Below is the IEMS Fall 2009 seminar series schedule, with a general area/topic of the talk. Please mark your calendar accordingly. All talks are from 4:00 to 5:00 in Tech Room M228.
Talk information and video presentations will be posted as soon as the content is made available.
For more information on the seminar series, please contact Benjamin Armbruster (armbruster@northwestern.edu).
Open All SeminarsClose All Seminars
- Tuesday, September 22rd, 2009 - Network and Combinatorial Optimization; Operations Research; Logistics
James Orlin, Massachusetts Institute of Technology - Title: A Faster and Simpler Algorithm for Computing Market Equilibrium
Abstract: Devanur, Papadimitriou, Saberi, and Vazirani recently published the first polynomial time algorithm for computing an equilibrium for the linear utilities case of the market model defined by Fisher. Here we provide a much faster and simpler algorithm. In addition, we provide the first strongly polynomial time algorithm for computing the equilibrium. (This latter algorithm is not as simple.) We describe our algorithms from an economic perspective in a way that is accessible to a general audience. This work is joint with Mehdi Ghiyasvand.
Bio: James Orlin specializes in network and combinatorial optimization. He has helped develop improved solution methodologies in airline scheduling, railroad scheduling, logistics, network design, telecommunications, inventory control, and marketing. Together with MIT Sloan colleague Thomas L. Magnanti and Ravindra K. Ahuja, he has written the award-winning text Network Flows: Theory, Algorithms, and Applications (Prentice Hall, 1993). To see video of this presentation, please click here. - Tuesday, September 29th, 2009 - Computational Epidemiology, Computational Economics
Madhav Marathe, Virginia Bioinformatics Institute - Title: Epidemics in Co-evolving Networks: The Role of Individual behaviors and Public Policies
Abstract: Complex Networks are pervasive in our society. Realistic biological, information, social and technical networks share a number of unique features that distinguish them from physical networks. Examples of such features include: irregularity, time-varying structure, heterogeneity among individual components and selfish/cooperative game-like behavior by individual components. Furthermore, the network structure, the dynamical process on the network and the behavior of constituent agents co-evolve over time. The size and heterogeneity of these networks, their co-evolving nature and the technical difficulties in applying dimension reduction techniques commonly used to analyze physical systems makes reasoning, prediction and controlling of these networks even more challenging. Recent quantitative changes in high performance and wireless computing capability have created new opportunities for collecting, integrating, analyzing and accessing information related to such large complex networks. The advances in network and information science that build on this new capability provide entirely new ways for reasoning and controlling these networks. Together, they enhance our ability to formulate, analyze and realize novel public policies pertaining to these complex networks. The talk will focus on elements of network and information science requires to support policy informatics as it pertains to epidemic processes in co-evolving social and wireless networks. Understanding these epidemiological processes is of immense societal importance. Additionally they serve as excellent "model organisms" for developing a theory of co-evolving complex networks. Perhaps more intriguing, recent advances in wireless communications provide compelling reasons for studying these networks together. I will discuss this possibility in my concluding remarks.
Bio: Madhav Marathe is a professor of Computer Science and deputy director of the Network Dynamics and Simulation Science Laboratory. He obtained his Bachelor of Technology degree in 1989 in Computer Science and Engineering from the Indian Institute of Technology, Madras, and his Ph.D. in 1994 in Computer Science from the University of Albany. Before coming to Virginia Tech in 2005, he worked in the Basic and Applied Simulation Science group (CCS-5) in the Computer and Computational Sciences division at Los Alamos National Laboratory where he was team leader in a theory-based, advanced simulation program to represent, design, and analyze extremely large socio-technical and critical infrastructure systems. He has publishsed more than 150 research articles in peer reviewed journals, conference proceedings, and books, and has over eight years of experience in project leadership and technology development, specializing in population dynamics, telecommunication systems, epidemiology, design and architecture of the data grid, design and analysis of algorithms for data manipulation, design of services-oriented architectures, and socio-technical systems.
- Tuesday, October 6th, 2009 - Computer Science
Michael Friedlander, University of British Columbia - Title: Algorithms for Large-scale Sparse Reconstruction
Abstract: Many signal-processing applications seek to approximate a signal as superposition of only a few elementary atoms drawn from a large collection. This is known as sparse reconstruction. The theory of compressed sensing allows us to pose sparse reconstruction problems as structured convex optimization problems. I will discuss the role of duality in revealing some unexpected and useful properties of these problems, and will show how they lead to practical, large-scale algorithms. I will also describe some applications of the resulting algorithms.
Bio: Michael Friedlander received his PhD in Operations Research from Stanford University in 2002, and his BA in Physics from Cornell University in 1993. His research is primarily in large-scale constrained optimization. From 2002-2004 he was the Wilkinson Fellow in Scientific Computing at Argonne National Laboratory. He is currently Associate Professor of Computer Science at the University of British Columbia, where he has been since 2004.
- Tuesday, October 20th, 2009 – The intersection of probability and optimization theory, and in particular in the optimization and control of large-scale stochastic systems
Ciamac Moallemi, Columbia University - Title: Approximate Dynamic Programming via a Smoothed Linear Program
Abstract: We present a novel linear program for the approximation of the dynamic programming cost-to-go function in high-dimensional stochastic control problems. LP approaches to approximate DP have typically relied on a natural`projection' of a well studied linear program for exact dynamic programming. Such programs restrict attention to approximations that are lower bounds to the optimal cost-to-go function. Our program---the `smoothed approximate linear program'---is distinct from such approaches and relaxes the restriction to lower bounding approximations in an appropriate fashion while remaining computationally tractable. Doing so appears to have several advantages: First, we demonstrate substantially superior bounds on the quality of approximation to the optimal cost-to-go function afforded by our approach. Second, experiments with our approach on a challenging problem (the game of Tetris) show that the approach outperforms the existing LP approach (which has previously been shown to be competitive with several ADPalgorithms) by an order of magnitude. This is joint work with Vijay Desai (Columbia) and Vivek Farias (MIT).
Bio: Ciamac Moallemi is an Assistant Professor at the Graduate School of Business of Columbia University, where he has been since 2007. He received SB degrees in Electrical Engineering & Computer Science and in Mathematics from the Massachusetts Institute of Technology (1996). He studied at the University of Cambridge, where he earned a Certificate of Advanced Study in Mathematics, with distinction (1997). He received a PhD in Electrical Engineering from Stanford University (2007). Prior to his doctoral studies, he developed quantitative methods in a number of entrepreneurial ventures, as a founder of a computer security software startup, in an early stage drug discovery startup, and as a partner of a fixed-income arbitrage hedge fund. He is a member of the IEEE and INFORMS. He is the recipient of a British Marshall Scholarship (1996) and a Benchmark Stanford Graduate Fellowship (2003).
- Tuesday, October 27th, 2009 - Integrated Topology Control and Routing in Wireless Sensor Network Design for Prolonged Network Lifetime
Halit Uster - Title: Integrated Topology Control and Routing in Wireless Sensor Network Design for Prolonged Network Lifetime
Abstract: A Wireless Sensor Networks (WSN) can be deployed in inhospitable environments, difficult-to-reach terrains, and wild habitats such as war zones, oceans, the poles, deserts, and rain forests to monitor and observe natural, environmental, and climate-related phenomena including temperature, pressure, humidity, vibration, light, chemical and biological agents, etc. The most distinguishing characteristic of a WSN is the fact that its sensors have finite and non-renewable energy resources. In this talk, we consider integrated topology control and routing issues in data gathering WSNs and address the problem of prolonging network lifetime until next sensor deployment. We suggest a policy to be employed in each period of a deployment cycle and computationally examine its efficiency in prolonging network lifetime against two other commonly employed approaches. We present mathematical models and a general efficient solution procedure that is applicable to all the three cases. We determine that the proposed policy has attractive properties and exhibits very good performance in extending network lifetime. For the proposed policy, we also discuss further parallelization of the solution approach along with additional cut inequalities as well as a Benders Decomposition approach. (Research supported by National Science Foundation under grant CMMI-0428831).
Bio: Halit Üster is an Associate Professor in the Department of Industrial and Systems Engineering at Texas A&M University. He holds a Ph.D. in Management Science/Systems from McMaster University, Ontario, Canada. His research interests are in the areas of logistics, design of networked systems, and applied optimization. His publications appeared in Computers and Operations Research, European Journal of Operational Research, Interfaces, IIE Transactions, Naval Research Logistics, and Transportation Science among others. He is visiting the IEMS Department at Northwestern University during 2009-2010 academic year where he has been named Eshbach Scholar in the McCormick School of Engineering & Applied Science for Fall Quarter 2009.
- Tuesday, November 3rd, 2009 – Supply Chain and Information Systems
Russell R. Barton, Penn State University - Title: A Framework for Monitoring Supply Chain Execution
Abstract: Real-time sensor data, particularly RFID and GPS data, provides new possibilities to monitor the execution of supply chain processes. This talk examines a basic framework for examining the timeliness and correctness of the movement of items and/or transactions through a sequence of supply chain events. The nature of the data poses special issues for statistical process control and process capability analysis.
Bio: Russell R. Barton is a professor in the Department of Supply Chain and Information Systems at Penn State University. He also serves at the Smeal College of Business Co-Director for a one-year Master of Manufacturing Management degree program that is jointly offered with the College of Engineering. He received a B.S. degree in Electrical Engineering from Princeton and M.S. and Ph.D. degrees in Operations Research from Cornell. Before entering academia, he spent twelve years in industry. He was program chair for the 2007 Winter Simulation Conference, and he serves on the Advisory Board for the Quality, Statistics and Reliability Section of INFORMS. His research interests include applications of statistical and simulation methods to system design and to product design, manufacturing and delivery.
- Tuesday, November 10th, 2009 – Financial Mathematics, Stochastic Control, Quantitative Finance
Thaleia Zariphopoulou, Oxford University - Title: A New Approach for Investment Performance Measurement
Abstract: A new method for measuring the performance of investment policies will be introduced. Optimality of investment strategies will be associated with a stochastic partial differential equation (spde). The novel concept of performance volatility, as the driver
to this spde, will be presented. Examples of performance volatility processes, modeling different numeraires, benchmarks and market views, will be presented.
Bio: Professor Thaleia Zariphopoulou is the first holder of the Man Professorship of Quantitative Finance at the University of Oxford. Previously, Professor Zariphopoulou was the VF Neuhaus Centennial Professor at the University of Texas at Austin. At UT-Austin, Professor Zariphopoulou held a joint appointment between the Department of Mathematics and the Department of Information, Risk and Operations Management at the Red McCombs School of Business. She is an expert in quantitative finance, portfolio management and stochastic optimization. She was President of the Bachelier Finance Society from 2006-2008; and also sits on the editorial board of five academic journals in financial mathematics and quantitative finance. To see video of this presentation, please click here. - Tuesday, November 17th, 2009 – The development of models and solution methods for decision making under uncertainty, in particular, stochastic programming models and algorithmic methods for their solution
Dr. Julie Higle, Ohio State University - Title: How Industrial Engineers Impact the World
Abstract: Industrial engineering isn't quite as tangible as other types of engineering. The industrial engineering function evolves from a viewpoint and training that transcends an amazing breadth of "businesses". This makes it hard to nail down a simplistic sound-bite-sized descriptor that can be applied to industrial engineering. Because of this, the contributions of industrial engineers often fail to capture the mainstream imagination. This seminar will explore the impact of industrial engineering -- past, present, and future.
Bio:Professor Julie Higle joined the faculty of Systems and Industrial Engineering at the University of Arizona in 1985, after completing her PhD in IOE at the University of Michigan. Her primary research has been in models algorithms for decision making under uncertainty, with an emphasis on stochastic programming. She is currently the chair of the department of Integrated Systems Engineering at THE Ohio State University, and Sr. Vice President for Academics of the Institute of Industrial Engineers. - Monday, November 23rd, 2009 -
Robert Smith, National Science Foundation - Title: A Fictitious Play Approach to Complex Systems Optimization
Abstract: Complex systems consisting of a large number of interacting components are in practice increasingly modeled through computer simulations rather than via traditional equation based approaches. The resulting model typically allows for little or no structural assumptions on the form of the objective function or constraints, thus posing a challenging optimization problem. We explore in this talk a novel optimization paradigm inherited from game theory that animates the components of the system within a non-cooperative game of identical interest. The optimizations take place though individual best replies of the players, thus vastly reducing the dimensionality of the optimization problems solved (the components joint interactions are reflected indirectly through their shared objective function). We will illustrate the approach by discussing an application to intelligent transportation systems. Opportunities for NSF funding in Operations Research will be discussed at the end of the talk.
Bio: Robert L. Smith is Director of the Operations Research Program at NSF. Dr. Smith received his Ph.D. in Engineering Science from the University of California at Berkeley where he held an NSF Fellowship. He holds a bachelors degree in Physics from Harvey Mudd College and an MBA from Berkeley. He is on leave from the University of Michigan in Ann Arbor where he is the Altarum/ERIM Russell D. O'Neal Professor of Engineering. He is the recipient of the first Altarum/ERIM Russell D. O'Neal Professorship of Engineering at the University of Michigan. He has also been honored with the Distinguished Faculty Achievement Award from the University of Michigan, the College of Engineering Research Excellence Award, the Industrial and Operations Engineering Award for Outstanding Accomplishment, an Outstanding Teacher Award from the Michigan Student Assembly, and National Science Foundation Fellowship. He is a Fellow of the Institute for Operations Research and the Management Sciences. Professor Smith teaches courses in dynamic programming and stochastic processes. He has supervised the doctoral research of twenty-eight students since 1984. At the University of Michigan, he serves as Director of the Dynamic Systems Optimization Laboratory. The Laboratory research is directed toward the modeling and analysis of dynamical systems over time. Dr. Smith worked earlier at Bell Laboratories in the Network Planning Department where he developed models and algorithms for optimal routing of communications traffic. He is an Associate Editor of Operations Research and past Associate Editor of Management Science, and is the author of nearly one hundred peer reviewed publications.
- Tuesday, December 1st, 2009 -
Ahmad Faruqui, Principal – The Brattle Group - Title: The Changing Architecture of Demand Response
Abstract:Virtually everyone concedes that demand response should play an important role in meeting future energy needs. But there is little agreement on what should be the nature of its role. Two key questions continue to be debated with much vigor in regulatory proceedings before state commissions and in executive boardrooms. First, how big of a role can demand response play in meeting future energy needs? And second, what should be the mix of programs in the demand response portfolio? <?xml:namespace prefix = o ns = "urn:schemas-microsoft-com:office:office" />
With a view to advancing this debate, we present the key findings from a comprehensive assessment of the potential for demand response that was submitted by the Federal Energy Regulatory Commission to Congress in June 2009. We also present evidence that points toward a major shift in both restructured and regulated states toward dynamic pricing programs. This shift may well change the architecture of demand response in the country. But before this can take place, several barriers need to be overcome.
Scientific experimentation can play a major role in overcoming several of the barriers. While many experiments have been conducted in the US, Canada, the European Union and Australia, and while they provide significant insights into customer behavior, their results either collectively or individually may not be transferable to other regions that have conducted such experiments.
The paper summarizes the results of prior experiments and discusses their strengths and weaknesses. It also lays out ways in which future experiments can be conducted. The authors, Ahmad Faruqui, Ryan Hledik and Sanem Sergici are economists with The Brattle Group. Ahmad Faruqui led the research team that produced the FERC assessment cited in this article. They are grateful to Lamine Akaba and Jennifer Palmer for research assistance.
Bio: Ahmad Faruqui is a principal with The Brattle Group. He led a state-by-state assessment of the potential for demand response for the Federal Energy Regulatory Commission and is assisting FERC in the development of a national action plan. Last year, he performed a national assessment of the potential for energy efficiency for the Electric Power Research Institute and wrote a report on quantifying the benefits of dynamic pricing for the Edison Electric Institute. He has worked on fostering economic demand response for the Midwest ISO and ISO New England and on load management standards for the California Energy Commission. Since the year 2000, he has been assisting utilities and commissions throughout the US and Canada assess the economics of dynamic pricing, demand response and advanced metering. This has often involved the design and evaluation of innovative pilot programs. Early in his career, he wrote an evaluation of 14 experiments with time-of-use pricing which is cited in Professor Bonbright’s text on public utility rates. The author of four books and more than a hundred papers on energy policy, he holds a doctoral degree in economics from the University of California at Davis. He is based in Brattle’s San Francisco, California office and can be reached via email at ahmad.faruqui@brattle.com or by phone at (925) 408-0149.
- Thursday, December 3rd, 2009 - Commodity Models: from Ags to Zinc
Alexander Eydeland, Morgan Stanley- Title: Commodity Models: From Ags to Zinc
Abstract: We will discuss various issues and challenges facing commodity quants and suggest a number of modeling methodologies designed to address these issues. We will also give a brief introduction of standard commodity structures as well as new products and recent developments in commodity markets.
Bio: Dr. Alexander Eydeland is Managing Director at Morgan Stanley in charge of global commodity strategists. His previous positions include Head of Research at Mirant Corp., vice president with Lehman Brothers and Fuji Capital Markets, and associate professor of mathematics at the University of Massachusetts. Eydeland holds a Ph.D. degree in Mathematics from Courant Institute of Mathematical Sciences. His papers on risk management, scientific computing, optimization and mathematical economics have appeared in a number of major publications and he has lectured extensively on these subjects throughout the United States, Europe, and Japan. Eydeland is a co-author (with K. Wolyniec) of thebook "Energy and Power Risk Management" published in 2002 by Wiley and Co.
- Tuesday, January 12th, 2010 - Conic Integer Programming
Alper Atamturk, University of California at Berkeley- Title: Conic Integer Programming
Abstract: In the last two decades we have experienced significant advances in conic programming. Polynomial interior point algorithms that have earlier been developed for linear programming have been extended to conic optimization problems such as convex quadratically constrained quadratic programs and semidefinite programs. Availability of efficient algorithms for conic programming spurred many optimization and control applications in diverse areas ranging from medical imaging to statistical learning, from finance to truss design. However, the advances in conic programming and linear integer programming have until recently not translated to major improvements in conic integer programming, i.e., conic optimization problems with integer variables. In this talk we will review the recent progress in computational methods for conic integer programming. We will discuss lifting methods, cuts, and conic reformulations for improving computations for general as well as special structured problems. We will present application of conic integer programming to logistic network design with risk pooling, portfolio optimization, value-at-risk minimization, scheduling with controllable process times.
Bio: Alper Atamturk is a Professor of Industrial Engineering and Operations Research at the University of California, Berkeley. He received his Ph.D. from the Georgia Institute of Technology in 1998 with a major in OperationsResearch and minor in Computer Science. His current research interests are in optimization, integer programming, optimization under uncertainty with applications to renewable energy, finance and operations interface, cancer therapy, and defense. He serves on the editorial boards of Discrete Optimization, Operations Research, and Networks. Previously, he was on the editorial board of Management Science. He served on the organizing committees of IPCO 2010, MIP 2009, INFORMS 2008, MIP 2005, among others. He is vice chair-integer programming of INFORMS Optimization Society. Dr. Atamturk is recently nominated to become a DoD national security fellow for his basic researchon networks and risk.
- Tuesday, January 19th, 2010 - Collaborative Supply Chain Games
Xin Chen, University of Illinois at Urbana-Champaign- Title: Collaborative Supply Chain Games
Abstract: As companies start exploring innovative collaboration strategies in effort to improve their supply chain efficiency, getting all players to agree on how to share costs and benefits is identified as one of the major barriers to collaborative commerce. To gain basic understanding of cost allocation in supply chain cooperation, we analyze several fundamental collaborative supply chain games: the single period inventory game with stochastic demand and quantity discount, the economic lot sizing game and the capacity investment game, by employing cooperative game theory. In such games, a number of retailers may form joint ventures to build production facilities, make joint production (or place joint orders) and/or keep centralized inventories by taking advantage of economies of scale and/or risk pooling effects. The analysis of cost allocation in these games gives arise to a variety of challenging non-convex minimization problems, for which we develop interesting optimization techniques by exploiting problem structures.
Bio: Xin Chen is an assistant professor at the University of Illinois at Urbana-Champaign. He obtained his PhD from MIT in 2003, MS from Chinese Academy of Sciences in 1998 and BS from Xiangtan University in 1995. His research interest lies in optimization and supply chain management. He received the Informs revenue management and pricing section prize in 2009. He is the coauthor of the book “The Logic of Logistics: Theory, Algorithms, and Applications for Logistics and Supply Chain Management (Second Edition).
- Friday, January 22nd, 2010 (This seminar will take place at 10:00 AM in Tech M228) - Use of Sub-sample Estimates to Reduce Errors in Optimization Models
John Birge, University of Chicago- Title: Use of Sub-sample Estimates to Reduce Errors in Optimization Models
Abstract:&Optimization software enables the solution of problems with millions of variables and associated parameters. These parameters are, however, often uncertain and represented with an analytical description of the parameter's distribution or with some form of sample. With large numbers of such parameters, optimization of the resulting model is often driven by mis-specifications or extreme sample characteristics, resulting in solutions that are far from a true optimum. This paper describes how asymptotic convergence results may not be useful in large-scale problems and how the optimization of problems based on sub-sample estimates may achieve improved results over models using full-sample estimates. A motivating example and numerical results from a portfolio optimization problem demonstrate the potential improvement. A theoretical analysis also provides insight into the structure of problems where sub-sample optimization may be most beneficial.
Bio: John R. Birge studies mathematical modeling of systems under uncertainty, especially for maximizing operational and financial goals using the methodologies of stochastic programming and large-scale optimization. He was first drawn to this area by a need to use mathematics in a useful and practical way. "My research has shown how special problem structure can allow for efficient solution of complex problems of decision making under uncertainty," Birge explains. This research has been supported by the National Science Foundation, the Ford Motor Company, General Motors Corporation, the National Institute of Justice, the Office of Naval Research, the Electric Power Research Institute, and Volkswagen of America. He has published widely and is the recipient of the Best Paper Award from the Japan Society for Industrial and Applied Mathematics, the Institute for Operations Research and the Management Sciences Fellows Award, and the Institute of Industrial Engineers Medallion Award. A former dean of the Robert R. McCormick School of Engineering and Applied Sciences at Northwestern University, he has worked as a consultant for a variety of firms including the University of Michigan Hospitals, Deutsche Bank, Allstate Insurance Company, and Morgan Stanley, and he uses cases from these experiences in his teaching. Birge earned a bachelor's degree in mathematics from Princeton University in 1977 and a master's degree and a PhD in operations research from Stanford University in 1979 and 1980, respectively. He joined the Chicago Booth faculty in 2004. He is a member of the Institute for Operations Research and the Management Sciences, the Mathematical Programming Society, the Mathematical Association of America, and Sigma Xi. He also speaks French, Russian, German, and English.
- Tuesday, January 26th, 2010 - Integrated Supply Chain Design and Management, Practical Mechanism Design
Max Shen, University of California at Berkeley- Title: Supply Disruptions and the Reverse Bullwhip Effect
Abstract: We postulate the existence of a “reverse bullwhip effect” (RBWE) that occurs during and immediately after supply disruptions. Whereas the classical bullwhip effect (BWE) describes an increase in demand/order volatility as one moves upstream in the supply chain, the RBWE describes the opposite. We motivate our analysis by an example involving gasoline-buying patterns following Hurricane Katrina in 2005. We then present theoretical and empirical evidence for a RBWE that occurs in a variety of situations involving supply uncertainty. RBWE has great implication in service system design and management, so we examine causes of RBWE, discuss its impact, and suggest strategies for mitigating it.
Bio: Zuo-Jun (Max) Shen received his Ph.D. from Northwestern University. He has been active in the following research areas: integrated supply chain design and management, market mechanism design, applied optimization, and decision making with limited information. He is currently on the editorial/advisory board for several leading journals. He received the CAREER award from National Science Foundation in 2003.- Tuesday, February 2nd, 2010 - Corner Polyhedra and Maximal Lattice-free Convex Sets: A Geometric Approach to Cutting Planes
Amitabh Basu, Carnegie Mellon University- Title: Corner Polyhedra and Maximal Lattice-free Convex Sets: A Geometric Approach to Cutting Planes
Abstract:Corner Polyhedra were introduced by Gomory in the early 60s and were further developed by Gomory and Johnson. The importance of the corner polyhedron is underscored by the fact that almost all ``generic'' cutting planes, both in the theoretical literature as well as ones used in practice, are valid for the corner polyhedron. Thus, the corner polyhedron can be viewed as a unifying structure from which many of the known cutting planes can be derived. Moreover, the corner polyhedron has recently attracted a lot of attention as a potential source of new kinds of cutting planes. It has also been observed that valid inequalities for the corner polyhedron have an intimate connection with maximal lattice-free convex sets. This connection provides a geometric way of understanding and analyzing these valid inequalities. Such an approach often yields new insights into properties of existing cutting planes, as well as provides a novel way of analyzing new families of cutting planes. This talk surveys results from my thesis which contribute to this line of research. The major focus will be on the so-called "multiple-constraint" cutting planes, which are being studied extensively at present by the Integer Programming community. The ultimate aim is to achieve the next leap in improving Integer Programming algorithms.
Bio:Amitabh is a PhD candidate at Carnegie Mellon University (expected graduation May 2010). His thesis work is on cutting planes for general mixed integer linear programs, advised by Gerard Cornuejols. Prior to CMU, Amitabh completed a Masters in Computer Science from Stony Brook University and he holds a Bachelor's degree in Computer Science from the Indian Institute of Technology, Delhi.
- Tuesday, February 9th, 2010 - Stochastic Combinatorial Optimization: Interdicting Diseases and Smugglers
Nedialko B. Dimitrov, University of Texas at Austin- Title: Stochastic Combinatorial Optimization: Interdicting Diseases and Smugglers
Abstract:Stochastic Combinatorial Optimization (SCO) has a tremendous variety of applications. In this talk, we focus on two application areas: infectious disease control and nuclear smuggle interdiction. We begin with a brief introduction to the mathematical modeling of infectious diseases, describing models ranging in complexity from simple compartmental models to complex agent-based simulations. We continue with novel uses of SCO in determining effective disease control strategies, both through simulation-based optimization and a Markov decision process model. In the second part of the talk, we present a brief introduction to mathematical modeling for nuclear smuggler interdiction. Modeling smuggler behavior is difficult as real nuclear smuggling events are rare. Thus, we compare and contrast a family of models of smuggler behavior and discuss the development of novel algorithms to compute interdiction strategies. Our discussion allows us to explore some modeling connections between the two areas. In addition, we illustrate how both areas give rise to abstractions in SCO that are amenable to theoretical analysis.- Tuesday, February 16th, 2010 - Alter Egos in Non-convex Optimization: The Curious Case of MIQCP
Anureet Saxena, Carnegie Mellon University- Title: Alter Egos in Non-convex Optimization: The Curious Case of MIQCP
Abstract:This talk concerns convex relaxations of mixed integer non-convex quadratically constrained problems (MIQCP). The extended formulation of MIQCP employs the system of non-linear equations Y=xxT which can be relaxed as a pair of SDP inequalities: Y-xxT: PSD and xxT-Y:PSD. While researchers have concentrated on exploring the strengths and weaknesses of the convex constraint Y-xxT: PSD, a detailed investigation of its non-convex alter ego had remained an unchartered territory. In this talk, we bring together ideas from disjunctive programming, spectral theory and lift-and-project methodology to generate strong valid cutting planes from the latter non-convex condition. These cutting planes are derived using a dynamic reformulation scheme that rotates the coordinate axes so as to amplify the hidden infeasibilities of the incumbent solution. As a byproduct of our research, we prove a generalization of the sequential convexification theorem of Balas for the case of MIQCP. Computational experiments conducted on 150 instances in the GlobalLib repository show that these cutting planes close around 80% of the duality gap on average; the SDP relaxations, on the other hand, are able to close only 25% of the duality gap on these instances. We discuss systematic techniques for projecting these extended formulations to the space of x variables, and demonstrate their practical utility through a series of computational experiments.
- Tuesday, February 23rd, 2010 - Adjustable Ordering Policies in Multi-period, Multi-echelon Robust Inventory Management
Dan Iancu, Massachusetts Institute of Technology- Title: Adjustable Ordering Policies in Multi-period, Multi-echelon Robust Inventory Management
Abstract:In this talk, we discuss the performance of a new class of inventory replenishment policies, in the context of robust models for multi-period, multi-echelon supply chains. For the case of a single echelon, using geometric ideas from polyhedral theory, we prove the optimality of ordering policies that are affine in the history of demands. Our approach leads to policies that can be computed by solving a single linear optimization problem, and which bear an interesting interpretation in terms of partial demand satisfaction. The result also underscores a key distinction between robust and stochastic models for dynamic optimization, with the former resulting in qualitatively simpler problems than the latter. For the case of a general supply chain network, we provide a hierarchy of polynomial policies that are also directly parameterized in observed demands, and that can be efficiently computed using semi-definite optimization methods. Empirical results for several supply chain configurations are very encouraging, suggesting that policies based on polynomials of degree two or three are near-optimal. This is joint work with Dimitris Bertsimas and Pablo Parrilo.
- Thursday, February 25th, 2010 (Please Note: This seminar is scheduled for a Thursday) - A Multifrequency Theory of the Interest Rate Term Structure
Liuren Wu, City University of New York- Title: A Multifrequency Theory of the Interest Rate Term Structure
Abstract: We develop a class of no-arbitrage dynamic term structure models that are extremely parsimonious. The model employs a cascade structure to provide a natural ranking of the factors in terms of their frequencies, with merely five parameters to describe the interest rate time series and term structure behavior regardless of the dimension of the state vector. The dimension-invariance feature allows us to estimate low and high-dimensional models with equal ease and accuracy. With 15 LIBOR and swap rate series, we estimate 15 models with the dimension going from one to 15. The extensive estimation exercise shows that the 15-factor model significantly outperforms the other lower-dimensional specifications. The model generates mean absolute pricing errors less than one basis point, and overcomes several known limitation of traditional low-dimensional specifications. Authors: Laurent E. Calvet, Adlai J. Fisher, and Liuren Wu- Bio: Liuren is a professor of economics and finance at Zicklin School of Business, Baruch College, City University of New York. Before he joined Zicklin in 2003, he was an assistant professor at Fordham University. Liuren's major research interests include option pricing, term structure modeling, credit risk, market microstructure, and general asset pricing. During the past ten years, Liuren has published over 30 articles, many of them in top finance journals such as the Journal of Finance, the Journal of Financial Economics, Review of Financial Studies, the Journal of Financial and Quantitative Analysis, Management Science, and Journal of Monetary Economics. Liuren has worked extensively as consultants in the finance industry, including Bloomberg, Morgan Stanley, Royal Bank of Canada, and several fixed income and equity hedge funds. As a consultant, he has developed statistical arbitrage strategies, risk management procedures, and quantitative models for pricing fixed income and equity derivative securities.
- Tuesday, March 2nd, 2010 – Cost-Sharing Contracts for Coordinating Semi-Centralized Global Production Networks
Jeanette Song, Fuqua School of Business, Duke University- Title:Cost-Sharing Contracts for Coordinating Semi-Centralized Global Production Networks
Abstract: We consider a semi-centralized global production network that is common to many global companies. The network comprises a home plant and a foreign branch. The foreign branch has the autonomy to make independent decisions on its local production and logistics. The home plant is the sole supplier of a key component, and is required (by the headquarters) to provide 100% service level to the foreign branch. Because of complex international logistics, the branch employs a long replenishment interval, causing high expediting costs at the home plant. We analyze two cost-sharing contracts between the two parties to reduce this inefficiency. The first contract assumes symmetric information on the branch's fixed order cost, while the second assumes asymmetric information. We show that optimal contracts exist under either information structure and can lead to significant improvement in supply chain performance. The division of the benefits between the two locations, however, depends highly on the cost structure of the home plant. These results can also provide guidelines for the headquarters to determine the degree of centralization in global production networks.
Bio: Professor Jeannette Song is a Professor and Area Coordinator in Operations Management of the Fuqua School of Business, Duke University, USA. Professor Song’s main research interests lie in the areas of supply chain management and operations management. Topics include supply chain logistics and coordination, supply-chain and outsourcing structures, multi-sourcing, impact of e-commerce, Vendor Managed Inventory programs, forecasting and inventory planning, Assemble-to-Order systems, delivery time quotation, and reverse logistics. She has published over 40 articles in leading academic journals such as Management Science and Operations Research. Professor Song teaches Operations Management, Supply Chain Management, Global Operations, and Global Academic Travel Experience in China in the MBA and Executive MBA programs. She also teaches Supply Chain Models, Optimization, and Stochastic Orders at the Ph.D. Level. Professor Song is the recipient of several research grants from the U.S. National Science Foundation. She is the current President of the Manufacturing and Service Operations Management Society of INFORMS and the Area Editor for Operations Research in the Manufacturing, Service and Supply Chain Operations area. Professor Song received a Ph.D. in Management Science & Operations Management from Columbia University, a M.S. degree in Operations Research from the Chinese Academy of Sciences, and a B.S. degree in Mathematics from Beijing Normal University. Before joining Duke University, she served on the faculties of University California, Irvine and Columbia University, and held a visiting position at the University of California, Berkeley.- Thursday, March 4th, 2010 - Option Market-Making, Pricing, and Risk Management
Sheldon Natenberg, Chicago Trading Company- Title: Option Market-Making, Pricing, and Risk Management
Abstract:The presentation will focus on the trading of options from the perspective of a professional market-making firm, including a discussion of how professional traders use theoretical pricing models to price options and manage risk, and the practical problems of adapting theory to the real world of the marketplace. (Copies of the PowerPoint presentation will be provided at the lecture)
Bio:Mr. Natenberg began his trading career in 1982 as an independent market maker in equity options at the Chicago Board Options Exchange. From 1985 to 2000 he traded commodity options, also as an independent floor trader, at the Chicago Board of Trade. Mr. Natenberg has also conducted seminars for option traders at major exchanges and professional trading firms in the United States, Europe, and the Far East. In 2000, he took over full-time management of education at Chicago Trading Company, a proprietary derivatives trading firm. Mr. Natenberg is the author of Option Volatility and Pricing: Advanced Trading Strategies and Techniques. The book addresses many of the special problems of the professional trader, with particular emphasis on the practical aspects of option evaluation and risk management.- Tuesday, March 9th, 2010 - Providing a Scientific Basis for Managing Illegal Drug Problems
Jonathan Caulkins, Carnegie Mellon University- Title: Providing a Scientific Basis for Managing Illegal Drug Problems
Abstract:Illegal drugs pose serious problems that vex policy makers throughout the world. The tools of operations research, industrial organizations, and economics can be harnessed to provide an empirical, scientific basis for drug policy making. Data are drawn from epidemiological studies, forensic laboratory analysis, undercover buys, and extensive interviews with incarcerated drug smugglers and dealers. This talk focuses on drug initiation (product diffusion), price responsiveness (elasticity of demand), and operation of the illegal supply chain, both during normal times and when the distribution network is disrupted. Resulting understanding provides the foundation for estimating the cost-effectiveness of different broad strategies for controlling drug use and associated social harms.
Time permitting the talk will also draw on observations from recent work advising the U.S. government on how to design counter-narcotics strategy in Afghanistan to support counter-insurgency efforts.- Bio:Jonathan P. Caulkins, Ph.D., is Professor of Operations Research and Public Policy at Carnegie Mellon University’s Heinz College and Qatar campus. Dr. Caulkins specializes in mathematical modeling and systems analysis with a particular focus on social policy systems pertaining to drugs, crime, terror, violence, and prevention – work that won the David Kershaw Award from the Association of Public Policy Analysis and Management. Other interests include software quality, optimal control, black markets, airline operations, and personnel performance evaluation. He has taught his quantitative decision making course on four continents to students from 49 countries at every level from undergraduate through Ph.D. and exec ed.
- Dr. Caulkins has published 7 books and monographs and over 85 journal articles in Operations Research, Management Science, JASA, Automatica, Decision Support Systems, JPAM, The American Journal of Public Health, Mathematical Biosciences, IEEE Security & Privacy, The Journal of Environmental Economics and Management, The Journal of Economic Dynamics and Control, and the Journal of Optimization Theory and Applications, among other outlets. At RAND he has been a consultant, visiting scientist, co-director of RAND’s Drug Policy Research Center (1994 – 1996), and founding director of RAND’s Pittsburgh office (1999-2001).
- Dr. Caulkins received a B.S., and M.S. in Systems Science from Washington University, an S.M. in Electrical Engineering and Computer Science and Ph.D., in Operations Research both from M.I.T.
- Tuesday, March 30th, 2010 - Modeling and Simulating Non-stationary, Non-poisson Arrival Processes
Barry L. Nelson, Northwestern University- Title: Providing a Scientific Basis for Managing Illegal Drug Problems
Abstract:Simulation models of real-life systems often assume stationary (homogeneous) Poisson arrivals. Therefore, when nonstationary arrival processes are required it is natural to assume Poisson arrivals with a time-varying arrival rate. For many systems, however, this provides an inaccurate representation of the arrival process which is either more or less variable than Poisson, and may exhibit dependence. We extend techniques that transform a stationary Poisson arrival process into a nonstationary Poisson arrival process by transforming a general stationary arrival process into a nonstationary, non-Poisson (NSNP) arrival process. We show that the desired arrival rate is achieved, and that certain variability and dependence properties of the base process are passed on to the transformed process. We also provide techniques for specifying the base process when presented with characteristics of, or data from, an arrival process and illustrate them by modeling e-mail arrival data. - Thursday, December 3rd, 2009 - Commodity Models: from Ags to Zinc
