Agent-based model of driver parking behavior as a tool for urban parking policy evaluation

Itzhak Benenson^1,2, Karel Martens³, Slava Birfir²

¹Dept of Geography and Human Environment, ²Environment Simulation Laboratory, Porter School of Environmental Studies, University Tel Aviv

³Department of Geography, Planning and Environment, Radboud University Nijmegen, the Netherlands

bennya@post.tau.ac.il

We construct an explicit Agent-Based Model of parking search in a city. The model agents – “drivers” – drive towards their destination, search for parking, park for a period of time, and leave the parking place. The city’s infrastructure is presented by a high-resolution GIS of the street network and parking lots, which includes information on traffic directions and turns, parking permissions along the streets, layers of off-street parking places and lots, and the destinations presented by layers of residential buildings and public places. Driver agents belong to one of four categories: residents and guests, with residential buildings as their destination, and employees and customers, with public places as their destination. Each agent has its own pre-defined destination, willingness to pay, time of arrival, and duration of stay, and agents of different groups differ in the typical values of these parameters.

The current version of the model ignores the stage of driving from the point of entrance into the network to their destination; instead, driver agents enter the network at a distance of 250 m from their destination (i.e. beyond the distance where the search for a parking place commences, but close to it) and drive towards the destination. First, a driver estimates the parking situation in the area and then (currently at ~100m from the destination) begins his/her search for a parking place; the maximal duration of the search can vary between drivers. During the search, a driver agent accounts for the availability of parking places, prices of various types of parking, parking enforcement efforts, and behavior of other drivers. The model results are given by: (1) the distributions of the search time, (2) the distribution of the distance between parking place and destination, (3) the distribution of the fees paid by the drivers; and (4) the revenues by parking facilities (public or private).

The model is implemented as an ArcGIS application and can work with a practically unlimited number of driver-agents simultaneously. The project is partially supported by the Municipality of Tel Aviv-Yaffo and is currently employed for a comparative investigation of alternative parking policies for Tel-Aviv.