Complex network representation of multiagent systems with cooperative and competitive interactions

Leonidas Facundo Caram, Cesar Federico Caiafa, Araceli Noemi Proto

Abstract


The dynamic behavior of Multi-Agent Systems (MAS) is analyzed in the context of a modified Lotka-Volterra model. The interaction strength is determined by the difference of agent sizes: as the difference increases, the interaction is weaker. Competitive and cooperative scenarios are analyzed, showing clusters of agents in the stationary state. However, meantime in the competitive scenario the agent sizes are constrained to be non greater than the capacity value (β = 1), in the cooperative scenario, they are allowed to exceed such capacity making clear the advantages of cooperation. The complex network representation is introduced in order to enhance the role of agent sizes and their one-on-one interactions in the dynamic behavior of the system.


Keywords


Dynamic Interaction, Multiagent System, Peer-to-Peer interaction.

Full Text:

PDF


DOI: http://dx.doi.org/10.1478/AAPP.92S1B2