On combining variable ordering heuristics for constraint satisfaction problems

被引:3
作者
Li, Hongbo [1 ]
Feng, Guozhong [1 ]
Yin, Minghao [1 ]
机构
[1] Northeast Normal Univ, Sch Informat Sci & Technol, Changchun 130117, Peoples R China
基金
中国国家自然科学基金;
关键词
Constraint satisfaction problem; Variable ordering heuristic; Activity-based search; Dom; wdeg; Impact-based search;
D O I
10.1007/s10732-019-09434-9
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Variable ordering heuristics play a central role in solving constraint satisfaction problems. Combining two variable ordering heuristics may generate a more efficient heuristic, such as dom/deg. In this paper, we propose a novel method for combining two variable ordering heuristics, namely Pearson-Correlation-Coefficient-based Combination (PCCC). While the existing combination strategies always combine participant heuristics, PCCC checks whether the participant heuristics are suitable for combination before combining them in the context of search. If they should be combined, it combines the heuristic scores to select a variable to branch on, otherwise, it randomly selects one of the participant heuristics to make the decision. The experiments on various benchmark problems show that PCCC can be used to combine different pairs of heuristics, and it is more robust than the participant heuristics and some classical combining strategies.
引用
收藏
页码:453 / 474
页数:22
相关论文
共 36 条
  • [1] Amadini R, 2015, PROCEEDINGS OF THE TWENTY-FOURTH INTERNATIONAL JOINT CONFERENCE ON ARTIFICIAL INTELLIGENCE (IJCAI), P232
  • [2] [Anonymous], 2013, INT C PRINC PRACT CO, P464
  • [3] Arbelaez A., 2009, ONLINE HEURISTIC SEL
  • [4] Bessiere C., 1996, Principles and Practice of Constraint Programming - CP96. Second International Conference - CP96. Proceedings, P61
  • [5] Boussemart F, 2004, FRONT ARTIF INTEL AP, V110, P146
  • [6] NEW METHODS TO COLOR THE VERTICES OF A GRAPH
    BRELAZ, D
    [J]. COMMUNICATIONS OF THE ACM, 1979, 22 (04) : 251 - 256
  • [7] Hyper-heuristics: a survey of the state of the art
    Burke, Edmund K.
    Gendreau, Michel
    Hyde, Matthew
    Kendall, Graham
    Ochoa, Gabriela
    Oezcan, Ender
    Qu, Rong
    [J]. JOURNAL OF THE OPERATIONAL RESEARCH SOCIETY, 2013, 64 (12) : 1695 - 1724
  • [8] Combine and conquer: an evolutionary hyper-heuristic approach for solving constraint satisfaction problems
    Carlos Ortiz-Bayliss, Jose
    Terashima-Marin, Hugo
    Enrique Conant-Pablos, Santiago
    [J]. ARTIFICIAL INTELLIGENCE REVIEW, 2016, 46 (03) : 327 - 349
  • [9] Collavizza H, 1998, LECT NOTES COMPUT SC, V1520, P147
  • [10] EXPERIMENTAL EVALUATION OF PREPROCESSING ALGORITHMS FOR CONSTRAINT SATISFACTION PROBLEMS
    DECHTER, R
    MEIRI, I
    [J]. ARTIFICIAL INTELLIGENCE, 1994, 68 (02) : 211 - 241