Journal of South China University of Technology (Natural Science Edition) ›› 2011, Vol. 39 ›› Issue (5): 113-119.doi: 10.3969/j.issn.1000-565X.2011.05.020

• Computer Science & Technology • Previous Articles     Next Articles

Sustainable Evolutionary Algorithm for Reusing Maximum Frequent Patterns

Yang Guan-ci1  Li Qin2  Li Shao-bo1,2  Zhong Yong1   

  1. 1. Chengdu Institute of Computer Applications,Chinese Academy of Sciences,Chengdu 610041,Sichuan,China; 2. Key Laboratory of Advanced Manufacturing Technology,Ministry of Education,Guizhou University,Guiyang 550003,Guizhou,China
  • Received:2010-10-21 Revised:2011-02-27 Online:2011-05-25 Published:2011-04-01
  • Contact: 杨观赐(1983-) ，男，博士生，主要从事计算智能研究． E-mail:guanci_yang@163.com
  • About author:杨观赐(1983-) ，男，博士生，主要从事计算智能研究．
  • Supported by:

    教育部新世纪优秀人才支持计划项目( NCET09-0094) ; 国家自然科学基金资助项目( 60975049) ; 贵州省科学技术基金资助项目( 黔科合J 字［2010］2095 号)

PDF

813

Abstract:

In order to make good reuse of the information precipitated in excellent individuals during the evolutionary process,a maximal frequent sequential pattern mining algorithm ( MFSPMA) is proposed,based on which a sustainable evolutionary algorithm for reusing the maximum frequent patterns is put forward and is abbreviated to MFPEA. In MFPEA,several subpopulations are adopted to provide survival space for the individuals with different
fitness levels,MFSPMA is used to extract excellent genes from the population,and new individuals with excellent gene schema are poured into the subpopulations to stabilize the inheritance of genetic information. Furthermore,a self-adaptive function is designed to adjust the population size for different problems,and a series of statistical data is used to investigate the parameters balancing the computation time and the evolutionary quality. Experimental results show that MFPEA performs good functions in maintaining information stability and avoiding premature convergence,and that it sets a new tour record,namely 3611. 496,for the xit1083 instance.

Key words: maximal frequent sequential pattern, sequence mining, gene reuse, sustainable evolutionary algorithm, traveling salesman problem