基于Logistic模型的大面积航班延误预测方法研究

A Forecasting Method for Large-scale Flight Delays Based on a Logistic Model

航班大面积延误发生时,对其后续发展做出准确的预测能够减少自身损失。研究从固定空域的角度出发对航班大面积延误进行预测,给出了基于Logistic模型的延误航班数量与航班累计延误时间的预测方法。考虑到航班数量的实时性,模型中采用时间变化量为参数,并借助2013年8月17日下午我国华北地区某空域航班大面积延误实际数据对Logistic曲线进行了拟合,对延误累积阶段和延误消散阶段分别进行计算,确定各自参数,最后得到预测数据。预测时刻距当前时刻越近,精度越高;预测的延误航班数量和航班累计延误时间与实际数据的最大相对误差分别为3.9%和3.6%。

When a large-scale flight delay occurs, it should be meaningful to accurately predict their progress and come up solutions to mitigate consequence. This study predicts large-scale flight delays from the perspective of fixed air- space. A Logistic model is developed based on the number of delayed flights and their cumulative delay time. Considering the number of flights in real-time, the variation of delay time is regarded as parameters in this model. A case study based on observed data of a large-scale flight delay in North China in the afternoon of August 17^th , 2013 is used to fit the Logistic curve. Based on which, the accumulation and dissipation phases of delay are computed to determine respective parameters. Finally, the forecasting delays are obtained. The closer to the present time, the higher the accuracy is. Moreover, the maximum relative errors between observed and the forecasted number of delayed flights and the accumulated delay time of flights are 3.9% and 3.6%, respectively.