基于修正延误波及树的航班延误传播研究

为了正确反映延误航班与机场随机因素之间的潜在关系,提出了1种基于贝叶斯网络的航班延误和机场随机因素分析方法.将机场随机延误因素的影响转化为时间参数.在综合考虑航班延误、航班计划的同时引入贝叶斯网络分析结果对延误波及树进行修正,在历史航班运营数据集上对修正延误波及树模型进行检验.结果表明,修正延误波及树模型提高了航班延误传播的预测精度,预测平均误差分别由修正前的11.76%和13.75%降低至3.45%和3.39%,多案例分析的准确性同时验证了修正延误波及树的普适性,丰富了航班延误传播问题的研究方法. 

A Study of the Spread of Flight Delay Based on the Modified Delay Propagation Tree

In order to illustrate the potential relationship between airport flight delay and random factors,a method is proposed to analyze the flight delay and stochastic factors based on the Bayesian network.First,the influence of random delay factors is transferred into time index.Secondly,after a comprehensive consideration of flight plans,airport flight delays,and the analysis result of Bayesian network,a modified delay propagation tree is proposed,and the historical flight data is used to test the modified delay propagation tree.The results show that,the modified model improves the prediction accuracy of flight delay propagation,and the average prediction error has been reduced from 11.76% and13.75%to 3.39%and 3.45%respectively.Meanwhile,the results from multi-case analysis also verify the usefulness of the modified delay propagation tree.This paper also provides an alternative method for studying flight delay propagation problem.