The Effect of Modern Traffic Information on Braess' Paradox

TL;DR

This paper investigates whether modern traffic information, like navigation apps, can mitigate Braess' paradox in urban networks by analyzing driver behavior with realistic traffic data.

Contribution

It introduces a microscopic traffic model to assess the impact of real-time information on the occurrence of Braess' paradox in classical network scenarios.

Findings

Traffic information can reduce but not eliminate Braess' paradox.

Driver behavior based on realistic info differs from theoretical user optima.

The paradox persists under certain conditions even with modern traffic info.

Abstract

Braess' paradox has been shown to appear rather generically in many systems of transport on networks. It is especially relevant for vehicular traffic where it shows that in certain situations building a new road in an urban or highway network can lead to increased average travel times for all users. Here we address the question whether this changes if the drivers (agents) have access to traffic information as available for modern traffic networks, i.e. through navigation apps and or personal experiences in the past. We study the effect of traffic information in the classical Braess network, but using a microscopic model for the traffic dynamics, to find out if the paradox can really be observed in such a scenario or if it only exists in some theoretically available user optima that are never realized by drivers that base their route choice decisions intelligently upon realistic traffic information. We address this question for different splits of the two information types.