Stochastic dynamics in curbside operations: Pricing, facility design, and access control

As demand for last-mile logistics and passenger transportation grows, the curbside has become a critical bottleneck linking transportation networks and service destinations. Efficient curbside management requires understanding stochastic arrival and service processes through underlying state dynamics. This talk presents rigorous, practical frameworks using continuous-time Markov decision processes (MDPs) and game-theoretic queues. First, I address cruising-for-parking congestion using game-theoretic queues with state-dependent retrials, employing computational approximations and Lyapunov bounds to evaluate welfare-maximizing pricing. Second, I analyse high-throughput pickup and drop-off facilities whose capacity is limited by conflicting vehicle movements. I introduce a microscopic MDP formulation to evaluate the impacts by layout design and access control. I conclude by discussing the transferability of these methods to broader urban and digital service systems.

Xinyu (Cindy) Liu holds a PhD in Operations Research from Georgia Tech, advised by Dr Anton J. Kleywegt, and a Bachelor of Engineering Systems Design, summa cum laude, from SUTD. She is currently a Schmidt AI in Science Fellow at the University of Michigan, mentored by Dr Yafeng Yin and Dr Vijay G. Subramanian. She studies stochastic dynamic resource allocation in urban and digital congestible systems, and applies interpretable machine learning for demand modelling. Xinyu was named a Rising Star in IE, OR, and management science by Georgia Tech ISyE, Stanford MS&E, and Michigan IOE. She received the Airport Cooperative Research Program Graduate Research Award from the Federal Aviation Administration. Her postdoctoral research is fully funded by Schmidt Sciences.

For more information about the ESD Seminar, please email esd_invite@sutd.edu.sg