Hybrid PSO‐GWO Algorithm for Design Optimization of Multi‐Repair Team Series‐Parallel System Under Failure Dependencies

This paper addresses the optimal design of a repairable series‐parallel system with multiple repair teams and component failure dependencies. Failure dependency among subsystems significantly influences system availability and complicates the optimization process due to its impact on subsystem state transitions. A Markov model is used to evaluate the state distribution of subsystems, while a dependency function captures the impact of failure relationships. The objective is to minimize total system cost, including component and repair team costs, while satisfying an availability constraint. Unlike previous studies that utilized genetic anlgorithms (GA) and particle swarm optimization (PSO), this work proposes the Hybrid PSO‐ Grey Wolf Optimization (GWO) optimization technique, a hybrid metaheuristic approach combining the exploration strengths of evolutionary algorithms with the exploitation capabilities of swarm intelligence. The Hybrid PSO‐GWO technique is applied to solve the resulting nonlinear integer optimization problem effectively. A series‐parallel system with six dependent subsystems is used as a case study. Four failure dependency scenarios—independence, linear dependence, weak dependence, and strong dependence—are considered in the analysis. Comparative analysis demonstrates that the Hybrid PSO‐GWO technique achieves superior performance in terms of solution quality, computational efficiency, and robustness across varying dependency scenarios including independent, weak, linear, and strong dependencies.