1997 Volume 14 Pages 671-678
In order to evaluate travel demand management measures prior to their implementation, it will be required to predict individual decision making in a continuous time process in response to TDM measures rather than evaluate only the final result of the process. This study aims at applying hazard-based duration models in order to understand the variations in individual travel behaviour over time after the introduction of a P&R policy. Multiple-spell proportional hazard models which deal with the recurring duration spells are developed to examine car commuters' behaviour attending to a one-month trial of Park-and-Ride system in Hiroshima. The models also simulate the changes in the future travel demand owing to the implementation in levels of services for the P&R system.