基于时间序列分析的航站楼安检旅客流量预测

对单位时间内通过安检的旅客流量进行预测是机场航站楼实时调控的重要依据,由此提出一种实时安检旅客流量预测方法,采用Wolf方法分析出安检旅客流量时间序列具有混沌特性;采用适用于混沌时间序列预测的遗传算法优化BP神经网络预测方法(GABP)预测安检旅客流量;分别设定时间尺度为2 min、5 min和10 min等,分析不同时间尺度对安检旅客流量预测精度的影响。基于北京首都国际机场T3航站楼实际安检旅客流量数据的实验结果表明,采用GABP神经网络对以2 min为时间尺度的安检旅客流量预测能取得较好的预测精准度。

Prediction of security check passenger flow in terminal building based on time series analysis

The passenger flow prediction through security check in unit time is an important part of the traffic system,a method of the security check passenger flow prediction was presented.The Wolf method was used to analyze the chaotic characteristics of the security check passenger flow time series.BP neural network optimized by genetic algorithm suitable for chaotic time series prediction was used to predict the security check passenger flow.Setting time scales of 2 min,5 min and 10 min respectively,the influence of different time scales on the prediction accuracy of the security check passenger flow was analyzed.Based on the actual passenger flow data of Beijing Capital International Airport T3 terminal,the experimental results show that the GABP neural network has a better prediction performance on the security check passenger flow prediction,in terms of taking 2 min as a time scale.