相似产品步进应力加速寿命试验的统计分析

本文将单组样本的步进应力加速寿命试验(SSALT)拓展到双样本情形.基于累积失效模型(CEM),对两组指数产品的联合Ⅱ型截尾数据进行综合分析.首先,计算未知参数的极大似然估计(MLE),并利用矩母函数(MGF)建立其条件概率分布.然后,采用两种方法,基于概率分布的精确分析方法和参数的Bootstrap方法,构建参数的置信区间(CIs).以覆盖率作为评价标准,对两种方法的区间估计效果进行分析.最后,通过数值算例验证了本文方法的有效性.

Statistical inference for step stress model with relative items

This paper makes a generalization to the situation when two competing populations with relative merits are simultaneously exposed to the same step stress accelerated life test (SSALT). Under the assumption of cumulative exposure model (CEM), we analyze a joint type-Ⅱ censoring data when the lifetime of two populations follow independent exponential distribution. Firstly, we obtain the maximum likelihood estimate (MLE) for unknown parameters, and establish the conditional distribution for MLE through the use of moment generating function (MGF). Then, two approaches, one based on exact distribution, and the other based on percentile Bootstrap method, are developed to construct the confidence intervals (CIs) for parameters. A comparison of the two methods is also made by Monte Carlo simulation in terms of coverage probabilities. And finally, we provide a numerical example to illustrate our methods.