基于高斯加权的GeesePSO改进算法

Improved Geese Swarm Optimization Algorithm Based on Gaussian Weighted Sum

为了提高粒子群算法的优化性能,通过观察和分析雁群结队飞行的智能群体现象,国内学者提出了基于雁群启示的粒子群优化算法(GeesePSO,GPSO)。该算法虽然在一定程度上提高了PSO算法的性能,但是在GPSO算法中存在着不合理的加权平均机制,即最小值寻优方面的加权缺陷。针对该问题,本文通过采用高斯加权方法对GPSO进行合理改进,提出一种基于高斯加权改进的粒子群优化算法(Gaussian-Weighted GPSO,GWGPSO)。实验结果表明:新算法在收敛精度、收敛速度和鲁棒性等指标上得到了提高,从而证明高斯加权方式是合理的和正确的。

In order to improve the optimization performance for PSO, through observation and analysis of the natural phenomenon of formation flight of geese, researchers at home proposed the geese swarm optimization algorithm（GPSO）. Although this algorithm has improved performance for PSO in some extent, the mechanism of average-weighted for GPSO is unreasonable, namely the defect of minimum optimization. For this issue, this paper proposed the geese swarm optimization algorithm based on gaussian-weighted（GWGPso） through reasonable improvements of GPSO. The experimental results show that the new algorithm has improved these indicators, such as convergence precision, convergence rate and robustness, which proves that the gaussian-weighted method is reasonable and correct.