Modelling parking search behaviour with an agent-based approach

Abstract

In the recent years several agent-based parking models have emerged, which provide policy makers with temporally and spatially rich data, what would be impossible to acquire using aggregated models. As we found out, all recent agent-based parking search models, lack the ability to account for the influence of parking shortages on mode or location choice. We close this modelling gap by proposing a new parking search model, which should be built on top of an existing agent-based traffic simulation called MATSim. We point out, that the parking strategy evaluation procedure of the previous parking search models delivers systematically too high search times which can be avoided using the proposed approach. Furthermore we also describe how individual valuation of parking search components (e.g. search time, cost and walk time) could be incorporated in the model. Currently an implementation of the proposed parking search model and an application for the city of Zurich is in progress.