一种自适应免疫记忆多克隆进化算法  

A Novel Function Optimization Algorithm Using Float Point Number Code

作　　者：刘传玉[1] 王兆青[1] 

机构地区：[1]浙江理工大学计算机技术教研部,杭州310018

出　　处：《浙江理工大学学报（自然科学版）》2009年第1期104-108,共5页Journal of Zhejiang Sci-Tech University(Natural Sciences)

基　　金：浙江省自然科学基金(Y106460)

摘　　要：基于免疫系统的克隆选择机制,提出一种用于函数优化的改进免疫克隆算法。该算法针对单克隆选择算法容易陷入局部最优的弱点以及算法在迭代后期易出现停止不前的现象,采用浮点编码,增加了两超变异、启发式交叉和错位交叉三种算子;对不同的抗体群采用不同的进化方法;自适应调节变异和交叉系数及抗体的克隆数量。对三个典型复杂函数的测试结果表明,该算法有效地克服了早熟问题,提高了全局寻优能力,收敛速度快,性能稳定。Based on clonal selection principle, a novel function optimization algorithm is proposed using float point number code. Directing at the shortcomings of SCSA that falls into local optimization and evolutionary stagnation easily, the three main operations of two kinds of hyper mutation, heuristic crossover and crossover in version are defined. To different antibody group, different evolution method is adopted; the scaling factors of mutation and crossover are adaptively adjusted, so is antibody＇s clone quantity. The simulation results, which have adopted 3 comprehensive benchmark functions, comparing to standard SCSA, demonstrate that the proposed algorithm has good performance.

关 键 词：克隆选择算法 高斯变异 超变异 启发式交叉 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]