A condition-based maintenance model

Condition-based maintenance (CBM) is very appealing because it enables one to make maintenance decisions based on the current information about the system. Various monitoring techniques have been developed to obtain this information, but there are still very few mathematical models capable of utilizing it for effective maintenance decision making. In this paper, we propose a CBM model for situations where only partial information is available through monitoring a signal process that does not necessarily exhibit monotone behaviour. The evolution of the signal process is determined by random factors and minor maintenance actions between inspections. The objective is to find the replacement policy that maximizes the total expected profit during the lifetime of the system. We will show that, under weak monotonicity assumptions, the optimal policy is of a control-limit type, and we will develop an algorithm for finding the limit for an ϵ-optimal policy.