Abstract
Stratified choice models for trips are presented for the analysis of travel patterns. Production variables include details about the person, his origin and destination, purpose of the trip and type of transport. The attraction end variable should reflect intensity of land use, e.g. employment density and parking costs. The technique of curve firing, by successive relaxation of one variable, are described. The calibration procedure is given, plus correction for trip length. Results are given for actual tests in Baltimore, Columbus and Detroit.