Abstract: To deal with the uncertain remanufacturing scheduling optimization problem with re-entrant operation and flexible machines,a new remanufacturing re-entrant flexible job-shop scheduling model was proposed with bi-fuzzy theory to describe the uncertainty of processing time and cost.To solve this model,an improved NSGA-II algorithm was proposed,which integrated reinforcement learning to dynamically adjust the crossover rate for maintaining population diversity.Besides,in the algorithm,a new two-dimensional coding scheme was proposed to improve the algorithmic efficiency,a hybrid population initialization strategy was embedded to improve the initial population quality,and a local search strategy was adopted to enhance the local search capability.Finally,simulation experiments demonstrated the performance and effectiveness of the proposed algorithm.

Key words: 再制造, 双模糊理论, 可重入柔性车间调度, NSGA-Ⅱ算法, 强化学习

摘要： 针对再制造过程中加工时间和成本的不确定性问题,引入双模糊理论进行二维模糊化处理,同时考虑了再制造操作可重入和机器柔性的特点,构建了再制造可重入柔性车间调度优化模型。在此基础上,提出了一种基于强化学习的改进型NSGA-Ⅱ算法(RLNSGA-Ⅱ),集成了强化学习技术对种群交叉率进行动态优化以提升种群多样性。此外,还提出了一种新的二维编码方案以提高算法效率,嵌入了混合种群初始化策略以提高初始种群质量,采用了局部搜索策略以增强局部搜索能力。最后,通过仿真实验验证了该算法的有效性和优越性。

关键词: TH165, TP18