人工鱼群算法收敛速度改进优化仿真

Artificial Fish Swarm Algorithm Convergence Speed Improvement Optimization Simulation

在基本人工鱼群算法更新过程中,人工鱼群通过觅食算子进化时,进化方向和步长都有一定的随机性,虽然有助于鱼群跳出局部最优,但是严重影响鱼群进化效率,增加算法运算量。针对上述问题,将最速下降法嵌入到基本人工鱼群算法中,得到改进的人工鱼群算法。当人工鱼利用聚群算子和追尾算子更新后,如果没有得到改善,此时利用最速下降法对人工鱼进行更新。保留聚群算子和追尾算子中的觅食算子,保证算法良好的全局搜索能力,同时嵌入具有较好局部搜索能力的最速下降法,增强人工鱼个体的局部寻优能力,加快人工鱼群算法收敛速度。数值仿真结果表明,所得改进人工鱼群算法在计算量减少的情况下,具有更快的收敛速度,同时收敛精度也得到一定提升。

ABSTRACT：In the process of basic artificial fish swarm algorithm updating, evolutionary direction and step size of artificial fish swarm are of a certain randomness when they are evolved by foraging operator. Atthough it would help the fish out of the local optimal, it seriously affects the evolution efficiency of fish populations and increases the com- putational complexity. To solve this problem, the steepest descent method is embedded into the basic artificial fish swarm algorithm, and an improved artificial fish swarm algorithm is proposed. When the artificial fish is updated with the clustering operator and the trailing operator, if there is no improvement, the artificial fish is updated with steepest descent method. The steepest descent method is embedded with better local search capability while preserving the for- aging operators in the clustering operator and the trailing operator to ensure the global searching ability of the algo- rithm and enhance the local searching ability of the artificial fish. The convergence rate of artificial fish swarm algo- rithm is accelerated. The results of numerical experiments show that the improved artificial fish swarm algorithm has faster convergence rate and lower convergence rate when the computational complexity is reduced.