Model of the O/D Matrix: Grid Driven Estimate of the O/D Matrices for a Car Sharing Service

Abstract

To plan a car sharing service and, in particular, to design the positions of the stations, it is fundamental to know the number of potential users corresponding to different scenarios. In this work, to answer the question: “how many potential users will take the car in Station A and will leave it in Station B?” a model has been designed and implemented to estimate the potential users of the car sharing system and consequently the Origin/Destination matrices of the service. A large amount of data was available, including cartographic data, census information, demand matrices and traffic flows. To be able to combine the necessary information, available in different formats and structures, a common grid has been considered as a reference for the computation and some hypotheses have been assumed, e.g. the census data have been considered homogeneously distributed within a grid cell. The available information has been referred to the cell to estimate the Origin/Destination matrices for the car sharing service with respect to different scenarios. The spatial data have been managed and displayed in a GIS environment, and an ad hoc algorithm has been developed to integrate the input data.