Hardy-type theorem for orthogonal functions with respect to their zeros. The Jacobi weight case (Q2480316)

From MaRDI portal
scientific article
Language Label Description Also known as
English
Hardy-type theorem for orthogonal functions with respect to their zeros. The Jacobi weight case
scientific article

    Statements

    Hardy-type theorem for orthogonal functions with respect to their zeros. The Jacobi weight case (English)
    0 references
    0 references
    0 references
    0 references
    31 March 2008
    0 references
    The work in this quite interesting paper is motivated by a result due to \textit{G. H. Hardy} [J. Lond. Math. Soc. 14, 37--44 (1939; Zbl 0021.01904 and JFM 65.0285.02)], stating that (under certain conditions) the only functions satisfying \[ \int_0^1\,f(\lambda_n t)f(\lambda_m t)dt=0,\;m\not= n, \] where \(\{\lambda_n\}_{n=1}^{\infty}\) are the real zeros of the function \(f\), are the Bessel functions. The authors introduce for \(\nu,\alpha\in\mathbb R\) with \(2\nu +\alpha >-1\) two classes of functions \(\mathcal A\): all entire functions of order less than \(2\) or of order \(2\) and minimal type of the form \[ f(z)=z^{\nu}\,\prod_{n=1}^{\infty}\,\left(1-{z^2\over \lambda_n^2}\right). \] \(\mathcal B\): all entire functions of the form \[ f(z)=z^{\nu} F(z), \] where \(F\) is entire with real (but not necessarily positive) zeros and of order less than \(1\) or of order \(1\) and of minimal type; moreover \(F(0)=1\). They introduce the following formula with a Jacobi weight on \((0,1)\) \[ \int_0^1\,f(\lambda_n t)f(\lambda_m t)t^{\alpha}(1-t)^{\beta}dt=0,\quad m\not= n,\;\alpha>-1-2\nu,\;\beta>-1.\tag{\(*\)} \] The main results then are: \textbf{Ia)} If \(f\in{\mathcal B}\) satisfies \((*)\) with \(\alpha>-1-2\nu,\;\nu\in\mathbb R,\beta>-1\) then it satisfies the integral equation \[ a\int_0^z\,u^{\nu +\alpha +1}(z-u)^{\beta}f(u)du=(az+1)\int_0^z\,u^{\nu +\alpha}(z-u)^{\beta}f(u)du + Az^{\nu +\alpha +\beta +1}f(z), \] where \(a=F'(0),A=-B(2\nu +\alpha +1,\beta +1)\;\;[B(\cdot,\cdot)\) the beta function]. \textbf{Ib)} If \(f\) satisfies the conditions of \textbf{Ia}, then \[ f(z)=\text{const.}\cdot z^{\nu}\,\sum_{n=0}^{\infty}\,a_nz^n,\;\;a_n={(a(\beta +1))^n\over \Gamma(\mu +\nu)}\,\prod_{j=1}^n\,{(\mu)_j\over (\mu +\beta +1)_j-(\mu)_j},\,n=0,1,\ldots \] Here \(\mu=2\nu +\alpha +1, \Gamma(\cdot)\) is Euler's Gamma function, \((b)_j\) is the customary Pochhammer symbol and an empty product is equal to \(1\). The series part represents an entire function of order \(\rho={1\over \beta +2}<1\) when \(\beta>-1\) (for \(\beta=0\) the original Hardy solution in terms of the Bessel function is recovered). \textbf{IIa)} If \(f\in{\mathcal A}\) satisfies \((*)\) with \(\alpha>-1-2\nu,\nu\in\mathbb R,\beta>-1\), then it satisfies the integral equation \[ a\int_0^z\,u^{\nu+\alpha+2}(z-u)^{\beta}f(u)du=(az^2+2)\,\int_0^z\,u^{\nu+\alpha} (z-u)^{\beta}f(u)du + Az{\nu+\alpha+\beta+1}f(z), \] where \(a=F''(0), A=-2B(2\nu+\alpha+1,\beta+1)\). \textbf{IIb)}: If \(f\) satisfies the conditions of \textbf{IIa}, then \[ f(z)=\text{const}\cdot z^{\nu}\,\sum_{n=0}^{\infty}\,a_{2n}z^{2n},\;\;a_{2n}=\left({a(\beta+1)\over 2}\right)^n\,\prod_{j=1}^n\,{[\beta+2(\mu+2j-1)](\mu)_{2(j-1)}\over (\mu+\beta+1)_{2j}-(\mu)_{2j}}, \] \(n=0,1,\ldots\). Here \(\mu=2\nu +\alpha+1\) and an empty product is equal to \(1\).
    0 references
    zeros of special functions
    0 references
    orthogonality
    0 references
    Jacobi weights
    0 references
    Mellin transform on distributions
    0 references
    entire functions
    0 references
    Bessel functions
    0 references
    hyper-Bessel functions
    0 references

    Identifiers