Optimal inspection policy for scheduled maintenance of aging systems

This research provides a flexible inspection policy in which the interval between the current and next inspection is determined on the basis of the current information about the system. A non-stationary partially observed Markov decision process model was used to formulate a decision-making problem for an aging system with imperfect inspection, and the optimal inspection policy which minimizes a long-run expected cost over an infinite time horizon was analytically investigated. Several struc-tural properties of the optimal expected cost function were obtained that establish the existence of an optimal policy in which there is at most one transition from keep-operating to replacement with respect to both the system's deterioration and age under certain conditions. A monotone property of the inspection interval was identified. Furthermore, the obtained conditions were discussed, and a numerical calculation was provided to show how to obtain the optimal inspection policy.