期刊文献+

一种基于混合策略的自适应多输出高斯过程响应面法

An adaptive multi-output Gaussian process response surface method based on mixed strategies
下载PDF
导出
摘要 针对工程结构数值模拟常常是矢量与函数(矢量函数)混合输出的特点,提出一种基于混合策略的含索引变量的矢量高斯过程响应面法,同时提出一种设计点自适应加密方法进一步提高效率。算例显示:本文方法能够优先拾取预测误差最大的点作为加密点;真实误差全部都落在最大可能误差范围之内,最大误差1.4%~1.6%远小于预设误差限5%,验证了本文方法的有效性。 In structural numerical simulations,many models give both vector type and function type of outputs,i.e.vector function type of outputs.For this reason,a new response surface method based on mixed strategies is presented,which is referred to as vector Gaussian process response surface method with index variables.To obtain higher efficiency,an adaptive process of thickening design points is also suggested together.An example is provided to show the validity of the new method.
出处 《应用力学学报》 CAS CSCD 北大核心 2011年第3期237-242,324-325,共6页 Chinese Journal of Applied Mechanics
基金 国家自然科学基金委员会-中国工程物理研究院联合基金(10876100)
关键词 混合策略 多输出模型 响应面法 高斯过程 贝叶斯统计 mixed dtrategies multi-output model tesponse surface method Gaussian process Bayesian statistics.
  • 相关文献

参考文献11

  • 1Oaldey J, O'Hagan A. Bayesian inference for the uncertainty distribution of compeer model outtmls[J]. Biometrika, 2002, 89(4): 769-784.
  • 2Oakley J, O'Hagan A. Probabilistic sensitivity analysis of complex model: a Bayesian approach[J]. Journal of the Royal Statistical Society B, 2004, 66(3): 751-769.
  • 3Kennedy M C, O'Hagan A. Bayesian calibration of computer model[J]. Journal ofthe Royal Stafislical Society B, 2001, 63(3):425-464.
  • 4Bayarri M J, Berger J O, Paulo R, et al. A framework for validation ofcomputermodels[J]. Technometrics, 2007;49(2): 138-154.
  • 5刘信恩,肖世富,莫军.复杂数值模拟的贝叶斯模型确认框架及其简化[C]//中国力学学会学术大会’2009,CCTAM2009-003789,郑州,2009.
  • 6刘信恩,肖世富,莫军.用于不确定性分析的高斯过程响应面模型设计点选择方法[J].计算机辅助工程,2011,20(1):101-105. 被引量:4
  • 7Higdon D, Gattiker J, Williams B, et al. Computer model calibration using high dimensional output[R]//LA-UR-05-6410. Los Alamos National Laboratory, 2005.
  • 8Bayarri M J, Berger J O, Cafeo J, et al. Computer modal validation with ftmctional output[R]. National Institute of Statisficai Sciences Technical Report, 2006.
  • 9Rougier J. Efficient emulators for multivariate deterministic fuctions[J]. Joumal of Computational and Graghical Statistics, 2008, 17(4): 827-843.
  • 10Conti S, O'Hagan A. Bayesian emulation of complex multi-output and dynamic computer models[J]. Journal of Statistical Planning and Inference, 2010, 140(3): 640-651.

二级参考文献7

  • 1OAKLEY J E, O' HAGAN A. Probabilistic sensitivity analysis of complex model: a Bayesian approach [ J ]. J R Stat Soc : Ser B : Stat Methodology, 2004, 66(3) : 751-769.
  • 2OAKLEY J E, O' HAGAN A. Bayesian inference for the uncertainty distribution of computer model outputs [ J ]. Biometrika, 2002, 89 (4) : 769-784.
  • 3KENNEDY M C, O' HAGAN A. Bayesian calibration of computer model[J]. J R Stat Soc: Ser B: Stat Methodology, 2001, 63(3) : 425-464.
  • 4BAYARRI M J, BERGER J O, PAULO R, et al. A framework for validation of computer models[ R]. North Carolina: Natl Inst Stat Sci Teeh Rep, 2005.
  • 5刘信恩,肖世富,莫军.复杂数值模拟的贝叶斯模型确认框架及其简化[C]//中国力学学会学术大会’2009,CCTAM2009-003789,郑州,2009.
  • 6MORRIS M D, MITCHELL T J. Exploratory designs for computational experiments [ J ]. J Stat Planning & Inference, 1995, 43 (3) : 381-402.
  • 7刘信恩,肖世富,莫军.高斯过程响应面法研究[J].应用力学学报,2010,27(1):190-195. 被引量:6

共引文献4

相关作者

内容加载中请稍等...

相关机构

内容加载中请稍等...

相关主题

内容加载中请稍等...

浏览历史

内容加载中请稍等...
;
使用帮助 返回顶部