How to Build the Best Macroscopic Description of your Multi-agent System? Application to News Analysis of International Relations

Robin Lamarche-Perrin 1 Yves Demazeau 2 Jean-Marc Vincent 3
2 MAGMA
LIG - Laboratoire d'Informatique de Grenoble
3 MESCAL - Middleware efficiently scalable
Inria Grenoble - Rhône-Alpes, LIG - Laboratoire d'Informatique de Grenoble
Abstract : The design and debugging of large-scale MAS require abstraction tools in order to work at a macroscopic level of description. Agent aggregation provides such abstractions by reducing the microscopic description complexity. Since it leads to an information loss, such a key process may be extremely harmful if poorly executed. This research report presents measures inherited from information theory (Kullback-Leibler divergence and Shannon entropy) to evaluate ab- stractions and to provide the experts with feedbacks regarding the generated descriptions. Several evaluation techniques are applied to the spatial aggregation of an agent-based model of international rela- tions. The information from on-line newspapers constitutes a complex microscopic description of agent states. Our approach is able to evalu- ate geographical abstractions used by experts and to deliver them with e cient and meaningful macroscopic descriptions of the world state.
Type de document :
Rapport
[Research Report] RR-LIG-035, 2013, pp.18
Liste complète des métadonnées


https://hal.inria.fr/hal-00947933
Contributeur : Sylvie Pesty <>
Soumis le : lundi 17 février 2014 - 15:06:03
Dernière modification le : samedi 17 septembre 2016 - 01:38:19
Document(s) archivé(s) le : dimanche 9 avril 2017 - 12:53:42

Fichier

RR-LIG-035_orig.pdf
Fichiers produits par l'(les) auteur(s)

Identifiants

  • HAL Id : hal-00947933, version 1

Collections

Citation

Robin Lamarche-Perrin, Yves Demazeau, Jean-Marc Vincent. How to Build the Best Macroscopic Description of your Multi-agent System? Application to News Analysis of International Relations. [Research Report] RR-LIG-035, 2013, pp.18. <hal-00947933>

Partager

Métriques

Consultations de
la notice

200

Téléchargements du document

153