Columbia University-IBM Conference
09/29/2006
September 29, 2006, 9:30-5:30
IBM T.J. Watson Research Center
1101 Kitchawan Rd.
Yorktown
Heights, NY
Please contact
for transportation details by
Wednesday, September 27, if
you are interested in attending.
Organizers
Guillermo Gallego, Department of Industrial Engineering and Operations Research, Columbia University
Giuseppe A. Paleologo,
Mathematical Sciences Department,
IBM Research
Attendance to the conference is
free. If you plan to attend, please
email
by
September 22.
World economies increasingly rely on their service sectors: services account for 80% of the U.S. economy and over 70% of the G.D.P. for OECD-area economies. The share of the service sector in China was 21% of G.D.P. in 1979; today, it is 33%. While services are changing the way we live, we have not yet changed the way we think about them, deliver them, and design them. We still model human supply chains after industrial supply chains and apply to service pricing the same categories and quantitative methods that were conceived from commodity pricing. In order to meet the challenges arising from this environment, we need a different intellectual arsenal. In this meeting, jointly organized by IBM Research and Columbia University, management scientists who are exploring the frontiers of the service economy will discuss problems they have met, and the novel ways they have devised to solve them.
SPEAKERS
DIMITRIS BERTSIMAS
Boeing Professor of Operations
Research, Sloan School of
Management, MIT
J. MICHAEL HARRISON
Adams Distinguished Professor of
Management, Graduate School of
Business, Stanford University
ROBERT L. PHILLIPS
Chief Technology Officer, Nomis
Solutions
BARRY SMITH
Chief Scientist and Senior Vice
President of Research and
Development, Sabre
LARRY ROSENBERGER
Vice President of Research and
Development, Fair Isaac
RICHARD R. WEBER
Churchill Professor of Mathematics
for Operational Research, Cambridge
University
DIMITRIS BERTSIMAS: PREDICTION OF HEALTH CARE COSTS VIA DATA-MINING AND ALGORITHMIC DISCOVERY OF MEDICAL KNOWLEDGE
ABSTRACT
Rising health care costs are one of the world's most important problems. Correspondingly, predicting such costs with accuracy is a significant first step in addressing this problem. Since the 1980s, there have been research efforts for predictive modeling of medical costs based on claims data that utilize heuristic rules and classical regression methods that have not been appropriately validated in populations that the methods have not seen. In this study, we utilize modern data mining methods, specifically classification trees and clustering algorithms, and claims data from close to four hundred thousand members over three years to provide a) predictions of health care costs in the third year, based on medical and cost data from the first two years, which we rigorously validate, and b) an illustration through two examples that our methods can lead to discovery of medical knowledge. We quantify the accuracy of our predictions using out of sample data from over one hundred thousand members. The key insights we obtain are: a) our data mining methods provide accurate predictions of medical costs and represent a powerful tool for prediction of health care costs, b) the pattern of past cost data are strong predictors of future costs, c) medical information is an accurate prediction of medical costs particularly on high risk members, and d) new medical knowledge can be obtained through our methods.
BIO
Dimitris Bertsimas is currently the Boeing Professor of Operations Research and the codirector of the Operations Research Center at the Massachusetts Institute of Technology. He has received a B.S. in Electrical Engineering and Computer Science at the National Technical University of Athens, Greece, in 1985, a M.S. in Operations Research at MIT in 1987, and a Ph.D in Applied Mathematics and Operations Research at MIT in 1988. Since 1988, he has been with MIT's Sloan School of Management. His research interests include optimization, stochastic systems, data mining, and their application. He is a member of the National Academy of Engineering, and he has received numerous research awards including the Erlang prize (1996), the SIAM prize in optimization (1996), the Bodossaki prize (1998) and the Presidential Young Investigator award (1991-96).
J. MICHAEL HARRISON: PROCESSING NETWORKS AND SERVICE PRODUCTS
ABSTRACT
This talk will focus on subscription-based services. Examples include outsourced business functions (IT support, basic accounting, benefits administration, etc.) and field support of complex equipment (routers and switches, military aircraft, mainframe computers, etc.). The three main topics to be discussed are: (1) modeling physical capabilities via the processing network paradigm; (2) modeling the heterogeneity of potential customers with regard to service sensitivity and price sensitivity; and (3) defining and pricing differentiated grades of service. With regard to topic (3), different grades of service are effectively defined by the system manager’s resource allocation policy; bid-pricing schemes are widely used by passenger airlines for dynamic resource allocation, and that same general approach has great appeal and wide potential application in the service arena more broadly.
BIO
J. Michael Harrison is the Adams Distinguished Professor of Management in the Graduate School of Business, Stanford University. He earned a B.S. degree in industrial engineering from Lehigh University, an M.S. in industrial engineering from Stanford, and a Ph.D. in operations research from Stanford before joining the faculty of the Graduate School of Business in 1970. He has developed and analyzed stochastic models in several different domains related to business, including mathematical finance and processing network theory. His current research is focused on call center management, dynamic pricing and revenue management. Professor Harrison has been honored by INFORMS with its Expository Writing Award (1998), the Lanchester Price for best research publication (2001), and the John von Neumann Theory Prize (2004). He is a fellow of INFORMS and of the Institute for Mathematical Statistics.
ROBERT PHILLIPS: OPTIMIZING CREDIT PRICING
ABSTRACT
Banks and other lenders have become highly sophisticated in calculating the risk–that is, the expected cost–of the loans they make to both consumers and to businesses. However, lenders have been less sophisticated in understanding how their customers respond to the prices of their offerings. I describe an approach to optimizing the price offered for credit in which the prospective borrower approaches potential lenders with a need and each lender has broad freedom to set prices based on customer, channel, and product characteristics. This characterizes a wide variety of credit markets including mortgages, auto loans, student loans, home equity lines and loans, and business loans. I present a proven methodology for optimizing the prices that lenders should offer to prospective customers. A unique feature of credit markets is the need to include adverse selection and complex expected profit models. In addition, financial service pricing often involves multi-stage pricing decisions. These complications have necessitated the development of pricing approaches specific to credit markets. Optimizing prices has resulted in substantial profitability gains across a wide variety of credit markets. I present results from applying the methodology in different settings including auto lending, home equity lines, and unsecured consumer loans.
BIO
Dr. Robert Phillips is founder and chief technology officer of Nomis Solutions, a company providing pricing and revenue optimization solutions to the financial services industry. Prior to founding Nomis, Dr. Phillips served as chief technology officer of Manugistics (NASDAQ: MANU) and as founder and chief executive officer of Talus Solutions, a pricing and revenue optimization company. Prior to that he was chief executive officer at Decision Focus Incorporated, a quantitative consulting firm. Dr. Phillips was the chief architect of the Talus Solutions pricing and revenue optimization software suite. He is author of the book, Pricing and Revenue Optimization, which was published by the Stanford University Press in 2005. Dr. Phillips holds a Ph.D. in Engineering-Economic Systems from Stanford University and B.A. degrees in Economics and Mathematics from Washington State University.
BARRY SMITH: HELPING
TRAVELOCITY TRANSITION TO TRAVEL
RETAILER
ABSTRACT
Sabre is the world’s largest travel distribution system, connecting consumers, travel agents and travel suppliers. In 1996 Sabre launched Travelocity, an online travel company that leverages Sabre’s travel experience and technology to begin selling travel directly to consumers. The foundation of Travelocity’s original business model depended on earning commissions from selling airline tickets and generating revenues from selling advertising space. Travelocity initially took the lead in Internet airline ticket sales, but as new competitors entered the space, Travelocity’s share of travel bookings steadily eroded. In May 2002, then-president of Travelocity, Sam Gilliland, asked the Sabre Research Group to help Travelocity adapt to this evolving, new environment and to increase their sophistication as retailers. The Sabre-Travelocity team developed the Enterprise Network Model (ENM) to improve decision-making in this environment. Since 2002, the ENM has yielded over $54 million of incremental value, with a current annual run-rate of $43 million.
In this presentation, we’ll describe our experience helping to transform Travelocity from an Internet company surviving on growth to a sophisticated on-line retailer. We’ll address the following topics: 1) customer data and modeling; 2) supply modeling; 3) marketing modeling; 4) modeling to support retail pricing and Low Fare Search (LFS). Finally we summarize the benefits of the ENM to Travelocity and review the transferability of these concepts and models to other applications.
BIO
Dr. Smith is Chief Scientist and Corporate Officer for Sabre Holdings, the world’s largest distributor of travel through traditional travel agencies as well as its on-line agency Travelocity. Sabre is also a leading supplier of airline reservation services and decision support software. Dr. Smith leads the Sabre Research Group and is responsible for developing new decision support applications for Sabre’s travel distribution, retailing and software lines of business. Dr. Smith developed many of the revenue management techniques used throughout the airline industry and pioneered the application of revenue management techniques in other industries. His team was awarded the 1991 Franz Edelman prize for airline yield management; and the 2005 INFORMS Revenue Management and Pricing Section Prize. He served as president of AGIFORS (Airline Group of the International Operational Research Societies); in 2004 he was named AGIFORS Fellow. Dr. Smith holds a Ph.D. in Industrial and Systems Engineering from Georgia Tech, a master of science in Operations Research from MIT and a bachelor of Aerospace Engineering from Georgia Tech.
LARRY E. ROSENBERGER
ABSTRACT
To be announced.
BIO
Larry Rosenberger leads Fair Isaac's Research and Development. Mr. Rosenberger joined Fair Isaac in 1974 and served as president and chief executive officer from 1991 to 1999. During that time, Fair Isaac experienced consecutive years of record growth, with annual revenues increasing from $31 million to over $276 million. Prior to that position, he managed the Engineering, Research, and Development Division. In that capacity, he was responsible for the technical development, production and marketing of the company's most advanced products. Mr. Rosenberger holds a B.S. in physics from the Massachusetts Institute of Technology and an M.S. in physics and operations research from the University of California, Berkeley. He is also chairman of the board of directors of the Marin Education Fund and board member of the Marin Community Foundation.
RICHARD WEBER: INCENTIVES FOR
LARGE PEER-TO-PEER SYSTEMS
ABSTRACT
The design of a peer-to-peer (P2P) system poses many interesting questions. What quality of service should it be designed to provide? How can peers contribute to its cost and how can the 'free-rider problem' be avoided? While it is almost never possible to calculate or implement a truly optimal mechanism design, we show that as the number of participants becomes large there is a very simple scheme that can do almost as well as an optimal one. It takes a fixed payment from any agent who wishes to join the system as a participating peer. We apply these ideas to designing two P2P systems.
BIO
Richard Weber is Churchill Professor of Mathematics for Operational Research in the Mathematics Department of the University of Cambridge. His research interests include the communications networks, network economics and pricing, online algorithms, queueing control, scheduling, search and rendezvous games. A recent book, with Costas Courcoubetis is titled "Pricing Communication Systems." He is presently involved in a European research project called GridEcon, which is concerned with the economic issues that arise in Grid computing business models.