Publication detail

Reachability Computation for Switching Diffusions:Finite Abstractions with Certifiable and Tuneable Precision

LAURENTI, L. ABATE, A. BORTOLUSSI, L. CARDELLI, L. ČEŠKA, M. KWIATKOWSKA, M.

Original Title

Reachability Computation for Switching Diffusions:Finite Abstractions with Certifiable and Tuneable Precision

Type

conference paper

Language

English

Original Abstract

We consider continuous time stochastic hybrid systems with no resets and continuous dynamics described by linear stochastic differential equations -- models also known as switching diffusions. We show that for this class of models reachability (and dually, safety) properties can be studied on an abstraction defined in terms of a discrete time and finite space Markov chain (DTMC), with provable error bounds. The technical contribution of the paper is a characterization of the uniform convergence of the time discretization of such stochastic processes with respect to safety properties. This allows us to newly provide a complete and sound numerical procedure for reachability and safety computation over switching diffusions.

Keywords

Switching diffusions; stochastic hybrid models; reachability and safety analysis; finite abstractions; time and space dis- cretisation; numerical computations 

Authors

LAURENTI, L.; ABATE, A.; BORTOLUSSI, L.; CARDELLI, L.; ČEŠKA, M.; KWIATKOWSKA, M.

Released

13. 3. 2017

Publisher

Association for Computing Machinery

Location

New York

ISBN

978-1-4503-4590-3

Book

Proceedings of the 20th ACM International Conference on Hybrid Systems: Computation and Control

Edition

ACM

Pages from

55

Pages to

64

Pages count

10

BibTex

@inproceedings{BUT144419,
  author="Luca {Laurenti} and Alessandro {Abate} and Luca {Bortolussi} and Luca {Cardelli} and Milan {Češka} and Marta {Kwiatkowska}",
  title="Reachability Computation for Switching Diffusions:Finite Abstractions with Certifiable and Tuneable Precision",
  booktitle="Proceedings of the 20th ACM International Conference on Hybrid Systems: Computation and Control",
  year="2017",
  series="ACM",
  pages="55--64",
  publisher="Association for Computing Machinery",
  address="New York",
  doi="10.1145/3049797.3049812",
  isbn="978-1-4503-4590-3"
}