Embedded Files
My research focuses on building sustainable and reliable transport and logistics systems with use of emerging analytics, sensing, and computing technologies.
My research focuses on building sustainable and reliable transport and logistics systems with use of emerging analytics, sensing, and computing technologies.
Specific areas include
Specific areas include
Sequential decision analytics
Multi-modal transport systems
Logistics operations
Reinforcement learning
Multi-agent systems
My list of work, publications, and co-workers could be found from my:
My talk on metro travel behaviour analytics with smart card data
(HKUST - "Did You Know?" series by alumni professors)
My current and recent research projects include:
Urban logistics with crowdsourcing and decentralisation
Urban logistics with crowdsourcing and decentralisation
We develop logistics operations with use of information and communication technology (ICT) and decentralised optimisation. It has been found that efficiency and cost effectiveness of the logistics system could be enhanced from traditional methods via use of crowdsourcing and collaborative production control. The optimisation algorithms have been validated and tested on different settings generated from real world scenarios in Hong Kong and mainland China.
We develop logistics operations with use of information and communication technology (ICT) and decentralised optimisation. It has been found that efficiency and cost effectiveness of the logistics system could be enhanced from traditional methods via use of crowdsourcing and collaborative production control. The optimisation algorithms have been validated and tested on different settings generated from real world scenarios in Hong Kong and mainland China.
Selected publications (*-corresponding author):
Deng, Y., Sheng, G., Chow, A.H.F., Zhou, Z., Bai, Q., Su, Z.* (2025) Collaborative production control and distributor selection via multi-agent reinforcement learning with differentiable communication. Expert Systems with Application, 282, 127539.
Deng, Y., Yan, Y., Chow, A.H.F.*, Zhou, Z., Kuo, Y.H. (2024) A proximal policy optimization approach for food delivery problem with reassignment due to order cancellation. Expert Systems with Application, 258, 125045.
Yan, Y., Deng, Y., Cai., S., Kuo, Y.H.*, Chow, A.H.F., Ying, C., (2023) A policy gradient approach to solving dynamic assignment problem for on-site service delivery. Transportation Research Part E, 178, 103260.
Project sponsors:
Hong Kong Research Grants Council (General Research Fund); National Natural Science Foundation of China (NSFC)
Public transport operations under disruptions
Public transport operations under disruptions
We develop reinforcement learning (RL) based methods for managing urban public transport systems under disruptions in real time. It has been found that latent robustness and resilience of transit operations could be unlocked with the computational effectiveness enhanced with appropriate use of approximation and parameterization techniques in the underlying dynamic optimisation framework. The optimisation algorithms have been validated and tested on different settings.
We develop reinforcement learning (RL) based methods for managing urban public transport systems under disruptions in real time. It has been found that latent robustness and resilience of transit operations could be unlocked with the computational effectiveness enhanced with appropriate use of approximation and parameterization techniques in the underlying dynamic optimisation framework. The optimisation algorithms have been validated and tested on different settings.
Selected publications (*-corresponding author):
Li, G., Chow, A.H.F.*, Ying, C. (2025) Robust optimization for adaptive bus service scheduling with adversarial reinforcement learning under demand uncertainties. Transportation Research Part C, 178, 105222. doi.org/10.1016/j.trc.2025.105222.
Ying, C., Chow, A.H.F.*, Yan, Y., Kuo, Y.H., Wang, S. (2024) Adaptive rescheduling of rail transit services with short-turnings under disruptions via a multi-agent deep reinforcement learning approach. Transportation Research Part B, 188, 103067. doi.org/10.1016/j.trb.2024.103067.
Project sponsors:
Hong Kong Research Grants Council (General Research Fund); National Natural Science Foundation of China (NSFC)
Reliability-based intelligent transportation systems
Reliability-based intelligent transportation systems
This is a joint research project with The University of Hong Kong and The Hong Kong Polytechnic University, where we aim to develop an intelligent transportation system that can generate reliable estimates of travel times and travel guidance for road users. The project explores innovative development and applications of advanced analytics and systems modelling techniques including heterogeneous data fusion, stochastic road traffic modelling, macroscopic fundamental diagrams, reliability-based routing algorithms, and deep learning framework. The proposed methodologies will be tested with real world scenarios in Hong Kong with support from government organisation, private and public sectors.
This is a joint research project with The University of Hong Kong and The Hong Kong Polytechnic University, where we aim to develop an intelligent transportation system that can generate reliable estimates of travel times and travel guidance for road users. The project explores innovative development and applications of advanced analytics and systems modelling techniques including heterogeneous data fusion, stochastic road traffic modelling, macroscopic fundamental diagrams, reliability-based routing algorithms, and deep learning framework. The proposed methodologies will be tested with real world scenarios in Hong Kong with support from government organisation, private and public sectors.
Selected publications (*-corresponding author):
Wu, X., Chow, A.H.F.*, Ma, W., Lam, W.H.K., Wong, S.C., (2025) Prediction of traffic state variability with an integrated modal-based and data-driven Bayesian framework. Transportation Research Part C, 171, 104953. doi.org/10.1016/j.trc.2024.104953.
Zhuang, L., Wu, X., Chow, A.H.F.*, Ma, W., Lam, W.H.K., Wong, S.C., (2025) Reliability-based journey time prediction with two-stream deep learning data fusion. Journal of Intelligent Transportation Systems, 29(2), 134-152. doi.org/10.1080/15472450.2023.2301707.
Wu, X., Chow, A.H.F.*, Zhuang, L., Ma, W., Lam, W.H.K., Wong, S.C., (2024) Estimation of vehicular journey time variability by Bayesian data fusion with general mixture model. IEEE Transactions on Intelligent Transportation Systems, 25(10), 13640 - 13652. 10.1109/TITS.2024.3401709.
Li, A., Lam, W.H.K., Ma, W.*, Chow, A.H.F., Wong, S.C., Tam, M.L. (2024) Filtering limited automatic vehicle identification data for real-time path travel time estimation without ground truth. IEEE Transactions on Intelligent Transportation Systems, 25(6), 4849-4861. 10.1109/TITS.2023.3336238.
Li, A., Lam, W.H.K., Ma, W.*, Chow, A.H.F., Wong, S.C., Tam, M.L. (2024) Real time estimation of multi-class path travel times using multi-source traffic data. Expert Systems with Applications, 237C, 121613. doi.org/10.1016/j.eswa.2023.121613.
Hu, Z., Lam, W.H.K., Wong, S.C., Chow, A.H.F., Ma, W.* (2023) Turning traffic surveillance cameras into intelligent sensors for traffic density monitoring. Complex & Intelligent Systems, 9. 7191-7195. doi.org/10.1007/s40747-023-01117-0.
Bai, L., Wong, S.C.,* Xu, P., Chow, A.H.F., Lam, W.H.K. (2021) Calibration of stochastic fundamental diagram with explicit consideration of speed heterogeneity. Transportation Research Part B, 150, 524-539. doi.org/10.1016/j.trb.2021.06.021.
Zhong , R, Luo, J, Cai, H, Sumalee, A., Yuan, F, Chow A.H.F.* (2017) Forecasting journey time distribution with consideration of abnormal traffic conditions. Transportation Research Part C, 85, 292-311, doi:10.1016/j.trc.2017.08.021.
Project sponsor:
Hong Kong Research Grants Council (Research Impact Fund)
Intelligent traffic control with reinforcement learning
Intelligent traffic control with reinforcement learning
We develop system-wide control systems with use of reinforcement learning (RL) for managing urban network traffic in real time. It is generally found that control performances and computational time can be significantly improved with appropriate use of approximation and parameterization techniques. The control algorithms have been validated on virtual SUMO microsimulation testbeds.
We develop system-wide control systems with use of reinforcement learning (RL) for managing urban network traffic in real time. It is generally found that control performances and computational time can be significantly improved with appropriate use of approximation and parameterization techniques. The control algorithms have been validated on virtual SUMO microsimulation testbeds.
Selected publications (*-corresponding author):
Su, Z, Chow, A.H.F.*, Fang, C.L., Liang, E.M., Zhong, R.X. (2023) Hierarchical control for stochastic network traffic with reinforcement learning. Transportation Research Part B, 167, 196-216. doi.org/10.1016/j.trb.2022.12.001.
Su, Z.C., Chow, A.H.F.*, Zhong, R.X. (2021) Adaptive network traffic control with an integrated model-based and data-driven approach and a decentralised solution method. Transportation Research Part C (Special issue of ISTTT 2021), 128, #103154. doi.org/10.1016/j.trc.2021.103154.
Su, Z, Chow, A.H.F., Zheng, N., Huang, Y, Liang, E., Zhong R* (2020) Neuro-dynamic programming for optimal control of macroscopic fundamental diagram systems. Transportation Research Part C, 116, #102628. doi.org/10.1016/j.trc.2020.102628.
Project sponsors:
Hong Kong Research Grants Council (General Research Fund); National Natural Science Foundation of China (NSFC)
Decentralised transit operations with multi-agent reinforcement learning
Decentralised transit operations with multi-agent reinforcement learning
We aim to develop a dynamic optimisation framework for transit networks with multiple control agents sharing information and decisions in real-time subject to prevailing service statuses and demand uncertainties. We will specifically investigate and develop a decentralised solution framework that can derive and accommodate asynchronous decisions made by connected local control agents. The proposed algorithms will be implemented and tested with real-world scenarios of Hong Kong transit networks. The vulnerability of the optimisation algorithms with respect to potential disruptions will also be investigated. This project contributes to dynamic rail transit network operations with advanced computing techniques.
We aim to develop a dynamic optimisation framework for transit networks with multiple control agents sharing information and decisions in real-time subject to prevailing service statuses and demand uncertainties. We will specifically investigate and develop a decentralised solution framework that can derive and accommodate asynchronous decisions made by connected local control agents. The proposed algorithms will be implemented and tested with real-world scenarios of Hong Kong transit networks. The vulnerability of the optimisation algorithms with respect to potential disruptions will also be investigated. This project contributes to dynamic rail transit network operations with advanced computing techniques.
Selected publications (*-corresponding author):
Wang, S., Chow, A.H.F.*, Ying, C., (2025) Adaptive and flexible rail transit network service dispatching as a partially observable Markov decision process. Transportation Research Part C, 179, 105286. doi.org/10.1016/j.trc.2025.105286.
Ying, C., Chow, A.H.F.*, Nguyen, H., Chin, K.S. (2022) Multi-agent deep reinforcement learning for adaptive coordinated metro service operations with flexible train composition. Transportation Research Part B, 161, 36-59. doi.org/10.1016/j.trb.2022.05.001.
Project sponsor:
Hong Kong Research Grants Council (General Research Fund)
Rail transit operations with approximate dynamic programming
Rail transit operations with approximate dynamic programming
We aim to develop adaptive control systems for rail transit operations with use of advanced dynamic programming techniques. The objective to reduce passengers' waiting time and operators' cost via adjusting service schedules, running speeds, and fleet sizes in real time. The proposed systems are validated and tested with scenarios configured by real world data collected from London Underground. In particular, this series of work has been presented as one of earliest effort applying reinforcement learning to adaptive mass urban transit system operations.
We aim to develop adaptive control systems for rail transit operations with use of advanced dynamic programming techniques. The objective to reduce passengers' waiting time and operators' cost via adjusting service schedules, running speeds, and fleet sizes in real time. The proposed systems are validated and tested with scenarios configured by real world data collected from London Underground. In particular, this series of work has been presented as one of earliest effort applying reinforcement learning to adaptive mass urban transit system operations.
Selected publications (*-corresponding author):
Nguyen, H., Chow, A.H.F.* (2023) Adaptive rail transit network operations with a rollout surrogate-approximate dynamic programming approach. Transportation Research Part C, 148, 104021. doi: 10.1016/j.trc.2023.104021.
Ying, C., Chow, A.H.F.*, Wang, Y., Chin, K. (2022) Adaptive metro service schedule and train composition with a proximal policy optimization approach based on deep reinforcement learning. IEEE Transactions on Intelligent Transportation Systems, 23(7), 6895-6906. doi: 10.1109/TITS.2021.3063399
Ying, C., Chow, A.*, Chin, K. (2020) An actor-critic deep reinforcement learning approach for metro train scheduling with rolling stock circulation under stochastic passenger demand. Transportation Research Part B, 140, 210-235. doi.org/10.1016/j.trb.2020.08.005.
Project sponsor:
Hong Kong Research Grants Council (Theme-based Research Scheme; General Research Fund)
Adaptive bus operations in urban networks
Adaptive bus operations in urban networks
We explore real time modelling and optimisation of bus service operations in urban road networks as an optimal control problem with use of heterogenuous traffic data. We started with using conventional gradient-based methods and have recently moved on use of reinforcement learning computational techniques. The study reveals various methods for improving bus service efficiency and reliability via adaptive traffic control, service scheduling and routing, and fleet assignment in congested traffic conditions.
We explore real time modelling and optimisation of bus service operations in urban road networks as an optimal control problem with use of heterogenuous traffic data. We started with using conventional gradient-based methods and have recently moved on use of reinforcement learning computational techniques. The study reveals various methods for improving bus service efficiency and reliability via adaptive traffic control, service scheduling and routing, and fleet assignment in congested traffic conditions.
Selected publications (*-corresponding author):
Chow, A.H.F.*, Li, G., Ying, C. (2024) Adaptive scheduling of mixed bus services with flexible fleet size assignment under demand uncertainty. Transportation Research Part C, 158, 104452. doi.org/10.1016/j.trc.2023.104452.
Yan, Y., Wen, H., Deng, Y., Chow, A.H.F., Wu, Q., Kuo, Y.H.* (2024) A mixed integer linear programming-based reinforcement learning for electric bus scheduling with multiple termini and service routes. Transportation Research Part C, 162, 104570.
Chow, A.H.F.*, Su, Z.C., Liang, E., Zhong, R.X. (2021) Adaptive signal control for bus service reliability with connected vehicle technology via reinforcement learning. Transportation Research Part C, 129, 103264. doi.org/10.1016/j.trc.2021.103264.
Project sponsor:
Hong Kong Research Grants Council (General Research Fund)
Decentralised optimisation for urban traffic networks
Decentralised optimisation for urban traffic networks
We look at development and implementation of decentralised optimisation techniques via Internet-of-Things (IoT) technologies for managing urban network traffic. We started with using conventional gradient-based methods and have moved on to multi-agent formulations with use of reinforcement learning. Apart from the algorithmic design, it is generally found that the performance of decentralised schemes depends on the levels and spatial distribution of demand flows.
We look at development and implementation of decentralised optimisation techniques via Internet-of-Things (IoT) technologies for managing urban network traffic. We started with using conventional gradient-based methods and have moved on to multi-agent formulations with use of reinforcement learning. Apart from the algorithmic design, it is generally found that the performance of decentralised schemes depends on the levels and spatial distribution of demand flows.
Selected publications (*-corresponding author):
Chow, A.H.F.*, Sha, R., Li, Y. (2020) Adaptive control strategies for urban network traffic via a decentralised approach with user-optimal routing. IEEE Transactions on Intelligent Transportation Systems, 21(4), 1697-1704, doi:10.1109/TITS.2019.2955425.
Chow, A.H.F.*, Sha, R., Li, S. (2020) Centralised and decentralised signal timing optimisation approaches for network traffic control. Transportation Research Part C (Special issue of ISTTT 2019), 113, 108-123. doi.org/10.1016/j.trc.2019.05.007.
Chow, A.H.F.* (2015) Optimisation of dynamic motorway traffic via a parsimonious and decentralised approach. Transportation Research Part C, 55, 69-84. doi:10.1016/j.trc.2015.01.009.0.
Project sponsors:
European Commission FP7 framework; Leverhulme Trust (UK)
Page updated
Google Sites
Report abuse