Weibull过程将来第k次失效时间的Bayes预测区间

BAYESIAN PREDICTION INTERVALS OF THE κ -TH FUTURE FAILURE TIME FOR A WEIBULL PROCESS

具Weibull强度函数的非齐次Poisson过程经常被用来分析可修系统的失效模式.基于极大似然估计,Engelhardt ＆ Bain(1978)导出了Weibull过程将来第k次失效时间的经典预测区间.在本文中,我们用无信息联合验前分布,根据Weibull过程的前n次失效时间,给出了建立将来第k次失效时间的Bayes预测区间的方法,并说明了如何应用这些方法。

A non-homogeneous Poisson process with Weibull intensity function has been often used to analyze the failure pattern of repairable systems. Classical prediction intervals, based upon maximum likelihood estimation, for the k -th future failure time of a Weibull process were derived by Engel-hardt & Bain (1978). In this paper, methods for constructing Bayesian prediction intervals for the k-th future failure time based on the first n failure times in a Weibull process are presented by using a noninformative joint prior distribution. Some application of these methods are illustrated.