项目“贵州省水资源综合承载力动态因子影响效应及耦合协调性研究”(LPS⁃SYSK202007)。

摘　　要：加速寿命试验是评估可靠性的重要手段。在产品寿命服从双参数指数分布时,试验施加步降应力水平,并且在观测到试验产品失效时有未失效产品被随机移走,试验采用定数截尾。通过对模型的分析,在Bayes估计中选择MH算法对参数做估计,并利用案例数据分析简单步降应力下不同α_(1)(接受建议分布生成值的概率)和m_(1)(应力水平S_(1)下失效样本数)对各参数估计精度的影响。发现m_(1)对参数的估计精度影响不大,但α_(1)对参数的估计精度影响较大。Accelerated life test is an important test method to evaluate reliability.It is assumed that the product life is ruled by Two-parameter exponential distribution,there are some products are randomly removed when anyone products fail,and the Type II censoring is adopted under step-down-stress level.Through the analysis of the model,MH algorithm is chose for parameter estimation,and analyze the influence of different α_(1)(the frequency of accept value which generated by proposal distribution)and m_(1)(the number of failure sample under S_(1))on the estimation accuracy under simple step-down stress model.It is found that m1 has a little influence on the estimation accuracy of parameters,but α_(1) has a great influence on the estimation accuracy of parameters.

关 键 词：双参数指数分布 定数截尾 MH算法 步降应力 优化设计 

分 类 号：O213.2[理学—概率论与数理统计]