ORIGAMI, une méthode organisation centrée de modélisation multi-agent de systèmes complexes

Inglés. In the field of agent-based modelling (ABM) for renewable resources management, actors (researchers, experts, stakeholders) are facing complex systems where social and physical dynamics take place within various interwoven levels of organization. Using conceptual modelling as a tool to formalize such systems facilitates their understanding as well as the sharing of the point of view of the various actors. In this context, we present an approach based on a multi-agent formalism called Agent-Group-Role. In order to implement this approach, we have developed a graphic methodology based mostly on UML and called ORIGAMI. This methodology was built extensively on the basis of a toy model and has been tested during a training experiment. Though some terminological problems occurred, the methodology has proven efficient in representing intervolved levels of organisation and their articulations and in explicating usually hidden assumptions.

Francés. Dans le champ de la gestion de ressources renouvelables, les acteurs (chercheurs, experts, partie prenantes) doivent appréhender des systèmes complexes où des dynamiques sociales et physiques s'inscrivent dans différents niveaux d'organisation imbriqués. Le recours à la formalisation des systèmes dans des modèles conceptuels permet de mettre en commun des points de vue et d'améliorer la compréhension du système. Dans ce contexte, nous proposons de recourir à une approche s'appuyant sur un formalisme multi-agent organisationnel nommé Agent-Groupe-Rôle. Afin de pouvoir mettre en ½uvre cette approche, nous avons conçu une méthodologie graphique basée principalement sur le langage UML et nommée ORIGAMI. Cette méthodologie a été développée extensivement sur la base d'un modèle-jouet et testée lors d'une expérience de formation. Le recours à cette méthodologie présente encore quelques difficultés notamment au niveau de la compréhension des terminologies utilisées, mais il permet d'expliciter des hypothèses d'habitude implicite et de mieux rendre compte des différents niveaux d'organisation imbriqués dans un système et de leurs articulations. / In the field of agent-based modelling (ABM) for renewable resources management, actors (researchers, experts, stakeholders) are facing complex systems where social and physical dynamics take place within various interwoven levels of organization. Using conceptual modelling as a tool to formalize such systems facilitates their understanding as well as the sharing of the point of view of the various actors. In this context, we present an approach based on a multi-agent formalism called Agent-Group-Role. In order to implement this approach, we have developed a graphic methodology based mostly on UML and called ORIGAMI. This methodology was built extensively on the basis of a toy model and has been tested during a training experiment. Though some terminological problems occurred, the methodology has proven efficient in representing intervolved levels of organisation and their articulations and in explicating usually hidden assumptions.