报告题目 | Designing a Forward-looking Matching Policy for Dynamic Ridepooling Service |
报告人（单位） | 王晓蕾 教授（同济大学） |
主持人（单位） | 李泳臻（bevictor伟德官网） |
会议时间 | 2022年10月14日 10:00 |
会议地点 | 腾讯会议ID：868-397-104 会议密码：2022 点击链接入会：https://meeting.tencent.com/dm/5yps8jyjzMeV 特别提醒：请参会者以真名进入，否则可能会被移出会议！ |
报告人简介 | |
王晓蕾，同济大学经济与管理学院教授。2008年本科毕业于中国科技大学，2012年博士毕业于香港科技大学，就读期间曾分别获得郭沫若奖学金和SENG PhD Research Excellence Award。 长期从事城市交通系统管理方向的研究，主要在城市交通需求管理政策设计和共享出行服务运营优化两个方面取得了一系列成果，发表SCI/SSCI论文20余篇，其中11篇发表于交通领域顶级期刊《Transportation Science》和《Transportation Research Part B》。由她与合作者杨海教授提出的可交易通行权机制克服了传统城市交通需求管理机制公平与效率难以兼顾的缺陷，在中国工程院发布的《全球工程焦点2017》中位列工程管理领域十大研究热点之一。主持和参与国家自然科学基金7项，2020年获自科优秀青年基金资助，创新群体“综合运输系统运营管理”项目骨干成员。 | |
报告内容提要 | |
The popularity of smartphones and the advent of GPS positioning and wireless communication technologies in recent years have facilitated large-scale implementations of dynamic ridepooling services, such as Uber Pool, Lyft Line, and Didi Pinche. In dynamic ridepooling services, service providers respond to on-demand mobility requests immediately, dispatch (vacant or partially occupied) vehicles in real-time, and keep searching for matching orders along the trip. Most existing dispatching strategies for dynamic ridepooling give matching priorities to partially occupied vehicles and form a ridepooling trip immediately when they find a match. Such matching strategies ignore forthcoming matching opportunities that may have higher matching qualities, therefore have short-sighted limitations. In this paper, to be forward-looking in vehicle dispatching, we propose a probabilistic matching policy under which every match is accepted with a certain probability. Assuming that each ridepooling passenger shares vehicle space with at most one another during the entire trip, and ridepooling orders between each OD pair appear following a Poisson process with a given rate, we develop a mathematical model to predict the mean ride/shared distance of ridepooling passengers between each OD pair given the acceptance probability of each match, and then propose a heuristic algorithm to optimize the matching policy (i.e., the acceptance probability of each match) for minimal total ride distance based on the model. With simulation experiments, we show that our model can generate fairly good predictions of the ride/shared distance of each OD pair under diverse matching policies, and the optimal matching policy generated by our method can lead to over 5% reduction of the total ride distance when demand is high. | |
