| 引用本文: | 冯树民,刘浩,李来成.雨雪天气下轨道交通客流预测模型[J].哈尔滨工业大学学报,2022,54(9):1.DOI:10.11918/202106043 |
| FENG Shumin,LIU Hao,LI Laicheng.Prediction model of rail transit passenger flow in rain and snow weather[J].Journal of Harbin Institute of Technology,2022,54(9):1.DOI:10.11918/202106043 |
|
| 摘要: |
| 为了完善雨雪天气下轨道交通客流预测模型,对哈尔滨市地铁1号线2017年12月份至2019年1月份的全线客流数据进行研究,引入客流基准值和客流偏差率的指标来量化轨道交通客流波动情况,研究雨雪天气下轨道交通客流波动规律,提出一种基于雨雪天气下轨道交通客流时空波动规律的短时客流预测模型WI-LSTM,以平均绝对误差（MAE）、均方根误差（RMSE）以及平均相对误差（MRE）作为预测模型的评价指标,与经典的SARIMA预测模型、支持向量回归（SVR）预测模型和未考虑雨雪天气的LSTM预测模型进行了对比。结果表明:考虑雨雪天气的WI-LSTM预测模型可以充分利用雨雪天气轨道交通客流波动规律,相比其他3种预测模型具有更高的准确性和可靠性。WI-LSTM预测模型进一步提升了雨雪天气下轨道交通客流预测精度,可为轨道交通企业运营管理提供数据支撑。 |
| 关键词: 城市轨道交通 雨雪天气 客流预测 客流波动规律 LSTM神经网络 |
| DOI：10.11918/202106043 |
| 分类号:U491.1+4 |
| 文献标识码:A |
| 基金项目:国家自然科学基金(71771062) |
|
| Prediction model of rail transit passenger flow in rain and snow weather |
|
FENG Shumin,LIU Hao,LI Laicheng
|
|
(School of Transportation Science and Engineering, Harbin Institute of Technology, Harbin 150090, China)
|
| Abstract: |
| For improving the rail transit passenger flow prediction models under rain and snow weather conditions, the passenger flow data of Harbin Metro Line 1 from December 2017 to January 2019 was studied, and the indicators of passenger flow benchmark and passenger flow deviation rate were introduced to quantify the passenger flow of rail transit. The fluctuation rule of rail transit passenger flow under rain and snow weather conditions was analyzed, and a WI-LSTM prediction model was proposed based on the temporal and spatial fluctuation of rail transit passenger flow in rain and snow weather. The mean absolute error (MAE), root mean square error (RMSE), and mean relative error (MRE) were used as the evaluation indexes of the prediction model. The proposed model was compared with SARIMA prediction model, support vector machine (SVR) prediction model, and the LSTM prediction model without considering rain and snow weather. Results show that the WI-LSTM model considering rain and snow weather could make full use of the fluctuation rule of rail transit passenger flow in rain and snow weather, and achieved higher accuracy and reliability than the other three prediction models. The proposed WI-LSTM model further improves the accuracy of rail transit passenger flow forecast in rain and snow weather, and can provide data support for the operation and management of rail transit enterprises. |
| Key words: urban rail transit rain and snow weather passenger flow prediction rail transit fluctuation rule LSTM neural network |
