基于佳点集交叉的粒子群算法  

Particle Swarm Algorithm Based on Good Point Set Crossover

作　　者：王为为[1] 程家兴[1] 贺晟[1] 

机构地区：[1]安徽大学计算智能与信号处理教育部重点实验室,安徽合肥230039

出　　处：《计算机技术与发展》2009年第12期32-35,共4页Computer Technology and Development

基　　金：国家自然科学基金(60273043);安徽大学研究生创新基金(20073053)

摘　　要：针对基本粒子群算法容易陷入局部极值、后期收敛速度慢和精度低等缺点,文中基于Lovbjerg,Rasmuwsen和Krink提出将进化算法中的交叉操作引入PSO的HPSO模型的启示和佳点集的优良特性,提出了基于佳点集交叉的粒子群算法,从试验上证明了将佳点集交叉算法用于粒子群算法的优化问题上无论从精度还是速度上都比其他算法好。算法整体上表现简单易行的特点,尽管相对基本粒子群算法增加了交叉概率这个参数,但这也给试验以更大的实施空间,是一种有效可行的方法。The basic particle swarm （bPSO） has some demerits, such as relapsing into local extreme, such convergence velocity and low convergence precision. Based on spiration from that Rasmuwsen and Krink in 2000 proposed that crossover operation in evolutionary, algorithms （genetic algorithm） can be used in the PSO＇s HPSO model and excellent characteristics of good point set, good point erossover PSO is proposed. Simulation result shows result＇s precision and velocity that apply good point crossover to optimization problems of the is better than other algorithms. This new algorithm has superiority in simplicity, despite of the algorithm adding relatively the crossover probability, hut it also gives implementation more space and is an effective method.

关 键 词：粒子群算法 佳点集 交叉算法 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]