As people's demand for flights has risen sharply, the problem of air traffic congestion has also followed. While economic loss has become a major problem, it has also brought corresponding safety hazards to people's travel. Therefore, scientifically establishing an optimized air traffic management system is an urgent problem to be solved. Based on this, firstly this paper introduces 30 flight sorting models of four different airline companies for comparative research. Secondly, in order to solve the problems of slow convergence speed and poor real-time performance of the model in the hyperparameter optimization process, this paper proposes the inbound and outbound flights are sorted based on the basic elements in the reinforcement learning algorithm. Finally, the proposed reinforcement learning model is applied to the flight sorting problems of different airlines for comparison. The experimental results show that the reinforcement learning algorithm proposed in this paper effectively solves the problems of slow convergence speed and low real-time performance in the hyperparameter optimization process, and at the same time minimizes the loss amount of flight delays and the variance of airline delay time , which shows that the algorithm proposed in this paper has certain practical significance and application value for solving the problem of ground waiting at a single airport.
Research on arrival aircraft sequencing based on reinforcement learning
Sixth International Conference on Computer Information Science and Application Technology (CISAT 2023) ; 2023 ; Hangzhou, China
Proc. SPIE ; 12800 ; 128004E
2023-10-11
Conference paper
Electronic Resource
English
Rule derivation for arrival aircraft sequencing
Online Contents | 2013
|Receding Horizon Control for Aircraft Arrival Sequencing and Scheduling
Online Contents | 2005
|