Speakers and Abstracts

Home Locations Program Speakers and Abstracts

Prof.Laurent El Ghaoui: "Managing energy in the digital age"

Abstract. Managing energy efficiently is one of the big challenges facing our societies. Data science technology offers a great promise in this area. In this talk I will review some applications I have explored in collaborative work with a large electricity provider, covering three distinct areas: a) “optimization-friendly” modeling of complex energy systems; b) agile management of electricity production; c) text analytics for predictive maintenance. These examples illustrate the power of an integrated view of machine learning and optimization for decision-making.


Bio. Laurent El Ghaoui is a Professor in the EECS and IEOR departments at UC Berkeley, and an expert in the field of optimization, particularly robust optimization, as well as in machine learning for large-scale text analytics.

Prof.Yueyue Fan: "Stochastic Estimation for Traffic Network Demand"

Abstract. In this talk, I will discuss a general network identification problem, its special applications in transportation, and how optimization approaches may be combined with statistics to address challenges brought by observability and stochasticity in this problem context. The general problem is defined as: how can one infer global network parameters (y) based on data measured on local parameters (x), with the relation between x and y built on complex network structure? A familiar example of such problem in transportation is origin-destination (O-D) demand estimation based on link traffic counts. With more traffic data becoming available through advanced information technologies, we face great opportunities as well as challenges in utilizing the data. In this talk, I will show how stochastic programming framework could be used to integrate heterogeneous data sources, which leads to better network identifiability and estimation quality.

Bio. Dr. Fan is a professor in Civil and Environmental Engineering at University of California, Davis. She is also a faculty advisor in the graduate programs in Applied Mathematics and Transportation Technologies and Policy. She received her PhD in Civil Engineering at University of Southern California in 2003. Dr. Fan’s research is on transportation and energy infrastructure systems modeling, with special interest in integrating applied mathematics and engineering domain knowledge to address challenges brought by system uncertainty, dynamics, and indeterminacy issues.

Prof.Michael Ferris: "Solution of equilibrium problems using extended mathematical programming"

Abstract. Equilibrium problems are widespread in the economics and physical literatures, and often involve interactions between competing players or physical systems. These include examples of generalized Nash Equilibria and Multiple Optimization Problems with Equilibrium Constraints. While Extended Mathematical Programming provides a convenient mechanism to describe these interactions, and to process them into models for solution by existing complementarity and optimization methods, they are often limited by an assumption of exclusive control or knowledge.

In many cases, shared constraints, linked multipliers, and shared implicit variables are necessary to model behavior within the equilibrium. We describe a framework that allows each of these features within competitive equilibrium problems and show how to exploit the structure in efficient computational schemes. Several large scale applications to electricity planning will be demonstrated that show the power of the framework and the speed of solution.


Bio. Michael C. Ferris is the Stephen C. Kleene Professor of Computer Sciences and leads the Optimization Group within the Wisconsin Institutes for Discovery at the University of Wisconsin, Madison, USA. He received his PhD from the University of Cambridge, England in 1989.

Dr. Ferris' research is concerned with algorithmic and interface development for large scale problems in mathematical programming, including links to the GAMS and AMPL modeling languages, and general purpose software such as PATH, NLPEC and EMP. He has worked on many applications of both optimization and complementarity, including cancer treatment planning, energy modeling, economic policy, traffic and environmental engineering, video-on-demand data delivery, structural and mechanical engineering.

Ferris is a SIAM fellow, an INFORMS fellow, received the Beale-Orchard-Hays prize from the Mathematical Programming Society and is a past recipient of a NSF Presidential Young Investigator Award, and a Guggenheim Fellowship. He serves on the editorial boards of Mathematical Programming, Transactions of Mathematical Software, and Optimization Methods and Software.

Prof.Alejandro Jofré: "Variance-based stochastic extragradient methods with linear search for stochastic variational inequalities

Abstract. We propose stochastic extragradient methods for stochastic variational inequalities with a linear search requiring only pseudo-monotonicity of the operator and no knowledge of the Lipschitz constant $L$. We provide convergence and complexity analysis, allowing for an unbounded feasible set, unbounded operator, non-uniform variance of the oracle and we do not require any regularization. We also prove the generated sequence is bounded in L$^p$. Alongside the stochastic approximation procedure, we iteratively reduce the variance of the stochastic error. Our methods cope with stepsizes bounded away from zero and attain the near-optimal oracle complexity $O(\log_{1/\theta}L)\cdot\epsilon^{-2}\cdot[\ln(\epsilon^{-1})]^{1+b}$ and an accelerated rate $O(1/K)$ in terms of the mean (quadratic) natural residual and the mean D-gap function, where $K$ is the number of iterations required for a given tolerance $\epsilon>0$ for arbitrary $\theta\in(0,1)$ and $b>0$. Explicit estimates for the convergence rate, oracle complexity and the $p$-moments are given depending on problem parameters and the distance of initial iterates to the solution set.


Bio. Alejandro Jofré is Director of the Center for Mathematical Modeling at University of Chile. He obtained the degree of Doctor and Habilitation on Applied Mathematics in France. He held appointments as professor at the Universities of Paris 1-Sorbonne and University of California-Davis, and visiting professor at several universities and centers, including Ecole Polytechnique, Princeton U., U. Washington-Seattle, UC Davis, UBC-Vancouver, HP Palo Alto, Bonn and NUS-Singapore.

Alejandro works on optimization, stochastic optimization, game theory, economic equilibrium and risk analysis, and electricity markets and mining. He has advised more than 20 PhDs and master theses. He has led more than twenty research projects and also developed optimization tools for pricing, planning and market behavior analysis for energy systems, risk analysis for network, telecommunication markets and sustainable exploitation of natural resources such as copper, energy and forestry. He has been plenary and invited speaker in several major international conferences such as International Conference on Stochastic Programming, Mathematical Programming, ICIAM and the Annual SIAM Conference in San Diego.

Alejandro Jofre was member of the Council of the Mathematical Optimization Society and member of the Scientific Board of three Research Centers/Initiatives of excellence in France, Japan and Ecuador. Furthermore, he has been associate editor of 8 mathematical and engineering journals, including Journal Optimization Theory and Application, Set-valued and Variational Analysis, Energy Systems, Optimization and Engineering, Mathematical and Financial Economics and Journal of Industrial and Management Optimization. Finally, Professor Jofre is member of SIAM, Mathematical Optimization Society, Econometric Society, IEEE Society and the Economic Theory Society.

Prof.Matthias Köppe: "Computer-assisted discovery of next-generation cutting planes for mixed integer programming"

Abstract. The revival of Gomory's classic (1960s) general purpose cutting planes that started in the mid-1990s enabled the development of powerful branch-and-cut software. Practitioners can nowadays solve many large-scale mixed integer optimization problems simply by competent modelization and running a black-box solver such as Gurobi, CPLEX, or SCIP. This is in contrast to an earlier paradigm, which called for strong (preferably, facet-defining) cutting planes that are specific to a class of problems, and which had led to the computational breakthroughs for hard combinatorial optimization problems such as the traveling salesman problem in the 1980s.

In the 2010s, challenging new applications and increased data sizes have prompted us to develop next-generation general-purpose cutting plane systems. We revisit another classic (early 1970s) idea, that of Gomory and Johnson's infinite-dimensional relaxations of integer programs. We propose to overcome the difficulties in studying the "facet-defining" valid inequalities of these relaxations by making use of computer-assisted discovery and proof techniques. The goal is to develop a large library of parametric families of such cutting planes as the foundation of a machine-learning-based cutting-plane system.

Bio. Matthias Köppe is a full professor of Mathematics at UC Davis, where he currently also serves as the chair of the Graduate Group in Applied Mathematics. He received his Ph.D. in Mathematics in 2002 from the University of Magdeburg, Germany, and joined the faculty of UC Davis in 2008. Among his professional duties, he serves as an associate editor for Math. Programming A.

Prof.Shu Lu: "Statistical inference for sample average approximation of constrainedo ptimization and variational inequalities"

Abstract. The sample average approximation is widely used as a substitute for the true expectation function in optimization and equilibrium problems. We study how to provide a confidence region or confidence intervals for the true solution, once the SAA solution is obtained. Our method is based on the asymptotic distribution of the SAA solution, and we handle polyhedral constraints by examining the nonsmooth structure of the asymptotic distribution.


Bio. Shu Lu received her B.S. and M.S. in Civil Engineering from Tsinghua University, and her M.A. in Mathematics and Ph.D. in Industrial and Systems Engineering from the University of Wisconsin-Madison. She is currently Associate Professor at the Department of Statistics and Operations Research, University of North Carolina at Chapel Hill. Her research interests include variational inequalities and variational analysis, optimization under uncertainty, and their applications.

Prof.David Morton: "Nested clustering on a graph"

Abstract. We study a clustering problem defined on a weighted, undirected graph. We remove a subset of edges with the goal of producing clusters, or connected components, on the residual graph. We wish to maximize the number of connected components while minimizing the weight of the removed edges. From a bi-criteria perspective, the parametric problem that results from optimizing a weighted sum of these two objectives can be solved in polynomial time and has the property that the parametric solutions are naturally nested.


Short Bio. Dave Morton is a Professor of Industrial Engineering and Management Sciences at Northwestern University. His research interests include stochastic and large-scale optimization with applications in security, public health, and energy systems. He received a B.S. in Mathematics and Physics from Stetson University and an MS and PhD in Operations Research from Stanford University.

Prof.Shmuel S. Oren:"Opportunities and Challenges for Optimization in Electricity Markets"

Abstract. Socio economic forces, technological developments and environmental considerations have led to restructuring of the electric power systems in part of the USA and in many systems worldwide, transforming them from vertically integrated regulated monopolies to competitive market based systems. Electricity markets represent, perhaps, the most challenging supply chain since the commodity is, practically, non-storable; demand is uncertain and highly correlated with weather, all the demand must be satisfied instantaneously with a high level of reliability (one day in ten years criteria for involuntary load curtailment). In addition service is provided over a network that is prone to congestion, flows over transmission lines cannot be directly controlled as in a transportation system (flows follow Kirchhoff’s laws) and the market is encumbered by numerous externalities and market power. The proliferation of smart grid technologies and the transition of the electric power infrastructure toward massive integration of renewable and distributed resources that are variable and uncertain, poses new challenges in planning and operation of the grid. In particular market mechanisms and optimization tools that have been employed in addressing such task must account for uncertainty and mobilize flexibility on the demand and supply side. This talk will review the basic elements and alternative approaches adopted in restructured power systems and the important role of optimization in system and markets operation. It will also discuss new challenges and opportunities for optimization in handling uncertainty and in exploiting embedded infrastructure flexibility as recourse options in response to contingency and diverse realizations of uncertainties.


Bio. Dr. Shmuel S. Oren is the Earl J. Isaac Chair Professor in the Department of Industrial Engineering and Operations Research at UC Berkeley He is a co-founder and the Berkeley site director, of PSerc, a multi-university Power Systems Engineering Research Center co-sponsored by the National Science Foundation and 40 Industry members. He has also been a member of the California ISO Market Surveillance Committee and a consultant to many private and public entities in the US and abroad. His Research has focused on nonlinear optimization, mechanism design, energy systems and on the design and analysis of electricity markets. He holds a B.S and M.S in Mechanical Engineering from the Technion, Israel and an M.S and Ph.D in Engineering Economic Systems from Stanford University. He is a Member of the US National Academy of Engineering, is a Life Fellow of the IEEE and Fellow of INFORMS.

Prof.Meisam Razaviyayn: "Duality and inference from low order marginals"

Abstract. In many modern inference problems, the task is to predict some target variable Y from some discrete feature vector X = (X1,X2,…,Xp). When the joint distribution of (X,Y ) is known, this task can be done “optimally" by employing the Maximum A-posteriori Probability (MAP) decision rule. However, when only some low order marginals of the joint distribution of (X,Y) is known, this task does not have a simple solution. A fundamental question in this setting is as follows: among all probability distributions satisfying the estimated low order marginals, which one should be used for prediction? We formulate this problem as a robust optimization problem; and suggest to use the Hirschfeld-Gebelein-Renyi (HGR) correlation principle for finding an approximate solution. The approximate solution can be shown to lead to a classier with mis-classication rate no larger than twice the mis-classication rate of the optimal classier. Under a certain “separability” condition, an efficient algorithm is proposed for finding the proposed solution.

Bio. Meisam Razaviyayn is an assistant professor at the department of Industrial and Systems Engineering at the University of Southern California. Prior to joining USC, he was a postdoctoral research fellow in the Electrical Engineering Department at Stanford University. He obtained his PhD degree in Electrical Engineering with a minor in Computer Science from the University of Minnesota in 2014. He is the recipient of the Signal Processing Society Young Author Best Paper Award in 2015 and the University of Minnesota Doctoral Dissertation Fellowship in 2014. He was among the three finalists of the Best Paper Prize for Young Researcher in Continuous Optimization in ICCOPT 2013 and 2016, and the finalist for the best student paper award in SPAWC 2010. His research interests include the design and study of data analysis algorithms and tools which can efficiently scale to modern big data problems.

Prof.R.Tyrrell Rockafellar: "Solving Stochastic Variational Inequalities by Progressive Hedging"

Abstract. Most of the research on stochastic variational inequalities has concentrated on models in which information about the uncertain future is revealed only once. Such models are inadequate to cover multistage stochastic programming, where information comes in stages that offer repeated opportunities for recourse decisions. That feature can be brought into stochastic variational inequalities by adapting them to a constraint of nonanticipativity. In that way not only stochastic programming but multistage multiagent games can be covered.

A particular advantage of this approach is that it generates information price vectors which can be used to decompose the overall problem into a separate problem for each scenario. This fits with solution approaches like the progressive hedging algorithm that owe so much to the ideas of Roger Wets.

Prof. Andrzej Ruszczynski: "Risk-Averse Control of Markov Systems"

Abstract. We shall focus on modeling risk in dynamical systems and discuss fundamental properties of dynamic measures of risk. Special attention will be paid to the local property and the property of time consistency. Then we shall focus on risk-averse control of discrete-time Markov systems. We shall refine the concept of time consistency for such systems, introduce the class of Markovian risk measures, and derive their structure. This will allow us to derive a risk-averse counterpart of dynamic programming equations. Then we shall extend these ideas to partially-observable systems and continuous-time Markov chains and derive the structure of risk measures and dynamic programming equations in these cases as well. In the last part of the talk, we shall discuss risk-averse control of diffusion processes and present a risk-averse counterpart of the Hamilton--Jacobi--Bellman equation. Finally, we shall review some solution methods for risk-averse control problems.


Bio. Andrzej Ruszczynski received his PhD and habilitation degrees in control engineering from Warsaw University of Technology in 1976 and 1983, respectively. He has been with Warsaw University of Technology (Poland), University of Zurich (Switzerland), International Institute of Applied Systems Analysis (Laxenburg, Austria), Princeton University, University of Wisconsin-Madison, and Rutgers University. Dr. Ruszczynski is one of the creators of and main contributors to the field of risk-averse optimization, author of "Nonlinear Optimization" (Princeton University Press, 2006), co-author of "Lectures on Stochastic Programming" (Society of Industrial and Applied Mathematics, 2009), "Stochastic Programming" (Elsevier, 2003),and author of more than 100 articles in the area of optimization. Dr. Ruszczynski was plenary speaker at several major international conferences and held positions in large scientific societies.

Prof.Suvrit Sra: "Geometric optimization: convex and nonconvex"

Abstract. In this talk, I will highlight some aspects of geometry and its role in optimization. In particular, I will talk about optimization problems whose parameters are constrained to lie on a manifold or in a specific metric space. These geometric constraints often make the problems numerically challenging, but they can also unravel properties that ensure tractable attainment of global optimality for certain otherwise non-convex problems.

We'll make our foray into geometric optimization via geodesic convexity, a concept that generalizes the usual notion of convexity to nonlinear metric spaces such as Riemannian manifolds. I will outline some of our results that contribute to g-convex analysis as well as to the theory of first-order g-convex optimization. I will mention several very interesting optimization problems where g-convexity proves remarkably useful. In closing, I will mention extensions to stochastic (non-convex) geometric optimization as well as some important open problems.

Bio. Suvrit Sra is a Research Faculty at the Laboratory for Information and Decision Systems (LIDS) at Massachusetts Institute of Technology (MIT), where he is also a part of the MIT-ML group. He obtained his PhD in Computer Science from the University of Texas at Austin in 2007. Before moving to MIT, he was a Sr. Research Scientist at the Max Planck Institute for Intelligent Systems, in Tübingen, Germany. He has also held visiting faculty positions at UC Berkeley (EECS) and Carnegie Mellon University (Machine Learning Department) during 2013-2014. His research bridges a number of mathematical areas such as metric and differential geometry, matrix analysis, convex analysis, probability theory, and optimization with machine learning. More broadly, his work also involves machine learning and optimization topics in several applications, including materials design. He has been a co-chair for OPT2008--2016, NIPS workshops on "Optimization for Machine Learning," and has also co-edited a book with the same name (MIT Press, 2011).

Prof.Phebe Vayanos: "Robust Wait-Time Estimation in Resource Allocation Systems with an Application to Kidney Allocation"

Abstract. In this paper we study systems that allocate different types of scarce resources to heterogeneous allocatees based on predetermined priority rules, e.g., the U.S. deceased-donor kidney allocation system or the public housing program. We tackle the problem of estimating the wait time of an allocatee who possesses incomplete system information with regard, for example, to his relative priority, other allocatees’ preferences, and resource availability. We model such systems as multiclass, multiserver queuing systems that are potentially unstable or in transient regime. We propose a novel robust optimization solution methodology that builds on the assignment problem. For first-come, first-served systems, our approach yields a mixed-integer programming formulation. For the important case where there is a hierarchy in the resource types, we strengthen our formulation through a drastic variable reduction and also propose a highly scalable heuristic, involving only the solution of a convex optimization problem (usually a second-order cone problem). We back the heuristic with a tight approximation guarantee that becomes tighter for larger problem sizes. We illustrate the generalizability of our approach by studying systems that operate under different priority rules, such as class priority. We conduct a wide range of numerical studies, demonstrating that our approach outperforms simulation. We showcase how our methodology can be applied to assist patients in the U.S. deceased donor kidney waitlist. We calibrate our model using historical data to estimate patients’ wait times based on their kidney quality preferences, blood type, location and rank in the waitlist.


Bio. Phebe Vayanos is an Assistant Professor of Industrial & Systems Engineering and an Associate Director of the Center for Artificial Intelligence in Society at the University of Southern California. Her research interests include optimization under uncertainty, data-driven optimization and analytics, with applications in healthcare, energy, security, and education. Prior to joining USC, she was lecturer in the Operations Research and Statistics Group at the MIT Sloan School of Management, and a postdoctoral research associate in the Operations Research Center at MIT. She holds a PhD degree in Operations Research and an MEng degree in Electrical & Electronic Engineering, both from Imperial College London.

Prof.Jon Wellner: "Bi-s-concave distributions"

Abstract. (pdf)

Bio. Wellner completed his undergraduate work at the University of Idaho in 1968 with B.S. degrees in Mathematics and Physics. After service in the U.S. Army from 1969 - 1971, he did graduate work in Statistics at the University of Washington, completing his Ph.D. under the supervision of Galen R. Shorack in 1975. His first academic appointment as an assistant Professor was at the University of Rochester (1975 - 1983) where he had the good fortune to work with W. J. (Jack) Hall. He returned to the University of Washington in 1983 and has been on the faculty of the UW Department of Statistics since 1983. Wellner visited universities in the Netherlands (Leiden, Amsterdam) and Germany (Munich, Heidelberg) during sabbatical leaves in 1980-1981, 1987-1988, 2004-2005, and 2011-2012. He has written two books on empirical process theory (Shorack and W (1986, 2009), and van der Vaart and W (1996)), one book on semiparametric statistical models (Bickel, Klaassen, Ritov, and W (1993, 1998)), and one book on statistical theory for maximum likelihood estimation methods (Groeneboom and W (1992)). Wellner has been working on statistical inference under shape restrictions for nearly 20 years, and continues to enjoy learning new inequalities and methods for asymptotic theory in statistics.