English

Monte Carlo integration of non-differentiable functions on $[0,1]^\iota$, $\iota=1,\dots,d$, using a single determinantal point pattern defined on $[0,1]^d$

Computation 2021-10-19 v2 Numerical Analysis Classical Analysis and ODEs Numerical Analysis Statistics Theory Statistics Theory

Abstract

This paper concerns the use of a particular class of determinantal point processes (DPP), a class of repulsive spatial point processes, for Monte Carlo integration. Let d1d\ge 1, Id={1,,d}I\subseteq \overline d=\{1,\dots,d\} with ι=I\iota=|I|. Using a single set of NN quadrature points {u1,,uN}\{u_1,\dots,u_N\} defined, once for all, in dimension dd from the realization of the DPP model, we investigate "minimal" assumptions on the integrand in order to obtain unbiased Monte Carlo estimates of μ(fI)=[0,1]ιfI(u)du\mu(f_I)=\int_{[0,1]^\iota} f_I(u) \mathrm{d} u for any known ι\iota-dimensional integrable function on [0,1]ι[0,1]^\iota. In particular, we show that the resulting estimator has variance with order N1(2s1)/dN^{-1-(2s\wedge 1)/d} when the integrand belongs to some Sobolev space with regularity s>0s > 0. When s>1/2s>1/2 (which includes a large class of non-differentiable functions), the variance is asymptotically explicit and the estimator is shown to satisfy a Central Limit Theorem.

Keywords

Cite

@article{arxiv.2003.10323,
  title  = {Monte Carlo integration of non-differentiable functions on $[0,1]^\iota$, $\iota=1,\dots,d$, using a single determinantal point pattern defined on $[0,1]^d$},
  author = {Jean-François Coeurjolly and Adrien Mazoyer and Pierre-Olivier Amblard},
  journal= {arXiv preprint arXiv:2003.10323},
  year   = {2021}
}