必发9988集团
必发9988集团首页English
活动讲座
当前位置: 首页 >> 活动讲座 >> 活动讲座 >> 正文
活动讲座

Optimizing Early Discharge: Trade-offs between Capacity and Readmissions

2024-04-11  

报告题目: Optimizing Early Discharge: Trade-offs between Capacity and Readmissions

报告人:朱桃增 东北财经大学现代供应链管理研究院 副教授

报告时间: 2024年426日(星期五) 14:00-15:30

报告地点:必发9988集团B309

邀请人:程春 教授

报告内容和摘要:

  In this work, we consider the ward capacity management problem with readmissions, where the decision-maker optimizes the elective schedule and early discharge policy, so as to minimize bed shortages. Existing research has shown that early discharge can lead to higher rates of readmission, and longer readmission length-of-stay. This sets up the need to balance the temporal trade-off between the immediate capacity freed up by early discharges and increased readmissions down the road. Such re-entry structure creates challenges when modelling via traditional methods. We appeal to the Pipeline Queues framework, and propose an optimization model where the early discharge policy is expressed as a state-dependent decision rule. The model has a reformulation, which can be solved as a sequence of convex programs with asymptotically linear constraints. In our numerical study, we identify an intermediate region of the probability of readmissions where time-invariant policies can lead to as much as 77% more shortages. Ignoring the effects of early discharge on readmissions can lead to at least 75% and 150% more bed shortages in time-homogeneous and non-time-homogeneous settings respectively, even against un-optimized elective admissions. Using optimal early discharge strategies without jointly optimizing elective admissions will lead to 20% more shortages.

报告人简介:

  朱桃增,中国科学技术大学管理学博士,新加坡国立大学联合培养博士。在加入东北财经大学前,曾于2019年至2021年在新加坡国立大学从事博士后研究工作。主要研究领域包括鲁棒优化、排队论、医疗管理等。其研究成果以第一作者身份发表在Management ScienceManufacturing & Service Operations ManagementFlexible Service and Manufacturing Journal等期刊。



上一条:A Penalized Sequential Convex Programming Approach for Continuous Network Design Problems

下一条:Data-Driven Optimization with Robust Stochastic Dominance Constraints

关闭