Computing with Intervals : the Ultimate Symbolic Computation ?

Yves Papegay 1 David Daney 1
1 COPRIN - Constraints solving, optimization and robust interval analysis
CRISAM - Inria Sophia Antipolis - Méditerranée , ENPC - École des Ponts ParisTech
Abstract : Intervals have been considered by mathematicians for a while, in the study of numerical functions and numerical analysis, But only in the 1960s Moore had the Idea of replacing real numbers by intervals in computations and to study the resulting alternative arithmetics. In the last few decades, it has been used for achieving certified computations, but more surprisingly by numericians designing a lot of methods and algorithms for root finding, systems solving, local and global optimization, etc. Even if they are not yet widely implemented nor well known, these methods have shown their efficiency by succeeding on some challenging problems. In Mathematica, a basic interval arithmetic is implemented. A few classical interval analysis algorithms, namely an extension of the Newton method for one variable root finding, are also available in MathSource in the NumericalMath'IntervalRoots' package. An extensive interface to a powerful C++ library of interval analysis was presented at the Wolfram Technology Conference 2005 by the authors.
Type de document :
Communication dans un congrès
Wolfram Technology Conference, 2007, Illinois, United States. 2007
Liste complète des métadonnées

https://hal.inria.fr/hal-00989873
Contributeur : David Daney <>
Soumis le : lundi 12 mai 2014 - 15:19:58
Dernière modification le : samedi 27 janvier 2018 - 01:31:32

Identifiants

  • HAL Id : hal-00989873, version 1

Collections

Citation

Yves Papegay, David Daney. Computing with Intervals : the Ultimate Symbolic Computation ?. Wolfram Technology Conference, 2007, Illinois, United States. 2007. 〈hal-00989873〉

Partager

Métriques

Consultations de la notice

120