机场交通需求概率密度预测方法研究  

Research on Probability Density Prediction Method of Airport Traffic Demand

作　　者：田文[1] 杨帆[1] 郭怡杏 宋津津 TIAN Wen;YANG Fan;GUO Yi-xing;SONG Jin-jin(College of Civil Aviation,Nanjing University of Aeronautics and Astronautics,Nanjing Jiangsu 211106,China)

机构地区：[1]南京航空航天大学民航学院,江苏南京211106

出　　处：《计算机仿真》2022年第1期56-59,共4页Computer Simulation

基　　金：国家自然科学基金项目(71971112、61903187);江苏省自然科学基金项目(BK20190416,K20190414);南京航空航天大学研究生创新基地(实验室)开放基金(kfjj20190717)。

摘　　要：针对基于进离场航空器统计数量的传统预测方法较难体现不确定性因素对机场交通需求预测结果影响的问题,采用核密度估计与神经网络分位数回归相结合的方法,构建了机场交通需求概率密度预测模型和方法,并基于广州白云机场运行数据得到机场交通需求在未来一定时间内的概率密度分布。预测结果相较传统BP神经网络预测,所得结果波动范围更小,误差更小。In view of the problem that the traditional prediction method based on the number of entry and departure aircraft is is difficult to reflect the impact of uncertainty factors on the prediction results of airport traffic demand, a probability density prediction model and the method of airport traffic demand were constructed by combining the nuclear density estimation with the quantile regression of neural network. Then, based on the operation data of ZGGG,the probability density distribution of this airport traffic demand in the future was obtained.Compared with the traditional BP neural network, the predicted results have smaller fluctuation range and smaller error.

关 键 词：分位数回归神经网络 核密度估计 概率密度预测 机场交通需求 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术]