Surrogate Based Evaluation of the Design of Experiments

Eva Myšáková; Matěj Lepš

doi:10.4028/www.scientific.net/AMR.1144.148

Paper Titles

Microstructure-Based Evolution of Compressive Strength of Blended Mortars: A Continuum Micromechanics Approach
p.121

Parameter Study on Subset Simulation for Reliability Assessment
p.128

Bayesian Updating of Aleatory Uncertainties in Heterogeneous Materials
p.136

Quasicontinuum Method Combined with Anisotropic Microplane Model
p.142

Surrogate Based Evaluation of the Design of Experiments
p.148

The Response of a SDOF System Driven by a Periodic Random Force
p.153

Adaptive Boundaries of Wang Tiles for Heterogeneous Material Modelling
p.159

Galerkin-Eshelby Meshless Approach to Multiple Inclusion Problem
p.167

Comparison of Semidefinite Solvers for Topology Optimization of Cantilever Trusses Subject to Fundamental Eigenvalue Constraint
p.172

HomeAdvanced Materials ResearchAdvanced Materials Research Vol. 1144Surrogate Based Evaluation of the Design of...

Surrogate Based Evaluation of the Design of Experiments

Abstract:

Meta-modeling also known as Surrogate Modeling is one of the commonly used tools foranalysis of complex systems' behavior. The meta-model is constructed based on training data whichconsist of the training points generated via Design of Experiments (DoE) and responses of the originalmodel evaluated in these training points. The positioning of the points is crucial for the approximationquality of the meta-model. Therefore it is appropriate to assess the DoE's quality not only usingthe common geometrical or statistical criteria but also from the point of view of its actual particularpurpose. Such testing is able to recognize the appropriate set of training points and also the possibleability of the individual meta-models for actual problem.

You might also be interested in these eBooks

Modern Methods of Experimental and Computational Investigations in Area of Construction II

View Preview

Info:

Periodical:

Advanced Materials Research (Volume 1144)

Pages:

148-152

DOI:

https://doi.org/10.4028/www.scientific.net/AMR.1144.148

Citation:

Cite this paper

Online since:

March 2017

Authors:

Eva Myšáková*, Matěj Lepš

Keywords:

Design of Experiments, Latin Hypercube Sampling, Meta-Modeling, Optimization

Export:

RIS, BibTeX

Price:

Permissions CCC:

Request Permissions

Permissions PLS:

Request Permissions

Сopyright:

Citation:

* - Corresponding Author

References

[1] E. Janouchová, A. Kučerová, Competitive comparison of optimal designs of experiments for sampling-based sensitivity analysis, Computers & Structures 124 (2013) 47-60.

DOI: 10.1016/j.compstruc.2013.04.009

Google Scholar

[2] R.T. Johnson, B. Jones, J.W. Fowler, D.C. Montgomery, Comparing designs for computer simulation experiments, In proceedings of Winter Simulation Conference (2008) 463-470.

DOI: 10.1109/wsc.2008.4736101

Google Scholar

[3] W. Schwarz, S. Schwub, K. Quering, D. Wiedmann, H.W. Höppel, M. Göken, Life prediction of thermally highly loaded components: modelling the damage process of a rocket combustion chamber hot wall, CEAS Space Journal 1 (2011) 83-97.

DOI: 10.1007/s12567-011-0007-9

Google Scholar

[4] R.L. Iman, W.J. Conover, Small sample sensitivity analysis techniques for computer models, with an application to risk assessment, Communications in Statistics: Theory and Methods A9 (1980) 1749-1842.

DOI: 10.1080/03610928008827996

Google Scholar

[5] B. Iooss, L. Boussouf, V. Feuillard, A. Marrel, Numerical studies of the metamodel fitting and validation processes, ArXiv e-prints (2010).

Google Scholar

[6] E. Myšáková, M. Lepš, Comparison of Maximin LHS Design Strategies, In proceedings of the 3rd Conference Nano & Macro Mechanics (2012) 163-170.

Google Scholar