Language-based abstraction refinement for hybrid system verification

  • Authors:
  • Felix Klaedtke;Stefan Ratschan;Zhikun She

  • Affiliations:
  • ETH Zurich, Computer Science Department, Zurich, Switzerland;Institute of Computer Science, Czech Academy of Sciences, Prague, Czech Republic;Max-Planck-Institut für Informatik, Saarbrücken, Germany

  • Venue:
  • VMCAI'07 Proceedings of the 8th international conference on Verification, model checking, and abstract interpretation
  • Year:
  • 2007

Quantified Score

Hi-index 0.01

Visualization

Abstract

The standard counterexample-guided abstraction-refinement (CEGAR) approach uses finite transition systems as abstractions of concrete systems. We present an approach to represent and refine abstractions of infinite-state systems that uses regular languages instead of finite transition systems. The advantage of using languages over transition systems is that we can store more fine-grained information in the abstraction and thus reduce the number of abstract states. Based on this language-based approach for cegar, we present new abstraction-refinement algorithms for hybrid system verification. Moreover, we evaluate our approach by verifying various non-linear hybrid systems.