Predicting whether, and if so, the extent to which, aircraft ground handling is delayed has become increasingly important in the last decade as the aviation industry aims to improve the punctuality of its flights. Timely identification of aircraft ground handling delays allows the operational partners to update their schedules and reallocate their resources. While obtaining an accurate prediction is important, understanding how a prediction comes about is at minimum equally important. After all, this yields insights into the complex and stochastic aircraft ground handling process system, consisting of many sequential and parallel activities such as fuelling, (de)boarding and baggage (un)loading, and allows the operational partners to establish mitigation and/or contingency measures using knowledge extracted from the model. This research delves into the prediction of scheduled ground handling end time adherence at intervals during an aircraft turnaround at Amsterdam Airport Schiphol (AAS). To accomplish this, the processes and variables at play in the aircraft ground handling process are first identified and assessed. Subsequently, the aircraft ground handling process is modelled using interpretable machine learning.
Graduated: August 2024