Deriving land use limits from infrastructure capacity: An artificial neural network approach

Given that it is extremely unlikely that the coming years will witness major capacity-expansion projects, transportation planners will now need to view the existing infrastructure as fixed, and start thinking about how much development the current system can sustain. This line of thinking, which involves deriving land-use limits from infrastructure capacity, requires solving the inverse of the typical transportation-planning problem. While there have been some reported studies in the literature that examined how to reverse the direction of the transportation planning process, the errors of the developed models were rather significant. In the current paper, we describe our efforts toward building on these earlier attempts in an effort to develop refined tools for solving the inverse of the transportation-planning problem. Specifically, our efforts have focused on developing artificial neural network (ANN) models for determining zonal trip ends from link volumes. The ANN models are developed and tested using real-world data from the Chittenden County region in Northwestern Vermont. Two different types of ANN models are developed, and their performance is compared. The study illustrates that ANNs are quite capable of solving the inverse of the transportation-planning problem.