基于模式基因提取的多值编码遗传算法性能改善  

Improving the Research Capability of Multi-value Coding Genetic Algorithms by Finding Schema Genes

作　　者：马炫[1] 乐天[1] 孙宁文 

机构地区：[1]西安理工大学自动化与信息工程学院,陕西西安710048 [2]宁夏新技术设计研究院,宁夏银川750001

出　　处：《微电子学与计算机》2005年第3期125-127,共3页Microelectronics & Computer

摘　　要：本文通过模式基因的提取,即根据种群个体的基因值在种群中的统计量确定统计模式基因,然后通过对统计模式基因的测试搜寻模式基因,并将其在种群中进行传播,有效改变种群搜索的方向,改善多值编码遗传算法的搜索性能。以组合优化的多重选择背包问题为对象验证了本方法的有效性。Aimed at the features of multi-value chromosomes in genetic algorithms, statistic schema genes is confirmed according to the statistics of chromosomal genes in population. Then expected schema genes are obtained by testing statistic schema genes. Through spreading it in population, the orientation of searching is changed and the research capability is improved. The simulation results for multiple choice knapsack problem show that in multi-value coding GA with this method statistic schema can be changed into the schema of approximate optimal solution and the global searching performance of population can be improved effectively.

关 键 词：遗传算法 多值编码 模式基因提取 背包问题 

分 类 号：TP301[自动化与计算机技术—计算机系统结构]