Yuhan Tang

Yuhan Tang is a graduate student in transportation tackling operational efficiency in transit systems through integer optimization, deep reinforcement learning, and urban mobility modeling. He developed a joint bus scheduling and vehicle assignment framework using mixed-integer quadratic programming that achieves a 10- to 50-fold computational speedup while maintaining solutions within 2% of global optimality, meaningful improvements for cities managing thousands of buses daily. For real-time adaptive control, Yuhan builds physics-informed environments incorporating macroscopic fundamental diagram surrogates, enabling efficient training of decentralized reinforcement learning agents that reduce passenger wait times by 13% on real-world data. As a MathWorks Fellow, Yuhan will further develop a unified computational framework that links optimization-based scheduling, data-driven operational models, and learning-based control. He will draw on optimization, machine learning, and data-analysis methods to prototype, test, and evaluate decision-support tools for transit and ride-hailing systems under realistic, time-varying operating conditions. Yuhan’s algorithms could help transit and ride-hailing systems operate more reliably and efficiently while accounting for real-world constraints.