Trip redistribution and calibration of the mode choice model of Catalunya

Abstract: This paper investigates the creation and calibration of a mode choice model for Catalunya. Trip redistribution was performed as the data from light rail trips displayed the number of trips made between stations and not between the actual origin and destination of the trip maker. The adjustment was made to distribute the trips so that zones around the stations where people would start and end their journey from would be given a part of the trips. A mode choice model estimates the probability of public or private transport being chosen for commuting depending on different variables. The model used is a logistic regression (logit) model, which is commonly used in transport modelling when dealing with two alternatives. The coefficients of the model are calculated based on previous data regarding the number of trips made between zones (Origins and Destinations) to find what proportion of trips were made by a particular mode of transport and relating these to the time it takes, monetary cost and other characteristics related to the journey, trip maker and mode of transport. The study found that the model which most closely matches the input data uses zonal data of trips between Origins and Destinations, taking into account both urban and rural areas. These zones are areas within every municipality in Catalunya. The study also found that the number of cars per inhabitant and the number of daily trips had a significant positive correlation on the probability of choosing private transport over public transport.