Abstract
A new approach to metamodeling is introduced whereby a sequential technique is used to construct and simultaneously update mutually dependent metamodels for multiresponse, high-fidelity deterministic simulations. Unlike conventional approaches which produce a single metamodel for each scalar response independently, the present method uses the correlation among different simulation responses in the construction of the metamodel. These dependent metamodels are solved as a system of equations to estimate all individual responses simultaneously. Since several responses contribute to the construction of each individual metamodel, more information from the computed responses is used, thus improving the accuracy of the obtained metamodels. Examples are used to explore the relative performance of the proposed approach and show that the new approach outperforms conventional metamodeling approaches in terms of approximation accuracy. The new method should be particularly useful in problems that require very computationally intensive simulations.
This is a preview of subscription content, log in via an institution to check access.
Similar content being viewed by others
References
Bedewi ZA, Kan A, Marzougui D (1996) Validation of a non-linear finite element vehicle model using multiple impact data. In: ASME winter annual congress and exposition, Atlanta, GA
Booker AJ, Dennis JE Jr, Frank PD, Serafini DB, Torczon V, Trosset MW (1999) A rigorous framework for optimization of expensive functions by surrogates. Struct Optim 17(1):1–13
Chiao CH, Hamada MS (2001) Analysis of experiments with correlated multiple responses. J Qual Technol 33:451–465
Chung H, Alonso JJ (2002) Using gradients to construct cokriging approximation models for high-dimensional design optimization problems. In: Proceedings of the 40th AIAA aerospace sciences meeting and exhibit, Reno, NV (CD-ROM)
Currin C, Mitchell M, Morris M, Ylvisaker D (1988) A Bayesian approach to the design and analysis of computer experiments. Technical report ORNL-6498, Oak Ridge National Laboratory
Currin C, Mitchell M, Morris M, Ylvisaker D (1991) Bayesian prediction of deterministic functions, with applications to the design and analysis of computer experiments. J Am Stat Assoc 86:953–963
Del Castillo ED, Montgomery DC, McCarthy DR (1996) Modified desirability function for multiple response optimization. J Qual Technol 28(3):337–345
Derringer G, Suich R (1980) Simultaneous optimization of several response variables. J Qual Technol 12(4):214–219
Farhang-Mehr A, Azarm S (2002) A sequential information theoretic approach to design of computer experiments. In: Proceedings of the 9th AIAA/ISSMO symposium on multidisciplinary analysis and optimization, Atlanta, GA (CD-ROM)
Farhang-Mehr A, Azarm S (2005) Bayesian meta-modeling of engineering design simulations: a sequential approach with adaptation to irregularities in the response behavior. Int J Numer Methods Eng 62:2104–2126
Farhang-Mehr A, Azarm S, Diaz A, Ravisekar A (2003) Bayesian approximation-assisted optimization applied to crashworthiness design of a pickup truck. In: Proceedings of the ASME 2003 design engineering technical conferences, Chicago, IL (CD-ROM)
Jin R, Chen W, Sudjianto A (2002) On sequential sampling for global metamodeling in engineering design. In: Proceedings of IDETC ASME design automation conference, Montreal, Canada (CD-ROM)
Koehler JR, Owen AB (1996) Computer experiments. Handbook of statistics. Elsevier, New York
Kodiyalam S, Sobieszczanski-Sobieski J (2001) Multidisciplinary design optimization—some formal methods, framework requirements, and application to vehicle design. Int J Veh Des (Spec Issue):3–22
Kohavi R (1995) A study of cross-validation and bootstrap for accuracy estimation and model selection. In: Proceedings of the IJCAI'95. Morgan Kaufmann, San Mateo, CA, pp 1137–1143
Khuri AI (1996) Multiresponse surface methodology. Handbook of statistics. Elsevier, Amsterdam
Lindley DV (1956) On a measure of the information provided by an experiment. Ann Math Stat 27:986–1005
Lucas JM (1994) How to achieve a robust process using response surface methodology. J Qual Technol 26(4):248–260
Myers RH, Kim Y, Griffiths K (1997) Response surface methods with the use of noise variables. J Qual Technol 29:429–440
Roux WJ, Stander N, Haftka RT (1998) Response surface approximations for structural optimization. Int J Numer Methods Eng 42:517–534
Sacks J, Welch WJ, Mitchell TJ, Wynn HP (1989) Design and analysis of computer experiments. Stat Sci 4(4):409–435
Shannon CE (1948) A mathematical theory of communication. Bell Syst Tech J 27:379–423, 623–656
Shewry MC, Wynn HP (1987) Maximum entropy sampling. J Appl Stat 14:165–170
Simpson TW, Booker AJ, Ghosh D, Giunta AA, Koch PN, Yang R-J (2002) Approximation methods in multidisciplinary analysis and optimization: a panel discussion. In: 3rd ISSMO/AIAA Internet conference on approximations in optimization
Ver Hoef JM, Cressie N (1993) Multivariable spatial prediction. Math Geol 5(2):219–240
Vining GG, Myers RH (1990) Combining taguchi and response surface philosophies: a dual response approach. J Qual Technol 22:38–45
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Li, G., Azarm, S., Farhang-Mehr, A. et al. Approximation of multiresponse deterministic engineering simulations: a dependent metamodeling approach. Struct Multidisc Optim 31, 260–269 (2006). https://doi.org/10.1007/s00158-005-0574-5
Received:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00158-005-0574-5