NANAMar 6, 2012

Quadratic Interval Refinement for Real Roots

arXiv:1203.122729 citationsh-index: 13
Originality Incremental advance
AI Analysis

It provides a faster root refinement method for polynomials and well-behaved functions, combining the robustness of bisection with the speed of Newton's method.

The paper introduces a hybrid Bisection-Newton algorithm for refining real root intervals that achieves quadratic convergence without requiring derivative evaluation, using arbitrary precision rational arithmetic.

We present a new algorithm for refining a real interval containing a single real root: the new method combines characteristics of the classical Bisection algorithm and Newton's Iteration. Our method exhibits quadratic convergence when refining isolating intervals of simple roots of polynomials (and other well-behaved functions). We assume the use of arbitrary precision rational arithmetic. Unlike Newton's Iteration our method does not need to evaluate the derivative.

Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes