Analysis of computer experiments using penalized likelihood in Gaussian kriging models

被引:93
|
作者
Li, RZ [1 ]
Sudjianto, A
机构
[1] Penn State Univ, Dept Stat, University Pk, PA 16802 USA
[2] Bank Amer, Risk Management Qual & Prod, Charlotte, NC 28255 USA
基金
美国国家科学基金会;
关键词
computer experiment; Fisher scoring algorithm; kriging; meta-model; penalized likelihood; smoothly clipped absolute deviation;
D O I
10.1198/004017004000000671
中图分类号
O21 [概率论与数理统计]; C8 [统计学];
学科分类号
020208 ; 070103 ; 0714 ;
摘要
Kriging is a popular analysis approach for computer experiments for the purpose of creating a cheap-to-compute "meta-model" as a surrogate to a computationally expensive engineering simulation model. The maximum likelihood approach is used to estimate the parameters in the kriging model. However, the likelihood function near the optimum may be flat in some situations, which leads to maximum likelihood estimates for the parameters in the covariance matrix that have very large variance. To overcome this difficulty, a penalized likelihood approach is proposed for the kriging model. Both theoretical analysis and empirical experience using real world data suggest that the proposed method is particularly important in the context of a computationally intensive simulation model where the number of simulation runs must be kept small because collection of a large sample set is prohibitive. The proposed approach is applied to the reduction of piston slap, an unwanted engine noise due to piston secondary motion. Issues related to practical implementation of the proposed approach are discussed.
引用
收藏
页码:111 / 120
页数:10
相关论文
共 50 条
  • [1] Bayesian Metamodeling for Computer Experiments Using the Gaussian Kriging Models
    Deng, Haisong
    Shao, Wenze
    Ma, Yizhong
    Wei, Zhuihui
    [J]. QUALITY AND RELIABILITY ENGINEERING INTERNATIONAL, 2012, 28 (04) : 455 - 466
  • [2] PENALIZED BLIND KRIGING IN COMPUTER EXPERIMENTS
    Hung, Ying
    [J]. STATISTICA SINICA, 2011, 21 (03) : 1171 - 1190
  • [3] Penalized composite likelihood for colored graphical Gaussian models
    Li, Qiong
    Sun, Xiaoying
    Wang, Nanwei
    Gao, Xin
    [J]. STATISTICAL ANALYSIS AND DATA MINING, 2021, 14 (04) : 366 - 378
  • [4] Local influence analysis for penalized Gaussian likelihood estimators in partially linear models
    Zhu, ZY
    He, XM
    Fung, WK
    [J]. SCANDINAVIAN JOURNAL OF STATISTICS, 2003, 30 (04) : 767 - 780
  • [5] Robust Kriging models in computer experiments
    Park, Taejin
    Yum, Bongjin
    Hung, Ying
    Jeong, Young-Seon
    Jeong, Myong K.
    [J]. JOURNAL OF THE OPERATIONAL RESEARCH SOCIETY, 2016, 67 (04) : 644 - 653
  • [6] Extended Gaussian Kriging for computer experiments in engineering design
    Shao, Wenze
    Deng, Haisong
    Ma, Yizhong
    Wei, Zhuihui
    [J]. ENGINEERING WITH COMPUTERS, 2012, 28 (02) : 161 - 178
  • [7] Extended Gaussian Kriging for computer experiments in engineering design
    Wenze Shao
    Haisong Deng
    Yizhong Ma
    Zhuihui Wei
    [J]. Engineering with Computers, 2012, 28 : 161 - 178
  • [8] Penalized maximum likelihood estimation for Gaussian hidden Markov models
    Alexandrovich, Grigory
    [J]. COMMUNICATIONS IN STATISTICS-THEORY AND METHODS, 2016, 45 (20) : 6133 - 6148
  • [9] A note on the choice and the estimation of Kriging models for the analysis of deterministic computer experiments
    Ginsbourger, David
    Dupuy, Delphine
    Badea, Anca
    Carraro, Laurent
    Roustant, Olivier
    [J]. APPLIED STOCHASTIC MODELS IN BUSINESS AND INDUSTRY, 2009, 25 (02) : 115 - 131
  • [10] A penalized blind likelihood Kriging method for surrogate modeling
    Yi Zhang
    Wen Yao
    Xiaoqian Chen
    Siyu Ye
    [J]. Structural and Multidisciplinary Optimization, 2020, 61 : 457 - 474