DRO PhD Seminar: Chun Ye (IEOR), Yunru Han (DRO), Kim Song-Hee (IEOR) & Lijian Lu (DRO)
Friday, October 4, 2013
12:30pm – 2:30pm EST
Warren 3rd Floor - 311
1125 Amsterdam Avenue, New York, NY 10025
Event Details
Speaker: Chun Ye Title: Cake Cutting Algorithms for Piecewise Constant Valuations Abstract: Cake cutting is a metaphor for the allocation of a heterogeneous divisible good among multiple agents with different preference. Chen et al. gave a deterministic cake cutting algorithm that is Pareto efficient, robust envy-free, robust proportional, and strategyproof when all agents have piecewise uniform valuations. In this paper, we give two extensions of this algorithm to the case when agents have piecewise constant valuations. One of our algorithms is Pareto efficient, envy-free and proportional, while the other is robust envy-free, robust proportional and satisfies a weaker notion of efficiency. We also give a randomized algorithm that is proportional, weakly efficient, and strategyproof in expectation. Finally, we derive two impossibility results showing that some of the aforementioned properties are incompatible with each other when the agents have piecewise constant valuations. Joint work with Haris Aziz and Jay Sethuraman Speaker: Yunru Han Title: On the Value of Demand Learning in a Supply Chain with Demand Censorship Abstract: Stockouts in retail industry could significantly drive down demand forecast due to lack of observations of excess demand, leading to self-reinforcing understocking. The existing literature that study the impact of demand censorship and tactically propose demand learning for the retailer often overlooks that retail stockouts also have serious implications for suppliers. We consider a supply chain with one supplier and one retailer subject to demand censorship, to compare the supply chain performances as a result of the supplier's wholesale price decisions and the retailer's inventory policies in three situations: fully observable demand ("O"), myopic retailer under censorship ("M"), and forward-looking retailer under censorship ("F"). Our study points out a few facts that contradict with the previous results without considering the supplier's strategic role. For example, a myopic retailer under demand censorship may entail a higher profit for himself and a better supply chain efficiency compared to the forward-looking retailer. Also, the supply chain under demand censorship could have a better efficiency than the system with fully observed demand. These phenomena present significant impact that the existing theory cannot account for, and identifies the possibility to increase supply chain efficiency through the strategic effects of demand censorship and demand learning. In a broader sense, the study also enriches the literature of collaborative learning in operations management, and lays the groundwork to understand the strategic issues emerging from large-scale experimentation in complex operational environments. Joint work with Fangruo Chen. Speaker: Kim (Hailey) Song-Hee Title: Should Hospitals Keep Their Patients Longer? The Impact of Medicare Penalties on Length of Stay and Readmissions Abstract: The Affordable Care Act financially penalizes hospitals with higher than expected 30-day readmissions for Medicare patients (one-third of the nation's hospitals were charged a total of $280 million in fiscal year 2013) to incentivize hospitals to reduce 'preventable' readmissions. Using retrospective data on patients enrolled in different Medicare programs (e.g., Fee-for-service and Medicare Advantage), we develop an understanding of the potential impact of the new legislation and how it relates to the interplay between financial and medical incentives in the provision of hospital care. Joint work with Carri Chan and Ann Bartel. Speaker: Lijian Lu Title: Management of a Hospital Step-Down Unit Abstract: We consider the optimization of patient routing from an intensive care unit (ICU) when there are step-down units (SDU). We provide the characterization of the optimal routing policy and construct some easily implemented and efficient heuristics. We also propose a theoretic framework to optimize the capacities of variant inpatients units. The empirical data is used to calibrate the theoretical results. Joint work with Carri Chan and Linda Green.
Where & When

Warren 3rd Floor - 311
1125 Amsterdam Avenue
New York, NY 10025

Friday, October 4, 2013, 12:30pm – 2:30pm

Download: Outlook · Google Calendar

Number Registered0 registered

Copy Link

Hosted By
Decision, Risk, and Operations

Contact the organizer