基于有限穿越可视图的进场航班流量波动特性研究

研究空中交通流量的波动特性是设计高效流量管理措施和控制策略的基础,掌握空中交通流量波动特性有利于空域资源配置与流量运行需求之间的均衡匹配。在3种时间粒度上,针对进场航班流量时间序列,一方面从复杂网络整体维度出发,采用有限穿越可视图对时间序列进行建网,利用k-core算法探究航班流量波动特性;另一方面从复杂网络局部维度出发,引入序模体方法,构造有限穿越可视图序模体,利用多元序模体类型转换规律来刻画流量的动态转移模式,进而掌握航班流量波动动态演化规律,为波动模式的预测提供了有效工具。研究结果表明：在有限穿越可视图方法映射得到的网络中,节点所属核阶数可以有效刻画流量波动强度,且与波动强度成正相关关系,即节点所属核阶数越大,波动强度越大,天津机场进场航班流量数据的强波动时段为16:50-17:30;序模体越长,波动特性刻画能力越强,但鉴于受到空中交通混沌特性影响,序模体过长对于流量预测意义不大,推荐使用5节点序模体;波动模式状态转移图在有效刻画流量波动动态演化的同时,也可以计算波动模式的转移概率,3种时间粒度下转移概率分别为12.315%、13.131%和10.638%,为波动模式的预测提供了有效工具。

Fluctuation Characteristics of Arrival Flight Flow Based onLimited Penetrable Visibility Graph

Studying the fluctuation characteristics of air traffic flow is the basis for designing efficient management andcontrol strategies. Understanding the fluctuation characteristics of air traffic flow is conducive to the balance betweenairspace resource allocation and demand. In three time granularities, this paper uses the limited penetrable visibilitygraph method to build the complex network for the time series and explores the fluctuation characteristics of the flightflow with the k-core kernel algorithm from the overall perspective of the complex network, based on the time series ofincoming flight traffic. The motif method is used to construct the sequence motif of the limited penetrable visibilitygraph, and the type conversion law of multivariate sequence motif is used to describe the dynamic transfer mode oftraffic flow, so as to grasp the regular pattern of the dynamic evolution of flight traffic fluctuation. The method providesan effective tool for the prediction of fluctuation mode. It is found that: (1) In the network mapped by the limitedpenetrable visibility graph method, the k-core order of the node can effectively describe the fluctuation intensity oftraffic flow, and has a positive correlative relationship with the fluctuation intensity. It means that the greater the k-coreorder of the node, the greater the fluctuation intensity, and the strong fluctuation period of arrival flight flow data ofTianjin airport is 16:50-17:30; (2) Although the longer the motif is, the more dynamic the motif can be, and the longermotif has no significance for the prediction of traffic flow under the influence of the chaotic characteristics of air trafficflow. For the research on the dynamic evolution of air traffic flow fluctuation, a 5-node motif is recommended. (3) The state transition diagram of fluctuation patterns can not only effectively describe the dynamic evolution of flowfluctuation, and it can also calculate the transition probability of fluctuation patterns. The transition probabilities underthe three time granularities are 12.315%, 13.131%, and 10.638%, respectively. The state transition diagram provides aneffective tool for the prediction of fluctuation patterns.