Performance indicator-based multi-objective reliability optimization for multi-type production systems with heterogeneous machines

The demand for product manufacturing is gradually leading the direction of production. Therefore, a production line that can produce diversified products is critical for meeting the requirements of modern systems. This study investigates multi-objective optimization of the reliability of a multi-type production system. In particular, machine deployment at each workstation can be associated with different reliability depending on the pur-chasing year or brand. Consequently, this study applies a component mixing strategy to multi-type production systems to estimate the reliability of an ensemble of heterogeneous machines. The goal of the studied problem is to maximize all production lines, and the cold-standby redundancy strategy based on the continuous-time Markov chain model is employed to calculate the exact reliability of the system. A performance indicator -based multi-objective algorithm, epsilon-MOABC, was employed to search for a near-optimal solution. The experi-mental results show that the multi-objective solution effectively converges to the Pareto front of the knee region and, more importantly, it can achieve high system reliability at a streamlined cost.