流量拥堵空域内-种基于 Q-Learning 算法的改航路径规划.

At present, the problem of airspace resource shortage caused by the surge of air traffic is becoming more and more prominent. In order to alleviate this situation, the study of aircraft rerouting based on the flow management was carried out. Firstly, the airspace environment was discretized by rasterization, and the airspace was divided into different types of raster regions according to the congestion degree of waypoint flow. Secondly, it was used to model by improving the reward function of Markov decision process in reinforcement learning. Based on the ε-greedy strategy, Q-Learning algorithm was used for iterative solution, and the corresponding parameter value was explored and compared to improve the applicability of the results. Finally, the optimal path and the corresponding performance indexes under different parameter assignments were calculated through simulation. The results show that the model and algorithm can be used to search for an appropriate rerouting path for the congested airspace in a certain period of time, so that the aircraft can avoid the congested waypoints, shorten the delay time in the air, and effectively improve the status of airspace congestion. [ABSTRACT FROM AUTHOR]