一种改进的粒子群人工鱼群算法

An improved particle swarm optimization-artificial fish swarm algorithm

针对人工鱼群算法在高维度求解中收敛速度慢且寻优结果有待于提高等问题,提出一种改进的粒子群人工鱼群算法(PSO-AFSA)。该算法基于精英策略改进,结合了人工鱼群算法(AFSA)的快速跳出局部极值的能力和粒子群算法(PSO)局部快速收敛的优点。此外,该算法将粒子的飞行速度和惯性权重属性与鱼群算法相结合。采用Matlab验证了PSO-AFSA算法在高维度寻优中比AFSA具有更快的收敛速度和更好的寻优结果。

The basic artificial fish swarm algorithm (AFSA) has a slow convergence rate and poor precision in the high-dimensional optimization.An algorithm called the particle swarm optimization-artificial fish swarm algorithm (PSO-AFSA) is proposed in the present paper.The PSO-AFSA is based on the elite strategy.It synthesizes the advantages of the convergent performance and the quick jump out of the local minima of the AFSA and the informational strategy and the quick local convergent performance of the particle swarm optimization algorithm (PSO).Furthermore,this algorithm inherits the velocity and inertia weight characteristics from the PSO.Through simulation,the PSO-AFSA has a faster convergence speed and a better precision than the AFSA,and has a stable performance in the high-dimensional optimization.