2019年1月2日 |
8:10 |
从国际学术交流中心（逸夫楼）大厅统一出发，由学生引导前往会场群贤二204室 |
8:25 |
会议开始 |
8:30-9:15 
|
Talk 1:Yichuan Ding (University of British Columbia) Title: An Achievable-Region-Based Approach for Kidney Allocation Policy Design with endogenous Patient Choice Abstract: The deceased-donor kidney transplant candidates in the US are ranked according to characteristics of both the donor and the recipient. We seek the ranking policy that optimizes the efficiency equity tradeoff among all such policies, taking into account patients’ strategic choices. Our approach considers a broad class of ranking policies, which provides approximations to the previously and currently used policies in practice. It also subsumes other policies proposed in the literature previously. As such it facilitates a unified way of characterizing good policies. We use a fluid model to approximate the transplant waitlist. Modeling patients as rational decision makers, we compute the resulting equilibria under a broad class of ranking policies, namely the achievable region. We then develop an algorithm that optimizes the system performance over the achievable region. Results. We show analytically that it suffices to restrict attention to priority scores that are affine in the patient’s waiting time. We also show through a numerical study that the total QALYs can be increased substantially by allowing patient rankings to depend on the kidney quality. Lastly, we observe that there is almost no improvement if only the healthier patients are prioritized for certain kidney types. Our results verify that ranking patients differently for kidneys of different quality can reduce the survival mismatch and the kidney wastage significantly. Consequently, the policy change in 2014, that implemented prioritizing the healthiest patients when allocating the highest 20% quality organs, is a step in the right direction. For further improvement, one may consider revising the current policy by also prioritizing the least healthy patients on the waitlist for the lowest-quality organs. Bio: Yichuan Ding is currently an assistant professor from the Sauder School of Business, University of British Columbia. He obtained his Ph.D. degree from the Department of Management Science and Engineering, Stanford University in 2012. He researches broadly in operations research methods and their applications in health care and other public sectors. He has published on Mathematics of Operations Research, Operations Research, and Manufacturing and Service Operations Management. He has participated in consulting projects with several agencies, including Department of Emergency Medicine in Metro Vancouver, BC Children’s Hospital, the Scientific Registration of Transplant Research, and BC Housing, etc. |
9:15-10:00 
|
Talk 2:Xuan Wang (Hong Kong University of Science and Technology) Title: Robust Optimization Approach to Process Flexibility Designs with Price Differentials Abstract: In the operations literature, the theoretical investigation of the effectiveness of limited flexibility has mainly focused on the performance metric that is based on the maximum sales in units. However, this could lead to substantial profit losses when the maximum sales metric is used to guide flexibility designs whereas the products have considerably large price differences. We address this issue by introducing price differentials into the analysis of process flexibility designs. We introduce the Profit Plant Cover Index (PPCI) and prove that a general class of robust measures can be expressed as functions of a design’s PPCIs and the given uncertainty set, and the PPCIs lead to a method to compare the worst-case performance of different designs. Applying these results, we prove that under a broad class of uncertainty sets and robust measures, the alternate long chain is optimal among all long chains with equal number of high profit products and low profit products. Finally, we develop a heuristic based on the PPCIs to generate effective flexibility designs when products exhibit price differentials. Joint work with Shixin Wang and Jiawei Zhang. Bio: Xuan Wang is an Assistant Professor in Operations Management at the School of Business and Management, The Hong Kong University of Science and Technology. Her main research interests lie in the broad area of decision-making under uncertainty with applications in supply chain optimization and revenue management. Xuan holds a Ph.D. in Operations Management from the Stern School of Business at New York University, and a Bachelor in Operations Research and Industrial Engineering from Tsinghua University in Beijing, China. |
10:00-10:20 |
Coffee Break |
10:20-11:05 
|
Talk 3:Yongbo Xiao(Tsinghua University) Title: Revenue Management of Probabilistic Selling of Multiple Substitutable Products Abstract: Probabilistic selling provides a new dimension to segment a market and benefits a seller from price discrimination. This paper extends the idea behind probabilistic selling to a more general setup, by developing a menu of probabilistic goods (PGs) based on multiple physical products. In particular, the seller has a fixed inventory level for each physical product at the beginning of a finite selling horizon. Faced with sequential customer arrivals, the seller dynamically controls the offering of multiple PGs in order to maximize expected revenue. Incorporating customer choice model, we formulate the seller's problem as a continuous-time, discrete-state, finite-horizon dynamic program (DP). Due to the complexity of solving this multi-dimensional DP, we resort to heuristics. We first study the deterministic version of the problem (i.e., the fluid control problem) and develop a time-based fluid policy. The policy is shown to be asymptotically optimal for the original stochastic problem. Further, the static nature of the fluid policy and its lack of flexibility in matching supply with demand motivate us to develop decomposition approximation which generates a dynamic control policy. Numerical experiments are conducted to evaluate the performance of the two heuristics and the benefits of selling a menu of PGs. Bio:Yongbo Xiao is an Associate Professor (with tenure) at School of Economics and Management, Tsinghua University, China. He received his Ph.D. and M.A. in Management Science and Engineering in 2006, and B.E. in Management Information Systems in 2000, all from Tsinghua University. He joined Tsinghua SEM as an assistant professor in Aug. 2008 after he completed his postdoctoral research in Department of Economics in Tsinghua SEM. He was awarded the “Science Fund for Excellent Young Scholar” under National Natural Science Foundation of China (NSFC) in 2012 and the “Young Scholar Award of Chinese Management Science” in 2014. He was elected as a Chang Jiang scholar in 2016. Dr. Xiao’s research interests include revenue and pricing management, service management, supply chain management, and healthcare management. |
11:05-11:50 
|
Talk 4:Ningyuan Chen (Hong Kong University of Science and Technology) Title: Nonparametric Learning and Optimization with Covariates Abstract: Modern decision analytics frequently involves the optimization of an objective over a finite horizon where the functional form of the objective is unknown. The decision analyst observes covariates and tries to learn and optimize the objective by experimenting with the decision variables. We present a nonparametric learning and optimization policy with covariates. The policy is based on adaptively splitting the covariate space into smaller bins (hyper-rectangles) and learning the optimal decision in each bin. We show that the algorithm achieves a regret of order $O(\log(T)^2 T^{(2+d)/(4+d)})$, where $T$ is the length of the horizon and $d$ is the dimension of the covariates, and show that no policy can achieve a regret less than $O(T^{(2+d)/(4+d)})$ and thus demonstrate the near optimality of the proposed policy. The role of $d$ in the regret is not seen in parametric learning problems: It highlights the complex interaction between the nonparametric formulation and the covariate dimension. It also suggests the decision analyst should incorporate contextual information selectively. Bio: Dr. Ningyuan Chen is currently an assistant professor at the Department of Industrial Engineering and Decision Analytics of the Hong Kong University of Science and Technology. He received his Ph.D. from the Department of Industrial Engineering and Operations Research at Columbia University. He was a postdoctoral associate at Yale School of Management from 2015 to 2016. His research interest includes revenue management, statistics, applied probability and networks. |
12:00-13:00 |
尝耻苍肠丑（厦大林梧桐楼） |
13:30-14:15 
|
Talk 5:Guanhua Yao (Director, Xiamen Municipal Health and Family Planning Commission) Topic:智慧健康、信息惠民 Abstract:厦门市创新“互联网+健康医疗”服务的新模式，通过区域电子健康档案系统实现检验检查结果共享，减少重复检查。居民可查阅个人电子健康档案并获得健康指导。通过移动互联网技术搭建了全市统一预约平台，市民通过电话、网站、手机端应用预约挂号、检查后，可直接到医疗机构就诊，平均等候时间缩短了2/3。家庭医生为居民提供互联网健康咨询问诊、转诊、慢病签约患者线上续方等服务。通过二维码技术实现各类就诊卡的多卡合一，患者在手机上就可以轻松实现就诊以及医疗保险的实时结算。通过信息技术推动医疗服务模式的深度改革，有效化解“看病烦”、“流程繁”，切实提高老百姓的获得感。 Short-Bio: Professor Guanhua Yao, MD, MBA, currently serves as the Director of the Xiamen Municipal Health and Family Planning Commission. His research focuses on disease control and prevention, community health, healthcare operations management and healthcare information management. He has rich experiences in the comprehensive management of chronic diseases, primary and community health, information construction of regional healthcare system.
|
14:15-15:00 
|
Talk 6: Zhankun (Kevin) Sun (City Univeristy of Hong Kong) Title: Who is Next: Dynamic Patient Prioritization in an Emergency Department Abstract:Upon arrival at an emergency department (ED), patients are triaged into different classes indicating their priority level. Using data from an urban teaching hospital in Alberta, Canada, we find that within the same priority level, the average waiting time (time to physician initial assessment) of discharged patients is shorter than that of admitted patients for middle-to-low acuity patients, suggesting that the order of patients being served deviates from first-come-first-served, and to certain extent, discharged patients are prioritized over admitted patients. This observation is intriguing as among patients of the same triage level, admitted patients - who need further care in the hospital - should be deemed no less urgent than discharged patients who only need treatment at the ED. To understand ED decision makers' patient prioritization behavior, i.e., how they choose the next patient for initial assessment, we develop a discrete choice framework and find that ED blocking level - measured by the percentage of ED beds occupied by boarding patients - is a key driver. We find that ED decision makers apply urgency-specific delay-dependent prioritization. Furthermore, we find that i) when ED blocking level is negligible, admitted patients are prioritized over discharged patients for high acuity patients; for middle-to-low acuity patients, FCFS is followed for patients from the same triage level; ii) as the blocking level increases, the chance of discharged patients being chosen for treatment is getting higher; and iii) the prioritization behavior varies across triage levels. We conclude that when the risk of ED being blocked becomes high, decision makers prioritize patients who are less likely to be admitted after treatment at ED, in an effort to avoid further blocking the ED. We then propose a Markov decision model and show that it is optimal for ED decision makers to alter their behavior in prioritizing patients when the ED faces increasing risk of being blocked by boarding patients. We derive policy implications based on the above findings. By testing and highlighting the central role of decision makers' prioritization behavior, this paper advances our understanding on ED operations and patient flow. Bio:Dr. Zhankun (Kevin) Sun is an Assistant Professor of Management Sciences in the College of Business, City University of Hong Kong. He holds a bachelor degree in Industrial Engineering from Tsinghua University and a Ph.D. in Statistics and Operations Research from the University of North Carolina Chapel Hill. He was an Eyes High postdoctoral fellow in the Haskayne School of Business, University of Calgary, Canada. He is the recipient of George Nicholson Award from UNC. His research interests are within the area of modeling, analysis, and control of stochastic systems. He is particularly interested in applications that arise from healthcare operations, such as emergency response management in the aftermath of mass-casualty events and patient flow management in hospital emergency departments. His work appears in reputed journals such as Management Science. |
15:00-15:20 |
Coffee Break |
15:20-16:05 
|
Talk 7:Hai Wang (Singapore Management University) Title: The Logic of Matching in Ride Sharing Markets: Revenues, Service Ratings or Pick-Up Times? Abstract: We study the multi-period multi-objective online ride-matching problem, where the dynamic ride-sharing platform needs to match passengers to available drivers in real time without observing future information, considering multiple objectives such as pick-up distance, driver service score, and platform revenue. We develop an efficient online matching policy that adaptively balances the trade-offs between multiple objectives in a dynamic and large-scale setting and provide theoretical performance guarantee for the proposed policy. We prove that the policy can achieve the "compromise solution'", i.e., the solution that minimizes the Euclidean distance to the utopia point, or achieves the solution that minimizes the Euclidean distance to any pre-determined target. Through numerical experiments and implementation using real data from a ride-sharing platform, we demonstrate that our approach is able to obtain a delicate balance of multiple objectives and bring value to all the stakeholders in the ride-sharing eco-system: (1) drivers with higher service scores (i.e., drivers who provide better service) are dispatched with more orders and earn higher incomes; (2) passengers are more likely to be served by drivers with higher service scores, and passengers with higher revenues are served with higher answer rates, at the expense of a tiny increase in pick-up distance; (3) the platform obtains a higher total revenue. Bio: Dr. Hai Wang is an Assistant Professor in the Area of Intelligent Systems & Optimization at Singapore Management University. He held a bachelor degree from Tsinghua University, dual Master's degree in operations research and transportation from MIT, and PhD in operations research from MIT ORC. His research has focused on the design and operation of future urban systems, including shared urban service, on-demand transportation, advanced logistics, and smart healthcare, using methodologies in operations research, machine learning, analytics and statistical inference. |
16:05-16:50 
|
Talk8:Renyu (Philip) Zhang (New York University Shanghai) Title:Recommender Systems on Platforms (Joint work with Xuan Wang and Dennis J. Zhang) Abstract:We study how a recommender system could shape the demand to maximize the number of matches between consumers and producers on a two-sided platform. To improve the long-term performance of a platform, the optimal recommender system design trades off expanding market demand by recommending high-quality producers and reducing congestion by recommending low-quality ones. The traditional recommender system has been focusing on identifying and recommending the producers of highest qualities. Albeit optimal when producers have unlimited capacities, such a recommender system may result in substantial optimality losses in the presence of producer capacity constraints. Such losses are most significant when the demand and supply of the platform is balanced. If, in addition, the quality of a producer improves upon working for the platform, the platform should recommend low-quality producers even more often to improve their qualities. A dynamic re-balancing recommendation policy which maximizes the number of matches myopically converges to an optimal policy in the long-run. We also extend our results to a setting with multiple quality levels and a setting with horizontally heterogeneous consumers and producers. Our results shed light on the optimal recommender system design for a broad range of two-sided platforms. Bio:Renyu (Philip) Zhang has been an Assistant Professor of Operations Management at New York University Shanghai since August 2016. His research addresses fundamental operations issues under the emerging trends in technology, marketplaces, and society. He is particularly enthusiastic about developing analytics techniques to study the operations problems in the context of online platforms and marketplaces, sharing economy, social networks, and sustainability. His research works have appeared in Operations Research and Manufacturing & Service Operations Management. Please visit his personal website for more about Philip:. Before joining NYU Shanghai, Philip obtained his doctoral degree in Operations Management at Olin Business School, Washington University in St. Louis in May 2016 under the supervision of Professor Nan Yang and Professor Fuqiang Zhang. |
16:50-17:35 
|
Talk 9:Jingqi Wang (Hong Kong University) Title: Impacts of Supplier Enforced Cross-licensing in a Supply Chain Abstract:Qualcomm, the largest cellphone chipmaker in the world, was recently fined RMB 6.088 billion (approximately $975 million) by the Chinese government for alleged anti-competitive conducts including requiring downstream phone manufacturers to cross-license their patents to Qualcomm and its customers. Qualcomm's cross-licensing practice has also received similar charges or scrutiny in South Korea, Japan, European Union, and the United States. Motivated by this practice, we study the impacts of cross-licensing in a supply chain in which an upstream supplier requires its downstream competing manufacturers to cross-license. We find that, contrary to common belief, cross-licensing may incentivize more innovation investment by the weak manufacturer. In addition, besides the weak manufacturer, even the strong one may benefit from cross-licensing under certain conditions. However, the supplier does not always benefit from conducting the cross-licensing practice. We show that cross-licensing does not always hurt social welfare or consumer surplus as it is accused for. We also find that allowing manufacturers to charge each other royalties benefit manufacturers at the cost of the supplier and consumers. Our results shed light on how cross-licensing affects innovation, profits and welfare, which have managerial implications to firms in high-tech industries, as well as to policy makers around the world. Bio:Jingqi Wang is an associate professor at the University of Hong Kong. He received his Ph.D. Degree in Operations Management from the Kellogg School of Management, Northwestern University, and obtains his Bachelor of Science degrees in Industrial Engineering from Tsinghua University. His research focuses on technology and innovation in supply chains, and empirical operations management. His work appears in journals such as Management Science, Manufacturing & Service Operations Management, and Production and Operations Management. |
17:35-19:30 |
顿颈苍苍别谤（厦大逸夫楼） |
2019年1月3日 |
小组讨论、业界参访 |
