Monitorability of stochastic dynamical systems

  • Authors:
  • A. Prasad Sistla;Miloš Žefran;Yao Feng

  • Affiliations:
  • University of Illinois at Chicago;University of Illinois at Chicago;University of Illinois at Chicago

  • Venue:
  • CAV'11 Proceedings of the 23rd international conference on Computer aided verification
  • Year:
  • 2011

Quantified Score

Hi-index 0.00

Visualization

Abstract

Monitoring is an important run time correctness checking mechanism. This paper introduces the notions of monitorability and strong monitorability for partially observable stochastic systems, and gives necessary and sufficient conditions characterizing them. It also presents important decidability and complexity results for checking these properties for finite state systems. Furthermore, it presents general monitoring techniques for the case when systems are modeled as quantized probabilistic hybrid automata, and the properties are specified as safety or liveness automata. Experimental results showing the effectiveness of the methods are given.