一种基于扰动项的混合粒子群优化算法

为了保持粒子种群的多样性而避免发生"早熟"的问题,本文提出一种基于扰动项混合粒子群优化算法(PSO),该方法通过提高粒子群多样性来提高PSO的收敛性能.首先用标准PSO来迭代,当粒子群失去多样性时,在包含粒子群的超球外随机设置一粒子对全局最优粒子干扰,并在PSO更新公式中加入扰动项来干扰每个粒子.最后将该改进的PSO应用于函数逼近,实验结果验证了本文提出的PSO性能优于几种经典的PSO算法.

An Improved Hybrid Particle Swarm Optimization Algorithm Based on Disturbance

In order to maintain the diversity of the swarm to avoid "premature",an improved hybrid particle swarm optimization(PSO) is proposed by adding disturbance.This method improves the convergence performance by increasing the particle swarm diversity.When the particle swarm losses the diversity,a particle outside the ultra-ball containing the swarm selected randomly is used to disturb the global best particle and introduce a disturbance to the basic PSO formula to update each particle.Finally,apply the improved PSO to function approximation,experimental results verify the proposed PSO outperforms several classical PSO algorithms.