On the Expressiveness of Parametric Timed Automata - Archive ouverte HAL
Communication Dans Un Congrès Année : 2016

On the Expressiveness of Parametric Timed Automata

Résumé

Parametric timed automata (PTAs) are a powerful formalism to reason about, model and verify real-time systems in which some constraints are unknown, or subject to uncertainty. In the literature, PTAs come in several variants: in particular the domain of parameters can be integers or rationals, and can be bounded or not. Also clocks can either be compared only to a single parameter, or to more complex linear expressions. Yet we do not know how these variants compare in terms of expressiveness, and even the notion of expressiveness for para-metric timed models does not exist in the literature. Furthermore, since most interesting problems are undecidable for PTAs, subclasses, such as L/U-PTAs, have been proposed for which some of those problems are de-cidable. It is not clear however what can actually be modeled with those restricted formalisms and their expressiveness is thus a crucial issue. We therefore propose two definitions for the expressiveness of parametric timed models: the first in terms of all the untimed words that can be generated for all possible valuations of the parameters, the second with the additional information of which parameter valuations allow which word, thus more suitable for synthesis issues. We then use these two definitions to propose a first comparison of the aforementioned PTA variants.
Fichier principal
Vignette du fichier
andre-FORMATS-16.pdf (360.58 Ko) Télécharger le fichier
Origine Fichiers produits par l'(les) auteur(s)
Loading...

Dates et versions

hal-02538061 , version 1 (09-04-2020)

Identifiants

Citer

Étienne André, Didier Lime, Olivier Henri Roux. On the Expressiveness of Parametric Timed Automata. 14th International Conference on Formal Modelling and Analysis of Timed Systems (FORMATS 2016), Aug 2016, Québec, Canada. ⟨10.1007/978-3-319-44878-7_2⟩. ⟨hal-02538061⟩
38 Consultations
85 Téléchargements

Altmetric

Partager

More