Condition-based reallocation and maintenance for a 1-out-of-2 pairs balanced system

In a system consisting of multiple functionally exchangeable units, differences in units’ degradation levels can significantly affect the system's performance. Dynamic reallocation of these units can improve performance and prolong the lifetime of the system. This study is the first to quantify the benefit of incorporating reallocation into a condition-based maintenance framework for a 1-out-of-2 pairs balanced system—a system with two pairs of units that functions if there is at least one functioning pair. The balance condition requires that the two units in the same pair should be active or inactive at the same time. Unit degradation is modeled as a Gamma process and is inspected periodically. A Markov decision process model is developed to determine the optimal integrated reallocation and maintenance policy that minimizes the long-run average cost per unit time. A numerical study illustrates that the proposed integrated reallocation and maintenance policy significantly outperforms other limited policies, including reallocate-only and maintain-only policies.