基于车站仿真和LSTM的轨道交通换乘站动态性能评估方法

doi:10.16182/j.issn1004731x.joss.21-1042

系统仿真学报 ›› 2023, Vol. 35 ›› Issue (3): 544-556.doi: 10.16182/j.issn1004731x.joss.21-1042

基于车站仿真和LSTM的轨道交通换乘站动态性能评估方法

何必胜¹^,²^,³(), 张宏翔¹, 朱永俊¹, 鲁工圆¹^,²^,³()

^1.西南交通大学交通运输与物流学院, 四川成都 611756
^2.综合交通运输智能化国家地方联合工程实验室, 四川成都 611756
^3.综合运输四川省重点实验室, 四川成都 611756

收稿日期:2021-10-13 修回日期:2021-12-06 出版日期:2023-03-30 发布日期:2023-03-22
通讯作者: 鲁工圆 E-mail:bishenghe@swjtu.edu.cn;lugongyuan@swjtu.edu.cn
作者简介:何必胜(1986-)，男，副教授，博士，研究方向为交通运输规划与管理，交通运输系统仿真。E-mail：bishenghe@swjtu.edu.cn
基金资助:
国家重点研发计划(2017YFB1200700);国家自然科学基金(61603317);中国铁路总公司科技研究开发计划(J2018z403)

Dynamic Performance Evaluation Method for Transfer in Rail Transit Station Based on Station Simulation and LSTM

Bisheng He¹^,²^,³(), Hongxiang Zhang¹, Yongjun Zhu¹, Gongyuan Lu¹^,²^,³()

^1.School of Transportation and Logistics, Southwest Jiaotong University, Chengdu 611756, China
^2.National United Engineering Laboratory of Integrated and Intelligent Transportation, Chengdu 611756, China
^3.Comprehensive Transportation Key Laboratory of Sichuan Province, Chengdu 611756, China

Received:2021-10-13 Revised:2021-12-06 Online:2023-03-30 Published:2023-03-22
Contact: Gongyuan Lu E-mail:bishenghe@swjtu.edu.cn;lugongyuan@swjtu.edu.cn

摘要/Abstract

摘要：

在轨道交通客流量快速增长的背景下，为有效评估换乘车站运营过程中动态性能，提出基于机器学习的换乘站动态性能评估方法。基于提出的有效换乘人数、换乘时间和拥挤度的动态性能评价指标，深入分析了车站动态性能的影响因素，利用考虑行车与行人的换乘车站仿真模型提供机器学习所需的时间序列数据，采用长短期记忆(long short-term memory，LSTM)的机器学习方法，建立评价指标的预测获取方法，动态获取车站在不同条件下的运营状况。以犀浦站为例，运用仿真模型构建的2.24万个样本来训练预测模型。预测结果证明了预测模型的精度，并量化了购票比例对车站动态性能的影响。所提方法能够为轨道交通换乘车站换乘组织和客运作业提供有效建议。

关键词: 轨道交通, 换乘车站, 动态性能, 时间序列, 长短期记忆(LSTM)

Abstract:

Given the boom increasing of rail transit passenger volume, the dynamic performance evaluation method for transfer in rail transit stations based on machine learning are proposed to effectively evaluate the performance of the transfer station in different scenarios. Based on the proposed dynamic performance evaluation indexes of effective transfer number, transfer time and congestion, the influence factors of station dynamic performance are analyzed. The simulation model integrated train operation and pedestrian movement is built to provide the time-series data for the machine learning method. The long short-term memory (LSTM) is implemented to forecast the evaluation indicators, and the station evaluation results can be obtained dynamically under different operational conditions. 22,400 samples generated by the simulation model as the train data are used to train the forecasting model with the Xipu station. The forecasting results demonstrate the accuracy of forecasting model. The impact of ticket purchase ratio on the dynamic performance of the station is quantified. The proposed station dynamic performance evaluation method is proved to be effective and efficient for the operation and organization of the transfer station.

Key words: rail transit, transfer station, dynamic performance, time series, long short-term memory (LSTM)

中图分类号:

TP391.9

何必胜, 张宏翔, 朱永俊, 鲁工圆. 基于车站仿真和LSTM的轨道交通换乘站动态性能评估方法[J]. 系统仿真学报, 2023, 35(3): 544-556.

Bisheng He, Hongxiang Zhang, Yongjun Zhu, Gongyuan Lu. Dynamic Performance Evaluation Method for Transfer in Rail Transit Station Based on Station Simulation and LSTM[J]. Journal of System Simulation, 2023, 35(3): 544-556.

图/表 11

表1

图1

表2

图2

图3

表3

不同类型时段下动态性能影响因素的取值范围

影响因素	高峰	平峰	低峰
$D 1 t r a$	384~448	320~384	256~320
$R 1 b$	0.25~0.55	0.25~0.55	0.25~0.55
$D 1 A$	560~640	480~560	400~480
$D 1 i n$	6 000	4 800	3 600
$D 2 t r a$	72~90	54~72	36~54
$R 2 b$	0.15~0.4	0.15~0.4	0.15~0.4
$D 2 A$	280~360	200~280	120~200
( $N 1 A G, r a i l$ , $N T V M$ , $N 1 A G, m e t r o$ )	(7, 8, 7), (7, 7, 6)	(6, 7, 6), (6, 6, 5), (5, 6, 5)	(5, 5, 4), (4, 5, 4)
( $N 2 A G, r a i l$ , $N 2 A G, m e t r o$ )	(7, 5), (6, 5)	(5, 5), (5, 4)	(4, 4), (4, 3)
$N I D$	(2/2)	(3/1)	(4/0)
( $H a r i_t r a i n$ , $T c h e c k$ )	(10, 8)	(12, 10), (20, 15)	(30, 25)
$H a r i_m e t r o$	2, 2.5	3, 3.5	4

表3

表4

部分仿真场景特征值取值及仿真结果

仿真场景序号	仿真场景特征取值						仿真运行结果			结果校核
仿真场景序号	$D 1 t r a$	$R 1 b$	$D 1 A$	$D 1 i n$	( $N 1 A G, r a i l$ , $N T V M$ , $N 1 A G, m e t r o$ )	$H a r i_m e t r o$	$P τ$	$A N t r a, τ$	$T t r a, τ$	$A N t r a, τ$ 相对偏差/%	$T t r a, τ$ 相对偏差/%
1	430	0.35	625	6 000	(7, 8, 7)	2	A	428	4.35	6.92	5.36
2	445	0.40	620	6 000	(7, 7, 6)	2	A	419	4.76	2.57	3.16
3	370	0.35	550	4 800	(5, 6, 5)	3	B	359	4.99	5.33	4.40
4	365	0.40	550	4 800	(6, 6, 5)	3	B	347	5.19	2.98	7.02
5	310	0.40	460	3 600	(5, 5, 4)	4	B	302	5.58	3.72	6.59

表4

表5

图4

图5

表6

购票比例变化下成灌铁路换乘地铁时评价指标的预测结果及变化情况

购票比例	50%			40%			30%			25%
评价指标	$P p r e τ$	$A N p r e t r a, τ$	$T p r e t r a, τ$	$P p r e τ$	$A N p r e t r a, τ$	$T p r e t r a, τ$	$P p r e τ$	$A N p r e t r a, τ$	$T p r e t r a, τ$	$P p r e τ$	$A N p r e t r a, τ$	$T p r e t r a, τ$
06:00—07:00	A	265	4.28	A	270	4.15	A	280	3.95	A	279	3.88
07:00—08:00	D	328	7.15	C	377	5.51	B	406	4.13	A	413	4.05
08:00—09:00	D	322	7.15	C	378	5.39	A	411	4.05	A	409	4.01
09:00—10:00	C	320	6.67	B	349	4.70	A	353	4.06	A	359	4.02
10:00—11:00	B	327	5.89	A	352	4.49	A	351	3.88	A	352	3.99
11:00—12:00	B	329	6.01	B	354	4.73	B	351	4.03	A	355	3.94
12:00—13:00	C	323	6.58	B	348	4.68	A	358	3.92	A	360	4.01
13:00—14:00	B	315	5.99	B	358	4.67	A	358	4.03	A	358	3.98
14:00—15:00	B	329	5.95	B	350	4.59	A	354	4.04	A	351	4.04
15:00—16:00	B	325	6.00	B	356	4.68	A	353	4.06	A	351	4.03
16:00—17:00	B	330	6.61	B	350	4.71	A	359	3.99	A	357	4.07
17:00—18:00	D	319	7.08	C	381	5.48	B	410	4.17	B	414	4.11
18:00—19:00	D	322	7.03	C	379	5.53	A	409	4.03	A	415	4.06
19:00—20:00	C	322	6.45	B	355	4.80	A	355	3.96	A	354	3.93
20:00—21:00	B	323	6.02	B	352	4.61	A	352	4.01	A	356	4.02
21:00—22:00	B	293	5.51	A	290	4.19	A	298	3.89	A	299	3.91

表6

参考文献 19

1	Corman F, D'Ariano A, Marra A D, et al. Integrating Train Scheduling and Delay Management in Real-Time Railway Traffic Control[J]. Transportation Research Part E: Logistics and Transportation Review (S1366-5545), 2017, 105: 213-239.
2	杨天阳, 朱志国. 基于AnyLogic的地铁车站通道设施设备规模与布局分析[J]. 交通运输工程与信息学报, 2017, 15 (1): 115-121.
	Yang Tianyang, Zhu Zhiguo. Scale and Layout Analyses of Subway Station Facilities Based on Anylogic Software[J]. Journal of Transportation Engineering and Information, 2017, 15(1): 112-121.
3	胡春平. 客运枢纽换乘效率及流线组织方法研究[D]. 北京: 北京交通大学, 2014.
	Hu Chunping. Transfer Efficiency Analysis and Passenger Flow Organization of Passenger Transportation Hubs[D]. Beijing: Beijing Jiaotong University, 2014.
4	Hänseler F S, Bierlaire M, Scarinci R. Assessing the Usage and Level-of-Service of Pedestrian Facilities in Train Stations: A Swiss Case Study[J]. Transportation Research Part A: Policy and Practice (S0965-8564), 2016, 89(7): 106-123.
5	Transportation Research Board. Highway Capacity Manual[K]. Washington, DC: Transportation Research Board, 2010.
6	胡清梅. 轨道交通车站客流承载能力的评估与仿真研究[D]. 北京: 北京交通大学, 2011.
	Hu Qingmei. Passenger Carrying Capacity Evaluation and Simualtion of Rail Transit Station[D]. Beijing: Beijing Jiaotong University, 2011.
7	黄令海, 李海鹰, 许心越. 城市轨道交通车站动态瓶颈识别方法研究[J]. 铁道学报, 2015, 37(7): 1-8.
	Huang Linghai, Li Haiying, Xu Xinyue. Identification of Dynamic Bottlenecks of Urban Rail Transit Stations[J]. Journal of the China Railway Society, 2015, 37(7): 1-8.
8	Van den Heuvel J, Voskamp A, Daamen W, et al. Using Bluetooth to Estimate the Impact of Congestion on Pedestrian Route Choice at Train Stations[C]// Traffic and Granular Flow '13. German: Jülich, 2013: 73-82.
9	高春霞, 董宝田. 交通枢纽内不确定情况下行人路径选择模型研究[J]. 交通运输系统工程与信息, 2012, 12(6): 170-176, 197.
	Gao Chunxia, Dong Baotian. Passenger Routing Selection Model with Uncertainty in Traffic Hub[J]. Journal of Transportation Systems Engineering and Information Technology, 2012, 12(6): 170-176, 197.
10	Pu Y S. Capacity Analysis of the Union Station Rail Corridor Using Integrated Rail and Pedestrian Simulation[D]. Toronto: University of Toronto, 2017.
11	Hänseler F S. Modeling and Estimation of Pedestrian Flows in Train Stations[D]. Lausanne: EPFL, 2016.
12	胡瑞, 徐传玲, 冯永泰, 等. 广深高速铁路列车分类型晚点预测[J]. 中国安全科学学报, 2019, 29(增2): 181-186.
	Hu Rui, Xu Chuanling, Feng Yongtai, et al. Prediction of Different Types of Train Delay of Guangzhou-Shenzhen High-speed Railway[J]. China Safety Science Journal, 2019, 29(S2): 181-186.
13	王岩韬, 陈冠铭. 基于时间序列模型的航班运行风险短期预测[J]. 中国安全科学学报, 2020, 30(5): 33-38.
	Wang Yantao, Chen Guanming. Short-time Prediction of Flight Operation Risk Based on Time Series Models[J]. China Safety Science Journal, 2020, 30(5): 33-38.
14	He Bisheng, Zhang Hongxiang, Wen Keyu, et al. Machine Learning based Integrated Pedestrian Facilities Planning and Staff Assignment Problem in Transfer Stations[C]// 8th International Conference on Railway Operations Modelling and Analysis. Sweden: IAROR, 2019: 579-600.
15	Zhang H X, He B S, Lu G Y, et al. A Simulation and Machine Learning Based Optimization Method for Integrated Pedestrian Facilities Planning and Staff Assignment Problem in the Multi-mode Rail Transit Transfer Station[J]. Simulation Modelling Practice and Theory (S1569-190X), 2022, 115: 102449.
16	丁小兵, 徐行方. 基于时段分布的市郊轨道交通车站滞留客流分布算法[J]. 交通运输系统工程与信息, 2016, 16(1): 116-122.
	Ding Xiaobing, Xu Xingfang. Algorithms of Station Passenger Flow Forecast of Suburban Rail Transit Based on Distribution Time[J]. Journal of Transportation Systems Engineering and Information Technology, 2016, 16(1): 116-122.
17	宋瑞, 邹彦雯, 李婷婷. 城市轨道交通线路运营计划优化方法[J]. 北京交通大学学报, 2015, 39(6): 32-38.
	Song Rui, Zou Yanwen, Li Tingting. Optimization Method of Operation Planning for Urban Rail Transit Line[J]. Journal of Beijing Jiaotong University, 2015, 39(6): 32-38.
18	中华人民共和国住房和城乡建设部. , 地铁设计规范 [S]. 北京: 中国建筑工业出版社, 2013.
	Ministry of Housing and Urban-Rural Development of the People's Republic of China. , Code for Design of Metro [S]. Beijing: China Architecture & Building Press, 2013.
19	Dong H R, Zhou M, Wang Q L, et al. State-of-the-Art Pedestrian and Evacuation Dynamics[J]. IEEE Transaction on Intelligent Transportation Systems (S1524-9050), 2020, 21(5): 1849-1866.

拥挤度等级	通行能力/(人l·(m·min)^-1)	行人速度/(m·s^-1)	人均空间/(m²/人)
A	≤16	>1.28	>3.85
B	16~27	1.11~1.28	2.15~3.85
C	27~42	0.91~1.11	1.37~2.15
D	42~68	0.69~0.91	0.86~1.37
E	68~97	0.55~0.69	0.54~0.86
F	波动	<0.55	<0.54

乘客因素	设备设施因素	行车组织因素
进站客流量Dⁱⁿ	流线布设方案F	换乘列车开行间隔H^tra
出站客流量D^out	楼梯宽度W	列车停站时间T^stop
到站客流量D^A	扶梯运行速度S^E	列车满载率(城市轨道交通)R^M
换乘客流量D^tra	安检设备开放数量N^SC	提前检票时长(铁路)T^check
乘客走行速度S^d	自动取/售票机开放数量N^TVM	到达列车开行间隔H^ari
使用单程票的比例R^b	检票闸机开放数量N^AG
	实名制验证开放数量N^ID

预测项	MAPE/%
拥挤度	3.60
平均换乘时间(换乘地铁)	1.90
有效换乘人数(换乘地铁)	4.38
平均换乘时间(换乘铁路)	3.23
有效换乘人数(换乘铁路)	0.70

[1]	陈伟, 李宗平, 刘灿, 鞠艳妮. 考虑拥堵传播的轨道交通车站站台承载人数研究[J]. 系统仿真学报, 2022, 34(7): 1582-1592.
[2]	胡剑鹏, 罗霞. 城市轨道交通动态客流分配仿真方法研究[J]. 系统仿真学报, 2022, 34(3): 512-526.
[3]	冯国亮, 卢伟, 杨建华. 用于时间序列数据建模的多模态模糊认知图[J]. 系统仿真学报, 2022, 34(3): 543-554.
[4]	李怡桐, 刘晓涛, 刘静, 吴凯. 基于加权局部复杂不变性的时间序列分类算法[J]. 系统仿真学报, 2022, 34(10): 2194-2203.
[5]	刘浩, 毛宏霞, 肖志河, 刘铮. 基于动态时间规整的火箭尾喷焰识别仿真[J]. 系统仿真学报, 2022, 34(1): 126-133.
[6]	刘枫, 邵春福, 贾洪飞, 董春娇. 城轨换乘站服务设施布局与流线优化仿真评价[J]. 系统仿真学报, 2021, 33(7): 1670-1681.
[7]	路雪鹏, 尚娇, 赵俊辉, 吕露露, 周丽. 基于系统动力学的新冠病毒传播过程预测[J]. 系统仿真学报, 2021, 33(7): 1713-1721.
[8]	李晓赫, 吴建平, 彭德品. 城轨站点高峰小时客流预测控制方法[J]. 系统仿真学报, 2021, 33(12): 2952-2958.
[9]	许慧, 田铖, 王永. 轨道交通换乘站密集客流应急疏散仿真研究[J]. 系统仿真学报, 2020, 32(3): 492-500.
[10]	杜恒, 郜春海, 黄勍, 仓怀明. 城市轨道交通全自动运行系统仿真研究及应用[J]. 系统仿真学报, 2020, 32(2): 157-163.
[11]	连晓峰, 叶璐, 王炎, 贾利民, 马慧茹. 基于FAST-ICA的城市轨道交通乘客路径选择方法[J]. 系统仿真学报, 2019, 31(8): 1692-1701.
[12]	冯振杰, 樊煜. 混沌时间序列的高斯过程混合模型预测[J]. 系统仿真学报, 2019, 31(7): 1387-1396.
[13]	许伊婷, 张蕊. 基于同伴群行人仿真的通道设置研究[J]. 系统仿真学报, 2019, 31(11): 2477-2484.
[14]	王剑, 沈丹青, 王子阳. 轨道交通与常规公交线路协调关系建模与仿真[J]. 系统仿真学报, 2019, 31(10): 1995-2009.
[15]	董兴辉, 马晓双, 程友星, 王帅. 风电机组轴承健康劣化趋势建模与仿真[J]. 系统仿真学报, 2019, 31(1): 151-158.

基于车站仿真和LSTM的轨道交通换乘站动态性能评估方法

Dynamic Performance Evaluation Method for Transfer in Rail Transit Station Based on Station Simulation and LSTM

RichHTML

PDF (PC)

可视化

摘要/Abstract

引用本文

使用本文

图/表 11

参考文献 19

相关文章 15

编辑推荐

Metrics

本文评价