2010 Seminars

Ben Lev

Department Head of the Department of Decision Sciences
Lebow College of Business, Drexel University

December 5, 2011

3:00pm - 4:00pm

Mohler Lab Room 451

ABSTRACT

Prof. Lev has been the Editor in Chief of OMEGA since 2002. Over that period there were 5,000 submission and 600 papers were published. In 2011 OMEGA Impact Factor is 3.467 and 5-year Impact Factor is 3.733. Acceptance rate in 2011 is about 10-15%. During his presentation Prof. Lev will discuss the reasons 4,400 were rejected and the characteristics of those 600 papers which were accepted for publication. Specific attention to Title, Abstract, Key Words, Paper, Examples, Conclusions, References, Appendices, Total length, and cover letter, Revised paper.

BIOGRAPHY

Professor Lev is the Head, Decision Sciences Department at Drexel University. Prior appointments include Dean of the School of Management, University of Michigan - Dearborn; Head, Management Department, Worcester Polytechnic Institute; Chairman, Management Department, Temple University. His Ph.D. degree is from Case Western Reserve University. Professor Lev published five books all by North Holland Publishing Company. His publications appeared in Journal of Operations Research, Journal of Management Science, Naval Research Logistic Quarterly, Omega The International Journal of Management Science, European Journal of Operations Research, Operational Research Quarterly, Interfaces, Computers and Operations Research, The Institute of Industrial Engineering – Transactions, The International Journal of Management Science and Engineering Management, and International Journal of Management. He was a guest editor of a special issue for European Journal of Operations Research. He is the Editor-in-Chief of OMEGA - The International Journal of Management Science; Co-Editor-in-Chief of The International Journal of Management Science and Engineering Management; and is/was an associate editor of: International Abstracts of Operations Research; Interfaces; Employee Responsibilities and Rights Journal; Institute of Industrial Engineering-Transactions; and Journal of Operations Research. Professor Lev served as INFORMS first Vice President-Meetings; Council member of TIMS; TIMS Vice President-Meetings. He organized these conferences: ORSA/TIMS 1990; INFORMS International 1998; IFORS 2002; INFORMS International Co - Chair, 2010.

Michael Fu

Title: Stochastic Gradient Estimation: Tutorial Review and Recent Research

Program Director for Operations Research, National Science Foundation (NSF)
Ralph J. Tyser Professor of Management Science
Department of Decision, Operations, and Information Technologies, University of Maryland, College Park

November 28, 2011

3:00pm - 4:00pm

Mohler Lab Room 453

ABSTRACT

Stochastic gradient estimation techniques are methodologies for deriving computationally efficient estimators used in simulation optimization and sensitivity analysis of complex stochastic systems that require simulation to estimate their performance. Using a simple illustrative example, the three most well-known direct techniques that lead to unbiased estimators are presented: perturbation analysis, the likelihood ratio (score function) method, and weak derivatives. Applications are discussed and then some recent research results in financial engineering and revenue management are presented.

Opportunities for NSF funding in the Operations Research Program will be discussed in the second half of the talk.

BIOGRAPHY

Michael Fu is Director of the Operations Research Program at NSF. He is on leave from the University of Maryland at College Park where he is Ralph J. Tyser Professor of Management Science in the Robert H. Smith School of Business, with a joint appointment in the Institute for Systems Research and affiliate faculty appointment in the Department of Electrical and Computer Engineering, both in the A. James Clark School of Engineering. He received degrees in mathematics and EE/CS from MIT in 1985, and a Ph.D. in applied mathematics from Harvard University in 1989. His research interests include simulation optimization and applied probability, with applications in supply chain management and financial engineering. At Maryland, he received the Business School's Allen J. Krowe Award for Teaching Excellence in 1995, the Institute for Systems Research Outstanding Systems Engineering Faculty Award in 2002, and was named a University of Maryland Distinguished Scholar-Teacher for 2004-2005. He has published four books: Conditional Monte Carlo: Gradient Estimation and Optimization Applications (1997, co-author J.Q. Hu), which received the INFORMS Simulation Society Outstanding Publication Award in 1998; Simulation-based Algorithms for Markov Decision Processes (2007, co-authors H.S. Chang, J. Hu, S.I. Marcus); Perspectives in Operations Research (2006, co-editors F.B. Alt, B.L. Golden); and Advances in Mathematical Finance (2007, co-editors R.A. Jarrow, J.-Y. Yen, R.J. Elliott). He served as Stochastic Models and Simulation Department Editor of Management Science from 2006-2008, as Simulation Area Editor of Operations Research 2000-2005, and also on the editorial boards of Mathematics of Operations Research, INFORMS Journal on Computing, IIE Transactions, and Production and Operations Management. He is a Fellow of INFORMS and IEEE.

Kees Roos

Title: Local self-concordance, a new paradigm for optimization?

Former Chair of Optimization Technology, Delft University of Technology

November 22, 2011

2:30pm - 3:30pm

Mohler Lab Room 453

ABSTRACT

We consider full-Newton-step interior-point methods (FNS-IPMs) for Linear Optimization based on kernel-function based barrier functions that are not self-concordant (SC). IPMs use the so-called central path of a given problem as a guideline to its set of optimal solutions. The iterates closely follow this central path.

We first provide a short introduction to FNS-IPMs based on SC barrier functions. The definition of a SC barrier function requires that two parameters, called $\kappa$ and $\nu$, are uniformly bounded on its domain. We call these the Nesterov-Nemirovski parameters of the barrier function. We recall the definitions of these parameters and discuss the main ideas underlying the complexity analysis of a FNS-IPMs based on a SC barrier function. The iteration bound turns out to depend on the product $\kappa\sqrt{\nu}$, which is called the complexity number of the barrier function.

The more recently introduced IPMs based on kernel-function based barrier functions have the advantage that for so-called large-update IPMs they admit better iteration bounds than methods based on SC barrier functions. For FNS-IPMs, however, we observed that the iteration bounds are essentially the same. Eventually we hope to nd an explanation for this observation.

As we show, in general kernel-function based barrier functions are not SC, since the parameters $\kappa$ and $\nu$ are not uniformly bounded. On the other hand, in the region where the iterates of a FNS-IPM occur, namely in a small neighborhood of the central path, these parameters are nicely bounded, and the complexity number may be even smaller than for SC barrier functions. We therefor call them locally SC. This phenomenon in itself is of interest. It is still an open question if it can be used to derive a good iteration bound for the new FNS-IPMs.

BIOGRAPHY

Kees Roos (1941) held the chair for Optimization Technology at Delft University of Technology until 2006, when he retired. From 1998 to 2002 he was a part-time professor at Leiden University. The past 25 years his research concentrated on interior-point methods for linear and convex optimization. He is a (co-)author of several books and more than hundred papers in refereed journals. He is (or was) member of the editorial board of several journals, among them the SIAM Journal on Optimization. He was secretary/treasurer of the SIAM Activity Group on Optimization. He supervised a large number of research projects, among them the Dutch nationwide NWO-project High Performance Methods for Mathematical Optimization, and the project Discrete Mathematics and Optimization of the (Dutch) Stieltjes Institute. He was also involved in a big project "Optimal safety of the dikes in the Netherlands", financed by the Dutch government. Earlier this month he shared the 2011 Khachiyan prize of the INFORMS Optimization Society with Jean-Philippe Vial.

Coralia Cartis

Title: Optimal Newton-type methods for nonconvex smooth optimization

Assistant professor at Edinburgh University in Scotland, United Kingdom

October 27, 2011

2:30pm - 3:30pm

Mohler Lab Room 453

ABSTRACT

We show that the steepest-descent and Newton's methods for unconstrained nonconvex optimization under standard assumptions may both require a number of iterations and function evaluations arbitrarily close to the steepest-descent's global worst-case complexity bound. This shows that the latter upper bound is essentially tight for steepest descent and that Newton's method may be as slow as the steepest-descent method in the worst case. Then the cubic regularization of Newton's method (Griewank (1981), Nesterov & Polyak (2006)) is considered and extended to large-scale problems, while preserving the same order of its improved worst-case complexity (by comparison to that of steepest-descent); this improved worst-case bound is also shown to be essentially tight. We further show that the cubic regularization approach is, in fact, optimal from a worst-case complexity point of view amongst a class of second-order methods. The worst-case problem-evaluation complexity of constrained optimization will also be discussed, time permitting. This is joint work with Nick Gould (Rutherford Appleton Laboratory, UK) and Philippe Toint (Namur University, Belgium).

BIOGRAPHY

Coralia Cartis has been a tenured assistant professor at Edinburgh University in Scotland, United Kingdom since 2007. Previously, she held postdoctoral appointments within the numerical analysis groups at Rutherford Appleton Laboratory and Oxford University. She pursued her PhD research in optimization under the supervision of Prof Mike Powell at Cambridge University (2005).

Coralia's research addresses the development, convergence and complexity analyses and implementation of algorithms for linear and nonlinear nonconvex smooth optimization problems, suitable for large-scale problems. She is also interested in the interconnections between dynamical systems and continuous optimization; and optimization aspects of compressed sensing and sparse approximation.

Timothy Chan

Assistant Professor in the department of Mechanical and Industrial Engineering at the University of Toronto

October 17, 2011

3:00pm - 4:00pm

Mohler Lab Room 453

ABSTRACT

The traditional approach to robust intensity-modulated radiation therapy treatment planning involves creating an appropriate uncertainty set to model the uncertain effect, solving a single treatment planning problem, and then delivering the same treatment over multiple weeks. In this talk, I will present an adaptive robust optimization approach to IMRT optimization, where information gathered in previous treatment sessions is used to update a model of uncertainty and guide treatment plan re-optimization for the next session. Such an approach allows for the estimate of the uncertain effect to improve as the treatment progresses. This approach involves solving a sequence of linear programs, and is therefore highly tractable. I will present computational results for a lung cancer case where the dominant uncertainty is in the patient’s breathing motion. Using this adaptive robust method, I demonstrate that it is possible to attain significant and simultaneous improvement in both tumor coverage and organ sparing over the non-adaptive approach. I also show that it is possible to closely approximate “prescient” solutions, and provide some theoretical insight as to why this occurs.

BIOGRAPHY

Timothy C. Y. Chan is an Assistant Professor in the department of Mechanical and Industrial Engineering at the University of Toronto. His primary research interests are in optimization under uncertainty and the application of optimization methods to radiation therapy, health care operations and sustainability. He received his B.Sc. in Applied Mathematics from the University of British Columbia, and his Ph.D. in Operations Research from the Massachusetts Institute of Technology. Before coming to Toronto, he was an Associate in the Chicago office of McKinsey and Company, a global management consulting firm. During that time, he advised leading companies in the fields of medical device technology, travel and hospitality, telecommunications, and energy on issues of strategy, organization, technology and operations.

Bahar Biller

Title: Improved Inventory Targets in the Presence of Limited Historical Demand Data

Associate Professor in the Tepper School of Business at Carnegie Mellon University

October 7, 2011

11:00 am - 12:00pm

Mohler Lab Room 453

ABSTRACT

Most of the literature on inventory management assumes that the demand distribution and the values of its parameters are known with certainty. We consider a repeated newsvendor setting where this is not the case and study the problem of setting inventory targets when there is a limited amount of historical demand data. Consequently, we achieve the following objectives: (1) to quantify the inaccuracy in the inventory-target estimation as a function of the length of the historical demand data, the critical fractile, and the shape parameters of the demand distribution; and (2) to determine the inventory target that minimizes the expected cost and accounts for the uncertainty around the demand parameters estimated from limited historical data. We achieve these objectives by using the concept of expected total operating cost and representing the demand distribution with the highly flexible Johnson translation system. We further consider the demand data that may be auto-correlated or intermittent as well as the data sets that may contain sales rather than demand realizations. Our procedures require no restrictive assumptions about the first four moments of the demand random variables, and they can be easily implemented in practical settings with reduced expected total operating costs.

BIOGRAPHY

Bahar Biller is an Associate Professor in the Tepper School of Business at Carnegie Mellon University. Her research focuses on three distinct yet related areas: (1) developing a comprehensive input modeling framework for stochastic system simulations with the ability to represent, fit, and generate multivariate time-series input processes with arbitrary marginal distributions and dependence structures; (2) accounting for the uncertainty of multivariate input-distribution parameters which are estimated from finite historical input data on simulation outputs; and (3) testing and proving the efficacy of the developed methods on a wide range of industrial applications in Operations Management and Finance. She received a National Science Foundation CAREER award in 2006 and the Presidential Early Career Award for Scientists and Engineers in 2007. She is a member of INFORMS and currently the Vice President and President-elect of the INFORMS Simulation Society.

Mikell P. Groover, Ph.D.

Title: Study Mission to Egypt

Professor Emeritus of Industrial and Systems Engineering, Lehigh University

Friday, April 8, 2011

2:30 p.m. - 3:30 p.m.

Mohler Lab Room 451

ABSTRACT

This presentation is based on a two-week vacation trip that Professor Mikell Groover and his wife took to Egypt in November, 2010. Our tour began in Cairo, and proceeded to Luxor (the ancient city of Thebes), where we boarded a river boat and sailed up the Nile River to the Aswan Dam. At Aswan we boarded another boat and sailed across Lake Nasser. Along the way, we saw the great pyramids of Giza (near Cairo), the Valley of the Kings (near Luxor), and many ancient Egyptian temples. The presentation will include many of the pictures we took and provide an overview of Egypt and the adventures we experienced there.

BIOGRAPHY

Mikell Groover is professor emeritus of Industrial and Systems Engineering at Lehigh University. He has the following degrees, all from Lehigh: BA in Arts & Science, BS in mechanical engineering, MS and PhD in industrial engineering. His teaching and research interests include manufacturing processes, automation, facilities planning, logistics, and work systems. During 44 years as a faculty member, his research has been sponsored by National Science Foundation, Advanced Research Development Agency, U.S. Department of Housing and Urban Development, and several companies. His publications include approximately 75 technical papers and 11 textbooks (including two co-authored). His 12th book is due to be released in September 2011.

Dr. Luis Nunes Vicente

Title: Direct Multisearch for Multiobjective Optimization

Professor of Mathematics, University of Coimbra, Portugal

Wednesday, February 2, 2011

2:30 p.m. - 3:30 p.m.

Mohler Lab Room 453

ABSTRACT

In practical applications of optimization, it is common to have several conflicting objective functions to optimize. Frequently, these functions are subject to noise or can be of black-box type, preventing the use of derivative-based techniques. We propose a novel multiobjective derivative-free methodology, calling it direct multisearch (DMS), which does not aggregate any of the objective functions. Our framework is inspired by the search/poll paradigm of direct-search methods of directional type and uses the concept of Pareto dominance to maintain a list of nondominated points (from which the new iterates or poll centers are chosen). The aim of our method is to generate as many points in the Pareto front as possible from the polling procedure itself, while keeping the whole framework general enough to accommodate other disseminating strategies, in particular when using the (here also) optional search step. DMS generalizes to multiobjective optimization (MOO) all direct-search methods of directional type.

We prove under the common assumptions used in direct search for single optimization that at least one limit point of the sequence of iterates generated by DMS lies in (a stationary form of) the Pareto front. However, extensive computational experience has shown that our methodology has an impressive capability of generating the whole Pareto front, even without using a search step.

This is joint work with A.L. Custódio (New Univ. Lisbon), J.F.A. Madeira (Polytechnic Inst. Lisbon), and A.I.F. Vaz (Univ. Minho).

BIOGRAPHY

Luis Nunes Vicente is a Professor of Mathematics at the University of Coimbra, Portugal. He obtained his PhD from Rice University, TX in 1996 under a Fulbright scholarship and was among the three finalists of the 94-96 A. W. Tucker Prize of the Mathematical Optimization Society. He held visiting positions at the IBM T.J. Watson Research Center and the IMA/University of Minnesota in 2002/2003 and at the Courant Institute of Mathematical Sciences/NYU and CERFACS in 2009/2010.

Luis’ research interests include the development and analysis of numerical methods for large-scale nonlinear programming, PDE constrained optimization problems, and derivative-free optimization problems, and applications in computational sciences, engineering, and finance. His research has been supported by the European Union, the European Space Agency, and the Portuguese Foundation for Science and Technology. He is a member of several editorial boards including SIAM Journal on Optimization and Journal of Global Optimization and he recently ended a six-year term as editor of the SIAM SIAG/Optimization Views-and-News.

L. N. Vicente is a co-author of the book Introduction to Derivative-Free Optimization, MPS-SIAM Series on Optimization, SIAM, Philadelphia, 2009. He participates in various boards and councils. He co-directs the UT Austin | Program in Mathematics (2007-2011) and was elected council member-at-large of the Mathematical Optimization Society (2009-2012).

Dr. Francisco Santos

Title: Counter-Examples to the Hirsch Conjecture

University of Cantabria, Spain

Wednesday, January 12, 2011

2:30 p.m. - 3:30 p.m.

Mohler Lab Room 453

ABSTRACT

The Hirsch conjecture, stated in 1957, said that if a polyhedron is defined by $n$ linear inequalities in $d$ variables then its combinatorial diameter should be at most $n-d$. That is, it should be possible to travel from any vertex to any other vertex in at most $n-d$ steps (traversing an edge at each step). The unbounded case was disproved by Klee and Walkup in 1967. In this talk, I describe my construction of the first counter-examples to the bounded case (polytopes).

The conjecture was posed and is relevant in connection to linear programming since the simplex method, one of the "mathematical algorithms with the greatest impact in science and engineering in the 20th century", solves linear programming problems by traversing the graph of the feasibility polyhedron.

BIOGRAPHY

Francisco Santos is a full professor at the University of Cantabria (Spain). His main research interest is the combinatorics of polytopes. In particular, he has recently published a monograph on Triangulations of Polytopes in the Springer series "Algorithms and Computation in Mathematics" (joint with J. A. De Loera and Jörg Rambau). He has published more than 50 research papers, in the Journal of the AMS (American Mathematical Society), Memoirs of the AMS, and Journal of the European Mathematical Society, among others. He was an invited speaker in the Combinatorics Section of the International Congress of Mathematicians in Madrid (2006). He has held visiting positions in the Ecole Normale Superieure de Paris, University of California Davis and MSRI (Mathematical Sciences Research Institute) Berkeley.