Abstract: In the process of flexible job shop scheduling under rolling production mode,the starting time of each process-level production sub-batch is comprehensively affected by practical constraints,including the production sub-batch completion time of the immediately preceding process containing the workpieces,the transmission sub-batch turnaround time of the immediately preceding process,the completion time of the previous production sub-batch in the same process and the occupation of equipment resources by existing tasks.To address this issue,a flexible job shop scheduling model that considered rolling production,external dynamic material arrivals and constraints imposed by existing tasks occupying equipment resources was proposed.Furthermore,an improved genetic algorithm was proposed and the corresponding crossover and mutation strategies were designed to avoid the generation of illegal solutions in the iteration process.During the equipment selection stage,three methods for selecting equipment were provided,and their selection probabilities were dynamically adjusted based on the optimal solution's stagnation generation in the iteration process,which  achieved a balance between inheriting excellent individual information and breaking through local optima during the iteration process,and improved the algorithm's optimization ability.Finally,based on standard examples and extended examples with additional constraints,we test the proposed model and algorithm were tested,and the results indicated that the proposed algorithm could optimize scheduling outcomes to obtain scheduling plans that satisfied the constraints of dynamic material arrivals and resource limitations existing in the actual production process of the workshop.

Key words: flexible scheduling, production sub-batch, rolling production, batch scheduling, equipment selection

摘要： 针对滚动生产模式下的柔性作业车间调度过程中,工序级生产子批的开工时间受到其所含盖工件所属紧前工序生产子批完成时间、所属紧前工序传输子批周转时间、同工序上一生产子批完成时间及设备资源已有任务占用等实际约束综合影响的问题,构建考虑滚动生产、外部动态来料及设备资源存在已有任务占用约束的柔性作业车间调度模型,提出一种改进遗传算法并设计了相适应的交叉、变异策略,避免迭代过程中非法解的产生。在设备选择阶段提供3种设备选取方式,基于迭代过程最优解停滞代数动态调整各自选取概率,实现了迭代过程中优异个体信息继承与局部最优突破能力的平衡,提升了算法寻优能力。最后,基于标准算例及在标准算例基础上增加相关约束后的扩展算例,对所提模型和算法进行测试,结果表明所提算法能够优化调度结果,获得满足车间实际生产过程中存在的动态来料、资源约束等约束关系的调度方案。

关键词: 柔性调度, 生产子批, 滚动生产, 批量调度, 设备选取