Simulating Carsharing Operations through Agent-based Modelling: An Application to the City of Lisbon, Portugal

Abstract

The increasing use of automobiles has had significant negative impacts on urban life: pollution, excessive energy use and time lost in traffic. The quick rise of auto externalities has policy makers facing the hard challenge of balancing demand for mobility on the one hand, and assuring sustainable urban life on the other. One strategy that can aid reducing these externalities is carsharing. Carsharing typically involves a fleet of vehicles in stations around a city, which clients may use on an hourly payment basis. Classical round-trip systems only address a niche market of urban trips such as shopping and errands, and few companies have risked the one-way carsharing option in the past due to vehicle stocks imbalance. Currently these systems are gaining new attention with important car builders investing in providing the one-way market in many cities in Europe and the United States. Nevertheless, there is still great uncertainty on the financial and economic viability of this type of carsharing. This results from a lack of realistic modeling tools that allow testing several operational parameters of this transportation alternative. In this paper we present a very detailed and realistic model to assess the potential of one way carsharing systems done through the use of agent based simulation. The simulation incorporates a stochastic demand model discretized in time and space and a detailed road network. It aims to assess the economic performance of the system both from the users’ perspective and the carsharing operator. The performance is a function of several planning and operational decisions which are included in the model: fleet size; station or free parking areas location decision; pricing policies. Some lower level configurations are also analyzed, such as an information system, vehicle reservation and maintenance operations scheduling. This paper focuses on the analysis of the spatial distribution and fleet size impacts on the demand for this transport option. This model was developed for the city of Lisbon, but its flexible architecture allows it to be adapted to different cities. Therefore this may become a tool that government authorities can use to rigorously estimate carsharing impacts and at the same time help private companies to manage their systems better.