Scholarlink

<p>Lot streaming is the most widely used technique to facilitate the overlapping of successive operations. It is particularly effective in manufacturing systems characterized by strict delivery deadlines and complex multi-stage processes, such as locomotive component production. Motivated by real-world scenarios where last stage waiting time constraints critically impact delivery performance, this paper investigates the hybrid flow shop scheduling problem with consistent sublots and last stage limited waiting time constraint (HFSP_CSLLWT), aiming to simultaneously optimize two conflicting objectives: the maximum delay (MD) and the total delay (TD). Considering the last stage limited waiting time constraint, a multi-objective mixed integer programming model is developed to evaluate the trade-off between MD and TD. Given the NP-hard property of the addressed problem, we propose a dynamic three-stage multi-objective algorithm based on decomposition (DTMOA/D), an evolutionary algorithm, for solving HFSP_CSLLWT. In DTMOA/D, each solution is associated with both a main exploration direction and a dynamically adjusted set of neighboring directions. The algorithm integrates: (1) Variable neighborhood search (VNS) with four local search methods for intensive exploitation; (2) Direction jumping for adaptive exploration; (3) Neighborhood collaboration with learning strategies for global search; (4) Dynamic disturbance to escape local optima. Numerical experiments on 300 instances demonstrate that DTMOA/D outperforms four existing algorithms in terms of solution quality and robustness. This research provides an efficient method for solving HFSP_CSLLWT, and offers valuable insights into handling similar complex scheduling challenges.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Dynamic three-stage decomposition-based multi-objective approach for solving hybrid flow shop scheduling problem with consistent sublots

  • 期刊论文

  • 发表日期：

    2025年12月10日

查看全文
  • Junjie Zhou,
  • Miaocong Shen,
  • Yuguo Wang,
  • Xiongbing Li

摘要

Lot streaming is the most widely used technique to facilitate the overlapping of successive operations. It is particularly effective in manufacturing systems characterized by strict delivery deadlines and complex multi-stage processes, such as locomotive component production. Motivated by real-world scenarios where last stage waiting time constraints critically impact delivery performance, this paper investigates the hybrid flow shop scheduling problem with consistent sublots and last stage limited waiting time constraint (HFSP_CSLLWT), aiming to simultaneously optimize two conflicting objectives: the maximum delay (MD) and the total delay (TD). Considering the last stage limited waiting time constraint, a multi-objective mixed integer programming model is developed to evaluate the trade-off between MD and TD. Given the NP-hard property of the addressed problem, we propose a dynamic three-stage multi-objective algorithm based on decomposition (DTMOA/D), an evolutionary algorithm, for solving HFSP_CSLLWT. In DTMOA/D, each solution is associated with both a main exploration direction and a dynamically adjusted set of neighboring directions. The algorithm integrates: (1) Variable neighborhood search (VNS) with four local search methods for intensive exploitation; (2) Direction jumping for adaptive exploration; (3) Neighborhood collaboration with learning strategies for global search; (4) Dynamic disturbance to escape local optima. Numerical experiments on 300 instances demonstrate that DTMOA/D outperforms four existing algorithms in terms of solution quality and robustness. This research provides an efficient method for solving HFSP_CSLLWT, and offers valuable insights into handling similar complex scheduling challenges.