摘要
基因表达式编程(GEP)的个体代表了问题的候选解。在缺乏先验知识的情况下,个体长度的设定是个"两难"问题,过长或过短都会降低GEP的效率。对此,分析了个体长度对GEP求解效率的影响;设计了开放阅读框(ORF)过滤算子根据最优个体的进化历程动态调节个体的有效编码区域;验证了ORF过滤算子的有效性,实验结果表明,在同样的进化代数内,引入ORF过滤算子,GEP能进化出更高适应度的最优解且减少平均运行时间17.0%。
The individuals of Gene Expression Programming (GEP) represent the candidate solutions of the problem. The determining of the length of an individual is a dilemma when there is no prior knowledge available. Either long or short length decreases the efficiency of GEP. The efficiency impact of the individual' s length was analyzed, Open Reading Frame (ORF) filter was designed to modify the coding region dynamically based on the evolution progress of the best individual, and the effectiveness of ORF filter was demonstrated. The experiments showed that solutions with higher fitness can be found and the average runtime is reduced by 17.0% in the same number of generations by introducing ORF filter into original GEP.
出处
《四川大学学报(工程科学版)》
EI
CAS
CSCD
北大核心
2007年第6期102-106,共5页
Journal of Sichuan University (Engineering Science Edition)
基金
国家自然科学基金资助项目(60473071)
高等学校博士学科点专项科研基金SRFDP(20020610007)
四川省青年软件创新工程项目(2005AA0807)
关键词
基因表达式编程
开放阅读框
个体
Gene Expression Programming
Open Reading Frame
individual