Chebyshev approximation of the second kind of modified Bessel function of order zero (Q2387501)

\textit{D. E. Amos} [ ''Computation of modified Bessel functions and their ratios'', Math. Comput. 28, 239--251(1974; Zbl 0277.65006)] presented an algorithm for calculating ratios \(r_{\nu}(x)=\frac{I_{\nu+1}(x)}{I_{\nu}(x)}, r\geq{0}, \nu\geq{0}\). \textit{M. Abramowitz} and \textit{I. Stegun} [Handbook of mathematical functions with formulas, graphs, and mathmetical tables, New York: John Wiley (1984; Zbl 0643.33001)] present tables where \(x^{\tfrac{1}{2}}e^{x}K_{i}(x), i=0,1\) are approximated by polynomials in \(\frac{1}{x}\) in accordance with the asymptotic formula (2) in the paper under review. Both \(I_{\nu}(x)\) and Macdonald's function \(K_{\nu}(x)\) are solutions to the modified Bessel equation (1). The authors of the paper under review compare the approximations of \(e^{x}K_{0}(x)\) obtained by a Chebyshev spectral method, Amos (loc. cit.) and \textit{Y. Luke}, [The special functions and their approximations. I. II. Mathematics in Science and Engineering. 53. New York-London: Academic Press (1969; Zbl 0193.01701)]. The conclusion is that the first two methods give the best approximations for small arguments. The paper concludes with a discussion of approximation of \(K_{0}(x)\) by exponential transformation.