基于EEMD-BP方法的城市轨道交通进站客流短期预测

城市轨道交通站点客流量是评价其服务水平和实现城市轨道交通资源有效配置的基础数据和依据。针对城市轨道交通站点进站客流量序列波动复杂的问题,构建基于EEMD-BP方法的城市轨道交通进站客流短期预测模型,通过对城市轨道交通站点的日间分时进站客流序列进行模态分解,并对分解的分量进行筛选和识别,探究进站客流的日间波动影响因素,实现对短期客流的合理预测。以广州珠江新城站短期客流预测为例,验证该组合模型在提高客流预测方面具有有效性,为城市轨道交通线网规划和运营管理提供客流预测依据。

Short-term Forecast of Passenger Flow into an Urban Rail Transit Station based on EEMD-BP

As urban rail transit becomes a key point for the development of smart transportation, its passenger flow forecast is of great significance to planning and operation management of urban rail transit network. In this paper, in view of the complex fluctuation of metro passenger flow sequence, showing nonlinear and non-stationary characteristics in spatial and temporal distribution, the prediction is constructed by Ensemble Empirical Mode Decomposition(EEMD) and BP neural network. The model is decomposed by the time-division passenger flow sequence of the metro station, and the components of the decomposition are screened and identified. The factors affecting the daytime fluctuation of the passenger flow of the metro station can be explored. Then, BP neural network is used to forecast short-term passenger flow. Taking Guangzhou Metro as an example, the results show that the combined model is effective in improving passenger flow forecasting.