English

Randomized sparse grid algorithms for multivariate integration on Haar-Wavelet spaces

Numerical Analysis 2022-02-11 v1 Numerical Analysis

Abstract

The \emph{deterministic} sparse grid method, also known as Smolyak's method, is a well-established and widely used tool to tackle multivariate approximation problems, and there is a vast literature on it. Much less is known about \emph{randomized} versions of the sparse grid method. In this paper we analyze randomized sparse grid algorithms, namely randomized sparse grid quadratures for multivariate integration on the DD-dimensional unit cube [0,1)D[0,1)^D. Let d,sNd,s \in \mathbb{N} be such that D=dsD=d\cdot s. The ss-dimensional building blocks of the sparse grid quadratures are based on stratified sampling for s=1s=1 and on scrambled (0,m,s)(0,m,s)-nets for s2s\ge 2. The spaces of integrands and the error criterion we consider are Haar wavelet spaces with parameter α\alpha and the randomized error (i.e., the worst case root mean square error), respectively. We prove sharp (i.e., matching) upper and lower bounds for the convergence rates of the NN-th mininimal errors for all possible combinations of the parameters dd and ss. Our upper error bounds still hold if we consider as spaces of integrands Sobolev spaces of mixed dominated smoothness with smoothness parameters 1/2<α<11/2< \alpha < 1 instead of Haar wavelet spaces.

Keywords

Cite

@article{arxiv.2001.08558,
  title  = {Randomized sparse grid algorithms for multivariate integration on Haar-Wavelet spaces},
  author = {Marcin Wnuk and Michael Gnewuch},
  journal= {arXiv preprint arXiv:2001.08558},
  year   = {2022}
}