Student Seminar by Mr. Kaihang Zhang on Apr 20, 2026, 4:00PM

Title: Let Drivers Roam: How ETA Slack Unlocks Network Effects in On-Demand Food Delivery Markets

Speaker: Mr. Kaihang Zhang (Department of Civil Engineering)

Date: Apr 20, 2026 (Monday)

Time: 4:00 pm - 5:00 pm

Venue: Room 8-28, Haking Wong Building, The University of Hong Kong

Registration Link: https://forms.cloud.microsoft/r/bMhERMSx61

ITS Student Committee will provide a beverage for registered participants.

Abstract: Platforms in on-demand delivery compete on speed by committing to tight estimated time of arrival (ETA) targets. Tight ETAs reduce drivers’ flexibility while busy, limiting order bundling and cross-neighborhood switching; this fragments the implicit service network and restricts pooling benefits. We develop a parsimonious steady-state model that links an ETA commitment to endogenous bundling, routing and repositioning workload for multi-stop tours, and a switching probability that captures whether drivers can relocate to adjacent restaurant clusters between bundles. These endogenous variables feed into a mean-field (aggregate level) queuing approximation for pooled matching, yielding tractable expressions for assignment-delay metrics. Within an operating range we call the active regime, relaxing ETA commitments increases bundling and the probability that a busy driver can accept a new assignment, reduces per-order service time, and lowers assignment delay. We show that tight ETAs fragment the cluster network, loosing pooling benefits. We then introduce a fleet-based price of urgency that quantifies the additional fleet required to meet a target assignment delay when operating at tighter ETAs relative to a benchmark, highlighting a nonlinear fleet cost near the connectivity threshold where switching becomes insufficient to maintain network connectivity.

Bios: Kaihang is a PhD student at the Department of Civil Engineering, HKU, supported by HKPF and HKUPS. His research interests include modeling of complex systems, optimization under equilibrium constraints, and behavioral analysis and operations. Prior to HKU, he worked on traffic state analysis and public transportation system design. He holds bachelor’s degrees from ZJU and UIUC, and a master’s degree from UC Berkeley.