Two formulae with nodes related to zeros of Bessel functions for semi-infinite integrals: extending Gauss-Jacobi-type rules (Q6141539)
From MaRDI portal
scientific article; zbMATH DE number 7780871
Language | Label | Description | Also known as |
---|---|---|---|
English | Two formulae with nodes related to zeros of Bessel functions for semi-infinite integrals: extending Gauss-Jacobi-type rules |
scientific article; zbMATH DE number 7780871 |
Statements
Two formulae with nodes related to zeros of Bessel functions for semi-infinite integrals: extending Gauss-Jacobi-type rules (English)
0 references
19 December 2023
0 references
In this interesting paper, the authors study approximations of integrals of the form: \(\int_0^{\infty}x^{\alpha}f(x)dx, \alpha>-1\) where \(x^{\alpha}f(x)\) decays sufficiently fast for large enough argument. Two formulae are given. They are constructed via a limiting process using a linear transformation from the Gauss-Jacobi (GJ) rule and its associated family rules: the Gauss-Jacobi-Radau (GR) and the Gauss-Jacobi-Lobatto (GL) rule on a finite interval. The limiting GJ (LGJ) formula uses as nodes the zeros of the Bessel function \(J_\alpha(x)\) squared after multiplied by a positive constant and for the limiting GR (LGR) and limiting GL (LGL) formula, the zeros of \(J_{\alpha+1}(x)\). The LGJ formula is also shown to be that derived from a formula developed by Frappier and Olivier for an integral on \([0,\infty)\). The authors show that for a function \(f(z)\) analytic on a certain domain in the complex plane satisfying some appropriate conditions, there exists an absolute constant \(d> 0\) such that the errors of both formulae are \(O(\exp(-2d/a))\) as \(a\to 0+\). The average of the formulae give smaller quadrature errors than each formula. Numerical examples confirm these behaviors and show that the formulae give asymptotically the same quadrature errors of opposite sign. Consequently, the LGR formula behaves like an anti-LGJ formula in the same way as the Lobatto rule for integrals on \([-1, 1]\) behaves like the anti-Gauss rule. The paper is well written with a good set of references.
0 references
integral on an unbounded interval
0 references
extension of Gauss-Jacobi-type rules
0 references
error analysis
0 references
anti-Gaussian rule
0 references