Convergence Analysis on Gene Expression Programming
The gene expression system works with mutation, inversion, transposition, recombination as well as fitness-proportional selection with elitism. As a novel geno-type/phenotype evolutionary algorithm, gene expression programming has many applications, but little is know about its convergence properties. This article deal with the convergence of two kinds of gene expression programming algorithms. By analytical techniques of homogeneous finite Markov chains, we perform proofs that the canonical gene expression algorithms could not converge to the global optimum with probability 1, while the modified algorithm do.
Gene expression programming Markov chain Convergence
Ming Chen Lixin Ding Jianping Yu
Slate Key Laboratory of Software Engineering Wuhan University,Wuhan 430072,China College of Mathemat State Key Laboratory of Software Engineering Wuhan University Wuhan 430072,China College of Mathematics and Computer Science Hunan Normal University Changsha 410081,China
国际会议
太原
英文
45-49
2011-02-26(万方平台首次上网日期,不代表论文的发表时间)