Developping Agents Populations with Ethogenetics

Samuel Landau (1) , Sébastien Picault (2, 3, 1)

1 OASIS - Objets et Agents pour Systèmes d'Information et Simulation
2 LIFL - Laboratoire d'Informatique Fondamentale de Lille
3 SMAC - Systèmes Multi-Agents et Comportements

This paper proposes the use of developmental and evolutionary metaphors to automatically design organization in adaptive multiagent systems. We propose therefore a novel evolutionary approach to design agent behaviors in situated MAS, called "Ethogenetics". We also describe ATNoSFERES, a framework implementing the concepts of Ethogenetics and discuss its specific properties.

Mots clés: Evolution; Artificial Life; adaptive multi-agent systems; Development; Ethogenetics

Domaines: Système multi-agents [cs.MA]; Apprentissage [cs.LG]; Intelligence artificielle [cs.AI]

Samuel Landau, Sébastien Picault. Developping Agents Populations with Ethogenetics. 1st International Workshop on Radical Agent Concepts (WRAC'2002), Jan 2002, McLean, VA, United States. pp.61-70, ⟨10.1007/978-3-540-45173-0_4⟩. ⟨hal-00860467⟩