不同映射的混沌免疫进化算法性能分析  

Performance Analysis of Chaos Immune Evolutionary Algorithm with Different Maps

作　　者：苏有良[1] 周德俭[1,2] 吴兆华[1] 万川[1] 

机构地区：[1]桂林电子科技大学机电工程学院,广西桂林541004 [2]广西工学院机械工程系,广西柳州545006

出　　处：《计算机工程》2010年第21期222-224,共3页Computer Engineering

基　　金：广西自然科学基金资助项目(桂科自0832242);广西制造系统与先进制造技术重点实验室主任课题基金资助项目(桂科能0842006_016_Z)

摘　　要：针对目前混沌免疫进化算法采用的混沌映射类型单一,并存在对混沌映射影响算法性能大小和机制缺乏深入研究等问题,分析和探讨基于不同混沌映射混沌免疫进化算法的性能。对几个典型测试函数的比较结果表明,Logistic-CIEA和Cubic-CIEA的性能易出现大波动,Kent-CIEA则具有相对稳定的收敛速度,表现出较强的鲁棒性。由此证明,混沌映射作为产生局部搜索轨迹的迭代函数,其混沌特性对算法性能影响较大。Considering single type of chaotic map and not deeply study of effect size and mechanism of chaotic map on the algorithm performance at present, this paper analyzes and dicusses performance of Chaos Immune Evolutionary Algorithm（CIEA） based on different chaotic maps. Four benchmark functions are employed to test performances of the three chaos immune evolutionary algorithms. Comparison results demonstrate that the performances of Logistic-CIEA and Cubic-CIEA are easy to cause large fluctuation. However, Kent-CIEA has relatively steady convergence rate and shows strong robustness. It proves that the performance of chaos immune evolutionary algorithm is deeply affected by the characteristic of chaotic map, which is applied to generate the local search trajectory.

关 键 词：混沌免疫进化算法 混沌映射 收敛速度 鲁棒性 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]