How Variables Graphs May Help to Correct Erroneous MAS Specifications - Archive ouverte HAL
Communication Dans Un Congrès Année : 2023

How Variables Graphs May Help to Correct Erroneous MAS Specifications

Résumé

This paper is situated in the context of multi-agent systems validation using theorem proving techniques. This document presents a preliminary case study, which is a part of a broader investigation aiming to explore whether such techniques could aid developers in detecting and characterizing errors in MAS specifications. Indeed, regardless of the verification system used (model checking or theorem proving), understanding the reason of a failure in order to correct the specification is most of the time rather difficult. In this article, we propose a method that may help in this task. This method relies on the variables (and their dependencies) that appear in proof obligations generated by GDT4MAS, a specification and verification method dedicated to Multi-Agent Systems. Graphs generated thanks to dependencies between variables occurring in an unproved theorem may indeed help to identify certain types of mistakes, giving a way to correct the specification.
Fichier principal
Vignette du fichier
articleIntelliSys2023.pdf (489.09 Ko) Télécharger le fichier
Origine Fichiers produits par l'(les) auteur(s)
Licence

Dates et versions

hal-04199918 , version 1 (08-09-2023)

Licence

Identifiants

  • HAL Id : hal-04199918 , version 1

Citer

Bruno Mermet, Gaële Simon. How Variables Graphs May Help to Correct Erroneous MAS Specifications. Intelligent Systems (Intellisys), Sep 2023, Amsterdam, France. ⟨hal-04199918⟩
41 Consultations
30 Téléchargements

Partager

More