Model checking for probability and time: from theory to practice

  • Authors:
  • Marta Kwiatkowska

  • Affiliations:
  • -

  • Venue:
  • LICS '03 Proceedings of the 18th Annual IEEE Symposium on Logic in Computer Science
  • Year:
  • 2003

Quantified Score

Hi-index 0.01

Visualization

Abstract

Probability features increasingly often in software andhardware systems: it is used in distributed coordinationand routing problems, to model fault-tolerance and performance,and to provide adaptive resource managementstrategies. Probabilistic model checking is an automaticprocedure for establishing if a desired property holds in aprobabilistic model, aimed at verifying probabilistic specificationssuch as "leader election is eventually resolvedwith probability 1", "the chance of shutdown occurring isat most 0.01%", and "the probability that a message willbe delivered within 30ms is at least 0.75". A probabilisticmodel checker calculates the probability of a given temporallogic property being satisfied, as opposed to validity.In contrast to conventional model checkers, which relyon reachability analysis of the underlying transition systemgraph, probabilistic model checking additionally involvesnumerical solutions of linear equations and linear programmingproblems. This paper reports our experience withimplementing PRISM (www.cs.bham.ac.uk/~dxp/prism/), a Probabilistic Symbolic Model Checker, demonstratesits usefulness in analysing real-world probabilisticprotocols, and outlines future challenges for this researchdirection.