Research
My research develops artificial intelligence methodologies for next-generation mobility systems, spanning learning-based decision making, foundation models for urban computing, human mobility behaviour, and AI-enabled simulation.
Research Themes
-
Decision Making & Optimization
Sequential decision making for large-scale mobility platforms: driver subsidy control under budget constraints, fleet repositioning for autonomous mobility-on-demand, order matching, and dynamic pricing.
-
Foundation Models & LLMs
Using large language and vision-language models as reasoning, augmentation, and agentic components in urban prediction pipelines — from class-incremental travel mode choice to iterative feature discovery and few-shot tabular generation.
-
Mobility & Behaviour
Empirical and learning-based analysis of how travellers and drivers actually behave, so that platform algorithms are designed against realistic human responses rather than idealised assumptions.
-
Simulation & Digital Twins
Open, scalable simulation infrastructure for on-demand and multimodal transport, providing reproducible testbeds for matching, repositioning, pricing, and reinforcement learning research.
Affiliations
-
HKU Smart Mobility Lab 2023 – presentDepartment of Civil Engineering, The University of Hong KongPrincipal Investigator: Prof. Jintao Ke
-
Urban Transport Systems Laboratory (LUTS) Mar – Sep 2026École Polytechnique Fédérale de Lausanne (EPFL)
Funded Projects 4 projects
-
Core MemberSmartSim: AI-assisted Simulation Software for Multimodal Transportation Operations
Smart Traffic Fund, Hong Kong SAR Government · PSRI/78/2311/RA
-
Core MemberA Study on Spatiotemporal Supply-Demand Regulation in the Ride-Hailing Scenario
CCF-DiDi GAIA Collaborative Research Fund · 202410
-
Project ManagerDevelopment of a Simulation Platform and Artificial Intelligence Algorithms for Optimising Operation and Management of Taxi E-hailing Services
Smart Traffic Fund, Hong Kong SAR Government · PSRI/29/2201/PR
-
Principal InvestigatorDigital Inheritance of Intangible Cultural Heritage Based on VR Technology — a Case Study of Tianjin's “Clay Figure Zhang” Culture
Tianjin Innovation and Entrepreneurship Training Project · 201910055373
Open-Source Software
-
Transportation Simulator
A multi-functional simulation platform for on-demand ride service operations, supporting order matching, idle-vehicle repositioning, dynamic pricing, and reinforcement learning experiments.
-
DL-MRD-Broadcasting
Deep multi-task learning models for dynamic matching radius decisions in on-demand ride services under the broadcasting mode.