基于剪枝策略的骨干粒子群算法

Pruning strategy based bare bones particle swarm optimization

为了优化算法的全局探索能力和局部开发能力,提出一种基于两方面改进的骨干粒子群算法.提出一种进化方程,通过即时搜索域的分析说明该方程可以改善粒子多样性.提出粒子群"剪枝"策略:每当粒子搜索到新的群体最优位置时,剪去该粒子,同时初始化一个新位置以安插该粒子.理论分析指出,在增强全局探索能力的同时,合适的剪枝策略能增加局部开发能力.实验结果表明,所提出算法的性能较几种经典PSO算法有显著的提升.

A bare bones particle swarm optimization（NPSO） algorithm is proposed to improve both global exploration and local exploitation.An evolution equation which obtains better swarm diversity is employed in the NPSO algorithm.Inspired by the apical dominance phenomenon in biology,a particle pruning strategy is introduced as follows：When a particle reaches a new best position of the swarm,it would be pruned and inserted to another position.Theoretical analysis shows that the pruning strategy can improve both global exploration and local exploitation.Finally,results of the experiments on benchmark problems show that the proposed algorithm obtains significant improvement when compared to some classical PSO algorithms.