Including time in a travel demand model using dynamic discrete choice

Activity based travel demand models are based on the idea that travel is derived from the demand to participate in different activities. Predicting travel demand should therefore include the prediction of demand for activity participation. Time-space constraints, such as working hours, restricts when and where different activities can be conducted, and plays an important role in determining how people choose to travel. Travelling is seen as a possibly costly link between different activities, that also implicitly leads to missed opportunities for activity participation. With a microeconomic foundation, activity based models can further be used for appraisal and for accessibility measures. However, most models up to date lack some dynamic consistency that, e.g., might make it hard to capture the trade-off between activity decisions at different times of the day. In this paper, we show how dynamic discrete choice theory can be used to formulate a travel demand model which includes choice of departure time for all trips, as well as number of trips, location, purpose and mode of transport. We estimate the model on travel diaries and show that the it is able to reproduce the distribution of, e.g., number of trips per day, departure times and travel time distributions.