基于种群多样性模糊控制的粒子群算法

关于优化粒子群算法问题,针对标准粒子群算法前期收敛速度过快,后期容易陷入局部最优解的问题,提出一种种群多样性模糊控制的粒子群算法。为了控制种群多样性的变化,提高算法跳出局部最优解的性能,在算法中加入模糊控制器和位置跳变策略,通过控制参数的变化来控制粒子的速度、位置和种群多样性的变化,使算法从全局探测平稳过渡到局部开采。仿真结果表明,改进算法能有效避免陷入局部最优解,且对高维函数优化时效果更为明显,是一种高效的优化算法。

Particle Swarm Optimization Based on Fuzzy Control of Population Diversity

To solve the problem that particle swarm optimization has a fast convergence rate in early stage,and is easy to fall into local optimal solution in late stage,a particle swarm optimization based on fuzzy control of population diversity was presented.In order to control the population diversity and improve the algorithm’s ability to jump out of local optimal solution,the fuzzy controller and location hopping strategy were added to the algorithm.By changing the parameters,we can control particle’s velocity,position and the population diversity,then make the algorithm to transit from global exploration to local exploitation smoothly.Simulation results show that the algorithm can effectively avoid falling into local optimum,and achieve better results compared to the BPSO.The effect is more obvious in high-dimensional function optimization.