Ranking abstractions

  • Authors:
  • Aziem Chawdhary;Byron Cook;Sumit Gulwani;Mooly Sagiv;Hongseok Yang

  • Affiliations:
  • Queen Mary, University of London;Microsoft Research;Microsoft Research;Tel Aviv University;Queen Mary, University of London

  • Venue:
  • ESOP'08/ETAPS'08 Proceedings of the Theory and practice of software, 17th European conference on Programming languages and systems
  • Year:
  • 2008

Quantified Score

Hi-index 0.02

Visualization

Abstract

We propose an abstract interpretation algorithm for proving that a program terminates on all inputs. The algorithm uses a novel abstract domain which uses ranking relations to conservatively represent relations between intermediate program states. One of the attractive aspects of the algorithm is that it abstracts information that is usually not important for proving termination such as program invariants and yet it distinguishes between different reasons for termination which are not usually maintained in existing abstract domains. We have implemented a prototype of the algorithm and shown that in practice it is fast and precise.