On irregular sampling in Bernstein spaces (Q2255250): Difference between revisions

For \(\sigma>0\), the Bernstein space \(B_\sigma\) consists of those entire functions \(f\) for which \(|f(x+iy)| \leq C(f) \exp(\sigma |y|)\) for \(x,y\in\mathbb{R}\). Given a set \(\Lambda\subset \mathbb{R}\), the sampling constant \(K(\Lambda, B_\sigma)\) is defined as the minimal \(K\) such that \[ \sup_{t\in\mathbb{R}} |f(t)| \leq K \sup_{\lambda\in\Lambda} |f(\lambda)|, \quad f\in B_\sigma. \] Beurling's theorem guarantees that \(K(\Lambda, B_\pi)<\infty\) if and only if \(D^{-}(\Lambda)<1\), where \(D^{-}(\Lambda)\) denotes the lower uniform density. The authors obtain sharp estimates for \(K(\Lambda, B_\sigma)\) when \(D^{-}(\Lambda)=1\) and \(\sigma<\pi\). In particular, let \(\Lambda\) be a uniformly discrete set with \(D^{-}(\Lambda)=1\). Then \[ K(\Lambda, B_\sigma) \geq C \log\frac{\pi}{\pi-\sigma}, \quad 0<\sigma<\pi. \] The above estimate is sharp because \[ K(\mathbb{Z}, B_\sigma) \geq \frac{2}{\pi} \log\frac{\pi}{\pi-\sigma}(1+o(1)), \quad \sigma\uparrow \pi, \] by a theorem of S. N. Bernstein.