A model that can estimate a truck-trip matrix for a large metropolitan area is under development for the New York City region. The model is part of a new suite of demand-forecasting tools intended to bring the analytical capabilities of the region up to best practice worldwide. At the heart of the model is a gradient-based optimization technique that seeks to create a trip matrix that best fits a set of input data. The model allows great flexibility in both the types of data that can be introduced and the functional form of the objective function. The model is supported by two submodels. One estimates trip generations and attractions by zone, and the other provides link-use probabilities based on a multiple-path traffic assignment. The model and its application to the New York City regional network are described.
Abstract