Joint Optimal Policy for Maintenance, Spare Unit Selection and Inventory Control Under a Partially Observable Markov Decision Process

This research investigates the joint optimization of maintenance and spare unit management for series systems composed of multiple heterogeneous units. With advancements in communication and sensing technologies, condition-based maintenance has gained attention as an integral aspect of spare unit management. Furthermore, the inherent interaction between maintenance activities and spare unit management underscores the necessity of their simultaneous optimization to enhance overall system performance. Based on uncertain information about the system’s deterioration state and spare unit inventory, decision-makers determine actions related to spare units and maintenance, such as replacements, selection of spare units for corresponding units, order quantities and inventory levels. Within the framework of a partially observable Markov decision process, this research proposes an optimal joint policy for maintenance and spare unit management with the objective of minimizing total expected costs. The proposed policy is demonstrated through a three-state, two-unit series system. Sensitivity analyses and comparisons with benchmark policies are also conducted to evaluate the performance of the proposed policy and to investigate the impact of various cost parameters on the proposed decisions.