基于动态反馈神经网络的城市轨道交通短期客流预测

城市轨道交通短期客流的预测是制定列车调度计划、车站客运组织工作等的关键。文章在研究城市轨道交通的断面客流特征的基础上,建立基于动态反馈神经网络的城市轨道交通短期客流预测模型,提出通过若干组连续历史断面客流数据训练动态连续的神经网络,以此对未来客流进行预测;并以北京轨道交通某日早高峰客流为例进行分析,验证了该模型与方法的有效性。

Short-term Passenger Flow Forecasting for Urban Rail Transit Based on Dynamic Feedback Neural Network

The short-term forecast of urban rail transit passenger flow is the key of developing and adjusting train plan and passenger organization of the stations. Based on the research of the characteristics of urban rail transit section passenger flow, the paper establishs the urban rail transit passenger flow short-term forecasting model based on dynamic feedback neural network. The future passenger flow is forecasted based on dynamic neural networks which trained by several groups of continuous historical section passenger flow data. Taking morning peak passenger flow of Beijing rail transit as an example, the effectiveness of the model and method are verified.