计算机工程与应用 ›› 2018, Vol. 54 ›› Issue (23): 131-136.DOI: 10.3778/j.issn.1002-8331.1708-0183

• 模式识别与人工智能 • 上一篇    下一篇

基因沉默机制的基因表达式编程

郭  勇1,张国锋1,刘丽萍2   

  1. 1.黔南民族师范学院 计算机与信息学院,贵州 都匀 558000
    2.黔南民族师范学院 生命科学与农学院,贵州 都匀 558000
  • 出版日期:2018-12-01 发布日期:2018-11-30

Gene expression programming based on gene silencing mechanism

GUO Yong1, ZHANG Guofeng1, LIU Liping2   

  1. 1.School of Computer and Information, Qiannan Normal University for Nationalities, Duyun, Guizhou 558000, China
    2.School of Life Science and Agriculture, Qiannan Normal University for Nationalities, Duyun, Guizhou 558000, China
  • Online:2018-12-01 Published:2018-11-30

摘要: 基因表达式编程(Gene Expression Programming,GEP)对多项式函数为目标的符号回归问题计算效果良好,而对包含多种运算目数、非多项式函数的计算效果欠佳。受转基因生物工程中基因沉默现象的启发,提出一种GEP拓展算法SFGEP(Gene Expression Programming of Symbol Field,SFGEP)。SFGEP染色体由表达因子域与表达基因域组成,按“深度优先”原则解释染色体,利用不同操作符目数,形成基因表达的抑制因子和位置效应,实现染色体解释中基因沉默的机制。实验结果表明,相较传统多基因染色体GEP,SFGEP既保持了一定多项式函数挖掘的能力,又在包含不同运算目数操作符的非多项式函数挖掘方面具有更好的效能,SFGEP的成功率更高、收敛速度更快。

关键词: 基因表达式编程, 染色体, 基因沉默, 符号回归, 多样性

Abstract: The computational effect of Gene Expression Programming(GEP) is better in the problem of symbolic regression with polynomial function as test function, while its computational effect in the more operator number and the non-polynomial function is poor. In this paper, it proposes a gene expression programming algorithm(SFGEP) based on symbolic field, which is based on gene silencing in transgenic biology engineering. The chromosome structure in the SFGEP algorithm is composed of the expression factor domain and the expression gene domain. The chromosomes are explained by the principle of “depth first”, and the gene silencing mechanism in the chromosome interpretation is implemented by using the operator number of different operators to form the inhibitory factor and the position effect of gene expression. The experimental results show that SFGEP, which is greater expression of linear space in comparison with the traditional polygene chromosome GEP, not only has the ability of mining polynomial functions, but also has better performance in non-polynomial function mining with different mesh operators. SFGEP has higher success rate and faster convergence speed.

Key words: gene expression programming, chromosome, gene silencing, symbol regression, diversity