×

zbMATH — the first resource for mathematics

Comparison of designs for computer experiments. (English) Zbl 1078.62085
Summary: For many complex processes laboratory experimentation is too expensive or too time-consuming to be carried out. A practical alternative is to simulate these phenomena by a computer code. This article considers the choice of an experimental design for computer experiments. We illustrate some drawbacks to criteria that have been proposed and suggest an alternative, based on the Bayesian interpretation of the alias matrix by N. R. Draper and I. Guttman [Ann. Inst. Stat. Math. 44, 659–671 (1992; Zbl 0772.62041)]. Then we compare different design criteria by studying how they rate a variety of candidate designs for computer experiments such as Latin hypercube plans, U-designs, lattice designs and rotation designs.

MSC:
62K99 Design of statistical experiments
62P30 Applications of statistics in engineering and industry; control charts
62J05 Linear regression; mixed models
68U20 Simulation (MSC2010)
PDF BibTeX XML Cite
Full Text: DOI
References:
[1] Bates, R.A.; Buck, R.J.; Riccomagno, E.; Wynn, H.P., Experimental design and observation for large systems, J. roy. statist. soc. B, 58, 77-94, (1996), Discussion 95-111 · Zbl 0850.62627
[2] Beattie, S.D., Lin, D.K.J., 1997. Rotated factorial design for computer experiments. Proceedings of SPES, American Statistical Association.
[3] Box, G.E.P.; Draper, N.R., A basis for the selection of a response surface design, J. amer. statist. assoc., 54, 622-654, (1959) · Zbl 0116.36804
[4] Box, G.E.P.; Draper, N.R., The choice of a second order rotatable design, Biometrika, 50, 335-352, (1963) · Zbl 0129.32103
[5] Bursztyn, D.; Steinberg, D.M., Rotation designs for experiments in high bias situations, J. statist. plann. inference, 97, 399-414, (2001) · Zbl 1015.62081
[6] Bursztyn, D.; Steinberg, D.M., Rotation designsorthogonal first-order designs with higher order projectivity, J. appl. stochastic models bus. ind., 18, 197-206, (2002) · Zbl 1010.62071
[7] Currin, C., Mitchell, T., Morris, M., Ylvisaker, D., 1988. A Bayesian approach to the design and analysis of computer experiments. ORNL-6498, available from National Technical Information Service.
[8] Currin, C.; Mitchell, T.; Morris, M.; Ylvisaker, D., Bayesian prediction of deterministic functions, with applications to the design and analysis of computer experiments, J. amer. statist. assoc., 86, 953-963, (1991)
[9] Draper, N.R.; Guttman, I., Treating bias as variance for experimental design purposes, Ann. inst. statist. math., 44, 659-671, (1992) · Zbl 0772.62041
[10] Fang, K.; Wang, Y.; Bentler, P.M., Some applications of number-theoretic methods in statistics, Statist. sci., 9, 416-428, (1994) · Zbl 0955.62620
[11] Fang, K.T.; Lin, D.K.J.; Winker, P.; Zhang, Y., Uniform designtheory and application, Technometrics, 42, 237-248, (2000)
[12] Hedayat, A.S.; Sloane, N.A.S.; Stufken, J., Orthogonal arrays, (2000), Springer Berlin
[13] Johnson, M.E.; Moore, I.M.; Ylvisaker, D., Minimax and maximin distance designs, J. statist. plann. inference, 26, 131-148, (1990)
[14] Kennedy, M.C.; O’Hagan, A., Bayesian calibration of computer models, J. roy. statist. soc. B, 63, 425-450, (2001), Discussion 450-464 · Zbl 1007.62021
[15] Koehler, J.R.; Owen, A.B., Computer experiments, (), 261-308 · Zbl 0919.62089
[16] Korobov, N.M., The approximate computation of multiple integrals, Dokl. akad. nauk SSSR, 124, 1207-1210, (1959) · Zbl 0089.04201
[17] Lindley, D.V., On a measure of the information provided by an experiment, Ann. math. statist., 27, 986-1005, (1956) · Zbl 0073.14103
[18] McKay, M.D.; Beckman, R.J.; Conover, W.J., A comparison of three methods for selecting values of input variables in the analysis of output from a computer code, Technometrics, 21, 239-245, (1979) · Zbl 0415.62011
[19] Mitchell, T.J., Computer construction of D-optimal first-order designs, Technometrics, 16, 211-220, (1974) · Zbl 0299.62040
[20] Morris, M.D.; Mitchell, T.J.; Ylvisaker, D., Bayesian design and analysis of computer experimentsuse of derivatives in surface prediction, Technometrics, 35, 243-255, (1993) · Zbl 0785.62025
[21] Owen, A.B., Orthogonal arrays for computer experiments, integration and visualization, Statist. sinica, 2, 439-452, (1992) · Zbl 0822.62064
[22] Robinson, G.K., That BLUP is a good thingthe estimation of random effects, Statist. sci., 6, 15-51, (1991) · Zbl 0955.62500
[23] Sacks, J.; Ylvisaker, D., Designs for regression problem with correlated errors, Ann. math. statist., 37, 66-89, (1966) · Zbl 0152.17503
[24] Sacks, J.; Ylvisaker, D., Designs for regression problem with correlated errors; many parameters, Ann. math. statist., 39, 49-69, (1968) · Zbl 0165.21505
[25] Sacks, J.; Ylvisaker, D., Designs for regression problem with correlated errors III, Ann. math. statist., 41, 2057-2074, (1970) · Zbl 0234.62025
[26] Sacks, J.; Welch, W.J.; Mitchell, T.J.; Wynn, H.P., Design and analysis of computer experiments, Statist. sci., 4, 409-435, (1989) · Zbl 0955.62619
[27] Shewry, M.C.; Wynn, H.P., Maximum entropy sampling, J. appl. statist., 14, 165-170, (1987)
[28] Steinberg, D.M., Model robust response surface designsscaling two-level factorials, Biometrika, 72, 513-526, (1985) · Zbl 0605.62086
[29] Steinberg, D.M., A Bayesian approach to flexible modeling of multivariable response functions, J. multivariate anal., 34, 157-172, (1990) · Zbl 0711.62027
[30] Steinberg, D.M., Bursztyn, D., 2004. Some data analytic tools for understanding random field regression models. Technometrics.
[31] Tang, B., Orthogonal array-based Latin hypercubes, J. amer. statist. assoc., 88, 1392-1397, (1993) · Zbl 0792.62066
[32] Welch, W.J.; Buck, R.J.; Sacks, J.; Wynn, H.P.; Mitchell, T.J.; Morris, M.D., Screening, predicting, and computer experiments, Technometrics, 34, 15-25, (1992)
This reference list is based on information provided by the publisher or from digital mathematics libraries. Its items are heuristically matched to zbMATH identifiers and may contain data conversion errors. It attempts to reflect the references listed in the original paper as accurately as possible without claiming the completeness or perfect precision of the matching.