Abstract: Aiming at the problem of Artificial Fish Swarm Algorithm(AFSA), such as long running time or being in local extremum, caused by inappropriate parameters setting, this paper uses the improved Global edition Artificial Fish Swarm Algorithm(GAFSA), optimizes three benchmark functions to conduct simulation studies with different parameters matching, and analyzes the change law of the optimization performances and the convergence speed of the artificial fish swarm algorithm under the influence of the main parameters, and gives the appropriate ranges of parameter setting. Experimental results show that reasonable parameter settings make algorithm get better optimization performances and quickly convergence to the global better solution.

Key words: Artificial Fish Swarm Algorithm(AFSA), swarm intelligence, parameter analysis, optimization performances

摘要： 针对人工鱼群算法由于参数选择不合理而导致算法运行时间长或陷入局部最优的问题，利用改进的全局版人工鱼群算法，采用不同参数匹配，以优化3个典型的测试函数为例进行仿真实验研究，分析人工鱼群算法在主要参数影响下，算法优化性能及收敛速度的变化规律，给出算法参数设置的适当取值。实验结果表明参数的合理设置使算法可以较快地收敛至全局较优解，并具有较好的性能。

关键词: 人工鱼群算法, 群体智能, 参数分析, 优化性能

CLC Number: 

• TP18