基于空间扩散的航班延误预测

Flight Delay Prediction Based on Spatial Diffusion

航班延误问题在航空网络中是有待解决的一大问题。航班运行的过程中,前序航班延误问题以及同一时间段的机场天气、资源调度、突发事件等多方面情况都会对后续航班或同一时间段的其他航班造成影响。因此文章分别将前序航班发生延误后的延误扩散和因机场容流量产生延误后的延误扩散定义为纵向和横向扩散,基于横向和纵向共同影响下的空间扩散的研究视角,使用梯度提升决策树方法进行航班延误预测。预测结果显示横纵向信息综合考虑后,预测准确率可达91.5%,分别优于纵向预测的准确率78.53%和横向预测的准确率80.2%,结论验证了模型的可行性和准确性。

The issue of flight delays is a major problem to be solved in the aviation network.In the course of flight operations,the problem of delayed preceding flights as well as various circumstances such as airport weather,resource scheduling and unexpected events during the same time period can affect subsequent flights or other flights during the same time period.Therefore,this paper defines the delay diffusion after the delay of the preceding flight and the delay diffusion after the delay due to airport capacity as the vertical and horizontal diffusion respectively,and uses the gradient boosting decision tree method for flight delay prediction based on the research perspective of spatial diffusion under the joint influence of horizontal and vertical.The prediction results show that the prediction accuracy can reach 91.5%after considering the horizontal and vertical information together,which is better than the accuracy of 78.53%for the vertical prediction and 80.2%for the horizontal prediction respectively,and the conclusion verifies the feasibility and accuracy of the model.