报告题目：Analytics for On-Demand Food Delivery: A Hurwicz Satisficing Perspective
报告人：孙清荷 助理教授
邀请人：边博旻 教授
报告时间及地点：2026.06.23 14:00-15:30 经管学院B220
报告人简况：
孙清荷博士现任香港理工大学工商管理学院物流及航运学系助理教授。她于2022年获得新加坡国立大学运筹学与商业分析学博士学位，于2017年获得南洋理工大学海事研究一等荣誉学位。她的研究聚焦于交通运输与航运管理，重点关注数据驱动优化与不确定环境下的决策分析。其研究成果发表于 Operations Research、Production and Operations Management、Transportation Research Part B/C/E 等期刊，并获得香港研究资助局（RGC）及国家自然科学基金委员会（NSFC）的资助。面向航运与交通领域的重要实践挑战，她长期与产业界保持合作，致力于开发兼具学术与实践价值的决策方法。其合作伙伴包括法国达飞海运集团（CMA CGM）和海洋网联船务（ONE）等航运企业。
报告内容摘要：
We study an on-demand food delivery (OFD) platform that dynamically coordinates customers, restaurants, and couriers under uncertainty to achieve multiple operational objectives, such as minimizing delivery delays and expirations, reducing costs, and enhancing value creation. The platform assigns orders and recommends courier relocations to balance supply and demand across spatial and temporal dimensions. This coordination is complicated by multiple sources of uncertainty—such as customer and courier arrivals, delivery times, and heterogeneous courier behaviors—whose underlying distributions are ambiguous and difficult to estimate reliably. To address these challenges, we propose a Hurwicz satisficing optimization (HSO) framework for decision-making under uncertainty that integrates multiple objectives with forward-looking considerations. The framework features the Hurwicz certainty equivalent (HCE) to balance optimism and pessimism in decision-makers’ attitudes toward ambiguity, relaxing the fully pessimistic assumption inherent in robust optimization. To enhance robustness, we incorporate a satisficing mechanism that determines the minimal optimism level required to meet performance targets, while maintaining computational tractability through the analytical properties of the HCE. The HSO model is implemented for OFD operations using a rolling-horizon mixed-integer program, evaluated through both simulation and a real-world case study based on Meituan data. Compared with benchmark models, the HSO model consistently achieves balanced and robust performance across multiple operational metrics under different scenarios, highlighting the value of combining robust satisficing and the Hurwicz criterion in addressing uncertainty.
上一条：Beyond Linear Decision Rules: LLM-Guided Basis Function Discovery for Data-Driven Optimization
下一条：基于零售细分的线下渠道营销优化分配方法
【关闭】
