作者投稿和查稿 主编审稿 专家审稿 编委审稿 远程编辑

计算机工程

• 人工智能及识别技术 • 上一篇    下一篇

海量种群基因表达式编程的内存删冗算法

钟坚成 1,彭 玮 2   

  1. (1. 湖南师范大学工程与设计学院,长沙410081;2. 昆明理工大学信息工程与自动化学院,昆明650093)
  • 收稿日期:2013-08-28 出版日期:2014-09-15 发布日期:2014-09-12
  • 作者简介:钟坚成(1981 - ),男,讲师、博士研究生,主研方向:机器学习,生物信息学;彭 玮(通讯作者),讲师、博士研究生。
  • 基金资助:
    湖南省教育厅优秀青年基金资助项目(12B080);湖南省科技计划基金资助项目(2010GK3023);湖南师范大学教学改革基金 资助项目(2013)。

Memory Reducing Redundant Algorithm of Gene Expression Programming with Mass Population

ZHONG Jian-cheng  1,PENG Wei  2   

  1. (1. College of Engineering and Design,Hunan Normal University,Changsha 410081,China;2. Faculty of Information Engineering and Automation,Kunming University of Science and Technology,Kunming 650093,China)
  • Received:2013-08-28 Online:2014-09-15 Published:2014-09-12

摘要: 在大样本、多种群、高进化代数的情况下,基因表达式编程(GEP)容易产生冗余个体染色体有效串,从而影响计算性能。为解决该问题,提出一种基于内存检测种群冗余的算法MPRRGEP。分析单基因、多基因对种群冗余性的影响,设计个体染色体有效性的测度方法。提出内存Hash 种群映射删冗算法,在内存中索引个体染色体数据,减少相同有效串的重复计算次数,大幅提高GEP 计算性能。实验结果表明,相比传统GEP 算法,MPRRGEP 算法平均减少60% 以上的计算时间。

关键词: 基因表达式编程, 删冗, 哈希表, 基因有效串, 多基因, 关键蛋白质预测

Abstract: Programming(GEP)is prone to produce redundant valid strings of chromosome,which impacts its performance dramatically. To address the problem,this paper proposes a new strategy named Memory Population Reducing Redundant GEP(MPRRGEP), which checks repeat valid strings of chromosome and reduces the redundant in memory. It analyses the influence of valid strings in both single-gene and multi-gene chromosome on the performance of GPE. And a method that can effectively measure the validity of individual chromosome is designed. By using Hash technique,the index of the data of valid individual chromosome is constructed in memory so as to reduce the amount of times that compute the same valid strings and improve the performance of GEP. Experimental results show that the method can averagely save the computing time for above 60% .

Key words: Gene Expression Programming (GEP), reducing redundant, Hash table, valid gene string, multi-gene, crucial protein prediction

中图分类号: