基于广义动态模糊神经网络的短时车站进站客流量预测

Short-term Entrance Passenger Flow Forecast at Urban Rail Transit Station Based on Generalized Dynamic Fuzzy Neural Networks

针对轨道交通车站短时进站客流的不均衡性、高度非线性和时变性特点,结合逻辑推理能力强的模糊技术与自学习能力强的神经网络,提出一种基于广义动态模糊神经网络(GD-FNN)的短时进站客流量预测方法。以北京轨道交通各车站的进站客流量数据为例,分析轨道交通车站的进站客流特征,确定影响短时客流分布的主要因素;然后采用GD-FNN方法构建车站短时进站量的预测模型,实现对北京轨道交通系统若干车站进站量的预测。预测结果表明:该方法与传统的神经网络相比,预测效果更准确(最大相对误差小于8%),稳定性好。

Directing against the characteristics of unbalanced, highly nonlinear and dynamic short-term entrance passenger flow of rail transit, the paper presents a short-term entrance passenger flow forecast method based on generalized dynamic fuzzy neural networks（GD-FNN）, which combines the logical reasoning ability of fuzzy technology with the self-learning ability of neural networks. The paper determines the main factors affecting the distribution of short- term passenger tlow by analyzing the features of passenger flow with the data of Beijing metro stations; and then a forecast model has been established by using GD-FNN to predict the short-term entrance passenger flow; finally, several stations in Beijing rail transit are used as numerical examples to confirm that this model can precisely approximate to the practical data （maximum relative error is less than 8% ） and has good stability compared with traditional neural networks.