区域分割的自适应变异粒子群算法

Regional-Segmentation Self-Adapting Variation Particle Swarm Optimization

为了提高粒子群算法(PSO)的收敛性及多样性,提出一种基于区域分割的自适应变异粒子群算法(RSVPSO).算法采用区域分割的思想,利用粒子间信息交叉,使粒子搜索区间快速缩小;同时在迭代后期与自适应变异策略相结合,提高粒子跳出局部最优陷阱的能力和增强粒子多样性,达到寻优的目的.将所提出的算法应用于8个测试函数,并与精英免疫克隆选择的协同进化粒子群等算法进行比较,结果表明,新算法在收敛速度、搜索精度及寻优效率等方面有较大提高.

To improve convergence and diversity of particle swarm optimization（ PSO）,an improved PSO which called regional-segmentation self-adapting variation particle swarm optimization（ RSVPSO） algorithm is introduced. Regional-segmentation is adopted in the algorithm,using information cross between particles,narrowsearch region quickly; combining with self-adapting variation strategy in late iterations at the same time,improved capacity of jumping out local optimum trap and enhanced the diversity of particles,reach the goal of optimization. The proposed algorithm is applied to eight test functions and compared with the elite immune clonal selection co-evolutionary particle swarm optimization and so on. The results showthat the proposed algorithm has considerable improvement in the convergence speed,search accuracy,optimum efficiency and so on.