RANA, a Real-Time Multi-agent System Simulator
Søren Vissing Jørgensen, Yves Demazeau, John C. T. Hallam · 2015
Multi-agent simulation of populations of acoustically-communicating animals presents challenges for existing MAS tools. We therefore present RANA, a new portable tool that enables large-scale simulation of precisely-timed models based on broadcast events. Agents are programmed in Lua, allowing for individualisation and abstraction while retaining efficiency. Events are managed by the C++ simulator core. Full run state can be recorded for post-processed visualisation or analysis. The new tool is demonstrated in three different cases: a mining robot simulation, which is purely action based; an agent-based setup that is verifies the high precision exhibited by RANAs simulation core; and a state-based firefly-like agent simulation that models real-time responses to fellow agents' signals, in which event propagation and reception affect the result of the simulation.