一种多值编码遗传算法的缺失基因复现方法  

Emergence of Lacked Allele in Multi-Value Coding Genetic Algorithm

作　　者：马炫[1] 乐天[1] 小仓久和[2] 

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]福井大学知能システム工学科

出　　处：《西安理工大学学报》2005年第3期249-252,共4页Journal of Xi'an University of Technology

基　　金：陕西省教育厅专项科研基金资助项目(05JK269)

摘　　要：与二值编码遗传算法相比,基因缺失问题对多值编码遗传算法的全局搜索性能影响比较大。提出一种缺失基因复现和存活的变异方法,即根据字符集中的字符在种群中的缺失作为预选变异基因集,然后对待变异个体集进行预变异;选择高适应度个体的对应变异基因和基因位,使其在原待变异个体集中扩散。将该方法应用到多重选择背包问题的遗传算法中,通过仿真实验验证了该方法的有效性。Comparing with two value coding genetic algorithm, the problem of allele lack may have a greater influence upon the performance of global searching of multi-value coding genetic algorithm. In this paper,a method of emergence and survival of lacked allele is proposed,that is,selecting the characters missed in current population as lack allele set by comparing the character set with genes of the population,implementing simulated mutation to the individuals selected from the population by mutation rate,choosing the gene just mutated and its locus of the best individual from mutated individuals by fitness,and using it to mutate the genes in the same locus of original individuals. The validity of this method is tested via the simulation expenriment by applying this method to the genetic algorithm for solving the multiple choice knapsack problem.

关 键 词：遗传算法 多值编码 基因缺失 变异 

分 类 号：TP301[自动化与计算机技术—计算机系统结构]