Flight time prediction accuracy and its effect on commercial airline delay management
No Thumbnail Available
URL
Journal Title
Journal ISSN
Volume Title
School of Business |
Bachelor's thesis
Electronic archive copy is available locally at the Harald Herlin Learning Centre. The staff of Aalto University has access to the electronic bachelor's theses by logging into Aaltodoc with their personal Aalto user ID. Read more about the availability of the bachelor's theses.
Authors
Date
2023
Department
Major/Subject
Mcode
Degree programme
Tieto- ja palvelujohtaminen
Language
en
Pages
31
Series
Abstract
Delays cost airlines billions of euros annually. From the operational point of view, there is often a tradeoff between fuel burn and flight time. Wrong decisions can lead to significant costs to the airlines and the goal is to find a balance between time and fuel related costs. To be able to do the right decisions in a limited timeframe, the airlines need accurate information on the arrival times of flights. Delay costs may reach hundreds of euros per minute, and thus, every minute is important and can improve airline profitability. Flight time prediction methods have been studied before, and both physics-based and machine learning methods have been proposed in the literature. The future will be in machine learning models. These models extract information from big data and consequently benefit from the best available set of features. Many of the proposed machine learning models use physics-based predictions as features. However, none of the reviewed models include human-made estimates, even though these are readily available for airlines. The human factor and pilots’ estimation accuracy are lacking research. With methods trying to utilize every piece of information in the data, it is worth analysing the performance of pilot-made predictions. In this thesis pilots’ ability to estimate remaining flight time on commercial airline operations is evaluated. The prediction performance is compared with methods in the literature and purely physics-based predictions produced by the aircraft flight management system. The results indicate that the pilot-made estimations provide better prediction accuracy during the late parts of the flight when compared to physics-based methods. No major differences in the accuracy were found for predictions made over 60 minutes before the estimated time of arrival.Description
Thesis advisor
Seppälä, TomiKeywords
airline operations, delay cost, delay management, flight time, prediction accuracy