Air traffic flow short and medium term prediction and flight assignment method based on bidirectional transformer
By using a bidirectional Transformer-based air traffic forecasting model and a multi-objective collaborative optimization algorithm, the problem of ignoring the spatial correlation characteristics of airspace traffic forecasting and balancing multiple demands of flight allocation schemes in traditional methods is solved. This results in more comprehensive forecasting and better allocation schemes, improving the adaptability and efficiency of air operations.
CN122114438APending Publication Date: 2026-05-29CIVIL AVIATION FLIGHT UNIV OF CHINA
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- CIVIL AVIATION FLIGHT UNIV OF CHINA
- Filing Date
- 2026-01-08
- Publication Date
- 2026-05-29
Smart Images
Figure CN122114438A_ABST
Abstract
The application provides a two-way Transformer-based air traffic short and medium-term prediction and flight allocation method, comprising the following steps: S1: air-related data acquisition and preprocessing; S2: constructing a two-way Transformer prediction model with local cut space alignment and fusion graph attention guidance; S3: constructing an improved butterfly optimization algorithm under multi-objective constraints; S4: model collaborative training and parameter optimization; S5: air traffic short and medium-term prediction implementation; S6: flight allocation implementation under multi-objective constraints; S7: two-way feedback optimization and iterative update between algorithms; S8: prediction and flight allocation report output and final output result verification. The application improves the intelligent level of air operation scheduling, and provides technical support for airspace resource efficient utilization and air operation safety guarantee.
Need to check novelty before this filing date? Find Prior Art