摘要
针对基本人工鱼群算法(AFSA)收敛精度不高的问题,提出在文化算法框架下,使粒子群(PSO)与基本人工鱼群相结合的改进人工合鱼群算法,利用文化算法框架,组成基于AFSA的群体空间和基于PSO的信仰空间。为了提高算法的运算速度,省略了拥挤度因子的计算,在觅食行为中,让人工鱼直接移动到搜索到的较好位置。为了增加粒子的多样性,在知识空间中引入变异操作。在人工鱼群算法的随机行为中添加随机数并动态调整算法中的最大试探次数以平衡整个算法的全局和局部搜索能力,提高算法的收敛精度。仿真分别通过固定进化代数比较精度,和指定收敛精度比较进化代数两个角度说明了改进方法具有更高的收敛精度和更快的计算速度,证明该算法具有较高的优化性能。
A new cuhural algorithm based on the hybrid of particle swarm optimization （PSO） and artificial fish swarm algorithm（AFSA） was proposed to overcome the low convergent precision of the basic AFSA. This algorithm model consists of a PSO based knowledge space and an AFSA based main population space. The congestion factor was omitted and the artificial fish was made to move directly to the superior position in the purpose of enhancing the computational speed of the algorithm. A special mutation operation was added into the knowledge space to improve the particle diversity. A random number was set in the random behaviour and the parameter ＂try number＂ in the al gorithm was dynamically adjusted so as to effectively balance the ability of searching the global and local extremum and also the stability and the convergent accuracy of the algorithm are greatly raised. The tests of benchmark func tions show that the algorithm has Rood optimization performance.
出处
《计算机仿真》
CSCD
北大核心
2014年第4期407-411,共5页
Computer Simulation
关键词
人工鱼群算法
文化算法
粒子群算法
全局最优
Artificial fish swarm algorithm
Cuhural algorithm
Particle swarm algorithm
Global optimum
