Reverse Assignment Formulation in Evacuation Simulation

To evaluate the performances of a transport system, many works in literature use traffic data (i.e. flow counts) observed on representative links of the transport network) together with transport model. The observed data are the input for the transport model update. Often the corrections in the costs are not accompanied by a corresponding correction in link flows and update in demand values in a whole methodology.In this paper, we apply a method for calibrating the parameters of cost functions and demand parameters and/or demand update at the same time. The data input are a (initial) demand matrix, the link cost function (i.e., a function for each link set), the model for users choices simulation. The outputs are an updated demand matrix, the optimal parameters of cost functions and the link flows. The method is able to calibrate the cost functions and update the demand values when, in a particular situation, the transport system is subject to a sudden load, as happens in the case of an evacuation. The flows and costs are observed on some characteristic links (with cameras and probe vehicles equipped with GPS). An application on the route performance of a fleet of vehicles that evacuates some users is reported.