考虑双资源约束多转速的绿色柔性作业车间调度研究

Research on green flexible job-shop scheduling considering dual-resource constraints and multiple-speed

中文关键词  双资源约束  多转速  绿色柔性车间调度  多目标优化  人工蜂群算法  Q学习

英文关键词  dual-resource constraints  multi-speed  green flexible job-shop scheduling  multiobjective optimization  artificial bee colony algorithm  Q-learning

中文摘要

      考虑实际生产车间机器不同转速产生能耗差异及精工序的生产需求,构建以最大完工时间和机器总能耗 为优化目标的双资源约束多转速绿色柔性作业车间调度模型,并提出一种动态学习人工蜂群算法进行求解.采用 混合初始化获取初始种群,提升算法的进化起点.在雇佣蜂完成搜索之后,引入新蜂种学习蜂,学习优秀蜜源的基因, 降低搜索的随机性,提高搜索精度,并采用Q学习算子对学习概率进行自适应优化,保证蜜源多样性的同时加强算 法的全局搜索能力.跟随蜂阶段设计一种动态邻域搜索策略,加入基于变速及平衡工人工作时长的邻域结构,提高 跟随蜂的局部搜索能力.通过不同算法对拓展算例的对比验证所提算法的优越性.

英文摘要

      Considering the energy consumption differences caused by different machine speeds in actual production workshops and the production requirements of fine processes, a dual-resource-constrained multi-speed green flexible job shop scheduling model is constructed, with the optimization objectives of maximizing completion time and total machine energy consumption. A learning bee colony algorithm is proposed to solve the model. Using hybrid initialization to obtain the initial population and improve the evolutionary starting point of the algorithm. After the employment bees complete the search, a new bee species is introduced to learn the genes of excellent honey sources, reduce the randomness of the search and improve the search accuracy. Adaptive optimization of learning probabilities is carried out by using the Q-learning operator to ensure the diversity of nectar sources while enhancing the global search capability of the algorithm. A dynamic neighborhood search strategy is designed for the following-bee stage, and a neighborhood structure based on variable speed and balancing the working hours of workers is incorporated to enhance the local search ability of the following bees. The superiority of the proposed algorithm is verified by comparing different algorithms on the extended standard examples.