Izvorni znanstveni članak
https://doi.org/10.7307/ptt.v34i6.4195
Multi-Flight Rerouting Optimisation Based on Typical Flight Paths Under Convective Weather in the Terminal Area
Shijin Wang
; Nanjing University of Aeronautics and Astronautics
Rongrong Duan
; Nanjing University of Aeronautics and Astronautics
Jiewen Chu
; Nanjing University of Aeronautics and Astronautics
Jiahao Li
; ZTE Corporation
Baotian Yang
; Nanjing University of Aeronautics and Astronautics
Sažetak
With the rapid growth of flight volume, the impact of convective weather on flight operations in the terminal area has become more and more serious. In this paper, the typical flight paths (TFPs) are used to replace flight procedures as the routine flight paths in the terminal area, and the TFP of each flight is predicted by Random Forest (RF), Boosting Tree (BT) and K-Nearest Neigh-bour (KNN) algorithms based on the weather and flight plan characteristics. A multi-flight rerouting optimisa-tion model by bi-level programming is established, which contains a flight flow optimisation model in the upper layer and a single flight path optimisation model in the lower layer. The simulated annealing algorithm and the bidirectional A* algorithm are used to solve the upper and lower models. This paper uses the terminal area of Guangzhou Baiyun Airport (ZGGG) and Wuhan Tianhe Airport (ZHHH) for case analysis. The RF algorithm has better performance in predicting TFPs compared with the BT and KNN algorithms. Compared to the historical radar trajectory, the flight path optimisation results show that for the Guangzhou terminal area, while meeting the Terminal Airspace Availability (TAA) as constraint, the flight flow increases and the flight distance reduces, ef-fectively improving the operational efficiency within the terminal.
Ključne riječi
terminal area; convective weather; typical flight path; terminal airspace availability; path optimisation; machine learning
Hrčak ID:
293674
URI
Datum izdavanja:
2.12.2022.
Posjeta: 251 *