考虑运输和机器预维护的柔性作业车间调度研究

PDF下载HTML

Title:

Flexible job shop scheduling considering transportation and machine pre-maintenance

作者:

王玉芳^1,2,3, 张毅¹, 姚彬彬¹, 陈凡¹, 葛师语¹

1. 南京信息工程大学 自动化学院, 江苏 南京 210044;
2. 南京信息工程大学 大气环境与装备技术协同创新中心, 江苏 南京 210044;
3. 南京信息工程大学 气象能源利用与控制工程技术研究中心, 江苏 南京 210044

Author(s):

WANG Yufang^1,2,3, ZHANG Yi¹, YAO Binbin¹, CHEN Fan¹, GE Shiyu¹

1. School of Automation, Nanjing University of Information Science & Technology, Nanjing 210044, China;
2. Jiangsu Collaborative Innovation Center of Atmospheric Environment and Equipment Technology (CICAEET), Nanjing University of Information Science & Technology, Nanjing 210044, China;
3. Engineering Research Center on Meteorological Energy Using and Control (C-MEIC), Nanjing University of Information Science & Technology, Nanjing 210044, China

关键词:

运输约束; 预维护; 航空制造; NSGA-II; 种群质量; 分组进化; 启发式初始化; 局部搜索

Keywords:

transportation constraint; pre-maintenance; aerospace manufacturing; NSGA-II; population quality; group evolution; heuristic initialization; local search

分类号:

TP18；TH165

DOI:

10.11992/tis.202405020

摘要:

航空制造业的快速发展，对高效率和低能耗生产模式的需求愈发迫切。通过综合分析考虑运输和预维护的航空复合材料柔性作业车间调度问题，建立了以最小化完工时间、瓶颈机器负载和总能耗为目标的模型，提出了一种基于种群质量的非支配排序遗传算法(nondominated sorting genetic algorithm II，NSGA-II)。采用启发式初始化方法产生高质量的初始种群，对个体进行分组进化：对优质种群进行局部搜索，深度挖掘种群的最优个体；中等种群通过交叉变异和机器负载操作改变自身部分基因来挖掘最优解；劣质种群则通过学习机制获取优质个体的优秀基因，提升个体优良率。通过测试算例与对比算法的比较验证了所提算法的有效性。最后，将算法应用于实际的航空制造系统，实现了实际生产活动的调度，验证了算法的可行性。

Abstract:

The rapid development of the aviation manufacturing industry has increased the demand for high-efficiency and low-energy modes of consumption and production. Here, a modeling and analysis approach was employed to address the green scheduling issues of an aerospace flexible job shop regarding transportation and pre-maintenance. Additionally, a model was established to minimize the completion time, bottleneck machine workload, and total energy consumption. Further, non-dominated sorting genetic algorithm II based on population quality was proposed to resolve the issues. Furthermore, heuristic initialization was employed to generate high-quality initial populations, and individuals were grouped to evolve. Thereafter, local search operations were conducted to comprehensively explore the optimal individuals. For the populations classified as moderate, crossover and mutation combined with machine load operations were applied to alter portions of their genetic makeup and achieve optimal solutions. Employing a learning mechanism, the inferior populations acquired superior genes from elite individuals to enhance their overall quality. Subsequently, the effectiveness of the proposed algorithm was verified by comparing it with those of other algorithms on test examples. Finally, the algorithm was applied to a real aerospace composite material manufacturing system to schedule actual production activities, thereby verifying its feasibility.