FOSSIL: a software tool for the formal synthesis of lyapunov functions and barrier certificates using neural networks

Alessandro Abate, Daniele Ahmed, Alec Edwards, M. Giacobbe, Andrea Peruffo

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

This paper accompanies FOSSIL: a software tool for the synthesis of Lyapunov functions and of barrier certificates (or functions) for dynamical systems modelled as differential equations. Lyapunov functions are formal certificates for stability analysis, whereas barrier functions are formal certificates for the safety of dynamical models. FOSSIL is sound and automatic thanks to a counterexample-guided inductive synthesis loop. This method exploits the flexibility of candidate functions generated by training neural network templates, the formal assertions provided by a verifier (namely, an SMT solver), and finally new procedures to ease the exchange of information between the two mentioned components. We endow the tool with features of usability, scalability, and robustness---all of which are showcased on benchmarks.
Original languageEnglish
Title of host publicationHSCC '21
Subtitle of host publicationProceedings of the 24th International Conference on Hybrid Systems: Computation and Control
PublisherAssociation for Computing Machinery (ACM)
Number of pages11
ISBN (Print)9781450383394
DOIs
Publication statusPublished - 19 May 2021

Keywords

  • Lyapunov functions
  • barrier certificates
  • neural networks
  • CEGIS
  • SAT modulo theory

Fingerprint

Dive into the research topics of 'FOSSIL: a software tool for the formal synthesis of lyapunov functions and barrier certificates using neural networks'. Together they form a unique fingerprint.

Cite this