计算机工程与应用 ›› 2007, Vol. 43 ›› Issue (11): 100-103.

• 产品、研发、测试 • 上一篇    下一篇

微粒群算法与郭涛算法在数值优化中的比较

贺毅朝 张翠军 王培崇 张巍   

  1. 石家庄经济学院 信息工程学院 北京科技大学信息工程学院
  • 收稿日期:2006-04-17 修回日期:1900-01-01 出版日期:2007-04-11 发布日期:2007-04-11
  • 通讯作者: 贺毅朝

Particle Swarm Optimization Compare with Guo Tao Algorithm on Function Optimization Problems

YiChao He   

  • Received:2006-04-17 Revised:1900-01-01 Online:2007-04-11 Published:2007-04-11
  • Contact: YiChao He

摘要: 粒子群优化(PSO)算法是由Kennedy和Eberhart于1995年提出的一种演化算法,其基本思想源于对自然界中生物群体觅食行为的仿真研究,在求解连续域优化问题时表现出较好的性能。郭涛算法(GuoA)是1999年由郭涛提出的一种基于子空间搜索和群体爬山法相结合的演化算法,在求解数值优化问题时取得了非常好的效果。文献[3,4]指出:PSO算法和GuoA算法均优于遗传算法,但这两个算法的优劣当前还没有比较结果。本文对于9个典型的复杂BenchMark测试函数,分别利用PSO算法和GuoA算法进行数值计算比较,大量的实验结果表明:GuoA算法更具有通用性,在全局收敛性方面更优,但是速度相对较慢;PSO算法的收敛速度很快,对于某些极难问题具有优越性,但成功率较低,且很容易早熟。

关键词: 粒子群优化算法, 郭涛算法, BenchMark函数

Abstract: Particle Swarm Optimization(PSO) is an Evolutionary Algorithm which is proposed by Kennedy and Eberhart in 1995. It is motivated by the social behavior of organisms,such as bird flocking, and it exhibits effective character for solving global optimization problems over continuous space. Guo-Tao Algorithm(GuoA) is an Evolutionary Algorithm which is proposed by Guo Tao in 1999. It is based on combining subspace searching with swarm climbing hill, and gain great effect for function optimization problems over continuous space too. In this paper, for 9 typical and complex BenchMark testing function, using the PSO and GuoA to calculate these problems. The result show that GuoA is more all-purpose and superiority on global convergence, but the velocity of convergence is more lower. The velocity of PSO is faster, and it can solve certain difficult optimization problem. But the success rate of PSO is more lower, and that PSO often occurs premature convergence is a great drawback.

Key words: Particle Swarm Optimization, GuoTao Algorithm, BenchMark Function