Abstract: A hybrid particle swarm Genetic Algorithm(GA) is presented to solve the classification question. Based on the relation between the individuals, the algorithm regulates the optimization with the “race” method and controls the individuals in a micro way with race crossover, meanwhile commixed the crossover operator based on the thought of Particle Swarm Optimization(PSO) in GA. With these operators, the speed of convergence and the diversity of the population are well balanced. According to the classification question’s characteristic, it designs the corresponding encoding method, the fitness function, and uses sowing seeds way to produce initial population to get better classification precision. Through classifying the international data sets and comparing with other algorithms classified effect, experimental results show the effectiveness of this algorithm.

Key words: Genetic Algorithm(GA), Particle Swarm Optimization(PSO), race crossover, classification, fitness function

摘要： 提出一种混合粒子群遗传分类算法，根据种群中个体的相互关系，采用“家族”思想对算法进行综合调控，利用家族交叉操作进行微调，并在各家族中引入粒子群思想的交叉算子，兼顾收敛速度和多样性2项指标。根据分类问题的特点，设计相应的编码方式和适应度函数，用播种的方式生成初始种群。对国际通用检验分类效果的数据集进行分类。实验结果证明，该算法的分类效果优于其他 算法。

关键词: 遗传算法, 粒子群优化, 族间交叉, 分类, 适应度函数

CLC Number: 

• TP311