For dynamic traffic (assignment) models an OD matrix is an important input. Normally, an OD matrix from a traffic survey or another (long-term planning) model is used. But these OD matrices are not dynamic. Using a departure profile from counts or other sources, the static OD matrix is transformed into a dynamic one. The dynamic matrix is calibrated with flow measurements and sometimes with travel time measurements. Can this process be improved? In other words: what is the best way to estimate a dynamic OD matrix from traffic data?
Other questions relate to the distribution of the demand to the different time periods. What is the best method to model departure time choice? Or, even in more detail: is it possible to allocate slots to traffic and does slot allocation improves traffic performance?