Group inspection and maintenance policy design for a heterogeneous multi-component system

Modern engineered systems contain components with heterogeneous degradation behaviors, lifetimes, and inspection and maintenance (IM) requirements, often requiring specialized IM crews. Existing IM strategies widely assume that system inspections cover all components and that all crews are mobilized at each inspection epoch, which may incur excessive labor and mobilization costs in practice. To address this issue, we study the long-term cost-rate optimization problem for a heterogeneous multi-component system subject to both gradual degradation and random external shocks, where the maintenance duration depends on the component's degradation level before maintenance. We first derive the optimal equidistant IM interval for each component when it is treated as a single-component system. Based on this result, we propose a Partial Periodic Inspection and Maintenance (PPIM) policy, under which only a subset of components is inspected and maintained at each scheduled epoch, and show that it can approximate an arbitrary IM policy for a multi-component system. We further identify that searching for a high-quality PPIM policy may involve potentially infinitely many unbounded decision variables, under which standard metaheuristics may converge to suboptimal solutions and exhibit low computational efficiency. Accordingly, we develop a problem-specific initial solution generation technique to improve both solution quality and convergence speed. Both the PPIM policy and the proposed solution procedure are further extended to more general system configurations, including series–parallel and heterogeneous k-out-of-n systems. A case study on a wind turbine bearing system demonstrates that the proposed policy and solution procedure significantly outperform benchmark alternatives.