演化元胞自动机函数优化算法案例研究  

A Case Study on Evolutionary Cellular Automata Algorithm for Function Optimization

作　　者：贾红伟[1] 何炎祥[2] 郑波尽[2] 

机构地区：[1]集美大学计算机工程学院,厦门361021 [2]武汉大学计算机学院

出　　处：《计算机工程与应用》2005年第26期54-57,共4页Computer Engineering and Applications

基　　金：集美大学校基金项目(编号:F01038)

摘　　要：BUMP是一个超多维,超多峰,超非线性的问题,被广泛应用于各种演化算法的性能比较。但最好解是未知的。基于元胞自动机的遗传算法报告了BUMP曾经发表过的最好解。该文设计了基于演化元胞自动机的新算法(ECAA)并获得了更好的结果。文中详细讨论了算法中各算子的设计方法及其在算法中扮演的角色,分析了该算法的极度并行,天然局部搜索等重要特性。BUMP is a multiple dimensions and multiple apices and super-nonlinear problem and widely used in performance comparison among variant evolutionary algorithms.But its best solutions are unknown to us,Cellular Automata Based Genetic Algorithm reports some good and ever published results of BUMP.In this paper,an instance of Evolutionary Cellular Automata Algorithm （briefly called ECAA）is proposed and achieves better results of BUMP,Every operator＇s and its roles in this algorithm are discussed in detail.At the same time,some important features such as extremely parallelism,and natural local search capacity of the ECAA.

关 键 词：演化计算 细胞自动机 函数优化 双算术交叉算子 自适应delta 变异算子 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]