A data-driven robust optimization method for the assembly job-shop scheduling problem under uncertainty

Abstract

This paper studies the production scheduling problem in an assembly manufacturing system with uncertain processing time and random machine breakdown. The objectives of minimizing makespan and the performance deviation of the actual schedule from the baseline schedule are simultaneously considered. Specifically, a boosting radial basis function network constructed using the data generated by Monte Carlo method, is used as the surrogate model to approximate the performance deviation. After that, a modified master-apprentice evolutionary algorithm (MAE) is developed for robust scheduling. In the design of MAE, we employ an extended adjacency matrix of subassemblies to cope with the sequential constraints of operations in AJSSP. Based on this, effective neighbourhood structures and distance metric of solutions are designed for tabu search and path relinking operators to generate feasible schedules. To evaluate the effectiveness of the proposed method, a series of computational experiments are conducted. The results indicate that, compared with several commonly used algorithms, the suggested method shows good performance in dealing with AJSSP under uncertainty.