Extended Methodology for the Estimation of a Zonal Origin-Destination Matrix: A Planning Software Application Based on Smartcard Trip Data

Abstract

A key planning consideration when designing transit services is the projected demand on each route, as this is a necessary input to determine frequencies and vehicle capacities. Predicting demand changes in complex networks requires a thorough modelling of the public transport system and its users’ decisions. Aiming to help in this process, a new transit planning tool (STEPTRANS) is being developed, seeking to predict the reactions of public transport users to network changes, allowing better design and operation of the transport system. This tool needs a reliable origin-destination (O-D) matrix for the city of interest, which is assigned to the transit network, using state of practice route choice models. This work presents an extension of the former methodology implemented in STEPTRANS, which allows a zonal O-D matrix for a city to be estimated, given a stop-to-stop matrix obtained from smartcard trip data. The methodology and extension were applied to estimate models and O-D matrices for Santiago de Chile’s integrated transit system, Transantiago. The extension focuses on three improvements: (1) addressing data limitations by using the latest large scale O-D survey available for Santiago; (2) updating the zonal system, including a new methodology for access links construction; and (3) enhancing the model formulation, incorporating trip pattern corrections and reformulating cost measures using route choice models. The inclusion of the resulting matrices in STEPTRANS will allow the evaluation of major changes to the transport system, as a zonal O-D matrix allows modelled users to choose their initial and final stops.