Efficient On-the-Fly Computation of Weak Tau-Confluence - Inria - Institut national de recherche en sciences et technologies du numérique Accéder directement au contenu
Rapport (Rapport De Recherche) Année : 2009

Efficient On-the-Fly Computation of Weak Tau-Confluence

Résumé

The notion of tau-confluence provides a form of partial order reduction of Labelled Transition Systems (LTSs), by allowing to identify the tau-transitions whose execution does not alter the observable behaviour of the system. Several forms of tau-confluence adequate with branching bisimulation were studied in the literature, ranging from strong to weak ones according to the length of tau-transition sequences considered. Weak tau-confluence is more complex to compute than strong tau-confluence, but provides better LTS reductions. In this report, we aim at devising an efficient detection of weak tau-confluent transitions during an on-the-fly exploration of LTSs. To this purpose, we define and prove new encodings of several weak tau-confluence variants using alternation-free boolean equation systems (BESs), and we apply efficient local BES resolution algorithms to perform the detection. The resulting reduction module, developed within the CADP toolbox using the generic OPEN/CAESAR environment for LTS exploration, was experimented on numerous examples of large LTSs underlying communication protocols and distributed systems. These experiments assessed the efficiency of the reduction and allowed us to identify the best variants of weak tau-confluence that are useful in practice.
Fichier principal
Vignette du fichier
RR-7000.pdf (717.18 Ko) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)
Loading...

Dates et versions

inria-00407381 , version 1 (24-07-2009)

Identifiants

  • HAL Id : inria-00407381 , version 1

Citer

Radu Mateescu, Anton Wijs. Efficient On-the-Fly Computation of Weak Tau-Confluence. [Research Report] RR-7000, INRIA. 2009, pp.42. ⟨inria-00407381⟩
122 Consultations
126 Téléchargements

Partager

Gmail Facebook X LinkedIn More