导弹多学科设计优化中响应面方法研究  

Research of RSM in Missile Multidisciplinary Design Optimization

作　　者：兰文博[1] 罗小云[1] 郑安波[1] 

机构地区：[1]中国航天空气动力技术研究院,北京100074

出　　处：《计算技术与自动化》2009年第3期111-115,共5页Computing Technology and Automation

摘　　要：使用响应面方法取代高精度工程模拟可以减少设计周期和成本。目前,有多种响应面模型应用到导弹多学科设计优化中,每种响应面模型有不同优点和不足。本文的目的是通过比较六种响应面模型来帮助设计者选择合适的响应面模型。七种测试函数用来比较响应面模型的三个方面:近似精度、鲁棒性和应用难易度。结果表明,Kriging响应面和增强径向基函数响应面对线性响应、二次响应和高阶非线性响应都有很好的近似效果,而二次多项式响应面和移动最小二乘响应面适合于线性和二次响应,径向基函数响应面适合于高阶非线性响应。神经网络响应面在使用更多的采样点时得到更精确的模型。Using response surface methodology in place of high fidelity engineering simulations can help reduce design cycle times and cost. Today, Many different response surface models have been applied to Multidisciplinary Design Optimization of missile, each with different capabilities and pitfalls. The goal of this research is to aid the designer in selecting the appropriate response surface model by comparing six popular modeling techniques. Seven test problems are used to compare model accuracy, robustness, and ease of use. It was found that Kriging and augmented Radial Basis Functions are generally appropriate for various responses including linear, quadratic, and higher - order nonlinear responses. Quadratic Polynomial and Moving Least Squares are appropriate for linear, quadratic responses. Radial Basis Functions are appropriate for higher - order nonlinear responses. Neural Network can produce more accurate model when larger sampling points were used.

关 键 词：导弹 多学科设计优化 响应面方法 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术]