Ground Delay Programs: Collaborative Planning Under Uncertainty

This paper will review two recently developed optimization models for planning a Ground Delay Program (GDP) under uncertainty in arrival capacity forecast at an airport. The first is a static model, which means ground delay decisions are made once at the beginning of planning horizon and are not updated later. In the second model, ground delay decisions are dynamically revised based on updated information on airport capacity. The performances of the two models are compared for a hypothetical GDP at Dallas Fort Worth International airport. By virtue of revising ground delay decisions, the dynamic model saves 10 – 30% delay cost compared to the static model. As a step forward towards implementing these models in practice, it is shown how the solutions obtained from the static and the dynamic stochastic models can be used under the collaborative decision making paradigm. Conditions are shown for when a slot-substitution between two flights is feasible, and a generalized procedure is shown that will allow collaborative planning of GDPs using the solutions from the existing stochastic models for allocating landing-slots to individual airlines. The experimental results indicate that the airlines can save 5 – 20% delay costs from the intra-airline substitutions, with the delay savings of airlines that operate many flights at an airport, i.e., hub-airlines, being on the higher side.