摘要
针对标准粒子群算法易出现早熟的问题,提出了一种带邻近粒子信息的粒子群算法。该算法中粒子位置的更新不仅包括自身最优和种群最优,还包括粒子目前位置最近粒子最优的信息。为了有效地平衡算法的全局探索和局部开发,并使其收敛于全局最优值,采用了时变加速因子策略,两个加速因子随进化代数线性变化。通过对5个经典测试函数优化的数值仿真实验并与其他粒子群算法的比较,结果表明了在平均最优值和成功率上都有所提高,特别是对多峰函数效果更加明显。
To overcome premature searching by standard Panicle Swarm Optimization(PSO) algorithm,a new modified PSO with information of the closest particle is proposed.In the algorithm,the particle is updated not only by the best previous position and the best position among all the panicles in the swarm,but also by the best previous position of the closest particle.To balance the trade-off between exploration and exploitation and convergence to the global optimum solution,a linearly varying acceleration coefficient over the generations is introduced.The simulation results show that the algorithm has better probability of finding global optimum and mean best value than others algorithm,especially for multimodal function.
出处
《计算机工程与应用》
CSCD
北大核心
2009年第18期40-42,共3页
Computer Engineering and Applications
基金
湖北省高等学校优秀中青年创新团队计划项目(No.T200803)
关键词
粒子群算法
优化
种群多样性
panicle swarm optimization
optimization
population diversity