AI-Driven Insights: Evaluating Hyperellipsoid, Hyperrectangle, and Neural Network Models in Emergency Landing Scenarios on Aviation

Abstract

This thesis examines the effectiveness of AI models in understanding pilots’ preferences during the airport selection process in mid-flight emergencies. When faced with an unexpected situation, pilots must quickly evaluate multiple factors to choose a new landing airport. Thus, employing AI for these multi-criteria decision-making scenarios can significantly improve their decision-making capabilities. This research compares the performance of two types of Learning Classifier Systems, hyperellipsoid and hyperrectangle, with Neural Networks, focusing on their accuracy as well as the interpretability and explainability of the models.