Optimal maintenance policy with degradation-shock dependence and a reliability constraint using simulation optimisation

This is the authors' abstract. We don't add key points for this paper.

Machine failures can generally be divided into soft failures caused by the normal usage of the equipment itself and hard failures caused by external impacts. Soft failures and hard failures may also affect each other, meaning that external impacts can accelerate the degradation of the equipment's performance and make it more prone to failures as the number of impacts increases. We introduce a new maintenance and reliability model which considers the dependencies between hard and soft failure modes for individual machine types, allowing hard failure thresholds to be dynamically adjusted. This model, based on a manufacturing system of general topology, not only incorporates degradation-shock dependence, but also includes a reliability constraint. The model is set up such that once the degradation level of any machine exceeds a predetermined threshold, maintenance and repair actions, both imperfect and perfect, will be executed. In order to determine the optimal thresholds for maintenance and upkeep strategy, we develop a simulation optimisation algorithm, known as the Jaya-Imperialist Competitive Algorithm (JICA). This method is based on the Imperialist Competitive Algorithm (ICA) and combines the Jaya Algorithm with an adaptive penalty function. Empirical analysis has confirmed that, compared to a set of benchmark algorithms, the algorithm proposed in this study can obtain a superior maintenance and repair strategy while demonstrating greater efficiency. We also perform sensitivity analysis on several important parameters, which can facilitate a deeper understanding of the optimisation model and support optimal decision-making.