Reinforcement learning based local search for grouping problems: A case study on graph coloring

被引:54
作者
Zhou, Yangming [1 ]
Hao, Jin-Kao [1 ,2 ]
Duval, Beatrice [1 ]
机构
[1] Univ Angers, LERIA, 2 Bd Lavoisier, F-49045 Angers, France
[2] Inst Univ France, Paris, France
关键词
Grouping problems; Reinforcement learning; Heuristics; Learning-based optimization; GENETIC ALGORITHM; OPTIMIZATION;
D O I
10.1016/j.eswa.2016.07.047
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Grouping problems aim to partition a set of items into multiple mutually disjoint subsets according to some specific criterion and constraints. Grouping problems cover a large class of computational problems including clustering and classification that frequently appear in expert and intelligent systems as well as many real applications. This paper focuses on developing a general-purpose solution approach for grouping problems, i.e., reinforcement learning based local search (RLS), which combines reinforcement learning techniques with local search. This paper makes the following contributions: we show that (1) reinforcement learning can help obtain useful information from discovered local optimum solutions; (2) the learned information can be advantageously used to guide the search algorithm towards promising regions. To the best of our knowledge, this is the first attempt to. propose a formal model that combines reinforcement learning and local search for solving grouping problems. The proposed approach is verified on a well-known representative grouping problem (graph coloring). The generality of the approach makes it applicable to other grouping problems. (C) 2016 Elsevier Ltd. All rights reserved.
引用
收藏
页码:412 / 422
页数:11
相关论文
共 54 条
[51]   Coloring large graphs based on independent set extraction [J].
Wu, Qinghua ;
Hao, Jin-Kao .
COMPUTERS & OPERATIONS RESEARCH, 2012, 39 (02) :283-290
[52]  
Wu W., 2013, 3 JOINT INT C P, P84
[53]  
Xin Y., 2013, EXPERT SYSTEMS APPL, V58, P10
[54]  
Xu Y., 2009, P LEARN INT OPT LION