Running Agent-based Models on a Discrete-event Simulator

One of the main obstacles hindering the use of large-scale agent-based simulation in practice is its scalability. The ability to run realistic and complex agent-based models is desirable to provide empirically and practically useful results. One of the suggested solutions is to run the agentbased models on top of a scalable parallel discrete-event simulation engine. This proposal raises a question whether an equivalent discrete-event model can be built for any agent-based model. This paper proves that an equivalent discrete-event model can be found for any agent-based model that conforms to a given specification. I show that a translator can be built to convert the agent-based model into an equivalent discrete-event model automatically and transparently. The advantage of this approach is that modellers do not need to change their modelling paradigm and at the same time a highly scalable parallel discrete-event simulator can be used to run the model.