Sensitivity of best recovery in the Sobolev spaces \(W^{r,d}_{\infty}\), \(\widetilde W^{r,d}_{\infty}\) for perturbed sampling (Q1569740)

This is a theoretical paper that deals with error bounds in Sobolev spaces. The author considers a linear operator \(S\) acting on a Sobolev space \(F\) which is either a set of functions defined on the \(d\)-dimensional unit cube \(I^d\) or of 1-periodic functions on \(\mathbb{R}^d\). He studies the sensitivity of the optimal recovery of \(S\) from \(n\) points forming a uniform grid on \(I^d\). His main result is that this sensitivity is bounded (in an appropriate norm) by a constant multiple of the bound to the data perturbations. This indicates that in this case the curse of dimensionality does not extend to sensivitity.