一种新颖的改进人工鱼群算法  

New Improved Artificial Fish Swarm Algorithm

作　　者：刘东林[1] 李乐乐[1] 

机构地区：[1]华东理工大学信息科学与工程学院,上海200237

出　　处：《计算机科学》2017年第4期281-287,共7页Computer Science

摘　　要：针对基本人工鱼群算法(AFSA)在函数优化问题中存在的后期收敛速度慢、求解精度低和易陷入局部最优等缺点,提出了一种新的改进人工鱼群算法(IAFSA)。首先,使用混沌变换来初始化鱼群个体的位置,使鱼群更加均匀地分布在有限的区域内,保证种群具有多样性,利于全局收敛;其次,对觅食行为中具有不同函数值的人工鱼个体采取不同的视野策略,不仅提高了算法的寻优速度,而且有效地降低了鱼群陷入局部最优的可能性;最后,根据运动和体能之间的关系构建体能变换模型,在鱼群觅食的后期,体能开始变弱,这时适时地减小鱼群觅食、聚群和追尾行为中移动的步长可有效提高算法收敛的速度和寻优的精度。通过标准测试函数和14个城市的TSP对算法进行验证,仿真实验结果表明,相比基本人工鱼群算法,改进后的算法具有更快的后期收敛速度和更高的求解精度。Aiming at the problems of easy to fall into the local optimum value,converging slowly in the later period and low solving accuracy that artificial fish swarm algorithm（AFSA）have,a new improved artificial fish algorithm（IAFSA）was proposed.Firstly,the new algorithm uses chaos transform to initialize the position of individual fish,and the fish is more evenly distributed in the specified area within the region,that keeps the fish population diversity,and it is conducive to global convergence.Secondly,the artificial fish with different function values in foraging behavior take different visual,and it not only improves the searching speed but also reduces the possibility of the artificial fish falling into local optimum.Finally,according to the relationship between physical and the activity,aphysical transformation model is construct,and the physical of artificial fish become weaken in the late stage of the algorithm,reducing the step timely is very important,which can improve the convergence speed and the accuracy of the algorithm.The algorithm is verified by standard test function and TSP of 14 cities,and the experimental results show that the improved algorithm has faster convergence speed and higher precision than the basic artificial fish swarm algorithm.

关 键 词：人工鱼群算法 混沌变换 觅食行为 体能变换模型 

分 类 号：TP181[自动化与计算机技术—控制理论与控制工程]