The technique of relaxed power series expansion provides an efficient way to solve so called recursive equations of the form , where the unknown is a vector of power series, and where the solution can be obtained as the limit of the sequence . With respect to other techniques, such as Newton's method, two major advantages are its generality and the fact that it takes advantage of possible sparseness of . In this paper, we consider more general implicit equations of the form . Under mild assumptions on such an equation, we will show that it can be rewritten as a recursive equation. Preprint version only: If we are actually computing with analytic functions, then recursive equations also provide a systematic device for the computation of verified error bounds. We will show how to apply our results in this context.

Author: Joris van der Hoeven

Keywords: implicit equation, relaxed power series, algorithm

A.M.S. subject classification: 68W25, 42-04, 68W30, 65G20, 30B10

View: Html, TeXmacs, Pdf, BibTeX

Revisited AAECC Version: Html, TeXmacs, Pdf, BibTeX