On a method for uniform summation of the Fourier-Jacobi series (Q2152599): Difference between revisions

On the interval \([ 0,1] \) consider the Jacobi weight \(\rho ( t) =t^{a}( 1-t) ^{b}\), where \(a,b>-1\). This weight introduces an inner product on \(C[ 0,1] \) by \[ \langle f,g\rangle _{a,b}=\int\limits_{0}^{1}f( t) g( t) \rho ( t)\, dt, \] for \(f, g\in C[ 0,1] \). Denote by \(p_{n}^{a,b}\) the orthogonal Jacobi polynomials of degree \(n\) attached to the weight \(\rho \). The formal Fourier-Jacobi series attached to a function \(f\in C[ 0,1] \) is: \[ f( x) \sim \sum\limits_{n=0}^{\infty }\langle f,p_{n}^{a,b}\rangle _{a,b}\cdot p_{n}^{a,b}( x) \] for \(x\in [ 0,1] \). Let \(A=( a_{n,m}) _{n,m}\) be a doubly infinite matrix with entries \(a_{n,m}\in\mathbb{R}\), \(n,m\geq 0\), with the property \(a_{n,m}=0\) for \(m>n\). We say that the matrix \(A\) defines a limitation method which attaches to each sequence \((s_{n}) _{n\text{ }}\) a generalized limit, denoted by \[ A\text{-}\underset{n\rightarrow \infty }{\lim }s_{n}:=\underset{n\rightarrow \infty }{\lim }t_{n}, \] where \[ t_{n}=\sum\limits_{m=0}^{n}a_{n,m}s_{m},\text{ }n\geq 0\text{,} \] under the hypothesis that \(\lim_{n\rightarrow \infty }t_{n}\) exists in the usual sense. When a limitation method given by a matrix \(A\) is applied to the sequence \( ( S_{n}) _{n}\) of the partial sums of a series \(\sum_{n=0}^{\infty }u_{n}\), \(S_{n}=\sum_{k=0}^{n}u_{k}\), one obtains a summation method of the series \(\sum_{n=0}^{\infty }u_{n}.\) The generalized sum given by a matrix $A$ of the series \(\sum_{n=0}^{\infty }u_{n}\) is denoted by \(A\text{-}\sum_{n=0}^{\infty }u_{n}\). If we have two limitation methods given by the matrices \(A\) and \(B\), respectively, we say that the method given by the matrix \(B\) is stronger than the method given by matrix \(A\) if any sequence which is generalized convergent by using the matrix \(A\) is also generalized convergent by using the matrix \(B\), with the same limit. Common matrix limitation methods are the Cesàro methods of order \(k\geq 1 \), denoted \(( C,k) \), which are attached to matrices \(C_{k}=( c_{n,m}^{k}) _{n,m}\), where \[ c_{n,m}^{k}=\binom{n+k}{k}^{-1}\binom{n+k-m-1}{k-1}. \] For \(a,b>-1\), consider the infinite matrix \(\Gamma ^{a,b}=( \gamma _{n,m}^{a,b}) _{n,m}\), where \[ \gamma _{n,m}^{a,b}=\begin{cases} \frac{n!\Gamma ( n+2+a+b) ( 2m+2+a+b) }{( n-m) !\Gamma ( n+m+3+a+b) }, & 0\leq m\leq n, \\ 0,& m>n\geq 0. \end{cases} \] The generalized limit of a sequence \(( x_{n}) _{n}\) given by the matrix \(\Gamma ^{a,b}\), when it exists, is denoted by \(\Gamma ^{a,b}\text{-} \lim{n\rightarrow \infty }x_{n}\) and is equal to \(\lim_{ n\rightarrow \infty }\sum_{m=0}^{n}\gamma _{n,m}^{a,b}x_{m}\). Also, the generalized sum of a series \(\sum_{n=0}^{\infty }u_{n}\), given by the matrix \(\Gamma ^{a,b}\), when it exists, is denoted by \(\Gamma ^{a,b}\text{-}\sum_{n=0}^{\infty }u_{n}\) and is equal to \(\lim_{ n\rightarrow \infty }\sum_{m=0}^{n}\gamma_{n,m}^{a,b}\sum_{k=0}^{m}u_{k}\). The main results of this paper are the following theorems. Theorem 1. For any \(f\in C[ 0,1] \), \(a,b>-1\), we have \[ \Gamma ^{a,b}\text{-}\sum\limits_{n=0}^{\infty }\langle f,p_{n}^{a,b}\rangle _{a,b}\cdot p_{n}^{a,b}( x) =f( x) ,\ x\in [ 0,1], \] and the convergence of the series is uniform with respect to~\(x\). Theorem 2. The summation method given by the matrix \(\Gamma^{a,b}\), with \(a,b>-1\), is stronger than all the Cesàro methods \((C,k)\) with \(k\in\mathbb{N}\) and \(k\geq 1\).