基于排队网络模型的机场航班延误预测

机场航班量不断增长,必然会带来机场高峰时段延误水平的增加。因此机场延误水平的科学预测对确保机场运行效率具有重要意义。论文首先根据航空器运行特性建立机场排队网络模型；然后利用Lempel-Ziv算法计算不同时间尺度的航班时间序列复杂度,确定刻画航班延误的小时间尺度,由此确定排队网络模型参数,并用实例进行验证；最后运用AirTop仿真软件以全天平均延误、高峰小时平均延误作为关键指标,仿真得出机场延误水平变化趋势。通过将仿真数据与计算结果进行对比分析发现,机场排队网络模型能够较好的反应真实情况下的机场延误水平；而选用小时间尺度15 min进行机场排队网络模型参数计算,结果更贴近真实机场延误水平。

Airport Flight Delay Prediction Based on Queuing Network Model

The increasing volume of airport flights inevitably brings about the increase of airport delay level during peak hours, so the scientific prediction of airport delay level is of great significance to ensure the efficiency of airport operation. The paper firstly established an airport queuing network model based on aircraft operation characteristics; then used the Lempel-Ziv algorithm to calculate the complexity of flight time series at different time scales, determined the small time scale to portray flight delays, Thus, the queuing network model parameters were determined and verified by example. Finally, the AirTop simulation software was used to simulate the trend of airport delay level with the average delay of the whole day and the average delay of the peak hour as the key indicators. By comparing the simulated data with the calculated results, it is found that the airport queuing network model can better reflect the real airport delays; And the small time scale of 15 min is chosen for the calculation of airport queuing network model parameters, the results are closer to the real airport delay level.