E-optimal designs for second-order response surface models

DOI10.1214/14-AOS1241zbMATH Open1310.62097arXiv1403.3805MaRDI QIDQ464200FDOQ464200

Authors: Holger Dette, Yuri Grigoriev 

Publication date: 17 October 2014

Published in: The Annals of Statistics (Search for Journal in Brave)

Abstract: $E$-optimal experimental designs for a second-order response surface model with $kgeq1$ predictors are investigated. If the design space is the $k$-dimensional unit cube, Galil and Kiefer [J. Statist. Plann. Inference 1 (1977a) 121-132] determined optimal designs in a restricted class of designs (defined by the multiplicity of the minimal eigenvalue) and stated their universal optimality as a conjecture. In this paper, we prove this claim and show that these designs are in fact $E$-optimal in the class of all approximate designs. Moreover, if the design space is the unit ball, $E$-optimal designs have not been found so far and we also provide a complete solution to this optimal design problem. The main difficulty in the construction of $E$-optimal designs for the second-order response surface model consists in the fact that for the multiplicity of the minimum eigenvalue of the "optimal information matrix" is larger than one (in contrast to the case $k=1$) and as a consequence the corresponding optimality criterion is not differentiable at the optimal solution. These difficulties are solved by considering nonlinear Chebyshev approximation problems, which arise from a corresponding equivalence theorem. The extremal polynomials which solve these Chebyshev problems are constructed explicitly leading to a complete solution of the corresponding $E$-optimal design problems.

Full work available at URL: https://arxiv.org/abs/1403.3805

Recommendations

• Non-Chebyshev E-optimal experimental designs and decompositions of positive polynomial. I
• scientific article; zbMATH DE number 2135710
• \(E\)-optimal designs for polynomial regression
• Algorithmic construction of \(R\)-optimal designs for second-order response surface models
• On the duality of the optimal design problem by the \(E\)-criterion

zbMATH Keywords

duality\(E\)-optimalityoptimal designsresponse surface modelsextremal polynomialnonlinear Chebyshev approximation

Mathematics Subject Classification ID

Optimal statistical designs (62K05) Response surface designs (62K20) Best approximation, Chebyshev systems (41A50)

Cites Work

• General equivalence theory for optimum designs (approximate theory)
• Asymptotic Properties of Non-Linear Least Squares Estimators
• Optimum Designs in Regression Problems
• Title not available (Why is that?)
• Title not available (Why is that?)
• Optimal Design of Experiments
• Title not available (Why is that?)
• Title not available (Why is that?)
• Response surface methodology. Process and product optimization using designed experiments
• An application of the Kiefer-Wolfowitz equivalence theorem to a problem in Hadamard transform optics
• A note on \(E\)-optimal designs for weighted polynomial regression
• Comparison of design for quadratic regression of cubes
• Comparison of rotatable designs for regression on balls. I. (quadratic)
• Canonical reduction of second-order fitted models subject to linear restrictions.
• Kiefer ordering of simplex designs for second-degree mixture models with four or more ingredi\-ents.
• Optimum Designs in Regression Problems, II
• Title not available (Why is that?)
• Title not available (Why is that?)
• Response surface design evaluation and comparison
• Optimal discrimination designs for multifactor experiments
• Title not available (Why is that?)
• Optimum experimental design for a regression on a hypercube- generalization of Hoel's result
• Functional approach to optimal experimental design.
• Efficient \(D_ s\)-optimal designs for multivariate polynomial regression on the q-cube
• \(E\)-optimal designs for polynomial regression
• Geometry of \(E\)-optimality
• D-optimum and quasi-D-optimum second-order designs on a cube
• E-optimal designs for linear and nonlinear models with two parameters

Cited In (10)

• Approximate \(I\)-optimal designs for polynomial models over the unit ball
• Adaptive grid semidefinite programming for finding optimal designs
• Design admissibility and de la Garza phenomenon in multifactor experiments
• Removal of the points that do not support an \(E\)-optimal experimental design
• Algorithmic construction of \(R\)-optimal designs for second-order response surface models
• Least squares polynomial chaos expansion: a review of sampling strategies
• Efficient response surface designs for the second-order multivariate polynomial model robust to missing observation
• Optimal designs for spline wavelet regression models
• D-optimal designs based on the second-order least squares estimator
• Optimal designs for non-competitive enzyme inhibition kinetic models

This page was built for publication: \(E\)-optimal designs for second-order response surface models

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q464200)