摘要
采用跳跃弹道空地巡航导弹为算例,研究了新型实验设计技术,即拉丁化近正交海默斯利序列取样法,在多学科设计优化中的应用.应用该实验设计方法进行了实验设计,并对取样结果与传统拉丁方取样进行了对比;基于支持向量机构建了学科近似模型,通过对验证数据的回归分析,检验了支持向量机响应的可靠性.应用该近似模型对算例进行多学科设计优化,并通过与精确模型的对比,检验了该近似模型的准确性.拉丁化近正交海默斯利序列取样法是一种高效的现代实验设计方法,在相似的实验条件下比传统拉丁方设计具有更好的正交性和均匀性,在复杂系统多学科优化设计中具有很大的应用价值.
By taking skip trajectory air-to-surface missile as example, the application of modern design of experiment (DOE) technology, namely latinized nearly orthogonal hammersley sequence sampling (LNOHSS) , in multidisciplinary design optimization (MDO) was studied. Design experiments were conducted using LNOHSS technology, sampling data was compared with the data of conventional latin hypercube sampling (LHS). Metamodels of exact disciplinary models were built based on support vector machine (SVM). The reliability of metamodel responses was checked by regression analysis of validate data set. The example system was optimized based on metamodels, the correctness of metamodels was validated by comparing optimal results with exact analysis. LNOHSS is a highly effective DOE technology. It provides better orthogonality and uniformity than conventional LHS. LNOHSS technology is of great application value to the MDO of complex systems.
出处
《北京航空航天大学学报》
EI
CAS
CSCD
北大核心
2008年第10期1121-1125,共5页
Journal of Beijing University of Aeronautics and Astronautics
关键词
实验设计
回归分析
近似方法
优化
design of experiments
regression analysis
approximation methodology
optimization