报告所属学科：管理科学与工程
报告人：Hossein Abouee Mehrizi（加拿大滑铁卢大学）
报告时间：2020年12月9日 09:00
报告地点：ZOOM：763 086 2328
报告摘要：
Using large scale data of MRI services, we measure service duration and show that a number of covariates, including the shift during which the procedure is performed, case mix of the proceeding patients, congestion, and batching of similar procedures affect service duration. We observe that the magnitude of the effect of various mechanisms on service duration may depend on the server type (teaching or community hospital), and customer type (high priority versus low priority patient). For instance, in general, service duration during the night shift is significantly shorter than the day shift, and the decrease is higher for low-priority patients or in community hospital settings. We also show that the inverted-U-shaped relationship observed between the service duration and workload in the literature depends on the priority level. Surprisingly, while this mechanism holds for high priority patients, it is completely vice-versa for low priority patients. We then provide a discrete-time data-driven model of the problem in a finite horizon setting. We analyze the proposed model and provide an analytical gap between the wait cost of different priority levels under the optimal offline policy and a class of online scheduling polices. We validate the results using real world data.
报告人简介：
Hossein Abouee Mehrizi，是加拿大滑铁卢大学管理科学系副教授，现任NRL和HCMS期刊的Associate Editor。他研究方向为服务系统（包括卫生服务系统）的实证分析研究、复杂随机系统的数据驱动建模、随机过程、统计学习。他先后在OR、POM、NRL、EJOR、MSOM、DS等国际著名期刊上发表了20多篇论文。他受邀在INFORMS年会等30余场会议上作专题报告，并担任OR、MS、MSOM、POM、NRL、EJOR等15个著名期刊的审稿人。
