The objective of this study was to investigate itinerary generation procedures and their effect on delay. In order for a National Airspace System (NAS) model to propagate delay, the flights within the demand set must be linked together. Currently, when modeling or analyzing the NAS, an analysis has two primary data sources to use in generating a demand set - ASQP and TFMS. Both of those datasets have their benefits and limitations. The ASQP dataset includes scheduled and actual flight data for scheduled nonstop passenger operations, as well as well as information about which aircraft operated a particular flight and causes of delay. However, the ASQP dataset only contains domestic flights operated by the largest U.S. carriers, currently only sixteen. The TFMS dataset contains a much larger set of flights, but does not contain a unique aircraft identifier or causes of delay. Linking individual flights is best done with a unique identifier. Because the ASQP dataset only contains domestic flights and the TFMS dataset does not include a unique aircraft identifier, tracking the daily movement of an aircraft becomes complicated. The ASQP recorded flights may suggest that an aircraft teleported from one airport to another or that it sat at an airport an unexpectedly long time, when the aircraft actually flew to an international destination and either moved on to another domestic airport (teleportation) or returned to the original airport (long aircraft turnaround time). Within this investigation, we developed a process for enhancing existing flight delay data by determining an appropriate aircraft tail numbers for domestic and international flight operations for a limited set of carriers. Our method uses a greedy algorithm to determine the possible international flights within the TFMS dataset that can fill the holes in the ASQP dataset created by a teleportation or a long sit time. We tested our process on one year of scheduled flights. We then compared the delay resulting from that set of itineraries with the delay resulting from a set of itineraries generated with a different methodology. The itineraries, generated using our new process, were more realistic than those generated with the other method. They also produced delays more similar to the actual delays.
Enhanced flight delay data for ASQP carriers
2012-04-01
178687 byte
Conference paper
Electronic Resource
English
Enhanced delay propagation tree model with Bayesian Network for modelling flight delay propagation
Taylor & Francis Verlag | 2018
|TIBKAT | 2012
|Flight delay-oriented airport group flight arrangement optimization method
European Patent Office | 2020
|