In this paper we consider two nested computer codes, with the first code output as one of the second code inputs. A predictor of this nested code is obtained by coupling the Gaussian predictors of the two codes. This predictor is non Gaussian and computing its statistical moments can be cumbersome. Sequential designs aiming at improving the accuracy of the nested predictor are proposed. One of the criteria allows to choose which code to launch by taking into account the computational costs of the two codes. Finally, two adaptations of the non Gaussian predictor are proposed in order to compute the prediction mean and variance rapidly or exactly.
Accepté le :
DOI : 10.1051/ps/2018011
Mots-clés : Nested computer codes, surrogate model, Gaussian process, uncertainty quantification, Bayesian formalism, sequential design, computer experiments
@article{PS_2019__23__245_0,
author = {Marque-Pucheu, Sophie and Perrin, Guillaume and Garnier, Josselin},
title = {Efficient sequential experimental design for surrogate modeling of nested codes},
journal = {ESAIM: Probability and Statistics},
pages = {245--270},
publisher = {EDP-Sciences},
volume = {23},
year = {2019},
doi = {10.1051/ps/2018011},
mrnumber = {3946295},
zbl = {1420.62355},
language = {en},
url = {http://archive.numdam.org/articles/10.1051/ps/2018011/}
}
TY - JOUR AU - Marque-Pucheu, Sophie AU - Perrin, Guillaume AU - Garnier, Josselin TI - Efficient sequential experimental design for surrogate modeling of nested codes JO - ESAIM: Probability and Statistics PY - 2019 SP - 245 EP - 270 VL - 23 PB - EDP-Sciences UR - http://archive.numdam.org/articles/10.1051/ps/2018011/ DO - 10.1051/ps/2018011 LA - en ID - PS_2019__23__245_0 ER -
%0 Journal Article %A Marque-Pucheu, Sophie %A Perrin, Guillaume %A Garnier, Josselin %T Efficient sequential experimental design for surrogate modeling of nested codes %J ESAIM: Probability and Statistics %D 2019 %P 245-270 %V 23 %I EDP-Sciences %U http://archive.numdam.org/articles/10.1051/ps/2018011/ %R 10.1051/ps/2018011 %G en %F PS_2019__23__245_0
Marque-Pucheu, Sophie; Perrin, Guillaume; Garnier, Josselin. Efficient sequential experimental design for surrogate modeling of nested codes. ESAIM: Probability and Statistics, Tome 23 (2019), pp. 245-270. doi : 10.1051/ps/2018011. http://archive.numdam.org/articles/10.1051/ps/2018011/
[1] , Cross validation and maximum likelihood estimation of hyper-parameters of Gaussian processes with model misspecification. Comput. Stat. Data Anal. 66 (2013) 55–69. | DOI | MR | Zbl
[2] , Parametric estimation of covariance function in Gaussian-process based Kriging models. Application to uncertainty 520 quantification for computer experiments. Ph.D. thesis. Université Paris-Diderot – Paris VII (2013).
[3] , The Numerical Treatment of Integral Equations. Clarendon Press, Oxford (1977). | MR | Zbl
[4] , , , and , Sequential design of computer experiments for the estimation of a probability of failure. Stat. Comput. 22 (2012) 773–793. | DOI | MR | Zbl
[5] , and , Objective Bayesian analysis of spatially correlated data. J. Am. Stat. Assoc. 96 (2001) 1361–1374. | DOI | MR | Zbl
[6] , , , and , Efficient global reliability analysis for nonlinear implicit performance functions. AIAA J. 46 (2008) 2459–2468. | DOI
[7] , , and , Fast parallel Kriging-based stepwise uncertainty reduction with application to the identification of an excursion set. Technometrics 56 (2014) 455–465. | DOI | MR
[8] and , Deep Gaussian processes, in Proceedings of the Sixteenth International Workshop on Artificial Intelligence and Statistics (AISTATS), AISTATS ’13, pages 207–215, JMLR W&CP 31, edited by and (2013).
[9] , and , AK-MCS: an active learning reliability method combining Kriging and Monte Carlo simulation. Struct. Saf . 33 (2011) 145–154. | DOI
[10] and , Uniform experimental designs and their applications in industry. Handb. Stat. 22 (2003) 131–178. | DOI | MR
[11] , and , Design and Modeling for Computer Experiments. Computer Science and Data Analysis Series. Chapman Hall, London (2006). | Zbl
[12] , and , Kriging is well-suited to parallelize optimization, in Computational Intelligence in Expensive Optimization Problems. Vol. 2 of Adaptation Learning and Optimization. Springer, Berlin, Heidelberg (2010) 131–162. | DOI
[13] and , Cases for the nugget in modeling computer experiments. Stat. Comput. 22 (2012) 713–722. | DOI | MR | Zbl
[14] and , Gaussian process single-index models as emulators for computer experiments. Technometrics 54 (2012) 30–41. | DOI | MR
[15] , and , Assessment of uncertainty in computer experiments, from Universal to Bayesian Kriging. Appl. Stoch. Model. Bus. Ind. 25 (2009) 99–113. | DOI | MR | Zbl
[16] and , Sequential design for ranking response surfaces. SIAM/ASA J. Uncertain. Quantif . 5 (2017) 212–239. | DOI | MR | Zbl
[17] and , Predicting the output from a complex computer code when fast approximations are avalaible. Biometrika 87 (2000) 1–13. | DOI | MR | Zbl
[18] and , Bayesian calibration of computer models. J. Royal Stat. Soc. Ser. B (Stat. Methodol.) 63 (2001) 425–464. | DOI | MR | Zbl
[19] , Regression and Kriging metamodels with their experimental designs in simulation: a review. Eur. J. Oper. Res. 256 (2017) 1–16. | DOI | MR | Zbl
[20] , Bayesian analysis of hierarchical multifidelity codes. SIAM/ASA J. Uncertain. Quantif . 1 (2013) 244–269. | DOI | MR | Zbl
[21] and , Recursive co-Kriging model for design of computer experiments with multiple levels of fidelity. Int. J. Uncertain. Quantif . 4 (2014) 365–386. | DOI | MR | Zbl
[22] and , Probability, Random Variables and Stochastic Processes. McGraw-Hill, Boston (2002).
[23] , Default priors for Gaussian processes. Ann. Stat. 33 (2005) 556–582. | DOI | MR | Zbl
[24] , , , and , Nonlinear information fusion algorithms for data-efficient multi-fidelity modelling. Proc. R. Soc. Lond. A: Math. Phys. Eng. Sci. 473 (2017) 20160751. | Zbl
[25] , Active learning surrogate models for the conception of systems with multiple failure modes. Reliab. Eng. Syst. Saf . 149 (2016) 130–136. | DOI
[26] and , A repulsion-based method for the definition and the enrichment of optimized space filling designs in constrained input spaces. J. Soc. Française de Stat. 158 (2017) 37–67. | Numdam | MR | Zbl
[27] , , and , Nested polynomial trends for the improvement of Gaussian process-based predictors. J. Comput. Phys. 346 (2017) 389–402. | DOI | MR | Zbl
[28] and , A nonstationary space-time Gaussian process model for partially converged simulations. SIAM/ASA J. Uncertain. Quantif . 1 (2013) 37–67. | DOI | MR | Zbl
[29] and , Gaussian Processes for Machine Learning. The MIT Press, Cambridge (2006). | MR | Zbl
[30] , The Bayesian Choice. Springer-Verlag, New York (2007). | MR | Zbl
[31] , , and , Design and analysis of computer experiments. Stat. Sci. 4 (1989) 409–435. | MR | Zbl
[32] , and , The Design and Analysis of Computer Experiments. Springer Series in Statistics. Springer, New York (2003). | DOI | MR | Zbl
[33] , Interpolation of Spatial Data: Some Theory for Kriging. Springer, New York (1999). | DOI | MR | Zbl
[34] , , , and , Sequential design of experiments to estimate a probability of exceedinga threshold in a multi-fidelity stochastic simulator, in 61th World Statistics Congress of the International Statistical Institute (ISI 2017), Marrakech, Morocco, July 2017 (2017).
[35] , and , Surrogate modeling of computer experiments with different mesh densities. Technometrics 56 (2014) 372–380. | DOI | MR
Cité par Sources :
