逐级删失定时截尾恒加寿命试验模型的抽样方案设计  

Design of sampling plans of accelerated life tests under progressive censoring

作　　者：杨春燕[1] 胡品 YANG Chun-yan;HU Pin(School of Mathematics and Statistics,Yunnan University,Kunming 650500,Yunnan,China)

机构地区：[1]云南大学数学与统计学院,云南昆明650500

出　　处：《云南大学学报（自然科学版）》2022年第3期444-452,共9页Journal of Yunnan University(Natural Sciences Edition)

基　　金：国家自然科学基金(11961078,11771384);云南省应用基础研究重点项目(202001BB050049).

摘　　要：研究了对数正态分布下伴有随机移走的逐级删失定时截尾恒加寿命试验模型的抽样方案设计.假定产品的寿命分布遵从对数正态分布,被随机移走的试验产品数服从二项分布,其中每一个试验产品被随机移走的概率为p.考虑了高于正常状态的2个应力水平,并假定对数寿命分布的参数与应力水平是线性关系,推导了模型的Fisher信息矩阵,并通过极小化广义渐近方差得到低应力水平的值和试验样品的分配比例,最后使用OC曲线来确定接收系数和样本量.利用数值模拟的方法,得到了最优的低应力水平值、试验样品的分配比例以及相应的抽样方案.We consider the design of acceptance sampling plans under TypeⅠprogressive censoring based on accelerated life-test for lognormal distribution with random removals.In particular,the random removal pattern follows a binomial distribution with removal probability p.Two levels of stress higher than the use condition are used.Assume that the location parameter of the log lifetime distribution is a linear function of stress.The Fisher information matrix is given in the text.The low stress level and the sample proportion are obtained by minimizing the generalized asymptotic variance of the MLEs of the model parameters.The sample size and the lot acceptability constant satisfy the requirements of producer’s risk and consumer’s risk are obtained.We get the optimal low stress level,the sample proportion allocated to each stress,as well as the sampling plans.

关 键 词：对数正态分布 定时截尾逐级删失 加速寿命试验 抽样方案 随机移走 

分 类 号：O213.2[理学—概率论与数理统计]