A deep reinforcement learning based approach for dynamic distributed blocking flowshop scheduling with job insertions
Abstract The distributed blocking flowshop scheduling problem (DBFSP) with new job insertions is studied. Rescheduling all remaining jobs after a dynamic event like a new job insertion is unreasonable to an actual distributed blocking flowshop production process. A deep reinforcement learning (DRL)...
Main Authors: | , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
Wiley
2022-09-01
|
Series: | IET Collaborative Intelligent Manufacturing |
Subjects: | |
Online Access: | https://doi.org/10.1049/cim2.12060 |
_version_ | 1811252130863906816 |
---|---|
author | Xueyan Sun Birgit Vogel‐Heuser Fandi Bi Weiming Shen |
author_facet | Xueyan Sun Birgit Vogel‐Heuser Fandi Bi Weiming Shen |
author_sort | Xueyan Sun |
collection | DOAJ |
description | Abstract The distributed blocking flowshop scheduling problem (DBFSP) with new job insertions is studied. Rescheduling all remaining jobs after a dynamic event like a new job insertion is unreasonable to an actual distributed blocking flowshop production process. A deep reinforcement learning (DRL) algorithm is proposed to optimise the job selection model, and local modifications are made on the basis of the original scheduling plan when new jobs arrive. The objective is to minimise the total completion time deviation of all products so that all jobs can be finished on time to reduce the cost of storage. First, according to the definitions of the dynamic DBFSP problem, a DRL framework based on multi‐agent deep deterministic policy gradient (MADDPG) is proposed. In this framework, a full schedule is generated by the variable neighbourhood descent algorithm before a dynamic event occurs. Meanwhile, all newly added jobs are reordered before the agents make decisions to select the one that needs to be scheduled most urgently. This study defines the observations, actions and reward calculation methods and applies centralised training and distributed execution in MADDPG. Finally, a comprehensive computational experiment is carried out to compare the proposed method with the closely related and well‐performing methods. The results indicate that the proposed method can solve the dynamic DBFSP effectively and efficiently. |
first_indexed | 2024-04-12T16:30:17Z |
format | Article |
id | doaj.art-9c3d591dc17941abaa888b7c4a37b613 |
institution | Directory Open Access Journal |
issn | 2516-8398 |
language | English |
last_indexed | 2024-04-12T16:30:17Z |
publishDate | 2022-09-01 |
publisher | Wiley |
record_format | Article |
series | IET Collaborative Intelligent Manufacturing |
spelling | doaj.art-9c3d591dc17941abaa888b7c4a37b6132022-12-22T03:25:10ZengWileyIET Collaborative Intelligent Manufacturing2516-83982022-09-014316618010.1049/cim2.12060A deep reinforcement learning based approach for dynamic distributed blocking flowshop scheduling with job insertionsXueyan Sun0Birgit Vogel‐Heuser1Fandi Bi2Weiming Shen3Department of Mechanical Engineering Chair of Automation and Information Systems Technical University Munich Munich GermanyDepartment of Mechanical Engineering Chair of Automation and Information Systems Technical University Munich Munich GermanyDepartment of Mechanical Engineering Chair of Automation and Information Systems Technical University Munich Munich GermanySchool of Mechanical Science and Engineering Huazhong University of Science and Technology Wuhan ChinaAbstract The distributed blocking flowshop scheduling problem (DBFSP) with new job insertions is studied. Rescheduling all remaining jobs after a dynamic event like a new job insertion is unreasonable to an actual distributed blocking flowshop production process. A deep reinforcement learning (DRL) algorithm is proposed to optimise the job selection model, and local modifications are made on the basis of the original scheduling plan when new jobs arrive. The objective is to minimise the total completion time deviation of all products so that all jobs can be finished on time to reduce the cost of storage. First, according to the definitions of the dynamic DBFSP problem, a DRL framework based on multi‐agent deep deterministic policy gradient (MADDPG) is proposed. In this framework, a full schedule is generated by the variable neighbourhood descent algorithm before a dynamic event occurs. Meanwhile, all newly added jobs are reordered before the agents make decisions to select the one that needs to be scheduled most urgently. This study defines the observations, actions and reward calculation methods and applies centralised training and distributed execution in MADDPG. Finally, a comprehensive computational experiment is carried out to compare the proposed method with the closely related and well‐performing methods. The results indicate that the proposed method can solve the dynamic DBFSP effectively and efficiently.https://doi.org/10.1049/cim2.12060deep reinforcement learningdistributed blocking flowshop scheduling problemdynamic schedulingjob insertionsmulti‐agent deep deterministic policy gradient |
spellingShingle | Xueyan Sun Birgit Vogel‐Heuser Fandi Bi Weiming Shen A deep reinforcement learning based approach for dynamic distributed blocking flowshop scheduling with job insertions IET Collaborative Intelligent Manufacturing deep reinforcement learning distributed blocking flowshop scheduling problem dynamic scheduling job insertions multi‐agent deep deterministic policy gradient |
title | A deep reinforcement learning based approach for dynamic distributed blocking flowshop scheduling with job insertions |
title_full | A deep reinforcement learning based approach for dynamic distributed blocking flowshop scheduling with job insertions |
title_fullStr | A deep reinforcement learning based approach for dynamic distributed blocking flowshop scheduling with job insertions |
title_full_unstemmed | A deep reinforcement learning based approach for dynamic distributed blocking flowshop scheduling with job insertions |
title_short | A deep reinforcement learning based approach for dynamic distributed blocking flowshop scheduling with job insertions |
title_sort | deep reinforcement learning based approach for dynamic distributed blocking flowshop scheduling with job insertions |
topic | deep reinforcement learning distributed blocking flowshop scheduling problem dynamic scheduling job insertions multi‐agent deep deterministic policy gradient |
url | https://doi.org/10.1049/cim2.12060 |
work_keys_str_mv | AT xueyansun adeepreinforcementlearningbasedapproachfordynamicdistributedblockingflowshopschedulingwithjobinsertions AT birgitvogelheuser adeepreinforcementlearningbasedapproachfordynamicdistributedblockingflowshopschedulingwithjobinsertions AT fandibi adeepreinforcementlearningbasedapproachfordynamicdistributedblockingflowshopschedulingwithjobinsertions AT weimingshen adeepreinforcementlearningbasedapproachfordynamicdistributedblockingflowshopschedulingwithjobinsertions AT xueyansun deepreinforcementlearningbasedapproachfordynamicdistributedblockingflowshopschedulingwithjobinsertions AT birgitvogelheuser deepreinforcementlearningbasedapproachfordynamicdistributedblockingflowshopschedulingwithjobinsertions AT fandibi deepreinforcementlearningbasedapproachfordynamicdistributedblockingflowshopschedulingwithjobinsertions AT weimingshen deepreinforcementlearningbasedapproachfordynamicdistributedblockingflowshopschedulingwithjobinsertions |