摘要
导出了指数分布场合基于逐次定数截尾步加寿命试验数据的最大似然估计和最好线性无偏估计.利用最大似然估计的渐近正态性导出了正常应力水平下平均寿命和失效率的近似无偏估计,利用最好线性无偏估计导出了正常应力水平下平均寿命和失效率的无偏估计.利用模拟方法研究了所给方法的精度.
The maximum likelihood estimations and the best linear unbiased estimations of the exponential distribution based on progressively type Ⅱ censored step-stress accelerated life test data are obtained. Using asymptotic normality of the maximum likelihood estimation, the approximately unbiased estimations of the mean and the failure rate at normal stress level are derived. The unbiased estimations of the mean and the failure rate at normal stress level are derived according to the best linear unbiased estimations of parameters in accelerated life equation. The precision of the maximum likelihood estimations, the approximately unbiased estimations and the unbiased estimations of the mean and the failure rate at normal stress level are compared by Monte Carlo method.
出处
《系统工程理论与实践》
EI
CSCD
北大核心
2005年第11期105-111,共7页
Systems Engineering-Theory & Practice
关键词
指数分布
步加试验
逐次截尾
最大似然估计
最好线性无偏估计
the exponential distribution
step-stress accelerated life test
progressively censored
maximum likelihood estimation
best linear unbiased estimation
作者简介
王炳兴（1965-），男，浙江绍兴人，副教授，研究方向为：可靠性统计。E-mail：wangbingxing@163．com。
