Developing air traffic flow management plays

Abstract

This publication describes the collection, processing, and analysis of historical data describing air traffic flow management initiatives. The goal of the work is the definition and exploration of traffic flow management plays, collections of initiatives frequently implemented together. Raw data consisting of advisories issued by the Federal Aviation Administration's Air Traffic Control System Command Center were processed to yield feature data describing initiatives implemented in the New York area. The features used are based on the available scientific literature and documents describing operational air traffic control. Example features describing an initiative include plan time, the difference between when advisories first mention the initiative and when the initiative goes into effect, and cancel time, the difference between when an initiative ends and the latest planned end time according to any relevant advisory. Individual initiatives were categorized according to their features and market basket analysis was applied to see which combinations of initiatives were frequently implemented together. Market basket analysis also provided association rules. For example the implementation, before 14:00 GMT, of a Ground Delay Program for flights destined for LaGuardia Airport greatly increases the chances of there being a Ground Stop at LaGuardia by 16:00 GMT.