Braess Paradox in Networks of Stochastic Microscopic Traffic Models
The Braess Paradox describes a counterintuitive situation that can arise in traffic networks which are used by selfish drivers who want to minimize their own traveltimes. For specific combinations of demand and traveltime functions of the roads in such networks the addition of a new road, resulting in a per se faster origin–destination connection, can lead to higher traveltimes for all network users. As an important addition to previous research on the paradox which focused on deterministic macroscopic models of traffic in road networks, we study its occurrence employing a stochastic microscopic traffic model—the totally asymmetric exclusion process (TASEP). We find that the paradox also occurs in these more realistic traffic models and that, depending on the degree of stochasticity, it dominates large parts of the phase space.
Financial support by the Deutsche Forschungsgesellschaft (DFG) under grant SCHA 636/8-2 and the Bonn-Cologne Graduate School of Physics and Astronomy (BCGS) is gratefully acknowledged. Monte Carlo simulations were carried out on the CHEOPS (Cologne High Efficiency Operating Platform for Science) cluster of the RRZK (University of Cologne).