Using GPU for Multi-agent Multi-scale Simulations

Guillaume Laville (1) , Kamel Mazouzi (1) , Christophe Lang (1) , Nicolas Marilleau (2) , Laurent Philippe (1)

1 FEMTO-ST - Franche-Comté Électronique Mécanique, Thermique et Optique - Sciences et Technologies (UMR 6174)
2 IRD - Institut de Recherche pour le Développement

Multi-Agent System (MAS) is an interesting way to create models and simulators and is widely used to model complex systems. As the complex system community tends to build up larger models to fully represent real systems, the need for computing power raise significantly. Thus MAS often lead to long computing intensive simulations. Parallelizing such a simulation is complex and it execution requires the access to large computing resources. In this paper, we present the adaptation of a MAS system, Sworm, to a Graphical Processing Unit.We show that such an adaptation can improve the performance of the simulator and advocate for a more wider use of the GPU in Agent Based Models in particular for simple agents.

Domaines: Réseaux et télécommunications [cs.NI]; Calcul parallèle, distribué et partagé [cs.DC]; Informatique ubiquitaire

Guillaume Laville, Kamel Mazouzi, Christophe Lang, Nicolas Marilleau, Laurent Philippe. Using GPU for Multi-agent Multi-scale Simulations. 9th International Conference on Distributed Computing and Artificial Intelligence (DCAI'12), Mar 2012, Salamanque, Spain. pp.197-204, ⟨10.1007/978-3-642-28765-7_23⟩. ⟨hal-01304681⟩