Multi-Objective MOBC-RL for Energy-Efficient DAPM Scheduling with Setup Time

Energy-efficient scheduling for distributed assembly parallel machines (DAPM) with sequence-dependent setup times (SDSTs) is critical for sustainable manufacturing. This work proposes a mixed-integer programming (MIP) model to minimize makespan (MCT) and total energy consumption (TEC). A reinforcement learning-enhanced multi-objective artificial bee colony algorithm (MABC-RL) is designed, incorporating: (1) crossover-augmented exploration, (2) evolutionary exploitation strategies, (3) problem-specific neighborhood structures, and (4) Q-learning-guided local search with simulated annealing. Comparative experiments against NSGA-II, MOEA/D, and SPEA2 demonstrate MABCRL’s superiority in convergence and stability, validated by inverted generational distance (IGD) and hypervolume (HV) metrics. The results offer practical insights for sustainable distributed assembly systems.