Abstract

Abstract : In general, it is not easy to determine any maintenance policy from the actual failure data. Thus, nonparametric methods to obtain directly the optimal maintenance policy without specifying the form of failure distribution functions are useful in practice. The purpose of this article is to introduce and explain several nonparametric methods to estimate the optimal replacement time, which minimizes a relevant expected cost function, in the typical age replacement problem. Further, we compare the estimators for the optimal policy on the methods and refer to the relation between the estimators and the number of the data in terms of precision.