结合聚类模型和自适应模型的遗传算法

在进化后期,自适应遗传算法有助于保存种群中的优秀模式;但在进化初期,对适应度值大的个体的保护,易降低种群的多样性、减弱算法的搜索性能。基于聚类的遗传算法可以提高遗传算法的收敛速度和搜索性能,但交叉概率和变异概率取定值,易使优秀模式在进化后期遭到破坏,难以收敛到全局最优。在遗传算法中同时引入聚类模型和自适应模型,有利于继承两类改进型遗传算法的优点,克服各自的不足。使用经典的测试函数对引入聚类模型和自适应模型的遗传算法进行测试,仿真结果表明:同时引入聚类模型和自适应模型的遗传算法比引入聚类模型或自适应模型的遗传算法具有更好的收敛速度和寻优能力。

Genetic Algorithm combined with Clustering Model and Adaptive Model

In the later stage of evolution, Adaptive genetic algorithm contributes to preserve excellent patterns in the population, but in the early stages of evolution, the protection of individuals with large fitness values may reduce the diversity of the population and weaken the search performance of the algorithm. The genetic algorithm based on clustering can improve the convergence speed and search performance of the genetic algorithm, but the cross probability and mutation probability are easy to be destroyed in the late evolution, and it is difficult to converge to the global optimal. Introducing the clustering model and adaptive model into genetic algorithm is beneficial to inheriting the advantages of the two improved genetic algorithms and overcome their shortcomings. The classical test function is used to test the genetic algorithm which introduces the clustering model and the adaptive model. The simulation results show that the genetic algorithm which introduces the clustering model and the adaptive model has better convergence speed and optimization ability than the genetic algorithm which introduces the clustering model or the adaptive model.