Effective real numbers


In our talk we will present a C++ library for the computation with effective real numbers (i.e. numbers which come with an algorithm that takes on input and compute a dyadic approximation with ) included in Mmxlib.

This library consists of three parts: MPFR multiprecision arithmetic, an extension for interval/ball arithmetic, and the high-level real number interface. We will present several strategies for computing with intervals and effective real numbers and analyze them from a complexity point of view. If time permits it, we will also present work in progress to extend the Mmxlib library with a class for “effective analytic functions”.

Occasions: Schloss Dagstuhl, January 10, 2006

Documents: slideshow, TeXmacs source