Error bounds for overdetermined and underdetermined generalized centred simplex gradients

DOI10.1093/IMANUM/DRAA089MaRDI QIDQ5024895zbMATH OpenOpenAlexFDO

Authors Gabriel Jarry-Bolduc, C. Planiden, Warren L. Hare

Publication date 27 January 2022

Published in IMA Journal of Numerical Analysis (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/2006.00742

zbMATH Keywords

error boundsgradient approximationsimplex gradients

Mathematics Subject Classification ID

Theory of matrix inversion and generalized inverses (15A09) Numerical solutions to overdetermined systems, pseudoinverses (65F20) Numerical optimization and variational techniques (65K10)

Abstract: Using the Moore--Penrose pseudoinverse, this work generalizes the gradient approximation technique called centred simplex gradient to allow sample sets containing any number of points. This approximation technique is called the emph{generalized centred simplex gradient}. We develop error bounds and, under a full-rank condition, show that the error bounds have order $O(Delt{a}^2)$, where $Delta$ is the radius of the sample set of points used. We establish calculus rules for generalized centred simplex gradients, introduce a calculus-based generalized centred simplex gradient and confirm that error bounds for this new approach are also order $O(Delt{a}^2)$. We provide several examples to illustrate the results and some benefits of these new methods.

Recommendations

• The calculus of simplex gradients
• Limiting behaviour of the generalized simplex gradient as the number of points tends to infinity on a fixed shape in \(\mathrm{IR}^n\)
• Limiting behavior of derivative approximation techniques as the number of points tends to infinity on a fixed interval in \(\mathbb{R}\)
• Calculus identities for generalized simplex gradients: rules and applications
• Efficient calculation of regular simplex gradients

Cited in

(10)

• Approximating the diagonal of a Hessian: which sample set of points should be used
• Limiting behavior of derivative approximation techniques as the number of points tends to infinity on a fixed interval in \(\mathbb{R}\)
• The calculus of simplex gradients
• Limiting behaviour of the generalized simplex gradient as the number of points tends to infinity on a fixed shape in \(\mathrm{IR}^n\)
• Calculus identities for generalized simplex gradients: rules and applications
• Scalable subspace methods for derivative-free nonlinear least-squares optimization
• Linear convergence of the derivative-free proximal bundle method on convex nonsmooth functions, with application to the derivative-free \(\mathcal{VU}\)-algorithm
• Adapting the centred simplex gradient to compensate for misaligned sample points
• On complexity constants of linear and quadratic models for derivative-free trust-region algorithms
• Using generalized simplex methods to approximate derivatives

This page was built for publication: Error bounds for overdetermined and underdetermined generalized centred simplex gradients

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q5024895)