摘要
针对标准粒子群算法在优化过程中受初始值影响较大且容易陷入局部极值的缺陷,将鱼群算法中聚群行为的基本思想引入粒子群算法中,据此建立了粒子中心的基本概念,并利用粒子的聚群特性调整粒子的飞行方向与目标位置,从而提出了一种新的混合粒子群算法,旨在改进原粒子群算法的全局收敛能力。为了检验混合粒子群算法的优化特性,采用三种典型的标准函数对五种现行智能算法进行了多方面的测试和比较。实验结果表明,新算法具有良好的搜索精度与速度,有效弥补了标准粒子群算法局部收敛和鱼群算法精度不高的双重缺陷,适用于解决复杂函数优化问题。
To overcome the drawbacks of sub-optimization and instability involved in standard PSO algorithm, this paper proposed a new hybrid PSO algorithm based on the swarm behavior of artificial fish. To show the searching performances of the new optimization algorithm, illustrated a series of comparing tests for demonstration on the basis of three typical-standard functions. The results of the experiments indicate that the searching precisions and speeds of the new hybrid PSO algorithm are much better than ones obtained from the other four current PSO and artificial fish algorithms.
出处
《计算机应用研究》
CSCD
北大核心
2009年第5期1700-1702,1705,共4页
Application Research of Computers
关键词
粒子群算法
鱼群算法
聚群行为
混合算法
particle swarm optimization
artificial fish search algorithm
swarm behavior
hybrid algorithm
