Large-scale Taxi Simulation Platform
We have developed a novel multi-functional open-source ride-hailing system simulation platform that can simulate various agent behaviors and movements on real traffic networks.
The platform provides users with several accessible entry points for training and testing various optimization algorithms, particularly reinforcement learning algorithms, suitable for multiple tasks including on-demand matching, idle vehicle repositioning, and dynamic pricing.
Open Source Platform
Our simulation platform is completely open source, allowing researchers and industry practitioners to freely access, use, and improve the code, promoting transparency and collaboration.
Real Road Network
The platform is built on real traffic network data, enabling accurate simulation of vehicle movements and passenger behavior in urban environments, ensuring practical and reliable results.
Our simulation platform offers a series of powerful features, making it an ideal tool for researching and optimizing ride-hailing systems.
Multi-agent Simulation
The platform can simultaneously simulate the behavior of thousands of vehicles and passengers, including driving decisions, route selection, waiting behavior, and matching preferences.
Algorithm Training and Testing
Provides standard interfaces for developing and testing various optimization algorithms, particularly suitable for the application and evaluation of reinforcement learning and other AI methods.
Visualization Tools
Built-in powerful data visualization tools help users intuitively understand simulation results and system dynamics, supporting decision analysis.
The platform is suitable for various ride-hailing system optimization tasks, providing comprehensive solutions and evaluation frameworks.
On-demand Matching
Optimize matching strategies between vehicles and passengers, reducing waiting time and empty mileage, improving overall system efficiency.
Idle Vehicle Repositioning
Develop intelligent strategies to guide idle vehicles to high-demand areas, balancing system supply and demand, improving service levels and vehicle utilization.
Dynamic Pricing
Test and optimize algorithms for adjusting prices based on real-time supply and demand, improving system revenue and service availability.
Our research results have been published in relevant academic journals and conferences, demonstrating the platform's effectiveness and practical value.
Related Papers
Feng S., Chen T., Zhang Y., Ke J.* & H. Yang, 2023. A Multi-functional Simulation Platform for On-demand Ride Service Operations. ArXiv:2303.12336 Preprint. https://doi.org/10.48550/arXiv.2303.12336
Open Source Code
The platform code is open-sourced on GitHub, welcoming researchers and practitioners to download and use:
https://github.com/HKU-Smart-Mobility-Lab/Transpotation_Simulator
Industry Applications
The platform has been adopted by multiple research institutions and enterprises for testing and optimizing real-world ride-hailing and taxi dispatch systems, significantly improving operational efficiency.
Total Number of Simulated Scenarios
Supported Matching Functions
Average Order Waiting Time at Equilibrium (s)
Release Year
Hu Jianping
Chief Executive Officer of Haylion TechnologiesFounder of eMaaS, Founder and Chairman of Haylion Technologies, Chief Expert of the Urban Passenger Transport Expert Committee of China Road Transport Association, Ph.D. in Road and Traffic Engineering from Tongji University. Previously served as General Manager and Chairman of Shenzhen Bus Group Co., Ltd., Executive Deputy General Manager of Shenzhen Metro Group Co., Ltd., Assistant to Director and Head of Personnel Division of Shenzhen Government Transport Bureau, Vice Chairman of China Road Transport Association, and Senior Vice President of US Microvast Power Systems.
Qiu Jiandong
Senior Engineer of SUTPCDirector of Transportation Information and Model Institute, Shenzhen High-level Talent. 17 years of experience in traffic modeling and big data. Led over 30 urban modeling and big data platform projects and 10 national and provincial research projects. Authorized 30 patents, co-authored 5 monographs and industry standards, published 35 papers, and received 24 national and provincial awards.
Ke Jintao
Assistant Professor of HKUServes as Youth Editor for the international top journal TRC and Editor for TRE in the transportation field. Led multiple research projects, including National Natural Science Foundation, Hong Kong Transport Department Smart Transport Fund, Hong Kong Research Grants Council Fund, and Hong Kong Environmental Protection Department Green Fund, with total funding exceeding HKD 10 million. Published over 50 SCI/SSC papers with more than 4,900 citations, ranked as a World Top 2% Scholar in Smart Transportation by Stanford in 2023.