变步长和拥挤度因子的自适应人工鱼群算法

为了改进传统的人工鱼群算法会随着迭代的深入而导致算法易陷入局部最优的问题,以及固定的参数导致算法收敛慢和求解精度不高的问题,提出了一种改进的人工鱼群算法.首先结合迭代次数,为移动步长引入一个权值; 然后以每条人工鱼的视野范围所构成的子群为小生境,结合子群最优解与当前人工鱼状态,为拥挤度因子引入一个变异策略.数值实验结果表明,本文提出的算法收敛速度快、精度高、鲁棒性强,优于传统的人工鱼群算法和文献4]提出的算法.

Self-adaptive improved artificial fish swarm algorithm with changing step and crowding factor

In order to improve the traditional artificial fish swarm algorithm, which is easy to fall into local optimum, and the fixed parameters lead to slow convergence and low precision of the algorithm, during the deepening of the iteration, an improved artificial fish swarm algorithm is proposed. Firstly, by combining the number of iterations, a weight value is introduced for the moving step. Then, taking the subgroup of each artificial fish as the niche, combining the optimal solution of the subgroup with the current state of artificial fish, a variation strategy was introduced for crowding factor. The results of numerical experiments show that the proposed algorithm has the advantages of fast convergence, high accuracy and strong robustness, and is superior to the traditional artificial fish swarm algorithm and the algorithm proposed in literature 4].