Organization Detection Using Emergent Computing

Cyrille Bertelle (1) , Antoine Dutot (1) , Frédéric Guinand (1) , Damien Olivier (1)

1 LITIS - Laboratoire d'Informatique, de Traitement de l'Information et des Systèmes

Organization is a central concept in systems. In this paper an ant algorithm for detecting organizations is presented. In a discrete-time context, at each time-step, an organization corresponds to a set of closely interacting entities in a system. This system is mapped to a graph where nodes represent entities and edges represent interrelations. Several colonies of ants compete, and inside each colony, ants collaborate in order to colonize the graph. Detected organizations emerge from the global behavior of the ants. The proposed approach is compared to other methods on a graph where the organizations are already known. It is then tested on two real world graphs studied in the related literature.

Mots clés: Organization; community; dynamic graph; dynamic network; ant algorithm

Domaines: Adaptation et Systèmes auto-organisés [nlin.AO]; Intelligence artificielle [cs.AI]

Cyrille Bertelle, Antoine Dutot, Frédéric Guinand, Damien Olivier. Organization Detection Using Emergent Computing. International Transactions on Systems Science and Applications, 2006, 2 (1), pp.61-69. ⟨hal-00430530⟩