Learning to Dispatch for Job Shop Scheduling via Deep Reinforcement Learning

被引:0
作者
Zhang, Cong [1 ]
Song, Wen [2 ]
Cao, Zhiguang [3 ]
Zhang, Jie [1 ]
Tan, Puay Siew [4 ]
Xu, Chi [4 ]
机构
[1] Nanyang Technol Univ, Singapore, Singapore
[2] Shandong Univ, Inst Marine Sci & Technol, Shandong, Peoples R China
[3] Natl Univ Singapore, Singapore, Singapore
[4] ASTAR, Singapore Inst Mfg Technol, Singapore, Singapore
来源
ADVANCES IN NEURAL INFORMATION PROCESSING SYSTEMS 33, NEURIPS 2020 | 2020年 / 33卷
基金
新加坡国家研究基金会; 中国国家自然科学基金;
关键词
DISJUNCTIVE GRAPHS; BENCHMARKS; ALGORITHM; RULES;
D O I
暂无
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Priority dispatching rule (PDR) is widely used for solving real-world Job-shop scheduling problem (JSSP). However, the design of effective PDRs is a tedious task, requiring a myriad of specialized knowledge and often delivering limited performance. In this paper, we propose to automatically learn PDRs via an end-to-end deep reinforcement learning agent. We exploit the disjunctive graph representation of JSSP, and propose a Graph Neural Network based scheme to embed the states encountered during solving. The resulting policy network is size-agnostic, effectively enabling generalization on large-scale instances. Experiments show that the agent can learn high-quality PDRs from scratch with elementary raw features, and demonstrates strong performance against the best existing PDRs. The learned policies also perform well on much larger instances that are unseen in training.
引用
收藏
页数:12
相关论文
共 50 条
[1]  
Amizadeh S., 2019, INT C LEARN REPR
[2]  
Bacciu D., 2020, NEURAL NETWORKS
[3]   AN ADDITIVE ALGORITHM FOR SOLVING LINEAR PROGRAMS WITH 0-1 VARIABLES [J].
BALAS, E .
OPERATIONS RESEARCH, 1965, 13 (04) :517-&
[5]  
Bengio Yoshua, 2020, EUROPEAN J OPERATION
[6]   The disjunctive graph machine representation of the job shop scheduling problem [J].
Blazewicz, J ;
Pesch, E ;
Sterna, M .
EUROPEAN JOURNAL OF OPERATIONAL RESEARCH, 2000, 127 (02) :317-331
[7]  
Chen XY, 2019, ADV NEUR IN, V32
[8]  
Dai HJ, 2017, ADV NEUR IN, V30
[9]   Benchmarks for shop scheduling problems [J].
Demirkol, E ;
Mehta, S ;
Uzsoy, R .
EUROPEAN JOURNAL OF OPERATIONAL RESEARCH, 1998, 109 (01) :137-141
[10]   A Review on Swarm Intelligence and Evolutionary Algorithms for Solving Flexible Job Shop Scheduling Problems [J].
Gao, Kaizhou ;
Cao, Zhiguang ;
Zhang, Le ;
Chen, Zhenghua ;
Han, Yuyan ;
Pan, Quanke .
IEEE-CAA JOURNAL OF AUTOMATICA SINICA, 2019, 6 (04) :904-916