This paper presents an ant colony optimization algorithm for allocating airport slots to flights. Slot allocation at origin and destination airports of each flight are not independent, and they must be coherent with flight durations. The algorithm is able to
deal effectively with large instances: in very short CPU time, it allows to fulfil almost all flight requests for slots. Conversely, an exact approach already proposed in the literature only solves small and medium size instances, and with longer CPU times. Our results suggest that ant colony optimization may be preferred to exact optimization when large instances are to be solved, or when only a short CPU time is available for the solution process.
