A Discrete Choice Approach to Define Individual Parking Choice Behaviour for the Parkagent Model

Abstract

PARKAGENT is an agent based model for simulating parking search in the city. In PARKAGENT, the agents choose a parking spot based on the expected number of free parking spaces, distance to destination and length of parking space. For a true representation of underlying parking choice behaviour of agents in PARKAGENT model, a behavioural model is required. Behavioural models are considered as the core of agent based simulations, therefore a behavioural model capable to exhibit parking choice process in PARKAGENT has been proposed in this paper. This model explains that parking choice is based on the principles of utility maximization. Several research studies have used discrete choice models to describe parking choice phenomena. Discrete choice models determine the utility associated with choice of services and products. It is assumed that individual make decisions rationally, it is very difficult to measure the actual utility associated with a parking space. For a realistic calculation of the utility, factors affecting parking choice such as (parking cost, distance to destination, etc.) are required. In this research, the choice of on-street parking is considered keeping in view the factors associated with the street situation (e.g. occupancy, security). The decision of an agent to choose a street for parking is based on the factors associated to street. The necessary data is collected through stated choice questionnaire. The collected data is analysed using a discrete choice model (multinomial logit model). The results indicate show that the identified attributes of streets significantly affect the parking choice behaviour of agents. Keywords: discrete choice model, agent based parking simulation model PARKAGENT, factors affecting on-street parking choice. INTRODUCTIONThe aim of urban mobility plans is to attain sustainable urban transport system that ensures accessibility and quality of urban environment. These goals can only be accomplished with the help of certain forecasting tools. These tools provide the assessment of the future situation and support in making well informed decisions. Traffic induced due to parking search has significantly grabbed the attention of policy makers. It is crucial to model parking behaviour of motorists, in order to identify the effect of change in parking behaviour on the overall travel behaviour of drivers. Understanding and modelling parking choice behaviour is essential for urban planning and decision making. It is necessary to find, measure and model all the relevant factors influencing individual parking choice behaviour and decision-making processes.Parking choice process involves decision making. An efficient parking model should represent drivers parking choice more accurately, therefore behavioural models are considered as the cornerstones of agent based simulations. The major challenge in developing an agent based parking simulation model is the realistic nature of the parking choice behaviour model. The concept of discrete choice modelling can be used to in...