摘要
针对粒子群优化算法(PSO)容易陷入局部极值、进化后期收敛速度慢和精度低等缺点,提出了一种改进的简化粒子群优化算法(YSPSO)。该算法采用黄金分割法平衡惯性与经验之间的相互影响;同时,为避免错过全局最优值,增加反向随机惯性权重,使粒子在一定程度上具有反向搜索的能力。最后,对几个经典基准测试函数进行实验,结果表明,YSPSO算法在提高算法收敛速度和精度的同时,降低了陷入局部极值的可能性,提高了PSO算法的实用性。
Aiming at some demerits of particle swarm optimization algorithm(PSO), such as relapsing into local extremum easily, slow convergence velocity and low convergence precision in the late evolutionary, an improved simple particle swarm optimization algorithm(YSPSO) was proposed. It employs golden section method to balance the mutual influence between inertia and experience. Meanwhile, in order to avoid missing the global optimal value, it adds reverse random inertia weights to make the particles have the ability to search reversely in a certain extent. Finally, the experiment results of several classic benchmark functions show that YSPSO improves the practicability of PSO via improving convergence velocity and precision, and reducing the possibility of relapsing into local extremum.
出处
《计算机科学》
CSCD
北大核心
2015年第B11期86-88,共3页
Computer Science
基金
黑龙江省教育厅科学技术研究项目(12511601)资助
关键词
群体智能
粒子群优化
黄金分割法
Swarm intelligence, Particle swarm optimization, Golden section method