Supply-Demand Matching Estimation with Machine Learning in Airline Planning Process

Airline companies operate in a highly competitive and uncertain environment where efficient resource allocation is essential for maximizing economic performance and supporting sustainability objectives. One of the major challenges in airline planning is matching aircraft capacity with uncertain passenger demand while maintaining operational efficiency. This study proposes a machine learning-based decision support framework that integrates demand uncertainty with aircraft tail assignment. Three nonlinear regression models—Support Vector Regression (SVR) with a Radial Basis Function (RBF) kernel, Polynomial Regression, and Random Forest Regression—were employed to predict flight-based economic returns using real-world airline operational data, including aircraft type, passenger demand, and cargo volume. In addition, a bounded prediction mechanism incorporating aircraft capacity constraints was introduced to ensure operational feasibility. Model performance was evaluated using R², Mean Squared Error (MSE), Root Mean Squared Error (RMSE), Mean Absolute Error (MAE), and cross-validation. At the end of the study, all three regression models achieved a high accuracy rate in representing the data, but the SVR-RBF model performed best. Furthermore, the proposed framework allowed for the determination of demand thresholds for optimal aircraft type selection, contributing to increased economic efficiency and reduced unnecessary capacity utilization. The findings demonstrate that machine learning-based predictive models can effectively support airline planning processes and indirectly contribute to sustainability goals.