中国机械工程 ›› 2015, Vol. 26 ›› Issue (16): 2179-2182.

• 智能制造 • 上一篇    下一篇

结合模糊聚类和合作竞争博弈的优化方法

陈加明1;宋小文2;徐浩2   

  1. 1.杭州职业技术学院,杭州,310000
    2.浙江大学,杭州,310027
  • 出版日期:2015-08-25 发布日期:2015-08-25

Optimization Method Based  on  Fuzzy Clustering and Cooperative Competition Game

Chen Jiaming1;Song Xiaowen2;Xu Hao2   

  1. 1.Hangzhou  Vocationaland  Technical  College,Hangzhou,310000
    2.Zhejiang University,Hangzhou,310027
  • Online:2015-08-25 Published:2015-08-25

摘要:

针对多目标优化问题转化为博弈问题的难点,即博弈问题的各博弈方战略集的确定问题,利用均匀设计试验法、回归分析和F检验为模糊聚类提供数据基础,将多目标优化设计与博弈分析结合起来,提出了一种多目标优化方法。以减速器优化设计为例,对该方法的性能进行了分析与验证,并与传统单目标优化方法进行对比分析。结果表明,该方法收敛速度快,计算效率高,无需人为确定各目标权重,具有工程应用价值。

关键词: 模糊聚类, 合作竞争博弈, 均匀设计, 回归分析

Abstract:

Aiming at the difficulty of transforming multi-objective optimization problem to game problem  which  was about determining the strategy set  for each partition of game players,a  multi-objective optimization method  was  proposed  by using uniform design  experimentation, regression analysis and F test statistic to obtain data bases for fuzzy clustering and combining cooperative competition game.Taking reducer design optimization as an example,performance of the method was analyzed and verified. Compared with the traditional single-objective  optimization  method,the experimental results show  that the  multi-objective method has fast convergence rate and high computational efficiency,and it is  unneccessary  to determine the weight of each target,which has engineering application values.

Key words: fuzzy clustering, cooperative competition game, uniform design, regression analysis

中图分类号: