Transportation network models are central to transportation operations and planning. There are two main components: network vehicle/passenger flow and travel behavior. The flow model describes flow propagation through the network arcs/nodes. The behavioral models encapsulate travelers’ time-of-day choices on routes, departure time, parking locations and traffic modes (such as solo-driving, carpool, transit, ride-sharing etc.). The idea is to infer/learn dynamic origin-destination (O-D) demands and behavioral choice models that altogether, if input into network simulations, would produce the spatio-temporal flows in the network consistent with the real-world multi-source data.

​