how important are updating schemes in multi-agent strongsystemsstrong an.pdfVIP

How important are updating schemes in multi-agent systems? An illustration on a multi-turmite model. Nazim Fatès Vincent Chevrier INRIA Nancy Grand-Est - LORIA Nancy University UHP - LORIA 615 rue du Jardin Botanique Campus Scientifique — B.P. 239 54 600 Villers-lès-Nancy ; France 54 506 Vandoeuvre-lès-Nancy, France nazim.fates@loria.fr vincent.chevrier@loria.fr ABSTRACT follows the influence-reaction principles [17]. This propo- It is to date an open question to know how the updating sition was then precised by a framework which represents methods affect the evolution of a multi-agent system. This the environment as “a dynamical system that encapsulates question has been tackled for various complex systems such and regulates its own dynamism” [14]. These propositions as cellular automata, Boolean networks, neural networks aim at removing the ambiguities of description that exist but little is known for multi-agent systems, especially for when a multi-agent system is described only in an informal the models with a complex behaviour which emerges from way, or when it uses a specific simulation platform where the simple local rules. This paper focuses on a multi-turmite simulation engine is only partially known. A first step was undertaken by Muller who proposed a formalism based on