双参数指数分布下可靠性增长试验的统计分析

A Better Analysis of Reliability Growth Tests

在充分利用可靠性增长试验中各阶段试验数据评定现阶段可靠性指标的基础上,对双参数指数模型试验下所得的抽样数据,运用陈家鼎提出的方法,给出了现阶段可靠度的点估计和在某种意义下最优的置信下限。文中的方法在计算上比较简便,适用于工程实际。

In our opinion, the two-parameter exponential distribution is a better mathematical distribution than the now generally used one-parameter exponential distribution for analyzing reliability growth tests; it is closer to describing accurately the life distribution of product and is therefore conducive to more accurately predicting reliability and leads to high reliability growth. In the full paper, we explain in much detail how to make better analysis using two-parameter exponential distribution; we emphasize that we prove Theorems 1 and 2, which form the core of our analysis. In place of full but cumbersome explanation ,here we give only a briefing. For a product which has undergone m stages of reliability growth tests,we take sample test data from these m stages and compute the point estimate and lower confidence bound of the m th stage. A simulation example shows preliminarily that our analysis, which is based on two-parameter exponential distribution, is simple,better and convenient for analyzing reliability growth tests.