JREP: Extending Repast Simphony for JADE Agent Behavior Components

Abstract

When modeling and simulating agent communities, one is usually interested in the macro effects that result at the system level from the interaction of individuals. However, when developing a scenario, designers need to specify the individual agents' behavior, their communication and organizational structure at the micro-level. Existing agent frameworks focus on either the macro or the micro perspective: Repast Simphony is an example of the former, JADE of the latter. We propose JREP, a novel integration of JADE and Repast Simphony that efficiently combines the macro and micro perspective with an interaction layer. It allows to see not only the overall system behavior, but also the individual together with its interests, goals and the communication to others for local coordination and cooperation. Scheduling of the agents, (time) synchronization and the registration of new agents with the environment has been solved. An agent-based airport scenario is introduced a proof of concept for modeling and simulating in the JREP platform, for validating scalability and performance properties, a simple coin flip scenario is described.