计算机科学与探索 ›› 2018, Vol. 12 ›› Issue (2): 332-340.DOI: 10.3778/j.issn.1673-9418.1610052

• 人工智能与模式识别 • 上一篇    

基于健康度的自适应过滤粒子群算法

袁  罗1,2,葛洪伟1,2+,姜道银1,2   

  1. 1. 轻工过程先进控制教育部重点实验室(江南大学),江苏 无锡 214122
    2. 江南大学 物联网工程学院,江苏 无锡 214122
  • 出版日期:2018-02-01 发布日期:2018-01-31

Partical Swarm Optimization Algorithm with Adaptive Filter Based on Health Degree

YUAN Luo1,2, GE Hongwei1,2+, JIANG Daoyin1,2   

  1. 1. Ministry of Education Key Laboratory of Advanced Process Control for Light Industry (Jiangnan University), Wuxi, Jiangsu 214122, China
    2. School of Internet of Things Engineering, Jiangnan University, Wuxi, Jiangsu 214122, China
  • Online:2018-02-01 Published:2018-01-31

摘要: 针对标准粒子群算法存在收敛速度慢和难以跳出局部最优等问题,提出了基于健康度的自适应过滤粒子群算法。首先,通过对粒子健康度的动态检测,区分粒子状态,处理并标记异常粒子,自适应过滤懒惰粒子位置,避免算法陷入局部最优;其次,利用引导因子更新全局最差粒子值,过滤异常粒子数,避免无效搜索,加快算法收敛速度。通过对11个标准函数进行仿真实验,并与标准粒子群和其他改进算法进行对比,结果表明,基于健康度的自适应过滤粒子群算法寻优精度高,收敛速度快。

关键词: 粒子群算法, 健康度, 自适应过滤, 懒惰粒子, 引导因子, 收敛速度

Abstract: Since the standard particle swarm optimization (PSO) algorithm converges slowly and is easy to fall into local optimization, this paper proposes the PSO algorithm with adaptive filter based on health degree (HAFPSO). Firstly, the algorithm can determine the particle's state by detecting particle health degree and mark ill particles at the same time. In order to avoid invalid search, the algorithm adapts to filter the ill particle's position and update a new position. In order to further improve the algorithm convergence speed and accuracy, the proposed algorithm adopts the guidance factor to update the global worst particle. Compared with standard particle swarm optimization and other optimization algorithms with 11 benchmark functions, the experimental results show that HAFPSO can improve convergence speed and accuracy significantly.

Key words: particle swarm optimization, health degree, adaptive filter, lazy particle, guidance factor, convergence speed