基于智能单粒子算法与支持向量机的稳健优化设计  

Robust Design Optimization based on Intelligent Single Particle Optimizer and Support Vector Regression

作　　者：刘茂福[1] 任丕顺[1] 

机构地区：[1]湖南机电职业技术学院机械工程系,湖南长沙410000

出　　处：《装备制造技术》2011年第7期35-38,共4页Equipment Manufacturing Technology

基　　金：湖南省科技计划项目(2010FJ3054)

摘　　要：为了求解混合离散变量优化设计问题,通过引入粒子位置矢量的离散化处理方法,对智能单粒子优化算法(ISPO)进行改进,用MATLAB语言设计了求解混合离散变量优化设计问题的ISPO算法程序,结合支持向量回归机的应用,研究了平面尺寸链离散公差的模糊稳健优化设计问题,给出了稳健设计实例。设计实例表明了本文所提方法的有效性和实用性。In order to solve the optimization problem with discrete variables,a method for the discretization of position vector of particle is adopted and the intelligent single particle optimizer（ISPO） is improved.A program,named as LS-ISPO written by MATLAB,is designed for solving optimization problem with discrete variables.The support vector regression（SVR） is used,and then a fuzzy robust problem is discussed for finding the optimal discrete tolerance of a dimension chain.An example of robust design is given and it shows that the proposed method is effective in engineering application.

关 键 词：智能单粒子算法 混合离散变量 支持向量机 稳健优化设计 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]