Abstract In the context of an automated assembly line and flexible manufacturing, good layout planning of the production facilities aids in improving production efficiency and reducing production costs. The current research on the facility layout problems mainly focuses on the static ones with the fixed flow between facilities, including the parallel row ordering problem (PROP). Based on existing research on the PROP, in this study, we first investigate the dynamic parallel row ordering problem (DPROP), considering the flexibility of the production line. This study proposes a DPROP model and utilises it to verify some PROP examples less than 15 facilities and some global optimal solutions are achieved. In addition, the performance of the model in solving DPROP is verified on some new small-scale examples designed based on the PROP. Owing to the limited performance of the exact methods for DPROP, this study proposes a hybrid algorithm of harmony search and tabu search to solve large-scale problems, and this new algorithm performs better than some existing ones so that some new better solutions for the PROP examples are obtained. The algorithm is first utilised to solve the small-scale PROP and DPROP and the results are compared with an exact methods. Moreover, by evaluating the large-scale PROP between 30 and 70 facilities, the results are compared with those of the improved genetic algorithm and the population-based simulated annealing algorithm, thereby 28.8 % of the current optimal solutions are improved. Finally, the hybrid algorithm of harmony search and tabu search is utilised to solve the large-scale DPROP with 30 and 40 facilities, and some optimal solution sequences are presented.

中文翻译：

---

动态并行行排序问题的和声搜索混合算法

摘要 在自动化流水线和柔性制造的背景下，良好的生产设施布局规划有助于提高生产效率和降低生产成本。目前对设施布局问题的研究主要集中在设施间流动固定的静态问题上，包括平行排序问题（PROP）。基于现有对 PROP 的研究，在本研究中，我们首先研究了动态并行行排序问题 (DPROP)，同时考虑了生产线的灵活性。本研究提出了一个 DPROP 模型，并利用它来验证一些少于 15 个设施的 PROP 示例，并实现了一些全局最优解。此外，该模型在求解 DPROP 方面的性能在一些基于 PROP 设计的新小规模实例上得到了验证。由于DPROP精确方法的性能有限，本研究提出了一种和声搜索和禁忌搜索的混合算法来解决大规模问题，这种新算法的性能优于现有的一些算法，从而为PROP提供了一些新的更好的解决方案得到的例子。该算法首先用于求解小规模PROP和DPROP，并将结果与​​精确方法进行比较。此外，通过对30~70个设施之间的大规模PROP进行评估，将结果与改进遗传算法和基于种群的模拟退火算法的结果进行比较，从而改进了当前最优解的28.8%。最后，利用和声搜索和禁忌搜索的混合算法解决了具有30和40个设施的大规模DPROP，