A perturbed Whittaker-Kotel'nikov-Shannon sampling theorem (Q541234): Difference between revisions
From MaRDI portal
Changed an Item |
Added link to MaRDI item. |
||
links / mardi / name | links / mardi / name | ||
Revision as of 07:58, 30 January 2024
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | A perturbed Whittaker-Kotel'nikov-Shannon sampling theorem |
scientific article |
Statements
A perturbed Whittaker-Kotel'nikov-Shannon sampling theorem (English)
0 references
6 June 2011
0 references
Suppose that \(P(x,t) = i e^{-itx}\int_{-\pi}^x r(\tau) e^{it\tau}\, d\tau \) for a given continuous real-valued function \(r\) on \([-\pi, \pi]\). The integro-differential equation \[ \ell_r (y) = i y' +r(x)\int_{-\pi}^\pi r(\tau) y(\tau) \, d\tau =ty\qquad (-\pi\leq x\leq\pi) \] has a general solution of the form \(\phi (x,t) =c(\varphi+ \langle \varphi, r\rangle\, P /C(t)) \) (\(\varphi(x,t)=e^{-itx}\)) where \(c\) is an arbitrary constant and it is assumed that \(C(t)\neq 0\) for some \(t\in \mathbb{C}\). Set \(\Delta(t) = \phi(\pi,t)- \phi(-\pi,t)\). Then \(t\) is an eigenvalue of the problem \(\ell_r (y)=ty\) precisely when \(\Delta (t)=0\). The collection of all eigenvalues is parametrized by \(\{t_n\}_{n\in\mathbb{Z}}\). The following perturbed sampling formula is proved for \[ f(t) = \int_{-\pi}^\pi g(x)\phi(x,t)\, dx = \int_{-\pi}^\pi g(x)\Bigl(e^{-itx}+\frac{\langle\varphi, r\rangle}{C(t)} P(x,t)\Bigr)\, dx, \] where \(g\in L^2(\mathbb{R})\) is given, under the assumptions that \(C(t)\neq 0\) for all \(t\in \mathbb{C}\) and \(\varphi\), \(P\) and \(\Delta\) are as above: The function \(f\) is entire and \[ f(t)=\sum_{n\in\mathbb{Z}} f(t_n) \frac{\Delta(t)}{(t-t_n)\Delta'(t_n)}. \] An additional formula involves the function \(\Phi(x,t)=\omega(t) H_0 (x,t)\) in which \(H_0(x,t)=H(x,\xi_0,t)\) for \(t\notin\{t_n\} \) with \[ H(x,\xi,t)= \sum_{n\in\mathbb{Z}} \sum_{\nu=1}^{\nu_n} \frac{\phi_{n,\nu} (x) \overline{\phi}_{n,\nu}(\xi)}{t_n-t} \] is given, where \(\phi_{n,\nu}\) are the eigenfunctions associated with \(t_n\). In this case, given \(g\in L^2[-\pi,\pi]\), the entire function \(F(t)=\int_{-\pi}^\pi \overline{g}(x) \Phi(x,t)\, dx\) has the representation \[ F(t) = \sum_{n\in\mathbb{Z}} F(t_n) \frac{\omega(t)}{(t-t_n)\, \omega' (t_n)} \] converging to \(F\) uniformly on \(\mathbb{R}\) and on compact subsets of \(\mathbb{C}\).
0 references
Green's function
0 references
eigenfunction expansions
0 references
rank-one perturbations
0 references
sampling
0 references