计算机工程与应用 ›› 2007, Vol. 43 ›› Issue (22): 53-55.

• 学术探讨 • 上一篇    下一篇

一种多目标微粒群算法及其收敛性分析

王俊年1,2,刘建勋2,陈湘州3   

  1. 1.湖南科技大学 信息与电气工程学院,湖南 湘潭 411201
    2.湖南科技大学 知识网格实验室,湖南 湘潭 411201
    3.湖南科技大学 商学院,湖南 湘潭 411201
  • 收稿日期:1900-01-01 修回日期:1900-01-01 出版日期:2007-08-01 发布日期:2007-08-01
  • 通讯作者: 王俊年

Multi-objective particle swarm optimization and it’s convergence analysis

WANG Jun-nian1,2,LIU Jian-xun2,CHEN Xiang-zhou3   

  1. 1.School of Information and Electrical Engineering,Hunan University of Science and Technology,Xiangtan,Hunan 411201,China
    2.Knowledge Grid Laboratory,Hunan University of Science and Technology,Xiangtan,Hunan 411201,China
    3.Economics School,Hunan University of Science and Technology,Xiangtan,Hunan 411201,China
  • Received:1900-01-01 Revised:1900-01-01 Online:2007-08-01 Published:2007-08-01
  • Contact: WANG Jun-nian

摘要: 在分析多目标优化问题的基础上,提出一种随机多目标微粒群算法,该算法采用在已经获得的Pareto解集中随机选取的两个Pareto解作为微粒更新公式中的pbest和gbest微粒,从而使微粒群的多样性增加,获得均匀分布的Pareto前沿。之后利用有限齐次马尔科夫理论给出了SMOPSO算法的收敛性进行了分析,证明SMOPSO算法以概率1收敛于极小元。最后通过对两个常用多目标函数的仿真实验,验证了算法的有效性。

关键词: 多目标优化, 微粒群算法, Pareto占优, 马尔科夫链, 收敛性

Abstract: A stochastic multi-objective particle swarm optimization is proposed based on analyzing the multi-objective optimization problem.In this algorithm,the varieties of swarm are increased and obtained homogeneous Pareto front by selecting pbest and gbest of particle updating formula randomly from Pareto set repository.Then the convergence of the algorithm is analyzed by using finite markov chain theory,and it is proved that the SMOPSO algorithm converged to the minimizer by probability of 1.

Key words: multi-objective optimization, particle swarm optimization, Pareto dominance, Markav chain, convergence