基于支持向量机的多目标模糊稳健优化设计  

Multi-Objective Fuzzy-Robust Design Optimization Based on Support Vector Regression

作　　者：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9期67-70,共4页Equipment Manufacturing Technology

基　　金：湖南省科技计划项目(2010FJ3054)

摘　　要：为了提高模糊稳健优化设计的计算效率,探讨了基于支持向量机回归机(SVR)的多目标模糊稳健设计方法,该方法以SVR作为非线性约束函数的替代模型,并采用SVR对模糊概率进行仿真计算,可显著降低模糊稳健优化设计的机时消耗;采用字典序优先级的目标规划法,建立了多目标稳健优化设计模型;把SVR与遗传算法相结合,构建了一种混合智能优化算法;通过多目标稳健设计实例,对所提出的方法进行了验证。In order to improve the computational efficiency of fuzzy-robust design optimization,a fuzzy-robust design method using support vector regression（SVR） is proposed.A support vector regression is used as a surrogate model of the stochastic simulation method for computing the fuzzy probability,and the computational cost of fuzzy-robust design optimization can be distinctly cut down.Lexicographic goal programming approach is used to model the problem of multi-objective robust design optimization.With genetic algorithm combined with SVR,an intelligent optimization algorithm is developed for effectively solving the established model.To illustrate the proposed method,an example is given,and it shows that the proposed method is effective.

关 键 词：多目标优化 模糊稳健设计 支持向量机 优化设计 

分 类 号：TH122[机械工程—机械设计及理论]