Air transportation systems display a rich phenomenology connected with several key topics in Complexity Science, such as complex networks, cascading failures and percolation. One example is flight delays that have usually the origin in primary events localized in limited areas of the network, but can later multiply and magnify as the daily operations go along. Given the large costs that delays convey, it is important to characterize their propagation and to model with predictive power the potential components or areas of the network affected. In this paper, we discuss the validation of an agent-based model, developed within the framework of the SESAR WP- E TREE project and aiming at simulating the propagation of delay in the ECAC airport network. Simulation outcomes are systematically confronted with empirical flight performance, the results show a good level of agreement with accuracies and precisions. Furthermore, we use the model to assess the effect on delay reduction in the network of two delay reduction scenarios: dropping passenger connections if the delay induced goes beyond a given threshold τ, or decreasing the service time of aircraft in the airports if they delay is larger than a given value ∆. Our results show how optimal values of τ and ∆ can be found within the simulation framework.
http://www.sesarinnovationdays.eu/2015/papersandpresentations