Experimentelle Evaluierung heuristischer Suchalgorithmen zur Optimierung von Abflugreihenfolgen im Air Traffic Flow Management

The SlotMachine project aims to optimize the allocation of air traffic management slots in the event of capacity bottlenecks. In particular, the project addresses the question of how the departure sequence of flights can be optimized in the event of capacity bottlenecks. To ensure the confidentiality of the data, the search for optimal solutions and the evaluation of the solutions are carried out by separate components. The evaluation of the found solutions is performed by means of multi-party computation. The search for optimal solutions is performed using heuristic search. In this work, different methods for the heuristic search of optimal departure sequences are experimentally investigated with respect to their suitability for the SlotMachine project. Experiments determine to what extent different optimization algorithms are suitable for the optimization of departure sequences. In particular, genetic algorithms and various local search algorithms will be investigated. In the experiments, existing open source frameworks with different settings are used to determine optimal departure sequences in different scenarios with respect to the nature of the preferences of the flights. The experiments conducted in this work show that the results found using heuristic search are generally close to the optimal solution as found by the (exact) Hungarian method.