基于Kriging模型的改进协同优化算法  

Improved Collaborative Optimization Algorithm

作　　者：张静[1,2] 李柏林[3] 张卫华[1] 刘永均[3] 

机构地区：[1]西南交通大学牵引动力国家重点实验室,四川成都610031 [2]成都理工大学核技术与自动化工程学院,四川成都610059 [3]西南交通大学机械工程学院,四川成都610031

出　　处：《西南交通大学学报》2010年第4期539-543,共5页Journal of Southwest Jiaotong University

基　　金：四川省科技支撑计划项目(2008GZ0149)

摘　　要：为了提高协同优化算法的求解效率,利用Kriging模型,构造系统级近似优化模型,提出了基于Kriging模型的改进协同优化算法.该算法采用置信域与均匀设计相结合的方法,完成近似模型的更新;采用序列二次规划算法,完成优化问题的求解.以经典函数和减速器设计为例,验证了改进协同优化算法.结果表明:该算法能提高计算效率,在减速器设计中,迭代次数减少50%左右.In order to improve the computational efficiency of the conventional collaborative optimization（CO）,an improved collaborative optimization algorithm based on the Kriging model（Kriging-CO for short） was proposed.In this algorithm,the approximate optimization model at the system level is constructed with the Kriging model,and is updated by uniform design combined with confidence regions.The Kriging-CO was verified through the optimization of a classical function and the design of a speed reducer,and the optimization models were solved using the sequential quadratic programming algorithm.Numerical results show that the Kriging-CO can improve the computational efficiency.For the complex speed reducer design,the Kriging-CO,compared with the CO,reduced the number of iterations in the optimization computation by about 50%.

关 键 词：多学科设计优化 协同优化 KRIGING模型 

分 类 号：TH123.1[机械工程—机械设计及理论]