Paul Brunet ; Damien Pous - Petri Automata

lmcs:3125 - Logical Methods in Computer Science, September 26, 2017, Volume 13, Issue 3 - https://doi.org/10.23638/LMCS-13(3:33)2017
Petri AutomataArticle

Authors: Paul Brunet ; Damien Pous

    Kleene algebra axioms are complete with respect to both language models and binary relation models. In particular, two regular expressions recognise the same language if and only if they are universally equivalent in the model of binary relations. We consider Kleene allegories, i.e., Kleene algebras with two additional operations and a constant which are natural in binary relation models: intersection, converse, and the full relation. While regular languages are closed under those operations, the above characterisation breaks. Putting together a few results from the literature, we give a characterisation in terms of languages of directed and labelled graphs. By taking inspiration from Petri nets, we design a finite automata model, Petri automata, allowing to recognise such graphs. We prove a Kleene theorem for this automata model: the sets of graphs recognisable by Petri automata are precisely the sets of graphs definable through the extended regular expressions we consider. Petri automata allow us to obtain decidability of identity-free relational Kleene lattices, i.e., the equational theory generated by binary relations on the signature of regular expressions with intersection, but where one forbids unit. This restriction is used to ensure that the corresponding graphs are acyclic. We actually show that this decision problem is EXPSPACE-complete.


    Volume: Volume 13, Issue 3
    Published on: September 26, 2017
    Accepted on: August 7, 2017
    Submitted on: August 3, 2017
    Keywords: Computer Science - Logic in Computer Science,68Q45,F.4.3,F.1.1
    Funding:
      Source : OpenAIRE Graph
    • PROJET AVENIR LYON SAINT-ETIENNE; Funder: French National Research Agency (ANR); Code: ANR-11-IDEX-0007
    • Coinduction for Verification and Certification; Funder: European Commission; Code: 678157

    1 Document citing this article

    Consultation statistics

    This page has been seen 1394 times.
    This article's PDF has been downloaded 729 times.