作　　者： ; ; ; ; ;

机构地区： 西安交通大学

出　　处： 《系统仿真学报》 2003年第8期1066-1068,1073,共4页

摘　　要： 为了在算法稳定性的基础上解决其收敛速度和全局收敛性之间的矛盾,提出了一种新的改进遗传算法。该改进算法设计了与进化代数相关的交叉概率,与个体适应度相关的变异概率,以及与早熟情况、进化代数和个体适应度有关的移民算法。将其应用于电能质量分类的计算结果表明,该改进遗传算法稳定性较好,且在收敛速度和获取全局最优解的概率两个方面都有很大的提高。 A new improved genetic algorithm (IGA) is proposed in this paper. In the IGA, it is designed a new probability algorithm of crossover depending on the number of generations, and a new probability algorithm of mutation depending on the fitness value. Furthermore, an improved immigration method depending on the convergence, evolution generation and fitness values is proposed also. All these methods can help to enhance the capability of the genetic algorithm, and solve the main conflict of the convergence speed with the global astringency. Finally, the IGA combined with LVQ Neural Network has been applied to classify the power quality signals, and the validity of the proposed method is verified by the results of the simulation results.

关 键 词： 遗传算法 交叉概率 自适应变异率 电能质量

领　　域： [自动化与计算机技术] [自动化与计算机技术] [电气工程]