摘要
空中交通系统作为典型复杂系统,其非线性聚合的动力学特征给延误预测带来挑战,使空中交通延误预测问题保持了开放性。近二十年来,针对此问题,利用传统机器学习、深度学习、复杂系统建模仿真、排队理论等,涌现出大量研究成果。首先系统性阐述了空中交通延误产生与传播的原因,并根据预测数值类型、预测对象以及预测时间尺度对现有研究中的延误预测类型进行分类;然后纵贯国内外研究成果,对空中交通延误预测方法进行分类,详细回顾了代表性方法的实现过程,并进行综合性对比。最后根据现有研究存在的问题,讨论了未来可能的发展方向。
The air traffic system is a typical complex system, and its nonlinear aggregate dynamic characteristics challenge the delay prediction, which keeps the problem of air traffic delay prediction open. In the past two decades, a large number of research results have emerged in response to this problem using traditional machine learning, deep learning, complex system modeling and simulation, and queueing theory. Firstly, the reasons for the occurrence and propagation of air traffic delays are explained systematically, and the types of delay predictions in existing research are classified according to the types of prediction values, prediction objects, and prediction time scales. Then, the relative studies at home and abroad are reviewed and the air traffic delays are analyzed. The forecasting methods are classified, the implementation process of representative methods is reviewed in detail, and a comprehensive comparison is made. Finally, based on the existing research problems, the possible future development directions is discussed.
作者
王春政
胡明华
杨磊
赵征
WANG Chunzheng;HU Minghua;YANG Lei;ZHAO Zheng(College of Civil Aviation,Nanjing University of Aeronautics and Astronautics,Nanjing 211106,China;National Key Laboratory of Air Traffic Flow Management,Nanjing 211106,China)
出处
《系统工程与电子技术》
EI
CSCD
北大核心
2022年第3期863-874,共12页
Systems Engineering and Electronics
基金
国家自然科学基金(61903187)
江苏省自然科学基金(BK20190414)资助课题。
关键词
空中交通
延误预测
机器学习
建模仿真
排队理论
air traffic
delay prediction
machine learning
modeling and simulation
queueing theory