Ambient Intelligence as a Never-Ending Self-Organizing Process: Analysis and Experiments
Résumé
Our team has been working for several years on building adaptive systems using self-organising mechanisms following a specific approach we called the AMAS1 Theory. Its main originality is that it enables artificial systems to show relevant emergent behaviours by focusing on local cooperative interactions among the agents. This article aims at showing the relevance of this approach specifically, and more generally of any self-organisation approach, for Ambient Intelligence. For this we analyse and discuss AmI problems in the light of our experience in complex systems, as well as show encouraging results in a first experiment in a kind of AmI system (a service providing network of agents).
Domaines
Système multi-agents [cs.MA]
Origine : Fichiers éditeurs autorisés sur une archive ouverte