A decomposition method for symbolic regression problems

被引：12
作者：
Astarabadi, Samaneh Sadat Mousavi [1 ]
Ebadzadeh, Mohammad Mehdi [1 ]
机构：
[1] Amirkabir Univ Technol, Dept Comp Engn & Informat Technol, Tehran, Iran
来源：
APPLIED SOFT COMPUTING | 2018年 / 62卷
关键词：
Genetic Programming; Symbolic regression; Performance estimation; Decomposition; Optimization;
D O I：
10.1016/j.asoc.2017.10.041
中图分类号：
TP18 [人工智能理论];
学科分类号：
081104 ; 0812 ; 0835 ; 1405 ;
摘要：
The purpose of this paper is to improve the efficiency of Genetic Programming (GP) by decomposing a regression problem into several subproblems. An optimization problem is defined to find subproblems of the original problem for which the performance of GP is better than for the original problem. In order to evaluate the proposed decomposition method, the subproblems of several benchmark problems are found by solving the optimization problem. Then, a 2-layer GP system is used to find subproblems' solutions in the first layer and the solution of the original problem in the second layer. The results of this 2-layer GP system show that the proposed decomposition method does not generate trivial subproblems. It generates subproblems that improve the efficiency of GP against when subproblems are not used. (C) 2017 Elsevier B.V. All rights reserved.
引用
收藏
页码：514 / 523
页数：10
相关论文
共 38 条
  • [1] Angeline P., 1993, Proceedings of the Second Annual Conference on Evolutionary Programming, P154
  • [2] [Anonymous], 2012, 2012 IEEE RIVF INT C, DOI DOI 10.1109/RIVF.2012.6169838
  • [3] Azad RMA, 2011, GECCO-2011: PROCEEDINGS OF THE 13TH ANNUAL GENETIC AND EVOLUTIONARY COMPUTATION CONFERENCE, P1315
  • [4] Multi-stage genetic programming: A new strategy to nonlinear system modeling
    Gandomi, Amir Hossein
    Alavi, Amir Hossein
    [J]. INFORMATION SCIENCES, 2011, 181 (23) : 5227 - 5239
  • [5] Gustafson SM, 2001, LECT NOTES COMPUT SC, V2038, P291
  • [6] Hoang TH, 2006, IEEE C EVOL COMPUTAT, P2119
  • [7] Hsu W.H., 2002, Genetic and Evolutionary Computation Conference, P764
  • [8] Jackson D, 2008, LECT NOTES COMPUT SC, V4971, P86, DOI 10.1007/978-3-540-78671-9_8
  • [9] Jackson D, 2008, LECT NOTES COMPUT SC, V4971, P170, DOI 10.1007/978-3-540-78671-9_15
  • [10] Jackson D, 2007, LECT NOTES COMPUT SC, V4445, P148
1234