当前位置: X-MOL 学术Flex. Serv. Manuf. J. › 论文详情
Our official English website, www.x-mol.net, welcomes your feedback! (Note: you will need to create a separate account there.)
Integrating preventive maintenance to two-stage assembly flow shop scheduling: MILP model, constructive heuristics and meta-heuristics
Flexible Services and Manufacturing Journal ( IF 2.7 ) Pub Date : 2021-03-03 , DOI: 10.1007/s10696-021-09403-0
Zikai Zhang , Qiuhua Tang

In this paper, flexible preventive maintenance (PM) activities are incorporated into two-stage assembly flow shop scheduling where m dedicated machines in the fabrication stage and one machine in the assembly stage. The operational status of each machine is described by a continuous variable, maintenance level. The maintenance level value is inversely proportional to the processing time. Once a PM activity is performed, this value will return to the initial value. Different from the PM at fixed predefined time intervals, flexible PM can be carried out at any time point, but the maintenance levels are not less than 0. Hence, a MILP model with maintenance level constraints is formulated to minimize the total completion time and maintenance time. Regarding the methods, a latest PM decision strategy is proposed to determine the execution time of PM activities. This new strategy is embedded into 15 constructive heuristics and 7 meta-heuristics (three variants of iterated local search, three variants of Q-learning-based ant colony system with local search and a Q-learning-based hyper-heuristics) to address this new problem. The final experimental analysis demonstrates the significance of the integrated model and the effectiveness of the proposed constructive heuristics and meta-heuristics.



中文翻译:

将预防性维护与两阶段装配流水车间调度相集成:MILP模型,建设性启发式方法和元启发式方法

本文将灵活的预防性维护(PM)活动纳入了两阶段的装配流水车间调度中,其中m在制造阶段使用专用机器,在组装阶段使用一台机器。每台机器的运行状态由连续变量维护级别来描述。维护级别值与处理时间成反比。一旦执行PM活动,该值将返回到初始值。与固定的预定时间间隔的PM不同,可以在任何时间点执行灵活的PM,但是维护级别不小于0。因此,制定了具有维护级别约束的MILP模型,以最大程度地减少总的完成时间和维护时间。关于这些方法,提出了一种最新的PM决策策略来确定PM活动的执行时间。这项新策略已嵌入15种建设性启发式方法和7种元启发式方法(迭代式局部搜索的三种变体,具有局部搜索的基于Q学习的蚁群系统的三种变体和基于Q学习的超启发式方法)新问题。最终的实验分析证明了集成模型的重要性以及所提出的建设性启发式方法和元启发式方法的有效性。

更新日期:2021-03-03
down
wechat
bug