China Mechanical Engineering ›› 2026, Vol. 37 ›› Issue (7): 1725-1733.DOI: 10.3969/j.issn.1004-132X.2026.07.022

Previous Articles    

Distributed Assembly Hybrid Flow Shop Scheduling with Dual Resource Constraints

ZHAO Cai1(), WU Lianghong2()   

  1. 1.School of Mechanical Engineering,Hunan University of Science and Technology,Xiangtan,Hunan,411100
    2.School of Information and Electrical Engineering,Hunan University of Science and Technology,Xiangtan,Hunan,411100
  • Received:2024-11-27 Revised:2026-04-06 Online:2026-07-25 Published:2026-08-18
  • Contact: WU Lianghong

具有双资源约束的分布式装配混合流水车间调度

赵才1(), 吴亮红2()   

  1. 1.湖南科技大学机电工程学院, 湘潭, 411100
    2.湖南科技大学信息与电气工程学院, 湘潭, 411100
  • 通讯作者: 吴亮红
  • 作者简介:赵才, 男, 1996年生,博士研究生。研究方向为生产调度、智能算法。发表论文10余篇。E-mail:zhaocai1996@163.com
    吴亮红*(通信作者), 男, 1977年生,教授, 博士研究生导师。研究方向为智能优化与调度、多目标优化。发表论文100余篇。E-mail:lhwu@hnust.edu.cn.
    第一联系人:韦进文*(通信作者),男,1976年生,教授。研究方向为机械电子工程。发表论文30篇。E-mail: my595@sina.com.cn
  • 基金资助:
    国家重点研发计划(2023YFC3011100);国家自然科学基金(62373146);湖南省自然科学基金(2022JJ30265);湖南省自然科学基金(2023JJ40286);湖南省人才托举工程青年人才项目(2022TJ-Q03)

Abstract:

This paper proposes a knowledge-driven iterative greedy (KDIG) algorithm for the distributed assembly hybrid flow shop scheduling problem with dual resource constraints. The algorithm adopts a knowledge-based NEH (Nawaz-Enscore-Ham) initialization strategy to generate the initial solution. Based on the problem characteristics, four local search operators are designed. Combined with a Q-learning mechanism, these operators enable individuals to dynamically select the optimal local search operator during iterative updates, thereby significantly improving search efficiency. Experimental results on 81 large-scale instances, comparing the KDIG algorithm with five other mainstream algorithms, demonstrate that the proposed KDIG algorithm outperforms all benchmark algorithms.

Key words: distributed assembly hybrid flow shop scheduling, knowledge-driven, iterative greedy algorithm, Q-learning

摘要:

针对具有双资源约束的分布式装配混合流水车间调度问题,提出一种基于知识驱动的迭代贪婪(KDIG)算法。采用基于知识的NEH(Nawaz-Enscore-Ham)初始化策略,生成初始解决方案。基于问题特征设计了4种局部搜索算子。局部搜索算子与Q学习机制的结合使个体在迭代更新时动态选择最优的局部搜索算子,从而显著提升个体在搜索过程中的效率。KDIG算法与5种主流算法在81个大型实例上的测试结果表明,KDIG算法优于其他对比算法。

关键词: 分布式装配混合流车间调度, 知识驱动, 迭代贪婪算法, Q学习

CLC Number: