Scheduling heterogeneous repair channels in selective maintenance of multi-state systems with maintenance duration uncertainty

Abstract

Selective maintenance with multiple repair channels has received increasing attention in recent years. The problem aims to jointly identify a subset of units to be maintained and assign the selected maintenance tasks to multiple repair channels. Existing works, however, all assumed that the time durations of maintenance tasks and breaks are deterministic. Due to the randomness of the starting time of future missions and the variation of efficiency of repair channels, these two quantities inevitably possess uncertainty. In this article, by taking account of the uncertainties associated with the time durations of maintenance tasks and breaks, a new selective maintenance model with multiple heterogeneous repair channels is formulated. The new selective maintenance problem decides not only a subset of units to be maintained and the corresponding maintenance tasks, but also the number of repair channels with their specific skill levels and the sequences of maintenance tasks in each repair channel. The objective is to maximize the probability of a system successfully completing the next mission subject to a limited maintenance budget. The resulting optimization problem is resolved via a double-loop algorithm embedded with an ant colony optimization. A five-unit system and a multi-state coal transportation system are presented to demonstrate the effectiveness of the proposed method.