Canonical Forms and Widening for Two Variables Per Inequality Systems
Brain, M. & Howe, J. ORCID: 0000-0001-8013-6941 (2025).
Canonical Forms and Widening for Two Variables Per Inequality Systems.
Mathematics in Computer Science,
Abstract
Program verification using abstract interpretation involves the symbolic calculation of fixpoints over lattices. Integral to these fixpoint calculations is widening, an operation that trades precision for guarantees of termination. Abstract interpretation often works with lattices of convex sets of points in n-dimensional space, represented by sets of linear inequalities. When the form of these inequalities is restricted these are known as weakly relational domains. This paper addresses weakly relational domains, the detail of their representation and the way in which widening is applied, including study of how the closure operators used in weakly relational domains interact with widening, and considers how sequences of constraints might be widened. Satisfiability checking for numeric constraints is one tool used in this work.
Publication Type: | Article |
---|---|
Additional Information: | This version of the article has been accepted for publication, after peer review and is subject to Springer Nature’s AM terms of use, but is not the Version of Record and does not reflect post-acceptance improvements, or any corrections. The Version of Record will be available online at http://link.springer.com/journal/11786 |
Publisher Keywords: | Abstract interpretation, widening, weakly relational domains |
Subjects: | Q Science > QA Mathematics Q Science > QA Mathematics > QA75 Electronic computers. Computer science |
Departments: | School of Science & Technology School of Science & Technology > Computer Science |
SWORD Depositor: |
![[thumbnail of main.pdf]](https://openaccess.city.ac.uk/style/images/fileicons/text.png)
This document is not freely accessible due to copyright restrictions.
To request a copy, please use the button below.
Request a copyExport
Downloads
Downloads per month over past year