Simultaneous point estimates for Newton's method (Q1864771): Difference between revisions

The primary aim of this paper is the establishment of a practical condition guaranteeing the convergence of the one-dimensional Newton iteration from \(n\) pairwise distinct approximations to \(n\) pairwise distinct zeros of a polynomial. The new condition is practical in that the error estimate used involves only function evaluations already computed in the course of computing the Newton iteration, as opposed to, for example, the error estimate found in \textit{X. Wang} and \textit{D. Han}'s work [Sci. China, Ser. A 33, 135-144 (1990; Zbl 0699.65046)]. Convergence is guaranteed to be simultaneous and quadratic, and is \(O(n)\) faster than the method proven in the author's thesis. The error estimate may be used for polynomial root approximation based on homotopy methods [see, e.g. \textit{M. Shub} and \textit{S. Smale}, J. Am. Math. Soc 6, 459-501 (1993; Zbl 0821.65035)]. Additionally, combining the techniques used by \textit{P. Tilli} [Computing, 59, 307-324 (1997; Zbl 0891.65051)] with the new estimate leads to a practical homotopy method using Newton iterations, rather than Durand-Kerner iterations.