基于拓扑结构与粒子变异改进的粒子群优化算法

为使粒子群优化算法(PSO)优化过程的多样性与收敛性得到合理解决,以提高算法优化性能,基于种群拓扑结构与粒子变异提出两种粒子群改进算法RSMPSO和RVMPSO.改进算法将具有信息定向流动的闭环拓扑结构与星型拓扑结构或四边形拓扑结构相结合,促使粒子在前期寻优过程中具有较高的多样性,保证搜索的广度,而在后期满足粒子群的整体收敛性,保证寻优的精度.同时,将布谷鸟搜索算法(CS)中的偏好随机游走变异策略引入改进算法中,增强粒子跳出局部最优的能力.对标准测试函数的仿真实验表明,所改进的PSO算法与其他6个对比算法相比不仅操作简单,优化精度高,而且在算法收敛性及稳健性方面都有着更出色的表现.

Modified particle swarm optimization algorithms based on topology and particle mutation

Two kinds of modified particle swarm optimization(PSO) algorithms called particle swarm optimization with ring-star topology and particle mutation(RSMPSO) and particle swarm optimization with ring-von neumann topology and particle mutation(RVMPSO) are presented based on topology and mutation of particle swarm, which enable the diversity and convergence of the particle swarm to be preserved in optimizing procedure to improve the optimization performance of the PSO algorithm. In the modified PSO algorithms, the ring topology with information directionally flowing is combined with star topology or square topology. This strategy enables the particle to have more diversity during the early-stage optimizing to guarantee the search extent. And in the later period, this strategy enables the particle swarm to be converged to make sure the accuracy of optimizing. Meanwhile, a mutation strategy with random walk in a biased way in cuckoo search(CS) algorithm is introduced to the modified PSO algorithm to improve the ability of particle of escaping the local optimum. The simulation experiment on the standard test function shows that, compared with other six algorithms, the modified PSO algorithm possesses the features of easy operation and high optimizing accuracy, and a better performance in diversity and robustness.