As economic globalization, large manufacturing enterprises build production centers in different places to maximize profit. Therefore, scheduling problems among multiple production centers should be considered. This paper studies a distributed hybrid flow shop scheduling problem (DHFSP) with makespan criterion, which combines the characteristic of distributed flow shop scheduling and parallel machine scheduling. In the DHFSP, a set of jobs are assigned into a set of identical factories to process. Each job needs to be through same route with a set of stages, and each stage has several machines in parallel and at least one of stage has more than one machine. For solving the DHFSP, this paper proposes two algorithms: DNEH with smallest-medium rule and multi-neighborhood iterated greedy algorithm. The DNEH with smallest-medium rule constructive heuristic first generates a seed sequence by decomposition and smallest-medium rule, and then uses a greedy iteration to assign jobs to factories. In the iterated greedy algorithm, a multi-search construction is proposed, which applies the greedy insertion to the factory again after inserting a new job. Then, a multi-neighborhood local search is utilized to enhance local search ability. The proposed algorithms are evaluated by a comprehensive comparison, and the experimental results demonstrate that the proposed algorithms are very competitive for solving the DHFSP.

随着经济全球化，大型制造企业在不同的地方建立生产中心以实现利润最大化。因此，应该考虑多个生产中心之间的调度问题。本文研究了基于makepan标准的分布式混合流水车间调度问题（DHFSP），结合了分布式流水车间调度和并行机调度的特点。在DHFSP中，将一组作业分配给一组相同的工厂进行处理。每个作业需要通过具有一组阶段的同一路线，并且每个阶段具有并行的多台机器，并且至少一个阶段具有多台机器。为了解决DHFSP问题，本文提出了两种算法：最小中型规则的DNEH算法和多邻域迭代贪婪算法。具有最小中型规则构造启发式的DNEH首先通过分解和最小中型规则生成种子序列，然后使用贪婪迭代将作业分配给工厂。在迭代贪婪算法中，提出了一种多搜索构造，该构造在插入新工作后将贪婪插入再次应用于工厂。然后，利用多邻域本地搜索来增强本地搜索能力。通过综合比较对所提出的算法进行了评估，实验结果表明所提出的算法在求解DHFSP方面具有很强的竞争力。插入新作业后，将贪婪插入再次应用于工厂。然后，利用多邻域本地搜索来增强本地搜索能力。通过综合比较对所提出的算法进行了评估，实验结果表明所提出的算法在求解DHFSP方面具有很强的竞争力。插入新作业后，将贪婪插入再次应用于工厂。然后，利用多邻域本地搜索来增强本地搜索能力。通过综合比较对所提出的算法进行了评估，实验结果表明所提出的算法在求解DHFSP方面具有很强的竞争力。