改进学习因子和约束因子的混合粒子群算法  

Hybrid particle swarm optimization algorithm of new learning factors and constraint factor

作　　者：张水平[1] 仲伟彪 

机构地区：[1]江西理工大学信息工程学院,江西赣州341000

出　　处：《计算机应用研究》2015年第12期3626-3628,3653,共4页Application Research of Computers

基　　金：国家自然科学基金资助项目(11062002);江西省教育厅科技项目(GJJ13426);江西省自然(青年)科学基金资助项目(2008GZS0030)

摘　　要：针对传统粒子群优化算法解决复杂问题时收敛速度太快、容易陷入局部最优解的问题,在全局—局部最优解粒子群算法的基础上,提出了一种改进学习因子和约束因子的混合粒子群优化算法。通过将粒子的邻域最优解加入到速度更新公式,使得粒子的速度更新同时受全局最优解和邻域最优解作用,提高了粒子的寻优能力。并改进了学习因子和约束因子,平衡粒子的全局搜索和局部开发能力。通过几个经典测试函数分析比较,该算法能大大提高粒子的寻优能力。In order to avoid traditional particle swarm optimization algorithm converges too fast and easy to fall into local optimal value, on the basis of particle swarm algorithm with global-local optimal value, this paper proposed a hybrid particle swarm algorithm with improved learning factor and constraint factor. It added the neighborhood optimal value to the updating formula of speed, the speed updated by the role of global optimal value and neighborhood optimal value together, and improved the search ability of particles. It improved the learning factor and constraint factor, balance global search capability and local development capacity of particles. Compared several classical test functions, the algorithm can greatly improve the search ability of the particles.

关 键 词：混合粒子群优化算法 全局—局部最优 学习因子 约束因子 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]