Abstract: Aiming at the reentrant and batch processing characteristics in the cold drawing production process of seamless steel tubes,the production process was modeled as a reentrant hybrid flow shop scheduling problem with batch processing machines considering machine breakdowns.To solve the problem,a scheduling optimization model was firstly developed,and the optimization objective was to minimize the makespan and the expectation of the difference between the process completion times before and after the breakdown,and then a proactive-reactive multi-objective evolutionary algorithm based on decomposition (PRMOEA/D) was proposed,which adapted the job-based encoding method and the effect of different batch methods on the makespan was investigated.It proactively responded to machine breakdowns by inserting redundancy time,and in the case of a machine breakdown that cannot be resolved by redundancy time,a multi-way tree was used to identify the affected operations and then a reactive method of shifting the operation to the right was proposed to respond.On the basis of MOEA/D algorithm,the main direction evolution was proposed to promote the individual evolution in the direction of each weight vector,and the problem of high individual similarity during iterations was solved by the neighbor solution diversity enhancement strategy.Four sets of experiments were designed to prove the effectiveness of the PRMOEA/D algorithm and the comparison experiments with other algorithms was conducted.The results showed that the PRMOEA/D algorithm could effectively solve the reentrant hybrid flow shop scheduling problem with batch processing machines considering machine breakdowns.

Key words: reentrant hybrid flow shop scheduling, machine breakdowns, batch processing machines, proactive-reactive scheduling, multi-way tree

摘要： 针对无缝钢管冷拔生产过程中的重入和组批加工特性,考虑加工过程的机器故障,将生产过程建模为考虑机器故障的带批处理机的可重入混合流水车间调度问题。为求解该问题,首先建立了调度优化模型,优化目标为最小化最大完工时间和故障前后工序完工时间差值的期望,然后提出了一种基于分解的主动反应式多目标进化算法(PRMOEA/D)。PRMOEA/D算法采用基于工件的编码方式,并在此基础上研究了不同组批方式对完工时间的影响;通过插入冗余时间主动应对机器故障,发生冗余时间无法解决机器故障的情况时,采用多叉树识别受影响工序并通过工序右移的反应式方法进行求解。在MOEA/D算法基础上,提出了主方向进化以促进每个权重向量方向上的个体进化,通过邻域解多样性增强策略来解决迭代过程中个体相似度高的问题。最后,设计了4组实验,证明了PRMOEA/D算法的有效性并与其他算法进行了对比实验。结果表明,PRMOEA/D算法能够有效解决考虑机器故障的带批处理机的可重入混合流水车间调度问题。

关键词: 可重入混合流水车间, 机器故障, 批处理机, 主动反应式调度, 多叉树