Dynamic Ground-Holding Policies for a Network of Airports

The yearly congestion costs in the U.S. airline industry are estimated to be of the order of $2 billion. In P. B. Vranas, Dimitris J. Bertsimas, and A. R. Odoni, The multi-airport ground-holding problem in air traffic control, Operations Research , Vol. 42, pp. 249–261, 1994, we introduced and studied generic integer programming models for the static multi-airport ground-holding problem (GHP), the problem of assigning optimal ground holding delays in a general network of airports, so that the total (ground plus airborne) delay cost of all flights is minimized. The present paper is the first attempt to address the multi-airport GHP in a dynamic environment. We propose algorithms to update ground-holding decisions as time progresses and more accurate weather (hence capacity) forecasts become available. We propose several pure IP formulations (most of them 0–1), which have the important advantages of being remarkably compact while capturing the essential aspects of the problem and of being sufficiently flexible to accommodate various degrees of modeling detail. For example, one formulation allows the dynamic updating of the mix between departure and arrival capacities by modifying runway use. These formulations enable one to assign and dynamically update ground holds to a sizeable portion of the network of the major congested U.S. or European airports. We also present structural insights on the behavior of the problem by means of computational results, and we find that our methods perform much better than a heuristic which may approximate, to some extent, current ground-holding practices.