Comparison of various temporal air traffic flow management models in critical scenarios

Overloads (i.e., critical imbalances between traffic demand and capacity) are currently resolved in the European air transportation network by activating air traffic flow management regulations. Flights subject to regulations are delayed on the ground by the computer assisted slot allocation system, which uses a first-planned, first-served principle to allocate the delays. In the last decades, many researchers have proposed alternative models based on more or less complex optimisation techniques, which could effectively resolve overloads with less delay. Despite the fact that these models were very promising, changing the current operational system and procedures -which trust is supported by a successful working history dating back to 1995-may not be realistic in the short term. This paper examines the gradual transition from current air traffic flow management to a futuristic model in which flights are assigned delays with massive cherry-picking measures that minimise the total delay in the network, without necessarily adhering to the first-planned, first-served rule. A reference model with greedy regulations representative of current operations is compared to a model where the Network Manager optimises regulations using hyper-heuristics, a model based on massive cherry-picking measures, and hybrid models that consider regulations to resolve major overloads and cherry-picking measures only for residual overloads. These models are compared in critical scenarios with major capacity issues caused by adverse weather conditions. Results suggest that hybrid models could be a viable option for reducing delays with minor changes in the current air traffic flow management paradigm.