English

Fast evaluation and root finding for polynomials with floating-point coefficients

Symbolic Computation 2023-02-14 v1

Abstract

Evaluating or finding the roots of a polynomial f(z)=f0++fdzdf(z) = f_0 + \cdots + f_d z^d with floating-point number coefficients is a ubiquitous problem. By using a piecewise approximation of ff obtained with a careful use of the Newton polygon of ff, we improve state-of-the-art upper bounds on the number of operations to evaluate and find the roots of a polynomial. In particular, if the coefficients of ff are given with mm significant bits, we provide for the first time an algorithm that finds all the roots of ff with a relative condition number lower than 2m2^m, using a number of bit operations quasi-linear in the bit-size of the floating-point representation of ff. Notably, our new approach handles efficiently polynomials with coefficients ranging from 2d2^{-d} to 2d2^d, both in theory and in practice.

Keywords

Cite

@article{arxiv.2302.06244,
  title  = {Fast evaluation and root finding for polynomials with floating-point coefficients},
  author = {Rémi Imbach and Guillaume Moroz},
  journal= {arXiv preprint arXiv:2302.06244},
  year   = {2023}
}
R2 v1 2026-06-28T08:38:35.491Z