Airport Capacity Prediction and Optimization Based on Semi-Varying Coefficient and Dynamic Programming

Air traffic congestion has emerged as a critical challenge in modern air traffic management. Adverse weather conditions and other factors frequently lead to reduced airport arrival capacity, resulting in flight delays. Accurate capacity prediction and optimization are essential for maximizing airport utilization and minimizing flight delays. This study introduces a capacity prediction model based on the semi-varying coefficient approach, utilizing real-world data from Guangzhou Baiyun Airport collected over a specific time period. Through 5-fold cross-validation, the semi-varying coefficient model is compared to the varying-coefficient model, demonstrating superior out-of-sample prediction capabilities. Subsequently, we establish and solve a flight arrival and departure flow allocation optimization model using dynamic programming to enhance traffic flow coordination. Analysis of case study results demonstrates that this method can generate optimal strategies for allocating arrival and departure traffic flows within airport capacity constraints, thereby minimizing overall flight delay costs.