Optimizing Energy-Efficient Flexible Job Shop Scheduling with Transportation Constraints: A Q-Learning Enhanced Quality-Diversity Algorithm

The Flexible Job Shop Scheduling Problem (FJSP), particularly one with transportation constraints, is prevalent in the intelligent manufacturing field. Leveraging the intricacies of these transportation constraints is recognized for its potential to enhance problem-solving efficacy. Despite this, there has been a dearth of research focusing on this approach. This paper posits that integrating transportation conditions into local search operators can significantly bolster the ability to solve such problems. To make well-informed decisions among local search operators, we have implemented a reinforcement learning technique known as Q-learning. Furthermore, we design a Quality-Diversity (QD) algorithm aimed at preserving solution diversity within a tailored feature space. This space is designed in accordance with the unique attributes of transportation constraints. The empirical results from testing on 20 instances indicate that our proposed algorithm shows great promise, achieving an average 6% reduction in the optimization objective when compared to existing state-of-the-art algorithms.