摘要
该文针对基本遗传算法(SGA)所存在的缺陷——早熟现象进行了分析,并在此基础上提出了基于种群多样度的变参数遗传算法(VPGA)。该算法从概率角度分析了遗传操作算子的作用,搜索范围以及多样性的影响,依据种群的多样度对遗传算法的参数进行自动调节,抑制早熟现象。并应用两种遗传算法对评价遗传算法性能的四个著名测试函数进行了仿真测试,仿真结果表明该算法相对于基本遗传算法的优越性和抑制早熟现象的有效性。
Aiming at the premature convergence of the simple genetic algorithm （SGA）, variance parameter genetic algorithm （VPGA） , which is based on the diversity of population, is proposed. The function of genetic operators,the range of search and the effect of diversity are analyzed. The parameters of VPGA are kept adjusting according to population diversity in order to restrain the premature convergence. In the light of the evaluating indicator of GA, these two algorithms are tested by using four different test functions and their corresponding fitness functions. The simulation results show the advantage of this method and the efficiency of premature convergence restraint.
出处
《计算机仿真》
CSCD
2006年第1期96-99,179,共5页
Computer Simulation
关键词
遗传算法
种群多样度
早熟现象
Genetic algorithm
Diversity of population
Premature convergence
