Distributed Dual-Resource Flexible Job Shop Scheduling Considering Multiple Speeds and Preventive Maintenance

Symmetry plays a crucial role in balancing production efficiency and energy consumption within distributed manufacturing systems. This study leverages symmetric decision-making structures in resource allocation and maintenance scheduling to achieve an equilibrium between productivity and sustainability. To address the multi-factory collaboration requirements for large-scale orders, a distributed dual-resource flexible job shop scheduling model considering multiple speeds and preventive maintenance on energy consumption is constructed. It aims to minimize the maximum completion time and total machine energy consumption. An artificial bee colony algorithm with adaptive scout bees is proposed to solve the model. An improved decoding method is designed according to the model characteristics to enhance convergence speed. Neighborhood structures based on preventive maintenance and machine speeds are designed, and a dynamic neighborhood search strategy is proposed to improve the local search capability. Three food source generation methods are defined as actions, and Q-learning is employed to dynamically select actions, ensuring population diversity while improving population quality. Extensive experiments are conducted to validate the effectiveness of the improved strategies, and the superiority of the proposed algorithm is verified through performance comparisons with state-of-the-art algorithms.