Automatic Calibration of Microscopic, Activity-Based Demand for a Public Transit Line

This work describes the methodology to get a more realistic transit route choice estimation, evaluated with occupancy counts comparison at bus stops in a real world scenario. First, a search of travel priorities for passengers was carried out with the selection of travel behavioral parameters whose values correction would help for the manual calibration. Moreover, the dynamic transport calibrator Cadyts was integrated on the transit simulation to help to estimate the trip generation and route choice. A bus line of the city of Berlin with real data of daily usage was taken as test scenario. The calibration experiments reduced the occupancy counts comparison error by 35% from about 50% to about 15%.