» Download Article

Authors
Bettinger, Régis; Pronzato, Luc; Duchêne, Pascal

Identifier
10.3182/20090706-3-FR-2004.00216

Index Terms
Grey Box Modelling; Maximum Likelihood Methods

Abstract
A design method presented in a previous paper for the sequential generation of observation sites used for the inversion of a prediction model is extended to cope with practical issues such as delayed observations and design of batches of imposed size. The final objective of the construction is to be able to associate with any target T in the output space a value xT of the input factors such that the response of the system at xT will be “close” to T (from an industrial point of view, xT corresponds to manufactory conditions that yield a product whose feature of interest is described by T). The problem is thus much different from the more standard one where one wishes to build a precise model over the whole input space: here the model only has to be precise over a set of values xT that permit to reach any target T, that is, the observation sites should not be spread over the entire admissible input space, but should rather concentrate in areas that cover the reachable output space when mapped by the system. Examples in low dimensions are presented that illustrate the behavior of the method and allow a comparison to be made with a standard sequential method for designing exploratory experiments.

References

R. Ababou, A.C. Bagtzoglou, and E.F.Wood. On the con-
dition number of covariance matrices arising in kriging,
estimation, and simulation of random fields. Mathemat-
ical Geology, 26(1):99–133, 1994.
R. Bettinger, P. Duchˆene, L. Pronzato, and E. Thierry. De-
sign of experiments for response diversity. In Proc. 6th
International Conference on Inverse Problems in Engi-
neering (ICIPE), Journal of Physics: Conference Series,
Dourdan (Paris), 2008. To appear. http://hal.archives-
ouvertes.fr/hal-00290418/fr/.
J.P. Chil`es and P. Delfiner. Geostatistics, Modeling Spatial
Uncertainty. Wiley, New York, 1999.
A.I.J. Forrester, A.J. Keane, and N.W. Bressloff. The
design and analysis of ‘noisy’ computer experiments.
AIAA Journal, 44:2331–2339, 2006.
S.N. Lophaven, H.B. Nielsen, and J. Sondergaard. DACE,
a Matlab kriging toolbox. Technical Report IMM-REP-
2002-12, Technical University of Denmark, 2002.
M.D. Morris and T.J. Mitchell. Exploratory designs for
computational experiments. J. Statist. Plann. Inference,
43:381–402, 1995.
L. Pronzato and E. Thierry. Robust design with nonpara-
metric models: prediction of second-order characteristics
of process variability by kriging. In Prep. 13th IFAC
Symposium on System Identification, Rotterdam, pages
560–565, August 2003.
J. Sacks, W.J. Welch, T.J. Mitchell, and H.P. Wynn.
Design and analysis of computer experiments. Statistical
Science, 4(4):409–435, 1989.
T. Sahama and N.T. Diamond. Sample size considera-
tions and augmentation of computer experiments. The
Journal of Statistical Computation and Simulation, 68:
307–319, 2001.
T.J. Santner, B.J. Williams, and W.I. Notz. The Design
and Analysis of Computer Experiments. Springer, Hei-
delberg, 2003.
M.L. Stein. Interpolation of Spatial Data. Some Theory
for Kriging. Springer, Heidelberg, 1999.
C. Tsallis. Possible generalization of Boltzmann-Gibbs
statistics. Journal of Statistical Physics, 52(1/2):479–
487, 1988.
M.P. Wand and M.C. Jones. Kernel Smoothing. CRC
Press, 1995.