Hostname: page-component-cd9895bd7-gxg78 Total loading time: 0 Render date: 2024-12-25T20:18:04.234Z Has data issue: false hasContentIssue false

Domain theory and differential calculus (functions of one variable)

Published online by Cambridge University Press:  16 November 2004

ABBAS EDALAT
Affiliation:
Department of Computing, Imperial College, London, UK Email: ae@doc.ic.ac.uk
ANDRÉ LIEUTIER
Affiliation:
Dassault Systemes Provence and LMC/IMAG, Aix-en-Provence and Grenoble, France

Abstract

We introduce a domain-theoretic framework for differential calculus. We define the set of primitive maps as well as the derivative of an interval-valued Scott continuous function on the domain of intervals, and show that they are dually related, providing an extension of the classical duality of differentiation and integration as in the fundamental theorem of calculus. It is shown that, for locally Lipschitz functions of a real variable, the domain-theoretic derivative coincides with the Clarke's derivative. We then construct a domain for differentiable real-valued functions of a real variable by pairing consistent information about the function and information about its derivative. The set of classical $C^1$ functions, equipped with its $C^1$ norm, is embedded into the set of maximal elements of this countably based, bounded complete continuous domain. This domain also provides a model for the differential properties of piecewise $C^1$ functions, locally Lipschitz functions and more generally of all continuous functions. We prove that consistency of function information and derivative information is decidable on rational step functions, which shows that our domain can be given an effective structure. We thus obtain a data type for differential calculus. As an immediate application, we present a domain-theoretic formulation of Picard's theorem, which provides a data type for solving differential equations.

Type
Paper
Copyright
© 2004 Cambridge University Press

Access options

Get access to the full version of this content by using one of the access options below. (Log in options will check for institutional or personal access. Content may require purchase if you do not have access.)