中国科学院机构知识库网格
Chinese Academy of Sciences Institutional Repositories Grid
首页 机构 成果 学者
  
  1. CAS IR Grid
  2. 计算技术研究所
  3. 中国科学院计算技术研究所
  4. 中国科学院计算技术研究所期刊论文
  5. 英文
Learning to schedule dynamic job-shop problems by graph attention network with reinforcement learning

文献类型：期刊论文

作者Huang, Chao2,3; Hu, Haibo1,2; Zhou, Yan1,2
刊名APPLIED INTELLIGENCE
出版日期2025-10-29
卷号55期号:16页码:16
关键词Job shop scheduling Reinforcement learning Graph attention network
ISSN号0924-669X
DOI10.1007/s10489-025-06949-6
英文摘要The job shop scheduling problem (JSSP) is the core of industrial production and manufacturing with strongly NP-hard complexity. Recently, deep reinforcement learning (DRL) has been shown to be an ideal technique for learning priority dispatching rules (PDRs) to solve complex scheduling problems. However, In real-world environments, scheduling is complex, stochastic, and dynamic, with stochastic job arrivals and random machine breakdowns, which is known as dynamic JSSP (DJSSP). Most DRL methods focus on JSSP problem. In this paper, we propose a framework which extends the DRL method based on graph neural networks and deep reinforcement learning to addresse DJSSP problem of dynamic jobs with random processing time. Specifically, we extend the graph attention network (GAT) as a common encoder between JSSP and DJSSP, encoding job and machine information into nodes and arcs represented by a disjunctive graph. Then we exploit the encoder to learn high-quality PDRs via an deep reinforcement learning agent with proximal policy optimization. Moreover, by the communality of the GAT encoder, the embeddings of nodes and arcs pretrained in JSSP can be transferred into DJSSP as a good prior knowledge. Thus, the learning framework can pay more attention to learn the multilayer perceptron (MLP) predictor of the policy, which combines GAT as a policy network and produces better performance, rather than randomly initializing the GAT, trained from the original environment. Experiments show that the framework outperforms the DRL methods without pretrained embeddings and is computationally efficient, even on instances of larger scales and different properties unseen in training.
资助项目Special Project of Strategic Research and Decision Support System Construction of the Chinese Academy of Sciences (CAS)-Strategic Research on Key Fields for the Research and Formulation of the 15th Five-Year Plan: Network Information[GHJ-ZLZX-2023-11] ; Kechuang Yongjiang 2035 key technology breakthrough plan of Zhejiang Ningbo grant[2025Z042] ; Kechuang Yongjiang 2035 key technology breakthrough plan of Zhejiang Ningbo grant[2024Z283] ; Kechuang Yongjiang 2035 key technology breakthrough plan of Zhejiang Ningbo[2024Z119] ; Entrusted Operation Project of Computing Power for Ningbo Artificial Intelligence Supercomputing Center
WOS研究方向Computer Science
语种英语
WOS记录号WOS:001603596500002
出版者SPRINGER
源URL[http://119.78.100.204/handle/2XEOYT63/41584]  
专题中国科学院计算技术研究所期刊论文_英文
通讯作者Huang, Chao
作者单位1.Univ Chinese Acad Sci, Beijing, Peoples R China
2.Chinese Acad Sci ICT CAS, Inst Comp Technol, Beijing, Peoples R China
3.Ningbo Inst Artificial Intelligence Ind, Ningbo, Peoples R China
推荐引用方式
GB/T 7714		 Huang, Chao,Hu, Haibo,Zhou, Yan. Learning to schedule dynamic job-shop problems by graph attention network with reinforcement learning[J]. APPLIED INTELLIGENCE,2025,55(16):16.
APA Huang, Chao,Hu, Haibo,&Zhou, Yan.(2025).Learning to schedule dynamic job-shop problems by graph attention network with reinforcement learning.APPLIED INTELLIGENCE,55(16),16.
MLA Huang, Chao,et al."Learning to schedule dynamic job-shop problems by graph attention network with reinforcement learning".APPLIED INTELLIGENCE 55.16(2025):16.

入库方式： OAI收割

来源：计算技术研究所

浏览0
下载0
收藏0
其他版本

除非特别说明，本系统中所有内容都受版权保护，并保留所有权利。