Microscopic Simulation Replicates the Capacity Drop Phenomenon

https://doi.org/10.1016/j.procs.2018.04.088Get rights and content
Under a Creative Commons license
open access

Abstract

Microscopic simulation is used widely to analyze traffic networks. The outputs of such models, such as travel time and delay, depend in large part on the assumed parameter values. These parameters determine the traffic capacity, whose interpretation is not straightforward as the maximum discharge flow of a bottleneck in a freeway can be lower under congested conditions than in an uncongested situation. This phenomenon is often referred to as the capacity drop, and must be accurately portrayed by microsimulation models to simulate traffic properly for long periods of time. This is specifically problematic when the total demand is between the maximum congested and uncongested flow rates as travel times may potentially be over- or underestimated depending on prevailing traffic conditions. Here, we simulate traffic flow in a bottleneck with a combined car-following and lane-changing model and use data from the 405 freeway near Irvine, CA to calibrate the model parameters. The calibrated model tracks closely the observed data and predicts accurately the capacity drop phenomenon.

Keywords

Microsimulation
Capacity Drop
Traffic Flow
Calibration

Cited by (0)