Affinity propagation clustering classification method for aircraft in arrival and departure sequencing

Abstract

With the increase of the fight flow on airport, the category of aircraft has a great influence on the effect of aircraft arrival and departure sequencing. The classification method based on trial whorl is a popular one, which divides all aircrafts into four types (Heavy, Large, Medium, Small). In this paper, a new classification method is presented, in which all aircrafts are classified with considering not only trial whorl but also some economical factors such as the arrival and departure time, flight mission, flight object, and flight priority. The fast and efficient affinity propagation clustering algorithm is applied to aircraft classification, which regards all the aircraft as the exemplars, and thus reduces the computing time for aircraft classification without iterative circulation. Finally, our new aircraft classification is applied to departure sequencing simulation, the results of which demonstrate our method is more rational, and economic benefit can be distinctly improved compared to the trial whorl method.