Using Queueing Networks to Approximate Pedestrian Simulations

Pedestrian or crowds simulation is a complex and expensive task that involves a plethora of technologies. In certain exigent environments, for example, whether we want to use a complex simulation in a machine learning system, in real-time decision making or when the user does not need the details of the simulation, this computational cost may not be desirable. Having simpler models is useful if you want to use these simulations in those exigent environments or we just want to obtain approximate calculations of these simulations quickly. In this paper, we propose a simplified model of simulation based on a network of configurable queues that helps us to approximate the results of a complex simulation in a very short time, while maintaining a high representativeness of the real simulation.