s'authentifier
version française rss feed
inria-00000524, version 1
Voir la fiche détaillée  BibTeX  EndNote  TEI  RefWorks
Extending Decentralized Discrete-Event Modelling to Diagnose Reconfigurable Systems
Alban Grastien () 1, Marie-Odile Cordier () a1, Christine Largouët 1
(2004)
Icone de reconf.ps
Icone de reconf.pdf
15 international workshop on principles of diagnosis (2004)
On-line reconfiguration is the ability to rearrange dynamically the elements of a system to accommodate failure events or new requirements. Due to the modular representation, decentralized discrete-event approach, recently proposed for the diagnosis of systems, is particularly well suited to the diagnosis of reconfigurable systems. The contribution of this article is to extend our decentralized approach to reconfigurable discrete-event systems. A first step in this direction is to extend the way a decentralized system is modelled. The idea consists in modelling separately the behavior of the components and the system topology. A second step is to formally define what is a reconfiguration. A property of reconfiguration, that we call \ital{safety}, is identified to be important. When satisfied, we show that our decentralized diagnosis approach can easily be extended to reconfigurable systems.
a –  Université Rennes I
1 :  DREAM (INRIA - IRISA)
CNRS : UMR6074 – INRIA – Institut National des Sciences Appliquées de Rennes – Université de Rennes 1
Informatique/Intelligence artificielle