基于重构变异算子遗传算法的研究

摘要

针对遗传算法存在早熟和局部搜索能力差的缺点,提出重构变异算子遗传算法( Reconstruction Mutation Operator Genetic algorithm,RMOGa). 该算法由速成算子和自适应算子组成. 首先,通过速成算子来平衡变异算子和交叉算子在遗传算法中的地位,以此来改善遗传算法中的早熟现象;其次,采用自适应算子来保留遗传过程中适应度大的个体,从而增强局部搜索能力;最后,引入"路由判断冶的方法来加快算法的收敛速度. 实验过程使用MaTLaB 7. 0仿真软件,选取四组典型的测试函数,采用基本遗传算法( Simple Genetic algorithms,SGa)、双变异率遗传算法( Double Mutation Genetic algo-rithm,DMGa)以及文中提出的基于重构变异算子遗传算法(RMOGa),分别对测试函数在收敛精度上进行对比. 结果表明,RMOGa算法能很好地解决遗传算法早熟与陷入局部最优解的问题.%For the problem of genetic algorithm in premature and poor local search capability,a Reconstruction Mutation Operator Genetic Algorithm ( RMOGA) is proposed. It consists of two parts:crash operator and adaptive operator. Firstly,use the crash operator to balance the proportion of mutation operator and cross operator in genetic algorithm,which can improve the property of premature in the genetic al-gorithm. Secondly,suitable elements are selected through the adaptive operator,which can enhance the local search capability of the algo-rithm. Finally,a routing method is introduced into the algorithm,which can be used to accelerate the convergence speed. The software of MATLAB 7. 0 is used in the simulation,all the algorithms of the Simple Genetic Algorithms (SGA),the Double Mutation Genetic Algo-rithm ( DMGA) ,RMOGA,have been adopted individually to test four groups of typical function and compare the effect of convergence. Experimental results show that the RMOGA algorithm presented in this paper is efficient for the problem of genetic algorithm in premature and local optimization.