Looking into the Crystal Ball—How Automated Fast-Time Simulation Can Support Probabilistic Airport Management Decisions

Abstract

Airport management plays a key role in the air traffic system. Introducing resources at the right time can minimize the effects of disruptions, reduce delays, and save costs as well as optimize the carbon footprint of the airport. Efficient decision-making is a challenge due to the uncertainty of the upcoming events and the results of the applied countermeasures. So-called "what-if" systems are under research to support the decision-makers. These systems consist of a user interface, a case management system, and a prediction engine. Within this paper, we evaluate different types of prediction engines (flow, event, and motion models) that can be used for airport management what-if systems by comparing them in terms of accuracy and calculation speed. Hence, two different operational situations are examined to evaluate the performance of the prediction engines. The comparison shows that accuracy and calculation speed are opposed. The flow model has the lowest accuracy but the shortest calculation time and the motion model has the highest accuracy but the longest calculation time. The event model lies between the other two models. The acceptable accuracy of a prediction tool is strongly dependent on the respective airport, whereas the calculation time is strongly dependent on the available decision time. Regarding airport management, this means that the selection of a prediction engine has to be made in dependence of the airport and the decision processes. The results show the advantages and disadvantages of each prediction engine and provide a first quantification by which a selection for what-if systems can happen.