Integrated optimization on production scheduling and imperfect preventive maintenance considering multi-degradation and learning-forgetting effects

This paper proposes a multi-objective integrated optimization model of production scheduling and machine maintenance to find the optimal production sequence and preventive maintenance (PM) decisions. This model considers the setup time, the learning-forgetting effects and the multi-degradation effects. The setup time and the learning-forgetting effects are associated with jobs' similarities and PM decisions. The multi-degradation effects including machine deterioration, failure rate and quality characteristic loss determine the stochastic nature of the objectives. A hybrid maintenance strategy combining imperfect PM and minimal repair (MR) is adopted to effectively reduce the failure frequency and improve the processing quality. Then, the multi-objective solution is simplified by normalizing cost, time and utilization. The local search and the elitism strategy are conducted to avoid the solutions falling into local optimum and losing the best chromosome during evolutions. Finally, a case study of automobile engine manufacturing shows that our proposed model can reduce the total maintenance costs by 27%, shorten the total processing time by 3%, as well as improve the machine utilization by 3%.