Full-scale pre-tactical route prediction: machine learning to increase pre-tactical demand forecast accuracy

The objective of this paper is to present an artificial intelligence-based methodology to predict the Flight plans that will be received during the pre-tactical phase of the Air Traffic Flow and Capacity Management (ATFCM) process. For this purpose, input features equivalent to those of EUROCONTROL’s PREDICT solution are fed to a Multinomial Logistic Regression algorithm over pre-clustered air routes in order to determine which route cluster is the most likely to be filed by an airspace user within each OD-pair. Results show that this procedure is capable of outperforming the current PREDICT solution in almost 40% of the 5,699 OD pairs considered and reducing current solution’s error by 11%, showing good and scalable prediction capabilities.This PhD study has received funding from the SESAR Joint Undertaking under the European Union’s Horizon 2020 research and innovation programme under grant agreement No 783287. Finally, the authors would like to acknowledge the support of the Spanish Centre for Industrial development (CDTI) through the PRETA project (Grant no. IDI-20190029)Peer ReviewedPostprint (published version