计算机工程与应用 ›› 2007, Vol. 43 ›› Issue (30): 47-49.

• 学术探讨 • 上一篇    下一篇

多阶段参数动态控制微粒群优化算法

曾 渊1,宋 涛1,王少波2,许家栋1   

  1. 1.西北工业大学 电子信息学院,西安 710072
    2.中国人民解放军,新疆 喀什 844000
  • 收稿日期:1900-01-01 修回日期:1900-01-01 出版日期:2007-10-21 发布日期:2007-10-21
  • 通讯作者: 曾 渊

Multiple phases coefficient dynamic control strategy in particle swarm optimization

ZENG Yuan1,SONG Tao1,WANG Shao-bo2,XU Jia-dong1   

  1. 1.School of Electronics and Information,Northwestern Polytechnical University,Xi’an 710072,China
    2.The Chinese People’s Liberation Army,Kashi,Xinjiang 844000,China
  • Received:1900-01-01 Revised:1900-01-01 Online:2007-10-21 Published:2007-10-21
  • Contact: ZENG Yuan

摘要: 为了平衡微粒群算法中全局搜索和局部开发之间的关系,多阶段参数动态控制机制被引入了标准的微粒群算法。在多阶段参数动态控制微粒群优化算法(MDPSO)中,微粒群的搜索过程在逻辑上被划分为三个阶段,每一个阶段都有各自的优化目标,对应着每一个搜索阶段,代表微粒个体经验、种群经验、全局经验和种群排斥力、全局排斥力的5个加速常数将会按照不同的规律变化,控制种群经验和全局经验对微粒的吸引与种群重心和全局重心对微粒的排斥,可以很好地避免在优化过程初期容易出现的早熟收敛现象和在优化过程末期容易出现的收敛放慢现象。通过对标准函数的测试,验证了该方法有效性和可靠性。

Abstract: In order to control the balance of global exploration and local exploitation efficiently,the concept of the Multiple phases coefficient Dynamic control strategy Particle Swarm Optimization(MDPSO) is introduced.In MDPSO algorithm,the search process is divided into three phases in logical,which has specific aim respectively.In each phase,five acceleration coefficients,which represent individual experience,swarm experinece,globe experience,swarm repulsion and globe repulsion,changes with different rules.The major consideration of this modification is to control the balance of attraction and repulsion effectively,which relates the particles and swarm and globe,and to avoid premature convergence in the early process and enhance convergence to the global optimum solution at the end of process.Four well known benchmark functions are used as testing functions for the MDPSO algorithm and the conclusion of testing is presented.