Computing Flight Departure Times Using an Advanced First-Come First-Served Scheduler

摘要

This paper presents a novel algorithm to compute departure times for aircraft traversing through a predefined node-link structure by combining two previously developed algorithms. Both the algorithms are based on the first-come first-served principle so that departure time is computed for one aircraft at a time assuming that once the schedule is computed it cannot be changed. The first algorithm, that has been previously used for traffic flow management studies, can impose capacity constraints both on the nodes and links. However, transit times cannot be changed, which resulted in poor utilization of resources. The second algorithm, developed for a study about arrival metering, imposed capacity constraints only on the nodes, but allowed variation of link transit times. The new algorithm efficiently computes required ground and airborne delays for each flight that minimizes the departure delay and ensures earliest possible arrival time for the given flight. Using this algorithm, a sensitivity study was performed to investigate the impact of allowing transit time variation by scheduling about 48,000 flights over the entire US using different ranges for maximum and minimum speeds. Simulation results show that the new scheduler can reduce the average per flight departure delay by 42% compared to fixed transit time scheduling when the transit time can vary between -3% and 15% of the nominal value. The average per flight delay was reduced by 30% even when the transit times are only permitted to increase by 5%.