基于共轭梯度法改进的人工鱼群算法

Hybrid AFSA with conjugate gradient methods

针对基本人工鱼群算法运算精度低和效率差的缺点,将共轭梯度法引入基本人工鱼群算法中,得到改进的人工鱼群算法。算法对每条人工鱼分别进行聚群算子和追尾算子,若更新结果没有得到改善,则利用共轭梯度法进行更新。在人工鱼群更新过程中引入共轭梯度法,减少随机性,增强人工鱼个体的局部寻优能力,确保人工鱼每次更新都会得到改善,从而加快人工鱼群算法收敛速度。数值实验结果表明,所得改进人工鱼群算法具有更快的收敛速度,同时收敛精度也得到一定提升。

The basic artificial fish swarm algorithm has the shortcomings of low precision and low efficiency. Aiming at this problem,this paper introduced the conjugate gradient method in the artificial fish swarm algorithm,and obtained the improved artificial fish swarm algorithm. The proposed algorithm performed clustering and trailing operators on each artificial fish. If the update result was not improved,the algorithm would be updated using the conjugate gradient method. This paper introduced the conjugate gradient method to updating the artificial fish swarm,which could reduce the randomness and enhance the local searching ability of the artificial fish. This ensured that the artificial fish would be improved at the same time,thus speeding up the convergence rate of the artificial fish swarm algorithm. The results of numerical experiments show that the improved artificial fish swarm algorithm has faster convergence speed,and the convergence accuracy is also improved.