Events

"Tipping in Service Systems: The Role of a Social Norm"

Presentation by Professor Laurens Debo
Tuck School of Business
Dartmouth College
 

Wednesday, March 6
3:15pm - Refreshments, Keller Hall 3-180
3:30pm - Graduate Seminar, Keller Hall 3-180

About:

We study the impact of tipping in a service facility on the server’s tipping wage, in the presence of an endogenously formed social norm. The problem is modeled as a two-stage tipping game: In the first stage, delay-sensitive customers arrive at an M/M/1 facility and decide whether to join or not, and if so, how much to tip. In the second stage customers who joined meet in a social market, compare their tip amounts, and are penalized on tipping differently from one another. All customers are homogeneous in terms of their valuation of the service and waiting cost rate. Yet, some customers visit the facility repeatedly and can therefore obtain a shorter waiting time by tipping more. Other customers visit only once and hence, cannot influence their waiting time via their tip. We find that in the presence of a social norm, the server cannot extract the optimal social welfare through tipping, because of rents that the repeat customers can accrue and the variability in tips, introduced by repeat customers, which causes the social costs to be strictly positive. Nevertheless, in some limiting cases, the tipping wage approaches the optimal welfare when the social norm is weak and there are many repeat customers such that no one-time customers join; or when the social norm is strong and there are few repeat customers who join.

Bio:

Laurens Debo is an Associate Professor of Operations Management at Tuck and the Harvey H. Bundy III T'68 Faculty Fellow. Previously, he was on the faculty of the Tepper School of Business of Carnegie Mellon University and the Booth School of Business at the University of Chicago. Professor Debo’s research focuses on the behavior of consumers and providers in different service settings. On the consumer side, he investigates how strategic consumer behavior shapes the demand for services. On the supply side, he studies the management of “discretionary services,” whose value to the consumer increases with the actual service time. His research has appeared in Manufacturing & Service Operations Management, Management Science, Operations Research and Production and Operations Management, among other journals. A part of his research has been funded by the NSF. He is an associate editor for Management Science, Manufacturing & Services Operations Management and Operations Research, a senior editor for Production and Operations Management, and serves on the editorial board of Service Science.

"Data Driven Forecasting and Revenue Management with Word of Mouth Dependent Reference Prices"

Presentation by Professor Opher Baron
Rotman School of Management
University of Toronto
 

Wednesday, February 27
3:15pm - Refreshments, Lind Hall 305
3:30pm - Graduate Seminar, Lind Hall 305

About:

We use data on the retail business of TMall to consider revenue management in the presence of reference prices. We consider questions as: Does reference price improves demand forecasts? And, does considering word of mouth (WOM) in the formation of reference price forecasts benefit revenue management? Revenue management considers the impact of prices on current and future sales via forecasts. The related literature uses reference price to study this impact empirically and theoretically, mostly focusing on sales of specific items to repeat customers. It ignores the effect of WOM on reference price. We (I) develop scalable, data driven methodologies to compare the effectiveness of different forecasts, (II) demonstrate these methodologies, (III) introduce models capturing the effect of WOM on reference price, and (IV) formulate and study (theoretically and numerically) the revenue management problem when forecasts are reference price-dependent. We provide a foundation for systematic implementation of revenue management in the presence of reference price effects. We (I) formulate the impact of WOM on reference price and investigate the effectiveness of different forecasts; and (II) demonstrate that revenue management could benefit from WOM-dependent reference price models. The improved accuracy and revenue management performance of our forecasting models with reference price affected by WOM support their usage in practice.

Bio:

Opher Baron is a Professor of Operations Management and the area coordinator for Operations Management and Statistics at the Rotman School of Management, the University of Toronto. He has a PhD in Operations Management from the Sloan school at the Massachusetts Institute of Technology along with an MBA and BSc in Industrial Engineering and Management from the Technion. On the teaching front, Opher is especially proud of the "Analytics for decision-making " MBA elective course he introduced and teach at Rotman. His research interest include queueing, applied probability, facility location, service operations (such as healthcare and call centers), inventory planning, and revenue management. Opher's published at leading journals such as Operations Research, and Manufacturing & Service Operations Management, and he has won several research awards and grants, including the 1000 Talent Plan Scholar of the Shanghai Municipal Government, 2016. Opher is active in the operations research and operations management community. He has chaired several conferences, clusters, and sessions and currently serves on the editorial board of the several journals including the Manufacturing & Service Operations Management and Service Science.

"Online Resource Allocation under Partially Predictable Demand"

Presentation by Professor Patrick Jaillet
Department of Electrical Engineering and Computer Science
Massachusetts Institute of Technology

Wednesday, December 12
3:15pm - Refreshments, Lind Hall 305
3:30pm - Graduate Seminar, Lind Hall 305

About:

There are many situations in which present actions must be made and resources allocated with incomplete knowledge of the future. Online optimization typically compares the performance of a strategy that operates with no knowledge of the future (on-line) with the performance of an optimal strategy that has complete knowledge of the future (off-line). In some cases, partial information about the future may be learnable and lead to provably better online algorithms. In this talk, we provide recent results obtained from that perspective on a simple online resource allocation problem where the sequence of arrivals (requests) contains both an adversarial component and a stochastic one.

Bio:

Patrick Jaillet is the Dugald C. Jackson Professor in the Department of Electrical Engineering and Computer Science and a member of the Laboratory for Information and Decision Systems at MIT. He is also co-Director of the MIT Operations Research Center and the Faculty Director of the MIT-France Program. Dr. Jaillet's research interests include online optimization and learning; real-time, dynamic, and data-driven problems; and networks. He is a Fellow of the Institute for Operations Research and Management Science Society (INFORMS) and a member of the Society for Industrial and Applied Mathematics (SIAM).

"The Application of Optimization Techniques in Power Systems"

Presentation by Lei Fan
Software Engineer
Siemens Industry, Inc.

Wednesday, November 28
3:15pm - Refreshments, Lind Hall 305
3:30pm - Graduate Seminar, Lind Hall 305

About:

In this talk, Dr. Fan will first introduce the operations of electricity grid systems. He will discuss the application optimization techniques and data analytics in electricity industry. Furthermore, he will talk about the software development experiences in power industry. Finally, he will discuss the hot topics in power industries.

Bio:

Dr. Lei Fan received the B.S. degree in electrical engineering from the Hefei University of Technology, Hefei, China, in 2009 and the Ph.D. degree in Industrial and Systems Engineering at the University of Florida, Gainesville, FL, USA in 2015. He was an application engineer in General Electric from 2015 to 2017. Currently, he is a software engineer in Siemens Industry, Inc. His research interests include the optimization of power system operations, planning, and energy market analysis.

"Sampling the City: Turning Operational Transit Data into Insights for Planners and Policymakers"

Presentation by Eric Lind
Research and Analytics Manager
Metro Transit

Wednesday, November 14
3:15pm - Refreshments, Lind Hall 305
3:30pm - Graduate Seminar, Lind Hall 305

About:

Every day Metro Transit implements transit service across the region, with the main purpose of providing 250,000 trips to destinations. As a secondary consequence of the service, a repeated time-varying network of data points is generated on how the municipal area functions: where people are going from and to, the speeds and delays of transit vehicles, occurrences of mobility-limited boardings, bicycle-bus connections, and more. These data can be in turn analyzed and modeled to create useable planning information today, and are envisioned to support real-time transportation optimization in the future of connected vehicles. In this talk I describe the advantages and challenges of acting to analyze and interpret these data for decision makers, focusing on concrete examples of fine-grained bus speed, people throughput in urban corridors, and ridership patterns.

Bio:

Eric M. Lind is Manager of Research and Analytics in the Strategic Initiatives group at Metro Transit, the largest public transportation provider in the Minneapolis-St. Paul metropolitan area. Eric has a Ph.D in Ecology from the University of Maryland and worked as a researcher in quantitative ecology at the University of Minnesota before joining Metro Transit in 2017. He uses the statistical and analytical skills he developed exploring ecosystem dynamics to aid Metro Transit in understanding the similarly complex systems of human interaction with a transportation network. His work has included models to predict new operator longevity with the agency, measure efficiencies of their internal systems, explain preferences and behavior of customers, and forecast trends in performance. He still considers himself an entomologist.

"Measuring the Benefit of Revenue Management Software Using Simulation"

Presentation by Betsy Enstrom
R&D Manager
IDeaS Revenue Solutions

Wednesday, October 31
3:15pm - Refreshments, Lind Hall 305
3:30pm - Graduate Seminar, Lind Hall 305

About:

Given the growing number of competitors and options in the hospitality revenue management software industry, there is a growing need for an unbiased, reliable way to measure the benefit of an automated revenue management solution. Not surprisingly, as revenue management has matured and increased in sophistication, so has its subscription costs. Existing and potential revenue management subscribers want to know how much benefit they could expect for their specific property. In this seminar, this benefit measurement analysis uses simulation to help address the need to quantify the benefit for properties with their own bookings, analysis period, inventory, market conditions, and analysis period.

Bio:

Betsy Enstrom works in revenue management at IDeaS Revenue Solutions as a R&D Manager in their Advanced Analytics Testing department in Bloomington, Minnesota. She has developed and tested projects in forecasting, optimization, statistical analysis, graphical data analysis, and quantifying revenue benefit from IDeaS revenue management solutions. She received her undergraduate degree from the University of Minnesota in Math, Physics and Astronomy and pursued her graduate degrees in Statistics at North Carolina State University and Duke University. She began her 11-year tenure at SAS headquarters in North Carolina in the Statistics Department. As a native Minnesotan, she is pleased to work at IDeaS for the past eight years in the hospitality industry revenue management research and development.

"Liver Transplant Equity through Redistricting Optimization"

Presentation by Professor Sommer Gentry
Department of Mathematics
United States Naval Academy

Research Associate
John Hopkins University School of Medicine

Wednesday, October 24
3:15pm - Refreshments, Lind Hall 305
3:30pm - Graduate Seminar, Lind Hall 305

About:

In some areas of the U.S., the sickest liver transplant candidates have an 82% chance of dying; in others, a 14% chance of dying, because of extreme disparities in the availability of livers for transplant. Historical precedent dictates organ sharing within a hierarchy of 50 small donor service areas grouped into 11 regions, and these areas have very different ratios of eligible liver donors to liver transplant candidates. We designed novel regions for liver allocation by partitioning the set of donor service areas according to an integer program redistricting model. Our work directly addressed the paramount clinical and ethical concern about geographic equity in transplantation. We validated our redistricted maps using the clinically detailed Liver Simulated Allocation Model that is the gold standard for testing allocation policy proposals, to compensate for the necessarily simplified and aggregated picture of liver allocation in an integer program. Trying to implement this solution led to an intense battle over the scarce livers that has had a number of twists and turns, and I will discuss a web of related studies our research group has undertaken to support the movement towards transplant equity.

Bio:

Sommer Gentry is a Professor of Mathematics at the United States Naval Academy, and is also on the faculty of the Johns Hopkins University School of Medicine. She is a senior investigator with the Scientific Registry for Transplant Recipients. She has a B.S. in Mathematical and Computational Science and an M.S. in Operations Research, both from Stanford University, and a Ph.D. in Electrical Engineering and Computer Science from MIT. She designed matching optimization methods used for nationwide kidney paired donation registries in both the United States and Canada, and is now redistricting liver sharing boundaries to help the Organ Procurement and Transplantation Network reduce geographic disparities in transplantation. Her work has attracted the attention of major media outlets including Time Magazine, Reader’s Digest, Science, National Public Radio, and the New York Times. Gentry has received the MAA’s Henry L. Alder award for distinguished teaching by a beginning mathematics faculty member, and was a finalist for the INFORMS Daniel H. Wagner prize for excellence in operations research practice.

"Bounding Optimality Gaps via Bagging"

Presentation by Professor Hentry Lam
Department of Industrial Engineering and Operations Research
Columbia University

Wednesday, October 17
3:15pm - Refreshments, Lind Hall 305
3:30pm - Graduate Seminar, Lind Hall 305

About:

We study a statistical method to estimate the optimal value, and the optimality gap of a given solution for stochastic optimization as an assessment of the solution quality. Our approach is based on bootstrap aggregating, or bagging, resampled sample average approximation (SAA). We show how this approach leads to valid statistical confidence bounds for non-smooth optimization, and demonstrate and compare its statistical efficiency and stability with some existing methods. We also present our theory by viewing SAA as a kernel in an infinite-order symmetric statistic, which leads to some generalizations of classical central limit results for SAA.

Bio:

Henry Lam is an Associate Professor in the Department of Industrial Engineering and Operations Research in Columbia University. He received his Ph.D. in statistics from Harvard University in 2011, and was on the faculty of Boston University and the University of Michigan before joining Columbia in 2017. Henry's research interests include Monte Carlo simulation, risk and uncertainty quantification, and optimization under uncertainty. His work has been recognized by several venues such as the NSF Career Award (2017) and the INFORMS JFIG Competition Second Prize (2016). He serves on the editorial boards of Operations Research and INFORMS Journal on Computing.

"Optimization in Lending: A Real-World Discussion"

Presentation by Jason Ford
Managing Consultant
AQN Strategies

Wednesday, October 10
3:15pm - Refreshments, Lind Hall 305
3:30pm - Graduate Seminar, Lind Hall 305

About:

The credit and lending industry is a fiercely competitive landscape with quickly thinning margins and significant downsides in the form of risk. Participants in this space optimize on many different axes – marketing, product, pricing, and so on – all with their own set of challenges that require careful modeling and frequent iteration. In this seminar, we will walk through a few projects in this space emphasizing real world realities – business constraints, low sample, and an environment that continues to move beneath your feet.

Bio:

Jason Ford works as a Managing Consultant with AQN Strategies, a boutique consulting firm specializing in consumer and small business lending. With AQN, Jason has led analytical projects for banks, fintechs, and private equity firms across their full spectrum of credit needs. His extensive analytical and modeling experience allows him the ability to critique, design and test credit and operations-related solutions through a practitioner’s lens. Prior to joining AQN, Jason performed roles in analytics at Capital One and fintech unicorn GreenSky. Jason holds a Bachelor of Science degree in Mathematics from the University of Minnesota.

"Dynamic, Personalized Surveys via Adaptive Matrix Sampling"

Presentation by Dr. Sean Taylor
Data Science
Facebook

Wednesday, October 3
3:15pm - Refreshments, Lind Hall 305
3:30pm - Graduate Seminar, Lind Hall 305

About:

Surveys represent the best chance online platforms have to efficiently gather feedback from users. In an era of declining response rates, we seek to efficiently gather the most meaningful data possible from respondents who are often unwilling to answer long surveys. One way to improve efficiency is to leverage rich sets of covariates from large data sources to estimate response models and borrow information across respondents and questions. We show how survey response models can be used shorten surveys without decreasing information, via an adaptive matrix sampling procedure that downsamples questions into the most informative subset for each respondent. Our proposed question selection method optimizes for variance reduction while incorporating side information about respondents. As a respondent answers questions, their posterior response distributions are updated in an online fashion. The efficiency gains of our approach enable the researcher to terminate sampling early when desired precision is reached or evolve a survey over time by adding and removing questions.

Bio:

Sean J. Taylor is a social scientist and applied statistician on Facebook’s Core Data Science team. Prior to Facebook, he earned his PhD in Information Systems from NYU’s Stern School of Business. He specializes in using machine learning methods and randomized experiments for measurement, prediction, and policy decisions. Sean’s research ranges from studying online social influence, viral marketing, and social networks to measuring how sports fans behave and the impact of data science on decision making in organizations. He is also an avid engineer who enjoys putting academic research into practice by building open source software for data scientists.