244 / 2025-06-15 17:27:07

Airport Adjustment Strategy Based on Air Traffic Flow Prediction Using EMD-BiLSTM Hybrid Mode

Air traffic flow prediction,Empirical mode decomposition,Bidirectional long short-term memory network,Margin analysis,Airport adjustment strategy,Reliability engineering,Intelligent Simulation System

Aiming at the problems of low runway utilization and congestion caused by airport flight scheduling relying on manual experience and insufficient flow prediction accuracy, this paper proposes an air traffic flow prediction method based on EMD-BiLSTM hybrid model and a dynamic margin driving strategy. Firstly, the Empirical Mode Decomposition (EMD) was used to adaptively decompose the non-stationary traffic data into multi-scale Intrinsic Mode Functions (IMFs) to eliminate noise interference. The bidirectional Long Short-Term Memory network (BiLSTM) was used to capture the bidirectional dependence of time series, and the fusion prediction model was constructed. Furthermore, the throughput margin equation was defined based on the prediction results and the maximum runway capacity, and a three-level response mechanism was established: when the margin was≥2, the flight timing was dynamically fine-tuned. When 0≤margin<2, adjust non-core flights and optimize resource allocation. When the margin is<0, the priority sorting and proactive postponement strategy are started. Based on the departure flow data of Hangzhou Xiaoshan Airport in the past year, the proposed model reduces the Mean Absolute Percentage Error (MAPE) to 7.18%, which significantly improves the prediction accuracy compared with traditional BiLSTM, GAUSS and other models. Combined with the margin strategy, the runway overload risk probability in peak hours is effectively reduced. The simultaneous construction of GIS airspace simulation and decision platform realizes the visualization of flow and strategy, provides a dynamic scheduling basis for air traffic control department, and effectively improves the efficiency of airport operation and resource utilization.