Dynamic train dwell time forecasting: a hybrid approach to address the influence of passenger flow fluctuations

Abstract

Train timetables and operations are defined by the train running time in sections, dwell time at stations, and headways between trains. Accurate estimation of these factors is essential to decision-making for train delay reduction, train dispatching, and station capacity estimation. In the present study, we aim to propose a train dwell time model based on an averaging mechanism and dynamic updating to address the challenges in the train dwell time prediction problem (e.g., dynamics over time, heavy-tailed distribution of data, and spatiotemporal relationships of factors) for real-time train dispatching. The averaging mechanism in the present study is based on multiple state-of-the-art base predictors, enabling the proposed model to integrate the advantages of the base predictors in addressing the challenges in terms of data attributes and data distributions. Then, considering the influence of passenger flow on train dwell time, we use a dynamic updating method based on exponential smoothing to improve the performance of the proposed method by considering the real-time passenger amount fluctuations (e.g., passenger soars in peak hours or passenger plunges during regular periods). We conduct experiments with the train operation data and passenger flow data from the Chinese high-speed railway line. The results show that due to the advantages over the base predictors, the averaging mechanism can more accurately predict the dwell time at stations than its counterparts for different prediction horizons regarding predictive errors and variances. Further, the experimental results show that dynamic smoothing can significantly improve the accuracy of the proposed model during passenger amount changes, i.e., 15.4% and 15.5% corresponding to the mean absolute error and root mean square error, respectively. Based on the proposed predictor, a feature importance analysis shows that the planned dwell time and arrival delay are the two most important factors to dwell time. However, planned time has positive influences, whereas arrival delay has negative influences.

References

[1.] Goverde RM. A delay propagation algorithm for large-scale railway traffic networks. Transp Res Part C Emerg Technol 2010, 18 3 269-287

[2.] Corman F, Meng L. A review of online dynamic models and algorithms for railway traffic management. IEEE Trans Intell Transp Syst 2014, 16 3 1274-1284

[3.] Parbo J, Nielsen OA, Prato CG. Passenger perspectives in railway timetabling: a literature review. Transp Rev 2016, 36 4 500-526

[4.] Li Z, Wen C, Hu R, Xu C, Huang P, Jiang X. Near-term train delay prediction in the Dutch railways network. Int J Rail Transp 2020, 9 6 520-539

[5.] Kecman P, Goverde RM. Predictive modelling of running and dwell times in railway traffic. Public Transport 2015, 7 295-319

[6.] Wen C, Mou W, Huang P, Li Z. A predictive model of train delays on a railway line. J Forecast 2020, 39 3 470-488

[7.] Huang P, Wen C, Fu L, Peng Q, Tang Y. A deep learning approach for multi-attribute data: a study of train delay prediction in railway systems. Inf Sci 2020, 516 234-253

[8.] Huang P, Li Z, Wen C, Lessan J, Corman F, Fu L. Modeling train timetables as images: a cost-sensitive deep learning framework for delay propagation pattern recognition. Expert Syst Appl 2021, 177 114996

[9.] Talebian A, Zou B. Integrated modeling of high performance passenger and freight train planning on shared-use corridors in the US. Transp Res Part B Methodol 2015, 82 C 114-140

[10.] Yang J, Shiwakoti N, Tay R (2019) Train dwell time models—development in the past forty years. In: Australasian Transport Research Forum 2019 Proceedings, Canberra.

[11.] Liu J, Schonfeld PM, Zhan S, Du B, He M, Wang K, Yin Y. The economic value of reserve capacity considering the reliability and robustness of a rail transit network. J Transp Eng Part A Syst 2023

[12.] Spanninger T, Trivella A, Büchel B, Corman F. A review of train delay prediction approaches. J Rail Transp Plan Manag 2022, 22 100312

[13.] Wen C, Huang P, Li Z, Lessan J, Fu L, Jiang C, Xu X. Train dispatching management with data-driven approaches: a comprehensive review and appraisal. IEEE Access 2019, 7 13 114547-114571

[14.] Kuipers RA, Palmqvist C-W, Olsson NO, Winslott Hiselius L. The passenger’s influence on dwell times at station platforms: a literature review. Transp Rev 2021, 41 6 721-741

[15.] Lessan J, Fu L, Wen C. A hybrid Bayesian network model for predicting delays in train operations. Comput Ind Eng 2019, 127 1214-1222

[16.] Li Z, Huang P, Wen C, Jiang X, Rodrigues F. Prediction of train arrival delays considering route conflicts at multi-line stations. Transp Res Part C Emerg Technol 2022, 138 103606

[17.] Huang P, Spanninger T, Corman F. Enhancing the understanding of train delays with delay evolution pattern discovery: a clustering and bayesian network approach. IEEE Trans Intell Transp Syst 2022, 23 9 15367-15381

[18.] Lessan J, Fu L, Wen C, Huang P, Jiang C. Stochastic model of train running time and arrival delay: a case study of Wuhan-Guangzhou high-speed rail. Transp Res Rec 2018, 2672 10 215-223

[19.] Li J, Li Z, Wen C . Train operation conflict detection for high-speed railways: a naive Bayes approach. Int J Rail Transp 2023, 11 2 188-206

[20.] Luo J, Peng Q, Wen C, Wen W, Huang P. Data-driven decision support for rail traffic control: a predictive approach. Expert Syst Appl 2022, 207 118050

[21.] Altinkaya M, Zontul M. Urban bus arrival time prediction: A review of computational models. Int J Recent Technol Eng 2013, 2 4 164-169

[22.] Dueker KJ, Kimpel TJ, Strathman JG, Callas S. Determinants of bus dwell time. J Public Transp 2004, 7 1 21-40

[23.] Coulaud R, Keribin C, Stoltz G. Modeling dwell time in a data-rich railway environment: with operations and passenger flows data. Transp Res Part C Emerg Technol 2023, 146 103980

[24.] Zhang C, Teng J. Bus dwell time estimation and prediction: a study case in Shanghai-China. Procedia Soc Behav Sci 2013, 96 1329-1340

[25.] Rashidi S, Ranjitkar P. Approximation and short-term prediction of bus dwell time using AVL data. J East Asia Soc Transp Stud 2013, 10 1281-1291

[26.] Lin W-H, Zeng J. Experimental study of real-time bus arrival time prediction with GPS data. Transp Res Rec 1999, 1666 1 101-109

[27.] Rajbhandari R, Chien SI, Daniel JR. (2003) Estimation of bus dwell times with automatic passenger counter information. Transp Res Rec 1841, 1 120-127

[28.] Glick TB, Figliozzi MA. Analysis and application of log-linear and quantile regression models to predict bus dwell times. Transp Res Rec 2019, 2673 10 118-128

[29.] Rashidi S, Ranjitkar P. Bus dwell time modeling using gene expression programming. Comput Aided Civil Infrastruct Eng 2015, 30 6 478-489

[30.] Meng Q, Qu X. Bus dwell time estimation at bus bays: a probabilistic approach. Transp Res Part C Emerg Technol 2013, 36 61-71

[31.] Padmanaban R, Vanajakshi L, Subramanian SC (2009) Estimation of bus travel time incorporating dwell time for APTS applications. In: 2009 IEEE Intelligent Vehicles Symposium, Xi’an, pp 955–959

[32.] Li F, Duan Z, Yang D. Dwell time estimation models for bus rapid transit stations. J Mod Transp 2012, 20 3 168-177

[33.] Li D, Goverde RM, Daamen W, He H (2014) Train dwell time distributions at short stop stations. In: 17th International IEEE Conference on Intelligent Transportation Systems (ITSC), Qingdao, pp 2410–2415

[34.] Karekla X, Tyler N. Reduced dwell times resulting from train–platform improvements: the costs and benefits of improving passenger accessibility to metro trains. Transp Plan Technol 2012, 35 5 525-543

[35.] Yang Y, Huang P, Peng Q, Jie L, Wen C. Statistical delay distribution analysis on high-speed railway trains. J Modern Transp 2019, 27 3 188-197

[36.] Pritchard J, Sadler J, Blainey S, Waldock I, Austin J. Predicting and mitigating small fluctuations in station dwell times. J Rail Transp Plan Manag 2021, 18 100249

[37.] Li D, Daamen W, Goverde RM. Estimation of train dwell time at short stops based on track occupation event data: a study at a Dutch railway station. J Adv Transp 2016, 50 5 877-896

[38.] Li D, Yin Y, He H. Testing the generality of a passenger disregarded train dwell time estimation model at short stops: both comparison and theoretical approaches. J Adv Transp 2018, 2018 1-16

[39.] Baee S, Eshghi F, Hashemi SM, Moienfar R (2012) Passenger boarding/alighting management in urban rail transportation. In: ASME/IEEE Joint Rail Conference, Philadelphia, pp 823–829

[40.] Jiang Z, Xie C, Ji T, Zou X. Dwell time modelling and optimized simulations for crowded rail transit lines based on train capacity. PROMET-Traffic Transp 2015, 27 2 125-135

[41.] Seriani S, Fernandez R. Pedestrian traffic management of boarding and alighting in metro stations. Transp Res Part C: Emerg Technol 2015, 53 76-92

[42.] Berbey A, Galan R, Bobi JS, Caballero R. A fuzzy logic approach to modelling the passengers’ flow and dwell time. WIT Trans Built Environ 2012, 128 359-369

[43.] Berbey Alvarez A, Merchan F, Calvo Poyo FJ, Caballero George RJ. A fuzzy logic-based approach for estimation of dwell times of panama metro stations. Entropy 2015, 17 5 2688-2705

[44.] Chu W-j, Zhang X-c, Chen J-h, Xu B. An ELM-based approach for estimating train dwell time in urban rail traffic. Math Probl Eng 2015, 2015 473432

[45.] Luo J, Huang P, Peng Q. A multi-output deep learning model based on Bayesian optimization for sequential train delays prediction. Int J Rail Transp 2022

[46.] Wang Y, Wen C, Huang P. Predicting the effectiveness of supplement time on delay recoveries: a support vector regression approach. Int J Rail Transp 2021, 10 3 375-392

[47.] Liu Q, Wang S, Li Z, Li L, Zhang J, Wen C. Prediction of high-speed train delay propagation based on causal text information. Railw Eng Sci 2022, 31 1 89-106

[48.] D’Acierno L, Botte M, Placido A, Caropreso C, Montella B. Methodology for determining dwell times consistent with passenger flows in the case of metro services. Urban Rail Transit 2017, 3 2 73-89

[49.] Cornet S, Buisson C, Ramond F, Bouvarel P, Rodriguez J. Methods for quantitative assessment of passenger flow influence on train dwell time in dense traffic areas. Transp Res Part C Emerg Technol 2019, 106 345-359

[50.] San HP, Masirin MIM (2016) Train dwell time models for rail passenger service. In: The 3rd International Conference on Civil and Environmental Engineering for Sustainability, Melaka

[51.] Cieslak DA, Chawla NV (2008) Learning decision trees for unbalanced data. In: Machine Learning and Knowledge Discovery in Databases: European Conference, Antwerp, pp 241–256

[52.] Simonyan K, Zisserman A (2014) Very deep convolutional networks for large-scale image recognition. arXiv preprint arXiv:14091556. https://doi.org/10.48550/arXiv.1409.1556

[53.] Ke G, Meng Q, Finley T, Wang T, Chen W, Ma W, Ye Q, Liu T (2017) Lightgbm: a highly efficient gradient boosting decision tree. In: Proceedings of the 31st International Conference on Neural Information Processing Systems, Long Beach, pp 3149–3157

[54.] Mason L, Baxter J, Bartlett P, Frean M (1999) Boosting algorithms as gradient descent. In: Proceedings of the 12th International Conference on Neural Information Processing Systems, Denver, pp 512–518

[55.] Breiman L. Classification and regression trees 2017 New York Routledge

[56.] Breiman L. Bagging predictors. Mach learn 1996, 24 123-140

[57.] Lundberg SM, Lee SI (2017) A unified approach to interpreting model predictions. In: Proceedings of the 31st International Conference on Neural Information Processing Systems, Long Beach, pp 4768–4777

Funding

National Natural Science Foundation of China(U1834209)