High order algorithms for approximatingnth roots

From MaRDI portal
Publication:4828667





The paper is concerned with the construction of high-order Newton-type methods of the form \[ t_{k+1}=t_{k}-p_{q}(L_{f}(t_{k}))\frac{f(t_{k})}{f'(t_{k})},\quad k\geq 0,\tag{1} \] which approximates the \(n\)th root or a real number, i.e., is associated to the particular scalar equation \[ f(t)\equiv t^{n}-R=0, \] where \(R\) is a real number, \(p_{q}(t)\) is given by \[ p_{q}(t)=\sum_{j=0}^{q-2}\binom {1/n}{j+1} (-1)^{j} \frac {n^{j+1}}{(n-1)^{j}} t^{j}, \] and \(L_{f}(t)\) is the dregree of logarithmic convexity, that is, \[ L_{f}(t)=\frac{f(t)f(t)}{(f'(t))^{2}}. \] The obtained methods are of \(q\)-order of convergence. Note that for \(q=3\), from \((1)\) we obtain Chebyshev's method, while Newton's method is obtained for \(q=2.\) Numerical examples are also given for \(R=35\), \(n=2\) and, respectively, \(n=5\) , which are empirically compared to other results in literature.











This page was built for publication: High order algorithms for approximatingnth roots

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4828667)