Abstract: Based on the Artificial Fish Swarm Algorithm（AFSA）, this paper proposes the improved AFSA, that the preying, following, swarming behavior is improved, and the vision of artificial fish is dynamically adjusted. The algorithm optimizes function value to conduct simulation studies with different parameters matching. The experiments analyze the algorithm optimization performance under the influence of the main parameters, and get that the appropriate ranges of parameter setting can solve the problem of the low optimization performance and the slow convergence speed of the AFSA. Finally, the improved AFSA is verified based on the different functions to have some advantages such as higher precision of solution, faster execution speed and higher stability.

Key words: Artificial Fish Swarm Algorithm（AFSA）, parameter value, optimization performance

摘要： 基于原始人工鱼群算法，进行觅食、追尾、聚群行为的改进，以及可视域的自适应调整，提出了改进的人工鱼群算法。算法采用不同的参数值进行匹配，以优化函数值为例进行仿真实验。实验分析研究了主要参数对该算法优化性能的影响，并得出了合理的参数取值，以解决人工鱼群算法寻优精度低、运行速度慢的问题；实验还通过不同函数验证了改进的人工鱼算法具有更高的求解精度、更快的执行速度、更高的稳定性等优点。

关键词: 人工鱼算法, 参数值, 优化性能