scientific article; zbMATH DE number 1076945

From MaRDI portal
Publication:4360089





The Hankel transform of \(f\) is defined by \[ F_\nu(x)= \int^\infty_0 f(t)(t/x)^{{3\over 2}} J_\nu(t/x)dt/t\quad (x>0). \] For \(\rho\in\mathbb{R}\), we write \(R_\rho\) for the class of functions regularly varying (at infinity) with index \(\rho\): the class of positive measurable \(f\), defined on some neighbourhood of infinity, such that \[ f(\lambda x)/f(x)\to \lambda^\rho\quad (x\to\infty)\;\forall\lambda>0. \] The upper order \(\rho(f)\) of a nonnegative function is defined by \[ \rho(f):= \limsup_{x\to\infty} {\log f(x)\over\log x}. \] Theorem 1. Let \(\nu\geq -{1\over 2}\), \(t^{\nu+{1\over 2}}f(t)\in L^1_{\text{loc}}[0, \infty)\), and \(f\) be ultimately decreasing to zero at infinity, with upper order \(\rho:= \rho(f)\) and Hankel transform \(F:= F\nu\). If \(-\nu-{3\over 2}<\rho< 0\) and \(F(x)/f(x)\to c(f)>0\) \((x\to \infty)\), then \[ c(f)= a_\nu(\rho)= 2^{\rho+{1\over 2}} {\Gamma(3/4+ \nu/2+\rho/2)\over \Gamma(1/4+ \nu/2-\rho/2)}. \] Theorem 2. Define \(\nu_0\in\left({1\over 2}, 1\right)\) by \[ \log 2+ {\Gamma'(\nu_0/2+ 1/2)\over \Gamma(\nu_0/2+ 1/2)}= 0 \] (\(\nu_0= 0.8660252365\dots\)). If in Theorem 1, we further assume \(\nu\leq \nu_0\) and \(\rho\neq -{1\over 2}\), then \(f\in R_\rho\).











This page was built for publication:

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4360089)