Convergence rates for Padé-based iterative solutions of equations (Q749159)

Let f(x) have a simple zero \(\omega\) and \(m+n+1\) bounded derivatives in a neighbourhood of \(x=\omega\), let \(d_ k\) be a simple zero of the \(r_{nm}\) Padé approximant of the Taylor series of \(f(\omega_ k+d)\) at \(d=0\), and \(\omega_ k\) converging to \(\omega\). Then the convergence is of order \(m+n+1\). This theorem is proved and generalized to the multidimensional case.