粒子群优化算法中惯性权重的研究

Research on Inertia Weight in Particle Swarm Optimization

惯性权重是粒子群优化算法中的关键参数,文章对惯性参数进行了系统的研究,在此基础上,分析了固定权重,典型的线性递减惯性权重,步长较小的线性递减惯性权重对收敛性能的影响。通过对4个测试函数的仿真实验,验证了它们各自的全局收敛性和收敛速度,说明了惯性权重在粒子群优化算法中有很大的自由度。

The inertia weight is the crucial parameter of the particle swarm optimization （PSO）. Systemic research is done on inertia weight in the paper, based on this, we analysis the impact of performance of the convergence based on regular weight, typical linear reduced inertia weight arid linear reduced inertia weight with small length of stride. The results on four benchmark functions proved that their own global convergence and convergence speed. It illustrates that the selection of inertia weight has more freedom in particle swarm optimization.