摘要
速度上限是影响微粒群算法性能的一个主要参数。针对现有调节策略存在参数统一设置、与微粒性能无关等缺点,本文提出一种基于性能反馈的调整策略,使得速度上限能随着个体性能的改变而动态调整,从而更加真实有效的模拟了鸟类觅食的群体行为特征。仿真结果表明该算法能较好地提高微粒群算法的计算效率。
Velocity threshold is an important parameter in particle swarm optimization. There are some disadvantages for previous proposed strategies such as the same values for the entire swarm, no relationship with particle’s performance. Therefore, this paper proposes a adjustment strategy based on performance feedback that the velocity threshold is "particle-dependent". This new method provides a deep insight for the birds seeking behavior model. Simulation results show the new proposed algorithm improves the performance greatly.
出处
《心智与计算》
2007年第2期256-263,共8页
Mind and Computation
基金
国家自然科学基金(No:60674104)资助
关键词
微粒群算法
速度上限
性能反馈
particle swarm optimization
velocity threshold
performance feedback
