The dynamics of large scale transportation systems result from: i) traffic flow in the system, a strongly nonlinear process, ii) traffic assignment, i.e. the route/departure time choice of travellers. Route and departure time choice occur on different time scales, real time and day-to-day. Travellers make their choices based on their assessment of their travel costs. Traffic assignment impacts the congestion patterns, thus retroacts with the travellers' choices. Available information and learning play a crucial role in this process. The aim of the paper is to analyse the day-to-day dynamics of the transportation system on a medium time scale. The questions of interest are: existence and unicity of equilibria, periodic orbits, possibly complex dynamics.