Dynamic Network Loading Model with Explicit Traffic Wave Tracking

Abstract

A dynamic network loading (DNL) model is presented: it can be used both for dynamic traffic assignment (DTA) and for an accurate description of traffic. The proposed DNL model is composed of (a) the link model based on the Lighthill-Whitham-Richards macroscopic first-order model solved with the wave-tracking method, (b) a new intersection model that generalizes the Daganzo macroscopic merge model to complex intersections, and (c) a traffic signal model that represents the mean effects of stage alternation on traffic in terms of delays and of capacity restrictions. Both the wave-tracking method and the traffic signal model are applied in a network context for the first time. The model can be consistently solved with various precision scales: low-precision scales to quickly provide a good estimation of travel times on the network for DTA purposes, and high-precision scales for accurate descriptions of traffic with all the desired modeling details.