Summary
We generalize the T-optimality criterion for discrimination between two rival multiresponse models with observations corrupted by normally distributed noise with zero mean and a covariance matrix which depends on unknown parameters. Calculation of optimum designs amounts to solving an optimal control problem whose discretization leads to a semi-infinite programming problem. It is then solved numerically using an exchange method. As an example we calculate T-optimum design for discrimination between two complex chemical kinetic reactions. The design variables are the temperature profile during the process run and the initial concentration of reactants.
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Preview
Unable to display preview. Download preview PDF.
References
A.C. Atkinson. Horwitz’s rule, transforming both sides and the design of experiments for mechanistic models. Journal of the Royal Statistical Society, Series C, Applied Statistics52:261–2782003.
A.C. Atkinson and B. Bogacka. Compound and other optimum designs for systems of nonlinear differential equations arising in chemical kinetics. Chemometrics and Intelligent Laboratory Systems61:17–332002.
A.C. Atkinson and R.D. Cook. D-optimum designs for heteroscedastic linear models. Journal of the American Statistical Association90(429):204–2121995.
A.C. Atkinson and A.N. Donev. Optimum Experimental Designs. Clarendon Press, Oxford1992.
A.C. Atkinson and V.V. Fedorov. The design of experiments for discriminating between two rival models. Biometrika62(1):57–701975.
A.C. Atkinson and V.V. Fedorov. Optimal design: Experiments for discriminating between two several models. Biometrika62(2):289–3031975.
A.L. Burke, T.A. Duever, and A. Pendilis. Model discrimination via designed experiments: Discriminating between the terminal and penultimate models on the basis of composition data. Macromolecules27:386–3991994.
D. Downing, V.V. Fedorov, and S. Leonov. Extracting information from the variance function: Optimal design. In A. C. Atkinson, P. Hackl, and W. G. Müller, editorsmODa 6 — Advances in Model-Oriented Design and Analysis (Puchberg/Schneeberg, June 25-29, 2001), pages 45–52. Physica-Verlag, Heidelberg2001.
V. V. Fedorov and P. Hackl. Model-Oriented Design of Experiments. Lecture Notes in Statistics. Springer-Verlag, New York1997.
V.V. Fedorov and V. Khabarov. Duality of optimal design for model discrimination and parameter estimation. Biometrika73(1):183–1901986.
K. Felsenstein. Optimal Bayesian design for discriminating among rival models. Computational Statistics & Data Analysis14:427–4361992.
R. Hettich and K.O. Kortanek. Semi-infinite programming: Theory, methods and applications. SIAM Review35(5):380–4291993.
W.G. Müller and A.C. Ponce de Leon. Discriminating between two binary data models: Sequentially designed experiments. Journal of Statistical Computation and Simulation55:87–1001996.
M.J. Pilling and P.W. Seakins. Reaction Kinetics. Oxford University Press, Oxford1995.
E. Polak. On the mathematical foundations of nondifferentiable optimization in engineering design. SIAM Review29(1):21–891987.
A.C. Ponce de Leon and A.C. Atkinson. Optimum experimental design for discriminating between two rival models in the presence of prior information. Biometrika78(3):601–6081991.
F. Pukelsheim. Optimal Design of Experiments. Probability and Mathematical Statistics. John Wiley & Sons, New York1993.
R. Reemtsen and S. Görner. Numerical methods for semi-infinite programming: A survey. In R. Reemtsen and J.-J. Rückmann, editorsSemi-Infinite Programming, pages 195–275. Kluwer Academic Publishers, Boston1998.
E.W. Sacks. Semi-infinite programming in control. In R. Reemtsen and J.-J. Rückmann, editorsSemi-Infinite Programming, pages 389–411. Kluwer Academic Publishers, Boston1998.
W.E. Stewart, Y. Shon, and G.E.P. Box. Discrimination and goodness of fit of multiresponse mechanistic models. AIChE Journal44(6):1404–14121998.
D. Uciński and B. Bogacka. Construction of T-optimum designs for multiresponse dynamic systems. In W. Härdie and B. Rönz, editorsCompstat 2002, Proceedings in Computational Statistics 2002, pages 267–272Berlin2002. Springer.
D. Uciński and B. Bogacka. T-optimum design for discrimination between two multiresponse dynamic models. Submitted for publication, available at http://www.maths.qmw.ac.uk/~bb/UcinskiBogackal.pdf, 2003.
É. Walter and L. Pronzato. Identification of Parametric Models from Experimental Data. Communications and Control Engineering. Springer-Verlag, Berlin1997.
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2004 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Ucinski, D., Bogacka, B. (2004). T-Optimum Designs for Multiresponse Dynamic Heteroscedastic Models. In: Di Bucchianico, A., Läuter, H., Wynn, H.P. (eds) mODa 7 — Advances in Model-Oriented Design and Analysis. Contributions to Statistics. Physica, Heidelberg. https://doi.org/10.1007/978-3-7908-2693-7_21
Download citation
DOI: https://doi.org/10.1007/978-3-7908-2693-7_21
Publisher Name: Physica, Heidelberg
Print ISBN: 978-3-7908-0213-9
Online ISBN: 978-3-7908-2693-7
eBook Packages: Springer Book Archive