DisSimJADE: a framework for the development of agent-based distributed simulation systems

The adoption of an agent-based approach that incorporates intelligence, adaptation and learning abilities has proved to significantly increase the realism and the accuracy of the simulation. Simulation systems of such a kind, however, require computational resources that might be considerable for a single agent, so to become unfeasible when the number of simulated agents scales up. A distributed environment is thus needed to allow the execution of such simulation systems, particularly in the case of scenarios populated by a large number of agents. Building an agent-based distributed simulation system, however, requires both specific expertise and knowledge of distributed simulation standards and a non-negligible amount of effort to develop ad-hoc components. This paper introduces a simulation framework named DisSimJADE, which enables the incorporation of distributed simulation facilities into existing agent-based systems. DisSimJADE is built on top of the popular agent-based framework JADE and allows to define agent-based simulation systems that can be transparently executed either in a local or distributed, therefore bringing significant savings in terms of effort and development time. In addition, DisSimJADE provides a uniform interface to the JADE framework, which further facilitates the production of distributed simulation systems to developers of JADE-based multi-agent systems.