City traffic jam relief by stochastic resonance

Abstract

We simulate traffic in a city by means of the evolution of a row of interacting cars, using a cellular automaton model, in a sequence of traffic lights synchronized by a “green wave”. When our initial condition is a small density jammed state, its evolution shows the expected scaling laws close to the synchronization resonance, with a uniform car density along the street. However, for an initial large density jammed state, we observe density variations along the streets, which results in the breakdown of the scaling laws. This spatial disorder corresponds to a different attractor of the system. As we include velocity perturbations in the dynamics of the cars, all these attractors converge to a statistically equivalent system for all initial jammed densities. However, this emergent state shows a stochastic resonance-like behavior in which the average traffic velocity increases with respect to that of the system without noise, for several initial jammed densities. This result may help in the understanding of dynamics of traffic jams in cities.