Multivariate approximation by translates of the Korobov function on Smolyak grids

Abstract

For a set $\mathbb{W} \subset L_p(\bT^d)$, $1 < p < \infty$, of multivariate periodic functions on the torus $\bT^d$ and a given function $φ\in L_p(\bT^d)$, we study the approximation in the $L_p(\bT^d)$-norm of functions $f \in \mathbb{W}$ by arbitrary linear combinations of $n$ translates of $φ$. For $\mathbb{W} = U^r_p(\bT^d)$ and $φ= κ_{r,d}$, we prove upper bounds of the worst case error of this approximation where $U^r_p(\bT^d)$ is the unit ball in the Korobov space $K^r_p(\bT^d)$ and $κ_{r,d}$ is the associated Korobov function. To obtain the upper bounds, we construct approximation methods based on sparse Smolyak grids. The case $p=2, \ r > 1/2$, is especially important since $K^r_2(\bT^d)$ is a reproducing kernel Hilbert space, whose reproducing kernel is a translation kernel determined by $κ_{r,d}$. We also provide lower bounds of the optimal approximation on the best choice of $φ$.