Point estimation of future failure times of a repairable system

In the reliability analysis of a repairable system a problem often encountered concerns making a forecast about the mth future failure time, given that the first n failure times have been observed. A very interesting mathematical idealization for the underlying failure pattern is the Non-homogeneous Poisson Process with a power intensity law. In this paper point estimators, based upon maximum likelihood principle, are derived for the future failure times of such a process. Their performances are compared via Monte Carlo simulation and tables are given to obtain the median point estimate of the mth future failure time in an easy way.