Day-to-day dynamic traffic assignment model with variable message signs and endogenous user compliance

TL;DR

This paper develops a comprehensive day-to-day traffic assignment model incorporating variable message signs and driver learning, revealing long-term adverse effects and equilibrium states in traffic flow and compliance.

Contribution

It introduces a dual-time-scale model that endogenizes traffic dynamics, route choices, and VMS compliance, integrating driver learning and adaptive behavior.

Findings

VMS can have a rebound effect increasing congestion in the long run

An equilibrium state exists where traffic flow and VMS compliance stabilize

VMS influences route choices and traffic dynamics significantly

Abstract

This paper proposes a dual-time-scale, day-to-day dynamic traffic assignment model that takes into account variable message signs (VMS) and its interactions with drivers' travel choices and adaptive learning processes. The within-day dynamic is captured by a dynamic network loading problem with en route update of path choices influenced by the VMS; the day-to-day dynamic is captured by a simultaneous route-and-departure-time adjustment process that employs bounded user rationality. Moreover, we describe the evolution of the VMS compliance rate by modeling drivers' learning processes. We endogenize traffic dynamics, route and departure time choices, travel delays, and VMS compliance, and thereby captur their interactions and interdependencies in a holistic manner. A case study in the west end of Glasgow is carried out to understand the impact of VMS has on road congestion and route choices in both the short and long run. Our main findings include an adverse effect of the VMS on the network performance in the long run (the 'rebound' effect), and existence of an equilibrium state where both traffic and VMS compliance are stabilized.