Memory-Efficient Tactics for Randomized LTL Model Checking

Kim Larsen, Doron Peled, Sean Sedwards

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review


We study model checking of LTL properties by means of random walks, improving on the efficiency of previous results. Using a randomized algorithm to detect accepting paths makes it feasible to check extremely large models, however a naive approach may encounter many non-accepting paths or require the storage of many explicit states, making it inefficient. We study here several alternative tactics that can often avoid these problems. Exploiting probability and randomness, we present tactics that typically use only a small fraction of the memory of previous approaches, storing only accepting states or an arbitrarily small number of “token” states visited during executions. Reducing the number of stored states generally increases the expected execution time until a counterexample is found, but we demonstrate that the trade-off is biased in favor of our tactics. By applying our memory-efficient tactics to scalable models from the literature, we show that the increase in time is typically less than proportional to the saving in memory and may be exponentially smaller.

Original languageEnglish
Title of host publicationVerified Software
Subtitle of host publicationTheories, Tools, and Experiments - 9th International Conference, VSTTE 2017, Revised Selected Papers
EditorsThomas Wies, Andrei Paskevich
PublisherSpringer Verlag
Number of pages18
ISBN (Print)9783319723075
StatePublished - 2017
Event9th International Working Conference on Verified Software: Theories, Tools, and Experiments, VSTTE 2017 - Heidelberg, Germany
Duration: 22 Jul 201723 Jul 2017

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume10712 LNCS


Conference9th International Working Conference on Verified Software: Theories, Tools, and Experiments, VSTTE 2017

All Science Journal Classification (ASJC) codes

  • Theoretical Computer Science
  • Computer Science(all)


Dive into the research topics of 'Memory-Efficient Tactics for Randomized LTL Model Checking'. Together they form a unique fingerprint.

Cite this