题目：Community Responder Crowdsourcing for Time-Sensitive Medical Emergencies
时间：2026年4月17日 13:30-15:00
地点：350vip8888新葡的京集团 F103会议室
邀请人：刘冉 教授（工业工程与管理系）
Biography
Weiliang Liu is an Assistant Professor at Department of Systems Engineering and Engineering Management, The Chinese University of Hong Kong (CUHK). His research focuses on service operations and applied probability, with interests in modern operations challenges, societally impactful applications, and data-driven queueing system design. His works have been published in top outlets, and have received notable recognitions, including the INFORMS SOLA Best Student Paper Award, and a finalist of the POMS-HK Best Student Paper Award. Weiliang received his B.E. from Shanghai Jiao Tong University in 2020 and his Ph.D. from the National University of Singapore in 2024. He spent a year as a principal researcher at the University of Chicago Booth School of Business before joining CUHK.
Abstract
In parallel with traditional ambulance dispatch, community first responder (CFR) systems use mobile applications to locate and alert nearby volunteer responders for rapid intervention in time-sensitive medical emergencies. A central operational challenge is determining whom to alert and when so as to minimize the time to first intervention, subject to alert volume constraints. Existing CFR systems rely on ad-hoc rules that fail to utilize alert budgets effectively. This talk introduces a dynamic programming (DP) framework for the alert decision problem, a model predictive control (MPC) algorithm that can generate high-quality solutions very quickly, and a score-based heuristic that closely matches MPC performance without the need for an optimization solver. Unlike existing rules that expand the alert radius over time, our proposed policies strategically narrow the responder pool for alert consideration as time progresses. Our numerical studies reveal further limitations of existing rules: 1) Many CFR systems issue all alerts at the start, ignoring different travel times, response probabilities, and real-time behavior of responders. Our models show that the optimal number of initial alerts varies according to responder characteristics; 2) The common practice of alerting the nearest responders first is not always optimal. An optimal strategy might alert a more distant responder if the higher probability of alert acceptance outweighs the longer travel time. Our simulations using real-world CFR system data validate the superiority of our proposed policies in practical situations.
学院OA系统
jAccount登录
ENGLISH
地址：上海市东川路800号350vip8888新葡的京集团闵行校区350vip8888新葡的京集团
邮编：200240
Email：sjtume@sjtu.edu.cn