Discrete Choice Modeling of Car Parking Search Behavior

Abstract

An understanding of parking choice and its associated effects is essential for modeling changes in parking supply. This study presents a model that uses a synthetic population and the results of a stated preference experiment to determine where car drivers park. The applied mixed logit regression model simulates the parking behavior of car commuters at RWTH Aachen University, performing an incremental assignment and reassigning drivers who were unable to find a parking space in previous iterations. The resulting model is capable of predicting shares of parking in public spaces, walking distances, and parking search traffic. Consequently, it enables the analysis of the benefits and disadvantages of modifications to the parking supply. The results indicate that a limited number of parking facilities generate the majority of car parking search traffic and that strategies of parking supply centralization are effective in reducing parking search traffic.