Saturation by the Fourier transform method for the sampling Kantorovich series based on bandlimited kernels (Q2285016)
From MaRDI portal
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | Saturation by the Fourier transform method for the sampling Kantorovich series based on bandlimited kernels |
scientific article |
Statements
Saturation by the Fourier transform method for the sampling Kantorovich series based on bandlimited kernels (English)
0 references
15 January 2020
0 references
Let \(L^{p}\left(\mathbb{R}\right) \), \(1\leq p\leq \infty \), be the space of all Lebesgue measurable functions \(f:\mathbb{R}\rightarrow \mathbb{R}\), for which the norm: \[\left\Vert f\right\Vert _{p}:=\left( \int\limits_{\mathbb{R}}\left\vert f\left( t\right) \right\vert ^{p}dt\right) ^{1/p}\text{, } \left\Vert f\right\Vert _{\infty }:=\underset{t\in\mathbb{R}}{ess\sup }\left\vert f\left( t\right) \right\vert\] is finite. Denote by \(B_{\sigma }^{p}\left(\mathbb{R}\right) \), the so-called Bernstein class (or Bernstein space), for \(\sigma\geq 0\) and \(1\leq p\leq \infty \), containing the functions of \(L^{p}\left(\mathbb{R}\right) \) which can be extended to an entire function \(f\left( z\right) \) \( \left( z=x+iy\in\mathbb{C}\right) \) of exponential type \(\sigma \), i.e., satisfying: \[\left\vert f\left( z\right) \right\vert \leq e^{\sigma \left\vert y\right\vert }\left\Vert f\right\Vert _{\infty }\text{, }z\in\mathbb{C}.\] Moreover, we denote by: \[\widehat{f}\left( v\right) :=\int\limits_{\mathbb{R}}f\left( u\right) e^{-iuv}du,\] \(v\in\mathbb{R}\), the Fourier transform of \(f\in L^{1}\left(\mathbb{R}\right) \). Let \(\chi:\mathbb{R}\rightarrow\mathbb{R}\) be a function belonging to \(L^{1}\left(\mathbb{R}\right) \), and bounded in a neighborhood of the origin. We will say that, a sequence \(\xi =\left( t_{k}\right) _{k\in\mathbb{Z}}\subset\mathbb{R}\), belongs to the class \(\prod \left( \chi ,\beta ,M,\delta ,\Delta \right) \), with some fixed \(\beta\) , \(M>0\) and \(0<\delta \leq \Delta \), if it satisfies the following assumptions: \(\left( \xi 1\right) \) \(\delta \leq \Delta _{k}:=t_{k+1}-t_{k}\leq \Delta \), for every \(k\in\mathbb{Z}\), and \(\underset{k\rightarrow \pm \infty }{\lim }t_{k}=\pm \infty \); \(\left( \xi 2\right) \) \(\sum\limits_{k\in\mathbb{Z}}\chi \left( u-t_{k}\right) =1\), for every \(u\in\mathbb{R}\); \(\left( \xi 3\right) \) the discrete absolute moment of order \(\beta \) is bounded by \(M\), i.e., \[m_{\beta ,\xi }\left( \chi \right) :=\underset{u\in \mathbb{R}}{\sup }\sum\limits_{k\in\mathbb{Z}}\left\vert u-t_{k}\right\vert ^{\beta }\left\vert \chi \left(u-t_{k}\right) \right\vert \leq M. \] In general, a function \(\chi \) which satisfies assumptions \(\left( \xi 2\right) \) and \(\left( \xi 3\right) \) for some sequences \(\xi =\left( t_{k}\right) _{k\in\mathbb{Z}}\) fulfilling \(\left( \xi 1\right) \) is called a kernel. Note that, if we choose the sequence \(t_{k}=k\) for \(k\in\mathbb{Z}\), we have a uniformly spaced sequence, belonging for instance to the class \(\prod \left( \chi ,\beta ,m_{\beta }\left( \chi \right) ,1,1\right) \), where \(\chi \in L^{1}\left(\mathbb{R}\right) \), is bounded in a neighborhood of the origin, satisfying: \[ \sum\limits_{k\in\mathbb{Z}}\chi \left( u-k\right) =1, \tag{1}\] for every \(u\in\mathbb{R}\), and \[ m_{\beta ,\xi }\left( \chi \right) =m_{\beta }\left( \chi \right) :=\underset{u\in\mathbb{R}}{\sup }\sum\limits_{k\in\mathbb{Z}}\left\vert u-k\right\vert ^{\beta }\left\vert \chi \left( u-k\right)\right\vert <+\infty. \tag{2}\] for some \(\beta >0\). Now, we recall the definition of the sampling Kantorovich operators. Let \(\xi =\left( t_{k}\right) _{k\in\mathbb{Z}}\) be a fixed sequence belonging to a class \(\prod \left( \chi ,\beta ,M,\delta ,\Delta \right) \). We denote by: \[ \left( S_{\omega }^{\chi ,\xi }f\right) \left( x\right) :=\sum\limits_{k\in\mathbb{Z}}\left[ \frac{\omega }{\Delta _{k}}\int\limits_{t_{k}/\omega }^{t_{k+1}/\omega }f\left( u\right) du\right] \chi \left( u\omega-t_{k}\right), \] \(x\in\mathbb{R}\) and \(\omega >0\), the sampling Kantorovich operators of \(f\), where \(f:\mathbb{R}\rightarrow\mathbb{R}\) is a locally integrable function, such that the above series is convergent for every \(x\in\mathbb{R}\). In case of the uniform spaced sequence \(t_{k}=k\) for \(k\in\mathbb{Z}\), and of a function \(\chi \) satisfying (1) and (2), the sampling Kantorovich series reduces to: \[ \left( S_{\omega }^{\chi ,\xi }f\right) \left( x\right) =\left( S_{\omega}^{\chi }f\right) \left( x\right) :=\sum\limits_{k\in\mathbb{Z}}\left[ \omega \int\limits_{k/\omega }^{\left( k+1\right) /\omega }f\left(u\right) du\right] \chi \left( u\omega -k\right) , \] \(x\in\mathbb{R}\) and \(\omega >0\). The main results of this paper are the following two theorems: Theorem. Let \(\chi \in B_{\mu \pi }^{1}\left(\mathbb{R}\right) \) be a kernel, with \(\mu >0\). Moreover, let \(f\in B_{\mu \pi \overline{\omega }}^{1}\left(\mathbb{R}\right) \), for some fixed \(\overline{\omega }>0\). Assume in addition that there exist \(c\geq 0\), and \(0<\gamma \leq 1\), such that: \[ \underset{\omega \rightarrow +\infty }{\lim }\frac{\omega ^{\gamma }}{\left\vert v\right\vert ^{\gamma }}\left[ 1-\widehat{\chi }\left( \frac{v}{ \omega }\right) \right] =c, \] \(v\in\mathbb{R}\), a. e. (with \(c>0\) if \(\gamma <1\)). If: \[ \left\Vert S_{\omega }^{\chi }f-f\right\Vert _{1}=o\left( \omega ^{-\gamma}\right) \text{, }\omega \rightarrow +\infty , \] then, it follows \(f\equiv 0\) in \(\mathbb{R}\). Theorem. Let \(f\in L^{1}\left(\mathbb{R}\right) \), and \(\chi \in B_{\mu \pi }^{1}\left(\mathbb{R}\right) \) be a fixed kernel, for some \(\mu >0\). Assume in addition that there exist \(c\geq 0\), and \(0<\gamma \leq 1\), such that: \[ \underset{\omega \rightarrow +\infty }{\lim }\frac{\omega ^{\gamma }}{\left\vert v\right\vert ^{\gamma }}\left[ 1-\widehat{\chi }\left( \frac{v}{ \omega }\right) \right] =c, \] \(v\in\mathbb{R}\), a. e. (with \(c>0\) if \(\gamma <1\)). If: \[ \left\Vert S_{\omega }^{\chi ,\xi }f-f\right\Vert _{1}=o\left( \omega^{-\gamma }\right) \text{, }\omega \rightarrow +\infty , \] uniformly with respect to \(\xi \in \prod \left( \chi ,\beta ,m_{\beta}\left( \chi \right) ,1,1\right) \), for some \(\beta >0\), then, it turns out that \(f\left( x\right) =0\), for almost all \(x\in\mathbb{R} \).
0 references
sampling Kantorovich series, Fourier transform method, bandlimited kernel, generalized sampling operator
0 references
singular integral, saturation theorem
0 references
0 references
0 references
0 references
0 references
0 references
0 references