<![CDATA[2015 Health & Humanitarian Logistics Conference (South Africa)]]> 27858 The 7th annual Conference on Health & Humanitarian Logistics will take place November 18-20th at the Gordon Institute (GIBS) in Johannesburg, South Africa. The conference draws practitioners from private industry, non-governmental organizations (NGOs), government, and military, who are active participants or interested in health and humanitarian operations,

*If you are interested in presenting a poster at the conference, please upload your submission online here as soon as possible or email msmithgall@isye.gatech.edu. A limited amount of funding may be available for accepted poster presenters.including both disaster response and long-term development. 

PLENARY PANEL topicsinclude:

SPEAKERS include representatives from the following organizations:
The Bill and Melinda Gates Foundation, BI Norwegian Business School, Clinton Health Access Initiative, The Coca-Cola Company, FHI 360, GAVI- The Vaccine Alliance, Georgia Tech, Humanitarian Logistics Association, ICRC, Imperial Health Sciences, IMRES Medical Solutions, INSEAD, Johnson & Johnson, Last Mile Health, MIT, McKinsey & Company, Médecins Sans Frontières, Medic Mobile, North Star Alliance, Northeastern University, Northwestern University, OXFAM, People that Deliver, Stanford University, The Global Fund to Fight AIDS, Tuberculosis and Malaria,  UNICEF, UN Mission on Ebola Emergency Response (UNMEER), UN World Food Programme, UPS, USAID, UTi Pharmaceuticals, and the World Health Organization.

THE GOAL of the conference is to identify challenges and solutions in logistics and supply-chain topics related to a broad range of health and humanitarian operations including preparation, response, and recovery from natural and man-made disasters and disease outbreak as well as ongoing humanitarian crises due to war, famine, infectious diseases, and chronic health problems. 

THE AGENDA features plenary panels, interactive workshops, and poster sessions on supply chain management and logistics in global health and humanitarian emergency response and long-term development. The program also includes ample opportunities for networking, health and humanitarian related site visits around Johannesburg, and leisure excursions in South Africa. 

OUR SPONSORS have made this year's conference possible through their generous support: UPS, IMRES Netherlands/Imperial Health Sciences, Walmart, Georgia Tech Stewart School of Industrial & Systems Engineering, and Johnson & Johnson. Sponsorship opportunities are available. Please contact Meghan Smithgaall at msmithgall@isye.gatech.edu if you are interested in joining our sponsors to support the conference.
 

Conference Organizers

]]> Meghan Smithgall 1 1443620606 2015-09-30 13:43:26 1652893545 2022-05-18 17:05:45 0 0 event Join Georgia Tech, INSEAD, MIT, and Northeastern as they host practitioners in global health and humaitarian logistics from across NGOs, government, private companies, and other organizations to identify challenges and solutions in health systems and emergency and long-term development.

]]>
2015-11-18T20:30:00-05:00 2015-11-21T14:59:00-05:00 2015-11-21T14:59:00-05:00 2015-11-19 01:30:00 2015-11-21 19:59:00 2015-11-21 19:59:00 2015-11-18T20:30:00-05:00 2015-11-21T14:59:00-05:00 America/New_York America/New_York datetime 2015-11-18 08:30:00 2015-11-21 02:59:00 America/New_York America/New_York datetime <![CDATA[]]> Meghan Smithgall
Marketing / Ext Affairs
HHS Center
404-381-1432

]]>
454151 454121 454131 454141 454151 image <![CDATA[Keynote speaker image]]> image/jpeg 1449256319 2015-12-04 19:11:59 1475895197 2016-10-08 02:53:17 454121 image <![CDATA[2015 Health & Humanitarian Logistics Conference]]> image/jpeg 1449256319 2015-12-04 19:11:59 1475895197 2016-10-08 02:53:17 454131 image <![CDATA[2015 HHL Conference hosts]]> image/png 1449256319 2015-12-04 19:11:59 1475895197 2016-10-08 02:53:17 454141 image <![CDATA[Ebola response photo- health workers]]> image/jpeg 1449256319 2015-12-04 19:11:59 1475895197 2016-10-08 02:53:17 <![CDATA[Health & Humanitarian Logistics Conference]]>
<![CDATA[Student Honors Luncheon]]> 27187 Anita Race 1 1406538364 2014-07-28 09:06:04 1492118534 2017-04-13 21:22:14 0 0 event 2015-04-16T12:00:00-04:00 2015-04-16T13:30:00-04:00 2015-04-16T13:30:00-04:00 2015-04-16 16:00:00 2015-04-16 17:30:00 2015-04-16 17:30:00 2015-04-16T12:00:00-04:00 2015-04-16T13:30:00-04:00 America/New_York America/New_York datetime 2015-04-16 12:00:00 2015-04-16 01:30:00 America/New_York America/New_York datetime <![CDATA[]]> <![CDATA[Spring Commencement]]> 27187 PhD Hooding Ceremony

]]> Anita Race 1 1406538600 2014-07-28 09:10:00 1492118534 2017-04-13 21:22:14 0 0 event 2015-05-01T11:00:00-04:00 2015-05-01T13:00:00-04:00 2015-05-01T13:00:00-04:00 2015-05-01 15:00:00 2015-05-01 17:00:00 2015-05-01 17:00:00 2015-05-01T11:00:00-04:00 2015-05-01T13:00:00-04:00 America/New_York America/New_York datetime 2015-05-01 11:00:00 2015-05-01 01:00:00 America/New_York America/New_York datetime <![CDATA[]]>
<![CDATA[Presidents Graduation Celebration]]> 27187 Anita Race 1 1406538700 2014-07-28 09:11:40 1492118534 2017-04-13 21:22:14 0 0 event 2015-05-01T16:00:00-04:00 2015-05-01T18:00:00-04:00 2015-05-01T18:00:00-04:00 2015-05-01 20:00:00 2015-05-01 22:00:00 2015-05-01 22:00:00 2015-05-01T16:00:00-04:00 2015-05-01T18:00:00-04:00 America/New_York America/New_York datetime 2015-05-01 04:00:00 2015-05-01 06:00:00 America/New_York America/New_York datetime <![CDATA[]]> <![CDATA[PhD/Master's Ceremony]]> 27187 Anita Race 1 1406538789 2014-07-28 09:13:09 1492118534 2017-04-13 21:22:14 0 0 event 2015-05-01T20:00:00-04:00 2015-05-01T22:00:00-04:00 2015-05-01T22:00:00-04:00 2015-05-02 00:00:00 2015-05-02 02:00:00 2015-05-02 02:00:00 2015-05-01T20:00:00-04:00 2015-05-01T22:00:00-04:00 America/New_York America/New_York datetime 2015-05-01 08:00:00 2015-05-01 10:00:00 America/New_York America/New_York datetime <![CDATA[]]> <![CDATA[Undergraduate AM Ceremony]]> 27187 Anita Race 1 1406538859 2014-07-28 09:14:19 1492118534 2017-04-13 21:22:14 0 0 event 2015-05-02T10:00:00-04:00 2015-05-02T12:30:00-04:00 2015-05-02T12:30:00-04:00 2015-05-02 14:00:00 2015-05-02 16:30:00 2015-05-02 16:30:00 2015-05-02T10:00:00-04:00 2015-05-02T12:30:00-04:00 America/New_York America/New_York datetime 2015-05-02 10:00:00 2015-05-02 12:30:00 America/New_York America/New_York datetime <![CDATA[]]> <![CDATA[Undergraduate PM Ceremony]]> 27187 Anita Race 1 1406538952 2014-07-28 09:15:52 1492118534 2017-04-13 21:22:14 0 0 event 2015-05-02T16:00:00-04:00 2015-05-02T18:30:00-04:00 2015-05-02T18:30:00-04:00 2015-05-02 20:00:00 2015-05-02 22:30:00 2015-05-02 22:30:00 2015-05-02T16:00:00-04:00 2015-05-02T18:30:00-04:00 America/New_York America/New_York datetime 2015-05-02 04:00:00 2015-05-02 06:30:00 America/New_York America/New_York datetime <![CDATA[]]> <![CDATA[PhD Thesis Defense]]> 27187 TITLE: Some New Ideas on Fractional Factorial Design and Computer Experiment

STUDENT: Heng Su

ADVISOR:  Dr. Jeff Wu

SUMMARY:

This thesis consists of two parts. The first part is on fractional factorial design, and the second part is on computer experiment. The first part has two chapters. In the first chapter, we use the concept of conditional main effect, and propose the CME analysis to solve the problem of effect aliasing in two-level fractional factorial design. In the second chapter, we study the conversion rates of a system of webpages with the proposed funnel testing method. The second part also has two chapters. In the third chapter, we use statistical models to calibrate the Perez model. In the last chapter, we propose a new Gaussian process that can jointly model both point and integral responses.

Ever since the founding work by Finney, it has been widely known and accepted that aliased effects in two-level regular designs cannot be “de-aliased” without adding more runs. A surprising result by Wu in his 2011 Fisher Lecture showed that aliased effects can sometimes be “de-aliased” using a new framework based on the concept of conditional main effects (CMEs). In the first chapter, this idea is further developed into a methodology that can be readily used. Some key properties are derived that govern the relationships among CMEs or between them and related effects. As a consequence, some rules for data analysis are developed. Based on these rules, a new CME-based methodology is proposed. Three real examples are used to illustrate the methodology. The CME analysis can offer substantial increase in the R-squared value with fewer effects in the chosen models. Moreover, the selected CME effects are often more interpretable.

Nowadays, internet has become an important source of revenue for various companies. How to design the webpages to maximize the conversions is now a hot topic in e-commerce. In the second chapter, we propose a new method called the funnel testing to simultaneously study a system of webpages and optimize its overall conversions. Directed graph is used to represent the system of webpages and identify its structure. Fractional factorial design is used to conduct the experiment systematically. A new method of analysis is proposed to maximize the total conversion rate of the system. A toy example is used to demonstrate the idea along the description of the method. Another more complicated simulated example is given to further illustrate the methodology.

Traditional uncertainty quantification (UQ) in the prediction of building energy consumption has been limited to the propagation of uncertainties in model input parameters. Models by definition ignore, at least to some degree, and, in almost all cases, simplify the physical processes that govern the reality of interest, thereby introducing additional uncertainty in model predictions that cannot be captured as input parameter uncertainty. Quantification of this type of uncertainty (which we will refer to as model form uncertainty) is a necessary step toward the complete UQ of model predictions. In the third chapter, we introduce a general framework for model form UQ and shows its application to the widely used sky irradiation model developed by Perez (1990), which computes solar diffuse irradiation on inclined surfaces. We collect a dataset of one-year measurements of solar irradiation at one location in the United States. The measurements were done at surfaces with different tilt angles and orientations, for a wide spectrum of sky conditions. A statistical analysis using both this dataset and published studies worldwide suggests that the Perez model performs non-uniformly across different locations and produces a certain bias in its predictions. Based on the same data, we then use a two-phase regression model, to express model form uncertainty in the use of the Perez model at this particular location. Using a holdout validation test, we demonstrate that the two-phase regression model considerably reduces the model bias errors and root mean square errors for every tilted surface. Lastly, we discuss the significance of including model form uncertainty in the energy consumption predictions obtained with whole building simulation.

In some computer experiments, the quantity of interest may be the average value of the responses over a specific region. One example from building energy simulation is the diffuse solar irradiance on a building façade representing the integral of the irradiance over the sky dome that the façade is exposed to. Treating this information as point responses will lead to estimation efficiency loss. In the last chapter, we extend the standard point Gaussian process framework so that it can handle both point and integral responses. This new methodology is called the point-integral Gaussian process model, which is abbreviated as the PIG process model. A generic expression of the PIG process model is given with its complicated covariance functions. Parameter estimation and prediction following the frequentist approach is shown. Closed-form expressions of the covariance functions are derived for axis-parallel rectangular regions, whose computational time are compared with the numerical integration using quadrature. Two examples are given to demonstrate the use and the performance of the new methodology. Two point GP models, one ignores the integral responses and the other transforms the integral into point responses, are compared with the PIG process. In all cases, the proposed PIG process model obtains higher prediction accuracy.

]]> Anita Race 1 1419852594 2014-12-29 11:29:54 1492118451 2017-04-13 21:20:51 0 0 event 2015-01-05T11:00:00-05:00 2015-01-05T11:00:00-05:00 2015-01-05T11:00:00-05:00 2015-01-05 16:00:00 2015-01-05 16:00:00 2015-01-05 16:00:00 2015-01-05T11:00:00-05:00 2015-01-05T11:00:00-05:00 America/New_York America/New_York datetime 2015-01-05 11:00:00 2015-01-05 11:00:00 America/New_York America/New_York datetime <![CDATA[]]>
<![CDATA[MLK Holiday]]> 27187 MLK Holiday

Campus Closed

]]> Anita Race 1 1420451228 2015-01-05 09:47:08 1492118451 2017-04-13 21:20:51 0 0 event 2015-01-19T00:00:00-05:00 2015-01-19T00:00:00-05:00 2015-01-19T00:00:00-05:00 2015-01-19 05:00:00 2015-01-19 05:00:00 2015-01-19 05:00:00 2015-01-19T00:00:00-05:00 2015-01-19T00:00:00-05:00 America/New_York America/New_York datetime 2015-01-19 12:00:00 2015-01-19 12:00:00 America/New_York America/New_York datetime <![CDATA[]]>
<![CDATA[ISyE/SCL January 2015 Supply Chain Day]]> 27233 ISyE students, please join us for our first Spring 2015 Supply Chain Day! The 3-hour session will host supply chain representatives from ChainalyticsDeloitte ConsultingIncomm and Logility who will be on campus to educate ISyE students about their organizations and available employment opportunities.Plus, enjoy a free pizza lunch!

EVENT DETAILS

Where: ISyE Main Bldg, Executive Classroom (Room 228) and ISyE Atrium

When: Thursday, January 22, 11:00AM-2:00PM

What: The session will include:

Please plan on staying for the duration of the event and bring copies of your resume and business cards. Dress is business casual.

REGISTER ONLINE by January 20! Seating is limited!

EVENT SPONSOR

The event is sponsored through the generosity and support of APICS - Atlanta Chapter. APICS is a non-profit educational organization addressing operations management and supply chain management issues, and providing professional development opportunities to our members. Visit http://www.apicsatlanta.org/ to learn more and make sure to stop by the APICS table at the event.

]]> Andy Haleblian 1 1420545213 2015-01-06 11:53:33 1492118447 2017-04-13 21:20:47 0 0 event ISyE students, please join us for our first Spring 2015 Supply Chain Day! The 3-hour session will host supply chain representatives from Chainalytics, Deloitte Consulting, Incomm and Logility who will be on campus to educate ISyE students about their organizations and available employment opportunities. Plus, enjoy a free pizza lunch!

]]>
2015-01-22T11:00:00-05:00 2015-01-22T14:00:00-05:00 2015-01-22T14:00:00-05:00 2015-01-22 16:00:00 2015-01-22 19:00:00 2015-01-22 19:00:00 2015-01-22T11:00:00-05:00 2015-01-22T14:00:00-05:00 America/New_York America/New_York datetime 2015-01-22 11:00:00 2015-01-22 02:00:00 America/New_York America/New_York datetime <![CDATA[]]> event@scl.gatech.edu

]]>
360981 360981 image <![CDATA[Supply Chain Day - January 22, 2015]]> image/png 1449245782 2015-12-04 16:16:22 1475895096 2016-10-08 02:51:36 <![CDATA[Register online(for ISyE students)]]>
<![CDATA[Ebola Modeling Workshop]]> 27187 TITLE:  Modeling the Spread and Control of Ebola in W. Africa

Begins at 8:30am on Jan 22 and concludes by 1:30pm on Jan 23.
Registration fee: $50 registration fee for all participants before Jan 8,
2015.  This fee includes access to all events, as well as breakfasts,
lunches and coffee breaks for the 1.5 day meeting.

Register online by January 8th at bit.ly/ebm_gt

The organizers include Pinar Keskinocak (ISYE) and Fred Vannberg (Biology)
as well as: Rustom Antia (Emory), John Drake (UGA), John Glasser (CDC),
Jonathan Dushoff (McMaster), and Lauren Meyers (UT-Austin).


The meeting is sponsored by the Burroughs Wellcome Fund and multiple Georgia
Tech partners including the College of Sciences, GT-FIRE, School of Biology,
Center for Health and Humanitarian Systems, with administrative support from
the Bioinformatics Program, School of Biology & IBB.

questions should be addressed to: ebola-modeling-workshop@gatech.edu

]]> Anita Race 1 1420548626 2015-01-06 12:50:26 1492118447 2017-04-13 21:20:47 0 0 event 2015-01-22T08:30:00-05:00 2015-01-23T14:00:00-05:00 2015-01-23T14:00:00-05:00 2015-01-22 13:30:00 2015-01-23 19:00:00 2015-01-23 19:00:00 2015-01-22T08:30:00-05:00 2015-01-23T14:00:00-05:00 America/New_York America/New_York datetime 2015-01-22 08:30:00 2015-01-23 02:00:00 America/New_York America/New_York datetime <![CDATA[]]>
<![CDATA[DOS Seminar]]> 27187 TITLE:  Multi-period portfolio optimization with alpha decay

SPEAKER: Kartik Sivaramakrishnan

ABSTRACT:

The traditional Markowitz mean-variance framework is based on a single-period portfolio model. Single-period portfolio optimization does not use any data and decisions beyond the rebalancing time horizon with the result that that its policies are "myopic" in nature. For long-term investors, multiperiod optimization offers the opportunity to make "wait-and-see" policy decisions by including approximate forecasts and long-term policy decisions beyond the rebalancing time horizon.


We consider portfolio optimization with a composite alpha signal that is composed of a short-term and a long-term alpha signal. The short-term alpha better predicts returns at the end of the rebalancing period but it decays quickly. On the other hand, the long-term alpha has less predictive power than the short-term alpha but it decays slowly. We develop a two-stage multi-period model that incorporates this alpha model to construct the optimal portfolio at the end of the rebalancing period. We compare this model with the traditional single-period MVO model on simulated and realistic backtests and show that the multi-period model generates portfolios with superior realized performance.

  

Short Bio:

Kartik Sivaramakrishnan is a Senior Research Associate at Axioma where he maintains Axioma's flagship portfolio optimizer that offers the most flexible construction tool in the market. Kartik is interested in developing second-order mixed-integer portfolio models, and interior point algorithms and associated software for solving these models in a high performance computing environment. He also researches the applications of these models in improving the investment process in practice. Most recently he was involved in developing the NASDAQ-Axioma commodity indexes that track Oil, Gold, and a basket of agricultural products. 

Prior to joining Axioma, Kartik was a tenure-track Assistant Professor in the Department of Mathematics and affiliated with the Operations Research program at North Carolina State University in Raleigh. Kartik has a Ph.D. in Applied Mathematics from Rensselaer Polytechnic Institute and a Masters degree in Electrical Engineering from the Indian Institute of Science in Bangalore. His research has been published in optimization, OR, and practitioner journals.



]]> Anita Race 1 1421165768 2015-01-13 16:16:08 1492118438 2017-04-13 21:20:38 0 0 event 2015-01-21T12:00:00-05:00 2015-01-21T13:00:00-05:00 2015-01-21T13:00:00-05:00 2015-01-21 17:00:00 2015-01-21 18:00:00 2015-01-21 18:00:00 2015-01-21T12:00:00-05:00 2015-01-21T13:00:00-05:00 America/New_York America/New_York datetime 2015-01-21 12:00:00 2015-01-21 01:00:00 America/New_York America/New_York datetime <![CDATA[]]>
<![CDATA[2015 Distinguished Scholarship Lecture]]> 27187 LECTURE: 3:00pm - 4:00pm

RECEPTION: 4:00pm - 5:00pm

TITLE: Multivariate statistics and machine learning under a modern optimization lens

SPEAKER: Dr. Dimitris Bertsimas

ABSTRACT:

Key problems of classification and regression can naturally be written as optimization problems. While continuous optimization approaches has had a significant impact in statistics, discrete optimization has played a very limited role, primarily based on the belief that mixed integer optimization models are computationally intractable. While such beliefs were accurate two decades ago, the field of discrete optimization has made very substantial progress.  

Dr. Bertsimas will discuss how to apply modern first order optimization methods to find feasible solutions for classical problems in statistics, and mixed integer optimization to improve the solutions and to prove optimality by finding matching lower bounds.

Specifically, he will report results for the classical variable selection problem in regression currently solved by LASSO heuristically, least quantile regression, and factor analysis.  He will also present an approach to build regression models based on mixed integer optimization. In all cases he will demonstrate that the solutions found by modern optimization methods outperform the classical approaches. Most importantly, he suggests that the belief widely held in statistics that mixed integer optimization is not practically relevant for statistics applications needs to be revisited.

BIO:

Dimitris Bertsimas is currently the Boeing Professor of Operations Research and the co-director of the Operations Research Center at MIT.  He has been with the MIT faculty since 1988.  His research interests include optimization, statistics and applied probability and their applications in health care, finance, operations management and transportation.

He has co-authored more than 150 scientific papers and three graduate level textbooks. His fourth book, The Analytics Edge, will be published this spring.  He is former department editor in Optimization for Management Science and in Financial Engineering in Operations Research. He has supervised 53 doctoral students and he is currently supervising 19 others. 

 He is a member of the National Academy of Engineering since 2005, and has received numerous research awards including the Morse prize (2013), the Pierskalla award (2013), the best paper award in Transportation (2013), the Farkas prize (2008), the Erlang prize (1996), the SIAM prize in optimization (1996), the Bodossaki prize (1998) and the Presidential Young Investigator award (1991-1996). He has co-founded several companies in the areas of financial services, health care, aviation and publishing. 

 He received his SM and PhD in Applied Mathematics and Operations Research from MIT in 1987 and 1988 respectively.

]]> Anita Race 1 1421835998 2015-01-21 10:26:38 1492118434 2017-04-13 21:20:34 0 0 event 2015-03-05T15:00:00-05:00 2015-03-05T17:00:00-05:00 2015-03-05T17:00:00-05:00 2015-03-05 20:00:00 2015-03-05 22:00:00 2015-03-05 22:00:00 2015-03-05T15:00:00-05:00 2015-03-05T17:00:00-05:00 America/New_York America/New_York datetime 2015-03-05 03:00:00 2015-03-05 05:00:00 America/New_York America/New_York datetime <![CDATA[]]> Lisa Tuttle  ltuttle@gatech.edu

404-385-2911

]]>
<![CDATA[Stat/SIAC Seminar]]> 27187 TITLE:  Internet of Sensed ServGoods: Considerations, Consequences and Concerns

SPEAKER:  James M. Tien

ABSTRACT:

In an earlier paper (Tien 2012), the author augurs that, in contrast to the first and second industrial revolutions which respectively focused on the development and the mass production of goods, the next – or third – industrial revolution is focused on the integration of services and/or goods; it began in this second decade of the 21st Century. The Third Industrial Revolution (TIR) is underpinned by the integration or mass customization of services and/or goods. The benefits of real-time mass customization cannot be over-stated as goods and services become indistinguishable and are co-produced as “ServGoods”, resulting in an overwhelming economic advantage to the industrialized countries where the consuming customers are at the same time the co-producing producers. Adding sensors to these ServGoods and letting them interact or communicate with each other and other components can result in an Internet of Things (i.e., sensed ServGoods). A number of considerations, consequences and concerns relating to such an Internet of Sensed ServGoods are discussed herein.

BIO:

In 2007, Dr. James M. Tien became a Distinguished Professor and the Dean of the College of Engineering at the University of Miami, Coral Gables, Florida.  He received the BEE from Rensselaer Polytechnic Institute (RPI) and the SM, EE and PhD from the Massachusetts Institute of Technology (MIT).  He has held leadership positions at Bell Telephone Laboratories, at the Rand Corporation, and at Structured Decisions Corporation (which he co-founded).  He joined the Department of Electrical, Computer and Systems Engineering at RPI in 1977, became Acting Chair of the department, joined a unique interdisciplinary Department of Decision Sciences and Engineering Systems as its founding Chair, and twice served as the Acting Dean of Engineering. Dr. Tien has published extensively, been invited to present dozens of plenary lectures, and been honored with both teaching and research awards, including being elected a Fellow in IEEE, INFORMS and AAAS and being a recipient of the IEEE Joseph G. Wohl Outstanding Career Award, the IEEE Major Educational Innovation Award, the IEEE Norbert Wiener Award, the IEEE Richard M. Emberson Award, and the IBM Faculty Award. He received a Doctor of Engineering (honoris causa) from Canada’s University of Waterloo and is also an Honorary Professor at over a dozen non-U.S. universities.  Dr. Tien is an elected member of the prestigious U. S. National Academy of Engineering.

]]> Anita Race 1 1421944192 2015-01-22 16:29:52 1492118433 2017-04-13 21:20:33 0 0 event 2015-01-27T11:00:00-05:00 2015-01-27T12:00:00-05:00 2015-01-27T12:00:00-05:00 2015-01-27 16:00:00 2015-01-27 17:00:00 2015-01-27 17:00:00 2015-01-27T11:00:00-05:00 2015-01-27T12:00:00-05:00 America/New_York America/New_York datetime 2015-01-27 11:00:00 2015-01-27 12:00:00 America/New_York America/New_York datetime <![CDATA[]]>
<![CDATA[Visiting Speaker Seminar]]> 27187 TITLE: Optimization and Learning for sequential decision making under uncertainty

SPEAKER:  Dr. Shipra Agrawal

ABSTRACT:

In this talk, I will present techniques that combine optimization and learning for  decision making in complex, uncertain, online environments. Much of this work is motivated by challenges faced in modern revenue management problems, namely, a) unknown or difficult to estimate demand distributions b) multiple complex nonlinear constraints and objectives c) the need for fast large-scale algorithms, and d) personalized data-driven decisions. Formulating these problem aspects into an "online stochastic convex programming" framework, we devise fast algorithms that combine primal-dual paradigm with online learning to achieve provably optimal performance bounds. When applied to the special case of online packing, our ideas yield simpler and faster algorithms with optimal competitive ratio for this widely studied problem.  An application area of our focus is internet advertising, where millions of ads need to be served every second to ensure long term revenues, without  knowing the future demand patterns. Our algorithms have been implemented and are in use for yield management in display advertising.  

 I will further discuss a "bandit" property present in many revenue management problems, where the uncertain demand at each point in time is determined by the decision, and can only be observed "after" taking the decision. For example, in pay-per-click revenue model in internet advertising, a click (or no click) on an ad can be observed only after the ad is selected and displayed in response to the user query. Similar situation occurs in many other applications including posted-price based revenue management, worker-task allocation problem in crowdsourcing, machine scheduling, sensor network management etc. Modeling this problem as a combination of the classic multi-armed bandit problem and online stochastic convex programming, we design algorithms that balance the inherent exploration-exploitation (i.e., learning vs. optimization) tradeoff to achieve asymptotically optimal decision policies. Our results significantly improve upon several known results in online linear programming, blind network revenue management and multi-armed bandits literature, and reveal many interesting connections between convex optimization algorithms, Fenchel duality, multi-armed bandits and online learning.

This talk is based on joint work with Nikhil R. Devanur and Navin Goyal.

 

Brief Bio:


Shipra Agrawal is a researcher at Microsoft Research, India. She received a PhD in Computer Science from Stanford University in 2011, under the direction of Professor Yinyu Ye, Department of Management Science and Engineering. Her research spans several areas of optimization and machine learning, including data-driven optimization under partial, uncertain, and online inputs, and related concepts in learning, namely multi-armed bandits, online learning, and reinforcement learning.  She is also interested in prediction markets and game theory. Application areas of her interest include, but are not limited to, internet advertising, recommendation systems, revenue management and resource allocation problems.

]]> Anita Race 1 1422355346 2015-01-27 10:42:26 1492118424 2017-04-13 21:20:24 0 0 event 2015-02-05T11:00:00-05:00 2015-02-05T12:00:00-05:00 2015-02-05T12:00:00-05:00 2015-02-05 16:00:00 2015-02-05 17:00:00 2015-02-05 17:00:00 2015-02-05T11:00:00-05:00 2015-02-05T12:00:00-05:00 America/New_York America/New_York datetime 2015-02-05 11:00:00 2015-02-05 12:00:00 America/New_York America/New_York datetime <![CDATA[]]>
<![CDATA[DOS Seminar]]> 27187 TITLE:  The Multiple-Job Repair Kit Problem with Forward Stocking Location Recourse

SPEAKER:  Brian Kues

ABSTRACT:

The multiple-job repair kit problem is concerned with choosing which spare parts a traveling technician should carry. The demand for parts at each job is not known until the technician diagnoses the problem on-site. The general objective is to understand and optimize the trade-off between low inventory cost and high customer service level, which is defined as the fraction of jobs completed in a single visit. Hence, jobs which require parts not stocked in the repair kit are considered service failures.


We introduce a new model for a multiple-job repair kit problem where a technician can retrieve needed parts not carried in the repair kit from a forward stocking location in order to complete a job successfully. However, such trips require additional time and hinder the technician’s ability to complete later jobs. This type of spare parts supply chain trades off low inventory cost and high technician productivity. We develop a model for the problem as well as an algorithm to find an inventory policy. We show that the algorithm is not always optimal and can be arbitrarily bad in the worst case but performs well in many computational experiments. We also perform factor screening experiments to examine the sensitivity of the heuristic solutions to the problem parameters.

 

]]> Anita Race 1 1422355512 2015-01-27 10:45:12 1492118424 2017-04-13 21:20:24 0 0 event 2015-01-28T12:00:00-05:00 2015-01-28T13:00:00-05:00 2015-01-28T13:00:00-05:00 2015-01-28 17:00:00 2015-01-28 18:00:00 2015-01-28 18:00:00 2015-01-28T12:00:00-05:00 2015-01-28T13:00:00-05:00 America/New_York America/New_York datetime 2015-01-28 12:00:00 2015-01-28 01:00:00 America/New_York America/New_York datetime <![CDATA[]]>
<![CDATA[TAG Panel - Global Supply Chain Visibility]]> 27233 Please join TAG Supply Chain & Logistics for a panel discussion and networking event featuring panelists from Elemica, UPS, Retalix and Software AG who will share real world examples.

Event Details

Where: UPS Auditorium (55 Glenlake Parkway, Ste 590, Atlanta, GA 30328 https://goo.gl/maps/OVmM3)

When: Tuesday, Feb 10, 2015, 6:00 p.m. – 8:00 p.m.

The Supply Chain and Logistics community has been discussing global supply chain visibility for 15+ years. In this panel discussion, TAG SC&L will discuss why global supply chain visibility is so elusive and present how new technologies and approaches are bringing innovative solutions to the problem.

In this panel, TAG SC&L will discuss why global supply chain visibility is so elusive and present how new technologies and approaches are bringing innovative solutions to the problem. The panelist will also discuss the evolution of global supply chain visibility and the components of a global supply chain visibility technology solution of today. They will elaborate on how visibility is not just seeing what is happening now - but predicting what will happen next.

 Panelists from: Elemica, UPS, Retalix and Software AG. Lee Blackstone (Blackstone and Cullen) will be moderating.

Register Today!

TAG members: FREE; Non-members $20

For more information and to register online, please visit http://bit.ly/taggscv.

 Sponsors

SilverPyramid Consulting, CatavoltGeorgia Tech Supply Chain & Logistics InstituteInsperityPrimus Software CorporationSoftware AG Suite

]]> Andy Haleblian 1 1422450360 2015-01-28 13:06:00 1492118424 2017-04-13 21:20:24 0 0 event Please join TAG Supply Chain & Logistics for a panel discussion and networking event featuring panelists from Elemica, UPS, Retalix and Software AG who will share real world examples.

]]>
2015-02-10T18:00:00-05:00 2015-02-10T20:00:00-05:00 2015-02-10T20:00:00-05:00 2015-02-10 23:00:00 2015-02-11 01:00:00 2015-02-11 01:00:00 2015-02-10T18:00:00-05:00 2015-02-10T20:00:00-05:00 America/New_York America/New_York datetime 2015-02-10 06:00:00 2015-02-10 08:00:00 America/New_York America/New_York datetime <![CDATA[]]> Please email julianne@tagonline.org with any questions.

]]>
369241 369241 image <![CDATA[TAG Panel - Global Supply Chain Visibility]]> image/png 1449245845 2015-12-04 16:17:25 1475894347 2016-10-08 02:39:07 <![CDATA[Register Online]]> <![CDATA[TAG Supply Chain & Logistics Society]]> <![CDATA[The Supply Chain and Logistics Institute at Georgia Tech]]>
<![CDATA[TAG Panel & Luncheon | The Path Forward - 3D Printing and Supply Chain Management]]> 27233 Please join TAG Supply Chain & Logistics for a networking lunch and panel discussion with members of the Georgia Tech Stewart School of Industrial and Systems Engineering, Georgia Tech Manufacturing Institute, Terminal Velocity Aerospace, and Coca-Cola Enterprises.

Event Details

Where: Georgia Tech Manufacturing Institute Auditorium (813 Ferst Drive, N.W. Atlanta, GA 30332-0560 | 404-894-9100 https://goo.gl/maps/avB3A). Please arrive at least 30 minutes early to give yourself ample time to find parking and to walk to the venue. We suggest using Visitor Lot 4: State Street & Ferst Drive or Visitor Lot 3: Ferst Drive & Regents. Parking map/instructions.

When: Wednesday, April 22, 2015, 11:30 a.m. – 1:30 p.m.

3D printing has captured the imagination of the media and the financial markets. Prognosticators are predicting a fundamental disruption in the manufacturing paradigm, from mass production to mass customization. Are we on the path to achieving the "lot size of one?” Will factories disappear as custom-designed items are created on 3D printers in the home? How will these changes affect end-to-end supply chains?

Our speakers will discuss the current state of 3D printing (application, limitations, examples), advances on the horizon and their potential impact on the supply chain, and a case study of how 3D printing is incorporated into the supply chain at a specific company.

*Participants are invited to join in a tour of a GTMI lab following the event.

Panelists from:  Georgia Tech Stewart School of Industrial and Systems Engineering, Georgia Tech Manufacturing Institute, Terminal Velocity Aerospace, and Coca-Cola Enterprises.

Register Today!

TAG members: $10; Non-members $30

For more information and to register online, please visit http://bit.ly/tag3dprinting.

 Sponsors

Gigabyte: Manhattan Associates

Megabyte: Blackstone+CullenInsperitySoftware AG Suite

Hosted by the Georgia Tech Supply Chain & Logistics Institute

]]> Andy Haleblian 1 1422457055 2015-01-28 14:57:35 1492118424 2017-04-13 21:20:24 0 0 event Please join TAG Supply Chain & Logistics for a networking lunch and panel discussion with members of the Georgia Tech Stewart School of Industrial and Systems Engineering, Georgia Tech Manufacturing Institute, Terminal Velocity Aerospace, and Coca-Cola Enterprises.

]]>
2015-04-22T12:30:00-04:00 2015-04-22T14:30:00-04:00 2015-04-22T14:30:00-04:00 2015-04-22 16:30:00 2015-04-22 18:30:00 2015-04-22 18:30:00 2015-04-22T12:30:00-04:00 2015-04-22T14:30:00-04:00 America/New_York America/New_York datetime 2015-04-22 12:30:00 2015-04-22 02:30:00 America/New_York America/New_York datetime <![CDATA[]]> Please email julianne@tagonline.org with any questions.

]]>
369461 369461 image <![CDATA[TAG Panel & Luncheon | The Path Forward - 3D Printing and Supply Chain Management]]> image/png 1449245845 2015-12-04 16:17:25 1475894382 2016-10-08 02:39:42 <![CDATA[The Supply Chain and Logistics Institute at Georgia Tech]]> <![CDATA[Register Online]]> <![CDATA[TAG Supply Chain & Logistics Society]]> <![CDATA[Event flyer and Parking Map]]>
<![CDATA[DOS Seminar]]> 27187 TITLE: Center-points: A link between discrete geometry and optimization

SPEAKER:  Timm Oertel

ABSTRACT:

n this talk, I will consider mixed-integer convex minimization problems. First, I will present optimality conditions for this class of optimization problems. Then, I will introduce the concept of center-points, a generalization of the median from the one dimensional space to vector spaces. Through the theory of center-points I will show how to extend the general cutting plane scheme from the continuous setting to the mixed-integer setting. Further, I will present several properties of center-points and how to compute them approximately.


Timm's webpage: http://www.ifor.math.ethz.ch/staff/toertel]]> Anita Race 1 1422871373 2015-02-02 10:02:53 1492118417 2017-04-13 21:20:17 0 0 event 2015-02-04T12:00:00-05:00 2015-02-04T13:00:00-05:00 2015-02-04T13:00:00-05:00 2015-02-04 17:00:00 2015-02-04 18:00:00 2015-02-04 18:00:00 2015-02-04T12:00:00-05:00 2015-02-04T13:00:00-05:00 America/New_York America/New_York datetime 2015-02-04 12:00:00 2015-02-04 01:00:00 America/New_York America/New_York datetime <![CDATA[]]>
<![CDATA[2015 Georgia Logistics Summit]]> 27233 Stop by the SCL booth at the 2015 Georgia Logistics Summit. The Georgia Logistics Summit is a unique event packed with valuable business information and networking opportunities like no other. Keynote speakers will address topics that are relevant to the operation and logistics success of your business.

On March 31 and April 1, 2015 at the Georgia World Congress Center, the event will celebrate its seventh anniversary. The summit continues to grow in attendance every year, hosting 2,200 attendees from 38 states and 11 nations in 2014. Most importantly, 85% of these participants come from across the private sector, and the agenda and activities are purposefully created with input directly from the logistics industry.

2015 includes:

For more information relating to the event and to register, please visit http://logistics.georgiainnovation.org/logistics-summit/.

]]> Andy Haleblian 1 1422979095 2015-02-03 15:58:15 1492118412 2017-04-13 21:20:12 0 0 event Stop by the SCL booth March 31 - April 1, 2015 at the seventh annual Georgia Logistics Summit to be held at the Georgia World Conference Center. The Georgia Logistics Summit is a unique event packed with valuable business information and networking opportunities like no other. Keynote speakers will address topics that are relevant to the operation and logistics success of your business.

]]>
2015-03-31T10:30:00-04:00 2015-04-01T15:30:00-04:00 2015-04-01T15:30:00-04:00 2015-03-31 14:30:00 2015-04-01 19:30:00 2015-04-01 19:30:00 2015-03-31T10:30:00-04:00 2015-04-01T15:30:00-04:00 America/New_York America/New_York datetime 2015-03-31 10:30:00 2015-04-01 03:30:00 America/New_York America/New_York datetime <![CDATA[]]> Georgia Center of Innovation for Logistics
http://www.georgialogistics.com/about-the-center/contact.aspx

]]>
372181 372181 image <![CDATA[2015 Georgia Logistics Summit]]> image/png 1449245867 2015-12-04 16:17:47 1475894265 2016-10-08 02:37:45 <![CDATA[2015 Georgia Logistics Summit website]]>
<![CDATA[Faculty Candidate Seminar]]> 27187 TITLE:  Two-sample hypothesis testing for random dot product graphs

SPEAKER:  Dr. Minh Tang

ABSTRACT:

Two-sample hypothesis testing for random graphs arises naturally in   neuroscience, social networks, and machine learning. The talk discusses   the nonparametric problems of whether two finite-dimensional   random dot product graphs have generating latent positions that are
independently drawn from the same distribution, or distributions   that are related via scaling or projection. A consistent test procedure wherein the graphs are first embedded into  Euclidean space via spectral decomposition of the adjacency matrices followed by a kernel-based distance measure between the resultant embeddings is then presented. The talk concludes with a discussion of how the proposed test procedure might be applied to the general problem of identifying and classifying local structure in big data graphs, e.g., the identification of repeated processing modules in the neocortex as suggested by the cortical column conjecture, and the challenges that it entails.


]]> Anita Race 1 1423041493 2015-02-04 09:18:13 1492118412 2017-04-13 21:20:12 0 0 event 2015-02-10T11:00:00-05:00 2015-02-10T12:00:00-05:00 2015-02-10T12:00:00-05:00 2015-02-10 16:00:00 2015-02-10 17:00:00 2015-02-10 17:00:00 2015-02-10T11:00:00-05:00 2015-02-10T12:00:00-05:00 America/New_York America/New_York datetime 2015-02-10 11:00:00 2015-02-10 12:00:00 America/New_York America/New_York datetime <![CDATA[]]>
<![CDATA[Home Delivery World 2015, Click & Collect USA, Etail Show USA]]> 27233 For 2015, Home Delivery World will be joined by and co-located with Click & Collect USA and Etail Show USA April 8-9 at the Atlanta Convention Center at AmericasMart. The events are open to anyone and "Exhibition-only" passes are free. See the respective websites for other options and program details.

Georgia Tech Students!

A limited number of "Retailers and Manufacturers" passes will be made available to students in exchange for a day of blogging (grants access to activities in addition to the expo…stop by the SCL suite - Groseclose 228).

Home Delivery World 2015
The USA's only event for the entire home delivery chain

The 3rd annual Home Delivery World will bring retailers and grocers together to discuss their strategies and learn from their peers. The shippers in attendance are here to identify new ways to improve customer experience and retention through fulfillment, delivery, and reverse logistics programs.

The event will be combined with the Click & Collect and the Etail Show USA 2015 to address new models of customer engagement, pickup solutions, and omnichannel opportunities.

Leading retailers from across the globe will be there to:

Register for the VIP conferenceAttend the expo for FREE

Click & Collect Show USA 2015
The leading click and collect event

The Click & Collect Show USA is the first and only event in the US market to gather the heads of delivery operations from the major retailers in the world to discuss how Click & Collect will shape the way consumers collect their goods, either in store or by driving to a collection point.

From in-store pickup to drive thru, lockers, and kiosks approaches, Click & Collect examines the strategies that retailers and carriers are starting to roll out in the US market and that may change the entire home delivery chain. Co-located with the Home Delivery and Etail Show USA 2015, this will be an opportunity to evaluate which delivery approach is best suitable for your business.

Leading retailers, etailers, and grocers from across the globe will be there to:

Register for the VIP conferenceAttend the expo for FREE

Etail Show USA 2015

The Etail Show USA is the number one event in America exploring the use of technologies by retail goods for online customers. This event gathers leading etailers to tackle innovation across online and multi-channel retail and discuss the challenges in transforming e-commerce into the customer’s preferred marketplace.

Held in conjunction with the Home Delivery and Click & Collect Show USA 2015, this will be a unique opportunity to meet with retailers and merchants, emerging etailers, and the risk-taking visionaries of the commerce industry to create a company-wide culture of innovation and revolutionize your business.

Register for the VIP conference | Attend the expo for FREE

About Attending

Attend any of the above expos for FREE. If you would like to attend the additional seminars and conference events not included in the expo, please see each respective conference website. Discounts are available for groups of 3 or more - for more information please contact Felipe Lima at Felipe.Lima@Terrapinn.com or +1 212 379 6320.

]]> Andy Haleblian 1 1423052908 2015-02-04 12:28:28 1492118412 2017-04-13 21:20:12 0 0 event For 2015, Home Delivery World will be joined by and co-located with Click & Collect USA and Etail Show USA April 8-9 at the Atlanta Convention Center at AmericasMart. The events are open to anyone and "Exhibition-only" passes are free. See the respective websites for other options and program details.

]]>
2015-04-08T09:00:00-04:00 2015-04-09T18:00:00-04:00 2015-04-09T18:00:00-04:00 2015-04-08 13:00:00 2015-04-09 22:00:00 2015-04-09 22:00:00 2015-04-08T09:00:00-04:00 2015-04-09T18:00:00-04:00 America/New_York America/New_York datetime 2015-04-08 09:00:00 2015-04-09 06:00:00 America/New_York America/New_York datetime <![CDATA[]]> Discounts are available for groups of 3 or more - for more information please contact Felipe Lima at Felipe.Lima@Terrapinn.com or +1 212 379 6320.

]]>
372501 372501 image <![CDATA[Home Delivery World 2015, Click & Collect USA, Etail show USA]]> image/png 1449245867 2015-12-04 16:17:47 1475894382 2016-10-08 02:39:42 <![CDATA[Home Delivery World 2015]]> <![CDATA[Click & Collect USA 2015]]> <![CDATA[Etail show USA 2015]]>
<![CDATA[FREE Inventory Management webinar, Wednesday, Feb 18 1:30 - 2:30 pm EST!]]> 27233 Attend this free webinar and receive a *discount code towards our upcoming Inventory Planning and Management course! REGISTER at http://bit.ly/invplngwebinar.

OVERVIEW
Supply chain management and logistics encompasses all the activities involved in getting products to consumers including planning, storing, moving, and accounting for inventory. Inventory availability is the most important aspect of customer service, and the cost of inventory is one of the most important entries on a company's balance sheet. In our webinar, we will explore how to go about minimizing inventory in your supply chain by first understanding your customer’s needs.

Join Dan Gardner as he touches upon the below issues:

Those who will benefit from this webinar include:

Attendees will receive a discount* to the course, “Inventory Planning and Management,” March 18-20, 2015 in Atlanta. Included in this course is a tour of the Federal Reserve Bank. (*Webinar discount cannot be combined with any other discount.) 


ABOUT THE PRESENTER

Dan Gardner is co-founder and president of supply chain consulting firm Trade Facilitators, Inc. Prior to founding TFI in 2007, Dan spent twenty three years in the Third Party Logistics business where he held senior level positions with firms that include Exel Global Logistics and DHL Global Forwarding. Dan is the instructor for the Georgia Tech Supply Chain & Logistics Institute course Inventory Planning and Management.

]]> Andy Haleblian 1 1423579442 2015-02-10 14:44:02 1492118407 2017-04-13 21:20:07 0 0 event Attend this free webinar to learn how to minimize inventory in your supply chain by first understanding your customer’s needs.

]]>
2015-02-18T13:30:00-05:00 2015-02-18T14:30:00-05:00 2015-02-18T14:30:00-05:00 2015-02-18 18:30:00 2015-02-18 19:30:00 2015-02-18 19:30:00 2015-02-18T13:30:00-05:00 2015-02-18T14:30:00-05:00 America/New_York America/New_York datetime 2015-02-18 01:30:00 2015-02-18 02:30:00 America/New_York America/New_York datetime <![CDATA[]]> webinar@scl.gatech.edu]]> 376631 376631 image <![CDATA[Understanding and Optimizing Inventory in Supply Chains]]> image/jpeg 1449246205 2015-12-04 16:23:25 1475894342 2016-10-08 02:39:02 <![CDATA[Register Online to Attend]]> <![CDATA[Inventory Planning and Management course]]>
<![CDATA[FREE Lean Problem Solver webinar, Tuesday, Feb 24 1:30 - 2:30 pm EST!]]> 27233 Attend this free webinar and receive a *discount code towards our upcoming Building the Lean Supply Chain Problem Solver course! REGISTER at http://bit.ly/leanpswebinar.

OVERVIEW
To become a lean supply chain professional, you first need to become a lean thinker and problem solver. In our live webinar, we will explore lean fundamentals and critical concepts needed to identify and eliminate waste at the root of the cause. Time will be allotted for questions and answers!

Join Brad Bossence as he touches upon the below issues:

Those who will benefit from this webinar include supply chain professionals, logistics professionals, material managers, production control managers, transportation managers, warehousing managers and purchasing managers.

Attendees will receive a discount* to the course, “Building the Lean Supply Chain Problem Solver,” March 10-12, 2015 in Atlanta. (*Webinar discount cannot be combined with any other discount.) 


ABOUT THE PRESENTER

Brad Bossence is Vice President of Customer Relations for LeanCor Supply Chain Group and instructor for SCL's Building the Lean Supply Chain Problem Solver course. Brad has over 14 years of third party logistics experience with a specific focus in Japanese production system environments such as Toyota USA, Toyota Canada, Toyota Europe, Kubota, Yamaha, Suzuki, and Subaru. His previous roles include General Manager of Logistics and Operations at Transfreight as well as various contract and operations management positions across the globe.

]]> Andy Haleblian 1 1423657612 2015-02-11 12:26:52 1492118406 2017-04-13 21:20:06 0 0 event Attend this free webinar to learn about becoming a lean thinker and problem solver. In our live webinar, we will explore lean fundamentals and critical concepts needed to identify and eliminate waste at the root of the cause.

]]>
2015-02-24T13:30:00-05:00 2015-02-24T14:30:00-05:00 2015-02-24T14:30:00-05:00 2015-02-24 18:30:00 2015-02-24 19:30:00 2015-02-24 19:30:00 2015-02-24T13:30:00-05:00 2015-02-24T14:30:00-05:00 America/New_York America/New_York datetime 2015-02-24 01:30:00 2015-02-24 02:30:00 America/New_York America/New_York datetime <![CDATA[]]> webinar@scl.gatech.edu]]> 377151 377151 image <![CDATA[FREE Lean Problem Solver webinar]]> image/jpeg 1449246205 2015-12-04 16:23:25 1475894388 2016-10-08 02:39:48 <![CDATA[Register Online to Attend]]> <![CDATA[Course webpage within the SCL website]]>
<![CDATA[PhD Thesis Defense]]> 27187 TITLE: Robust Optimization with Applications in Maritime Inventory Routing

STUDENT: Chengliang Zhang

ABSTRACT:

In recent years, the importance of incorporating uncertainty into planning models for logistics and transportation systems has been widely recognized in the Operations Research (OR) and transportation science communities. Maritime transportation, as a major mode of transport in the world, is subject to a wide range of disruptions at the strategic, tactical and operational levels. This thesis is mainly concerned with the development of robustness planning strategies that can mitigate the effects of some major types of disruptions for an important class of optimization problems in the shipping industry.

The problem is motivated by an application in the design and negotiation of an Annual Delivery Plan (ADP) involving a single vendor and multiple customers in the Liquefied Natural Gas (LNG) business. The overall ADP planning activity is to develop contractual agreements of delivery plans that specify delivery dates (or time windows) and the corresponding delivery quantities over a long-term horizon. In the first part of the thesis, we study a maritime inventory routing problem with given time windows for deliveries with uncertain disruptions. We propose a Lagrangian heuristic scheme to obtain robust solutions by incorporating soft constraints, whose satisfaction can aid robustness, into the objective function with Lagrange multipliers. By simulating random disruption events, we show that the actual operational costs in case of disruptions can be significantly reduced when robust plans are implemented. In addition, the simulator enables us to determine the cost of achieving the robustness and to generate recovery solutions under various disruption events with different lead times.

In the second part, we study a more general robust maritime inventory routing problem with time windows, where the length and placement of the time windows are also decision variables. The vendor must simultaneously decide routes for all the vessels and time windows at all the customers. We formulate the problem as a two-stage stochastic mixed-integer program and propose a two-phase solution approach that considers a sample set of disruptions as well as their recovery solutions. In the first phase, we introduce two planning strategies to generate robust routes, and in the second phase, we propose a multi-scenario construction heuristic to obtain good feasible solutions. We also investigate an iterative procedure between updating the routes and re-optimizing the time windows by coupling the Lagrangian heuristic approach proposed in the first part.

Finally, we study a robust single-item uncapacitated lot-sizing problem with backlogging and random machine breakdowns. The objective is to optimize the costs of production, inventory and backlogging against the worst-case scenario. By identifying the solution characteristics of the worst-case disruptions, we show that the optimal solutions to the robust model can be characterized by a set of stationary production plans.

]]> Anita Race 1 1424074399 2015-02-16 08:13:19 1492118403 2017-04-13 21:20:03 0 0 event 2015-02-19T10:30:00-05:00 2015-02-19T10:30:00-05:00 2015-02-19T10:30:00-05:00 2015-02-19 15:30:00 2015-02-19 15:30:00 2015-02-19 15:30:00 2015-02-19T10:30:00-05:00 2015-02-19T10:30:00-05:00 America/New_York America/New_York datetime 2015-02-19 10:30:00 2015-02-19 10:30:00 America/New_York America/New_York datetime <![CDATA[]]>
<![CDATA[ISyE/SCL March 2015 Supply Chain Day]]> 27233 ISyE students, please join us for our second Spring 2015 Supply Chain Day! The 3-hour session will host supply chain representatives from AmericoldAndritzExelGraphic PackagingIsotrak and Pacesetter Steel Service who will be on campus to educate ISyE students about their organizations and available employment opportunities. Plus, enjoy a free pizza lunch!

EVENT DETAILS

Where: ISyE Main Bldg, Executive Classroom (Room 228) and ISyE Atrium

When: Tuesday, March 3, 11:00AM-2:00PM

What: The session will include:

Please plan on staying for the duration of the event and bring copies of your resume and business cards. Dress is business casual.

REGISTER ONLINE by March 1st! Seating is limited!

EVENT SPONSOR

The event is sponsored through the generosity and support of APICS - Atlanta Chapter. APICS is a non-profit educational organization addressing operations management and supply chain management issues, and providing professional development opportunities to our members. Visit http://www.apicsatlanta.org/ to learn more and make sure to stop by the APICS table at the event.

]]> Andy Haleblian 1 1424098977 2015-02-16 15:02:57 1492118403 2017-04-13 21:20:03 0 0 event ISyE students, please join us for our second Spring 2015 Supply Chain Day! The 3-hour session will host supply chain representatives from Americold, Andritz, Exel, Graphic Packaging, Isotrak and Pacesetter Steel Service who will be on campus to educate ISyE students about their organizations and available employment opportunities. Plus, enjoy a free pizza lunch!

]]>
2015-03-03T11:00:00-05:00 2015-03-03T14:00:00-05:00 2015-03-03T14:00:00-05:00 2015-03-03 16:00:00 2015-03-03 19:00:00 2015-03-03 19:00:00 2015-03-03T11:00:00-05:00 2015-03-03T14:00:00-05:00 America/New_York America/New_York datetime 2015-03-03 11:00:00 2015-03-03 02:00:00 America/New_York America/New_York datetime <![CDATA[]]> event@scl.gatech.edu

]]>
378831 378831 image <![CDATA[Supply Chain Day - March 3, 2015]]> image/gif 1449246214 2015-12-04 16:23:34 1475894388 2016-10-08 02:39:48 <![CDATA[Register online(for ISyE students)]]>
<![CDATA[ISyE Seminar]]> 27187 TITLE: Countably infinite linear programs: theory, algorithms, and applications

SPEAKER: Archis Ghate

ABSTRACT:

We will consider linear programs (LPs) with a countably infinite number of variables and a countably infinite number of constraints. These countably infinite linear programs (CILPs) arise in several applications including countable state Markov decision processes (MDPs), infinite-stage minimum cost network flow problems, and non-stationary infinite-horizon planning problems. Standard results, intuition, and interpretations from finite-dimensional LPs may not extend to CILPs. For example, weak and strong duality may not hold, extreme points may not be equivalent to basic feasible solutions, dual variables may not have a shadow price interpretation, and a finitely implementable Simplex algorithm is not known. In this talk, we will explore sufficient conditions under which such theoretical results and algorithms can be developed for CILPs. Several examples and counterexamples will be discussed to explain key ideas. Non-stationary infinite-horizon MDPs will be employed as a flagship example where everything works out nicely. Time permitting, an inverse optimization framework and a robust optimization approach for CILPs will be presented briefly.

 Speaker bio:  Archis Ghate is an Associate Professor of Industrial and Systems Engineering at the University of Washington in Seattle. His research focuses on stochastic and dynamic optimization. Archis received a PhD from the University of Michigan in 2006, an MS from Stanford University in 2003, and completed his undergraduate education at the Indian Institute of Technology, Bombay in 2001. He is a recipient of the NSF CAREER award and the award for Excellence in Teaching OR from the Institute of Industrial Engineers. His doctoral students have won the Dantzig dissertation award and the Bonder scholarship from INFORMS, as well as other competitive awards from the University of Washington.

 

 

]]> Anita Race 1 1424250604 2015-02-18 09:10:04 1492118402 2017-04-13 21:20:02 0 0 event 2015-02-27T12:00:00-05:00 2015-02-27T13:00:00-05:00 2015-02-27T13:00:00-05:00 2015-02-27 17:00:00 2015-02-27 18:00:00 2015-02-27 18:00:00 2015-02-27T12:00:00-05:00 2015-02-27T13:00:00-05:00 America/New_York America/New_York datetime 2015-02-27 12:00:00 2015-02-27 01:00:00 America/New_York America/New_York datetime <![CDATA[]]>
<![CDATA[PhD Thesis Defense]]> 27187 TITLE:  Residual Life Prediction and Degradation-Based Control of Multi-Component Systems

STUDENT:  Li Hao

ABSTRACT:

The condition monitoring of multi-component systems utilizes multiple sensors to capture the functional condition of the systems, and allows the sensor information to be used to reason about the health information of the systems or components. This thesis focuses on modeling the relationship between multi-sensor information and component-level degradation, so as to prediction both system-level and component-level lifetimes. In addition, this thesis also investigates the dynamic control of component-level degradation so as to control the failure times of individual components based on real-time degradation monitoring.

The research topic that Chapter 3 focuses on is identifying component degradation signals from mixed sensor signals in order to predict component-level residual lives. Specifically, we are interested in modeling the degradation of systems that consist of two or more identical components operating under similar conditions. The key challenge here is that a defect in any of the components will excite the same defective frequency, which prevents an effective separation of the degradation signals of defective and non-defective components. To the best of our knowledge, no existing methodologies have investigated this research topic. In Chapter 3, we propose a two-stage vibration-based prognostic methodology for modeling the degradation processes of components with identical defective frequencies. The first stage incorporates the independent component analysis (ICA) to identify component vibration signals and reverse their original amplitude. The second stage consists of an adaptive prognostics method to predict component residual lives. In the simulated case study, we investigate the performance of the signal separation stage and that of the final residual-life prediction under different conditions. The simulation results show reasonable robustness of the methodology.

In Chapter 4, we focus on characterizing the interactive relationship between product quality degradation and tool wear in multistage manufacturing processes (MMPs), in which machine tools are considered as components and the product quality measurements are considered as condition monitoring information. Due to the sequential structure of MMPs, the degradation status of a tool affects the product quality current stage, which, on the other hand, may affect the degradation of tools at subsequent stages. To the best of our knowledge, although existing literature has modeled the impact of product quality on the tooling catastrophic failure, no published work has targeted on the impact of product quality on the actual process of tool wear. To address this research topic, we propose an high-dimensional stochastic differential equation model to capture the interaction relationship between the process of tool wear and product quality. We then leverage real-time quality measurements to on-line predict the residual life of the MMP as a system. In the simulation study, we conclude that our methodology consistently performs better than a benchmark methodology that does not consider the impact of product quality on the process of tool wear or utilize real-time quality measurements.

Chapter 5 explores a new research direction, which is the dynamic control of component-level degradation in the parallel multi-component system, in which each component operates simultaneously to achieve an engineering objective. This parallel configuration is usually designed with some level of redundancy, which means when a small portion of components fails to operate, the remaining components can still achieve the engineering objective by increasing their workloads up to the designed capacities. Consequently, if the component degradation can be controlled, we can achieve better utilization of the redundancy to ensure consistent system performance. To do this, Chapter 5 assumes that the degradation rate of a component is directly related to its workload and develops a strategy of dynamic workload adjustment in order to on-line control the degradation processes of individual components, and thus to control their failure times. The criterion of selecting the optimal workloads is to prevent the overlap of component failures. We conduct a simulated case study to evaluate the performance of our proposed methodology under different conditions.

]]> Anita Race 1 1424263191 2015-02-18 12:39:51 1492118402 2017-04-13 21:20:02 0 0 event 2015-03-05T10:30:00-05:00 2015-03-05T10:30:00-05:00 2015-03-05T10:30:00-05:00 2015-03-05 15:30:00 2015-03-05 15:30:00 2015-03-05 15:30:00 2015-03-05T10:30:00-05:00 2015-03-05T10:30:00-05:00 America/New_York America/New_York datetime 2015-03-05 10:30:00 2015-03-05 10:30:00 America/New_York America/New_York datetime <![CDATA[]]>
<![CDATA[Seminar]]> 27187 TITLE: When Data Analytics Meets Promotion Planning

SPEAKER:  Georgia Perakis

ABSTRACT:

 In many important settings, promotions are a key instrument for driving sales and profits. Important examples include promotions in grocery retail among others. The Promotion Optimization Problem (POP) is a challenging problem as the retailer needs to decide which products to promote, what is the depth of price discounts, when to schedule the promotions and how to promote the product. In this talk we discuss our ongoing collaboration over the past few years with Oracle RGBU where we introduce and discuss how optimization can play a key role in determining promotion planning and capture several important business requirements for grocery retail.

An important consumer behavior that is a direct consequence of promotions in grocery retail is that consumers stockpile the products on promotion and then experience promotion fatigue after the promotion ends. Therefore, as a first step, we propose and study two general classes of demand functions that capture this effect and can be directly estimated from data. Using these demand functions, we model and study the promotion planning problem. Unfortunately, the underlying formulation even for a single product is NP-hard and highly nonlinear (with neither a concave nor a convex objective). We propose a linear approximation and by showing the integrality of the feasible region in the formulation, we are able to solve the problem efficiently as a linear programming (LP) problem. For the classes of demand we introduce, we develop analytical bounds on the accuracy of the LP relative to the optimal (but intractable) POP solution. We also consider a graphical representation of the problem which allows us to employ a Dynamic Programming (DP) solution approach as an alternative. We discuss the tradeoffs between the two approaches (LP vs DP).

Furthermore, we illustrate how our approach generalizes to consider multiple products within a category that are substitutes and/or complementary. We discuss the tradeoffs when there are cross product effects.

 Together with our industry collaborators from Oracle Retail, our framework allows us to develop a tool which can help supermarket managers to better understand promotions by testing various strategies and business constraints. We show that the formulation we propose solves fast in practice using actual data from a grocery retailer and that the accuracy is high. We calibrate our models using actual data and determine that they can improve profits by 3% just by optimizing the promotion schedule and up to 5% by slightly modifying some business requirements.

Joint work with Lennart Baardman (ORC PhD student), Maxime Cohen, (ORC PhD student), Swati Gupta (ORC PhD student), Jeremy Kalas (EECS Undergraduate), Zachary Leung (recently graduated ORC PhD student), Danny Segev (Visiting Scholar ORC/MIT Sloan from U. Haifa) as well as Kiran Panchamgam (Oracle RGBU) and Anthony Smith (formerly from Oracle RGBU)

BIO: Georgia Perakis is the William F. Pounds Professor at the Sloan School of Management at MIT since 1998. She received an M.S. degree and a PhD in Applied Mathematics from Brown University and a BA from the University of Athens in Greece.  

 Georgia Perakis' research interests include the role of operations  in many areas such as pricing, energy, supply chain management and transportation among others. She has widely published in journals such as Operations Research, Management Science, POM, Mathematics of Operations Research and Mathematical Programming among others. She has received the CAREER award from the National Science Foundation and subsequently, the PECASE award from the office of the President on Science and Technology awarded to the 50 top scientists and engineers in the nation. She has also received an honorable mention in the TSL Best Paper Award, the second prize in 2011 and the first prize in 2012 in the Best Paper competition of the Informs Service Science Section for two of her papers, the Graduate Student Council Teaching Award  as well as the Jamieson Prize for excellence in teaching and the Samuel M. Seegal prize for “inspiring students to pursue and achieve excellence”. Perakis was the recepient of the Sloan Career Development Chair and subsequently of the J. Spencer Standish Career Development Chair. In 2009, Perakis received the William F. Pounds chair that she currently holds. Perakis has passion supervising her students and builds lifelong relationships with them. So far she has graduated sixteen PhD and twenty two Masters students.

Perakis is currently serving as the co-director from the MIT Sloan School side for the Leaders for Global Operations (former LFM) program. She is also currently the group head of the Operations Management Group at MIT Sloan. She serves as an Associate Editor for the journals Management Science, Operations Research and Naval Logistics Research and a senior editor for POM. Perakis has served as a member of the INFORMS Council. She served as the chair of the Pricing and Revenue Management Section of INFORMS and in 2009-2010 as the VP for Meetings of the MSOM Society of INFORMS. She has co-organized the MSOM 2009 conference and served in the organizing committee of the 2010 MSOM conference. She has also been the co-chair and co-organizer of the Annual Conference of the INFORMS Section on Pricing and Revenue Management for several years and the chair of the cluster on the same topics for the annual INFORMS and ISMP conferences for several years.

]]> Anita Race 1 1424355724 2015-02-19 14:22:04 1492118400 2017-04-13 21:20:00 0 0 event 2015-03-06T11:00:00-05:00 2015-03-06T11:00:00-05:00 2015-03-06T11:00:00-05:00 2015-03-06 16:00:00 2015-03-06 16:00:00 2015-03-06 16:00:00 2015-03-06T11:00:00-05:00 2015-03-06T11:00:00-05:00 America/New_York America/New_York datetime 2015-03-06 11:00:00 2015-03-06 11:00:00 America/New_York America/New_York datetime <![CDATA[]]> Anton Kleywegt

anton@isye.gatech.edu

]]>
<![CDATA[ISyE Seminar]]> 27187 TITLE:  Pricing non-convexities in an electricity pool: An non-orthodox primal-dual approach

SPEAKER: Antonio Conejo

ABSTRACT:

Electricity pools are generally cleared through auctions that are conveniently formulated as mixed-integer linear programming problems. Since a mixed-integer linear programming problem is non-continuous and non-convex, marginal prices cannot be easily derived. However, to trade electricity, prices are needed. Thus, a relevant question arises: how to generate appropriate prices? This paper addresses this important issue and proposes a primal-dual approach to derive efficient revenue adequate uniform prices that guarantee that dispatched producers are willing to remain in the market. Such prices may not significantly deviate from the marginal prices obtained if integrality conditions are relaxed in the original mixed-integer linear programming problem.

 Speaker bio: Antonio J. Conejo, professor at The Ohio State University, OH, US, received the M.S. from MIT and the Ph.D. from the Royal Institute of Technology, Sweden. He has published over 150 papers in SCI journals and is the author or coauthor of books published by Springer, John Wiley, McGraw-Hill and CRC. He has been the principal investigator of many research projects financed by public agencies and the power industry and has supervised 18 PhD theses. He is the Editor-in-Chief of the IEEE Transactions on Power Systems and an IEEE Fellow.

]]> Anita Race 1 1424425879 2015-02-20 09:51:19 1492118399 2017-04-13 21:19:59 0 0 event 2015-02-26T11:00:00-05:00 2015-02-26T11:00:00-05:00 2015-02-26T11:00:00-05:00 2015-02-26 16:00:00 2015-02-26 16:00:00 2015-02-26 16:00:00 2015-02-26T11:00:00-05:00 2015-02-26T11:00:00-05:00 America/New_York America/New_York datetime 2015-02-26 11:00:00 2015-02-26 11:00:00 America/New_York America/New_York datetime <![CDATA[]]> Andy Sun

andy.sun@isye.gatech.edu

]]>
<![CDATA[Statistics Seminar]]> 27187 TITLE:  Propensity Score Estimation with Boosted Regression

SPEAKER:  Beth Ann Griffin

ABSTRACT:

The theory of propensity score analysis (PSA) is elegant. Provided strong ignorability holds, the single propensity score is all that is required to control for pretreatment differences between two treatment groups or a treatment and a control group. In practice, use of propensity scores is more complicated because the propensity score function and its functional form are unknown and must be estimated from the data. Logistic regression has been the standard approach to estimating propensity scores. In this talk, I will demonstrate the use of the generalized boosting model (GBM) as an alternative to logistic regression for estimating propensity scores. GBM is a machine learning approach used primarily for predicting dichotomous outcomes. It combines many simple regression trees to provide a smooth and flexible propensity score model. It automatically conducts variable and feature selection as part of its iterative estimation procedure. Tools for implementing these methods are available in R, SAS, and Stata. I will also summarize recent comparisons of GBM to alternative methods for propensity score estimation, namely the covariate balance propensity scores (CBPS) estimation methods of Imai and Ratkovic (2014). I will contrast the methods in terms of covariate balance, and the bias and mean squared error of the treatment effects estimated by propensity score weighting. CBPS generally outperforms GBM in terms of covariate balance and bias in the absence of the need of non-linear transformation of the covariates. However, we find that in terms of mean squared error, GBM appears to be advantageous in the commonly encountered situation of propensity score model building in the presence of many candidate confounders, some of which may not actually be related to the outcomes of interest.

]]> Anita Race 1 1424428396 2015-02-20 10:33:16 1492118399 2017-04-13 21:19:59 0 0 event 2015-02-25T11:00:00-05:00 2015-02-25T11:00:00-05:00 2015-02-25T11:00:00-05:00 2015-02-25 16:00:00 2015-02-25 16:00:00 2015-02-25 16:00:00 2015-02-25T11:00:00-05:00 2015-02-25T11:00:00-05:00 America/New_York America/New_York datetime 2015-02-25 11:00:00 2015-02-25 11:00:00 America/New_York America/New_York datetime <![CDATA[]]> Ben Haaland

bhaaland3@gatech.edu

]]>
<![CDATA[PhD Thesis Defense]]> 27187 TITLE: A Modeling Framework for Analyzing the Education System as a Complex System

STUDENT: Pratik Mital

ABSTRACT:

In this thesis, the Education System Intervention Modeling Framework (ESIM Framework), is introduced for analyzing interventions in the K-12 education system. This framework is the first of its kind to model interventions in the K-12 school system in the United States. Techniques from systems engineering and operations research, such as agent-based modeling and social network analysis, are used to model the bottom-up mechanisms of intervention implementation in schools. By applying the ESIM framework, an intervention can be better analyzed in terms of the barriers and enablers to intervention implementation and sustainability.  The risk of failure of future interventions is thereby reduced through improved allocation of resources towards the system agents and attributes which play key roles in the sustainability of the intervention.


In the first part of this thesis, a case study is modeled which helped in the development of the framework. This case study was of an extracurricular school intervention, an Engineers Without Borders chapter, implemented in a magnet school setting through a partnership with Georgia Institute of Technology, Atlanta (Georgia Tech) as part of a National Science Foundation (NSF) GK-12 grant. This case study is ideal for the development of the framework and as its first application because it had two different outcomes over two different years, which helped in developing insights about the success of this intervention. Also, the scale of this intervention was small enough to test the development and application of the framework. With the help of this case study, a more generalized framework is developed which is applicable across a broad range of education system interventions.
In the second part of this thesis, the ESIM framework is developed. The framework developed is divided into four phases: model definition, model design, model analysis, and model validation. In the model definition phase, the overview of the problem to be modeled is documented. Then, detailed descriptions about the agents, attributes, and the environment being modeled are provided. Other modeling decisions, such as scale and time horizons, are also made in this phase. Finally, the criteria for a sustainable intervention is defined along with a method to analyze the risk of implementing this intervention in the particular school system. In the model design phase, the conceptual model is built using agent-based modeling, social network analysis, and discrete-time Markov chains. Then, the conceptual model is validated with the help of subject matter experts (SMEs) using Pace’s 4C’s framework for conceptual model validation. After that, the computer simulation model is implemented and verified. In the model analysis phase, simulation results are generated and analyzed. While simulating outcomes that are consistent with reality is helpful, the real contributions of this framework are two-fold: the sensitivity analysis of the model, and the determination of factors that are likely to affect the intervention outcomes.  The latter is accomplished using the Method of Morris, a factorial sampling technique. Finally, in the model validation phase, verification and validation techniques are applied. This step is critical in developing confidence in the model amongst its users.


In the third part of this thesis, the ESIM framework is applied to a case study of a curriculum intervention, Science Learning: Integrating Design, Engineering and Robotics, involving the design and implementation of an 8th-grade, inquiry-based physical science curriculum across three demographically varying schools. This intervention was also implemented in collaboration with Georgia Tech as part of an NSF DRK-12 grant. This was a five year intervention from Sep, 2009 to Oct, 2014. This case study provides a good comparison of the implementation of the intervention across different school settings because of the varied outcomes at the three schools.

]]> Anita Race 1 1424771483 2015-02-24 09:51:23 1492118398 2017-04-13 21:19:58 0 0 event 2015-03-09T09:30:00-04:00 2015-03-09T09:30:00-04:00 2015-03-09T09:30:00-04:00 2015-03-09 13:30:00 2015-03-09 13:30:00 2015-03-09 13:30:00 2015-03-09T09:30:00-04:00 2015-03-09T09:30:00-04:00 America/New_York America/New_York datetime 2015-03-09 09:30:00 2015-03-09 09:30:00 America/New_York America/New_York datetime <![CDATA[]]>
<![CDATA[DOS Seminar]]> 27187 Title: A Polynomial Time Algorithm to Solve a Class of Optimization Problems with a Multi-linear Objective Function and Affine Constraints


Speaker: Hadi Charkhgard
Abstract: 

In this talk, I will present the first polynomial-time linear programming based algorithm for a class of optimization problems with a multi-linear objective function and affine constraints. This class of optimization problems arises naturally in a number of settings in game theory, such as the bargaining problem, linear Fisher markets, and Kelly capacity allocation markets, but has applications in other fields of study as well. The algorithm computes an optimal solution by solving at most O(p^3) linear programs, where p is the number of variables in the multi-linear objective function.

]]> Anita Race 1 1424785874 2015-02-24 13:51:14 1492118396 2017-04-13 21:19:56 0 0 event 2015-02-27T13:00:00-05:00 2015-02-27T14:00:00-05:00 2015-02-27T14:00:00-05:00 2015-02-27 18:00:00 2015-02-27 19:00:00 2015-02-27 19:00:00 2015-02-27T13:00:00-05:00 2015-02-27T14:00:00-05:00 America/New_York America/New_York datetime 2015-02-27 01:00:00 2015-02-27 02:00:00 America/New_York America/New_York datetime <![CDATA[]]>
<![CDATA[DOS Seminar - Andres Iroume]]> 27187 TITLE: Some Lower Bounds on Sparse Outer Approximations of Polytopes

SPEAKER: Andres Iroume

ABSTRACT:

Trying to understand the use of sparse cutting-planes in integer programming solvers, a recent paper by Dey, Molinaro and Wang studied how well polytopes are approximated by using only sparse valid-inequalities. In this talk, we consider "less-idealized" questions such as: effect of sparse inequalities added to linear-programming relaxation, effect on approximation by addition of a budgeted number of dense valid-inequalities, sparse-approximation of polytope under every rotation and approximation by sparse inequalities in specific directions.

]]> Anita Race 1 1425310941 2015-03-02 15:42:21 1492118395 2017-04-13 21:19:55 0 0 event 2015-03-04T12:00:00-05:00 2015-03-04T12:00:00-05:00 2015-03-04T12:00:00-05:00 2015-03-04 17:00:00 2015-03-04 17:00:00 2015-03-04 17:00:00 2015-03-04T12:00:00-05:00 2015-03-04T12:00:00-05:00 America/New_York America/New_York datetime 2015-03-04 12:00:00 2015-03-04 12:00:00 America/New_York America/New_York datetime <![CDATA[]]>
<![CDATA[FREE Material Handling 101 webinar, Thursday, March 12 1:30 - 2:30 pm EDT!]]> 27233 Attend this free webinar and receive a *discount code towards our upcoming Material Handling 101: Fundamentals, Analysis and Selection course! REGISTER at http://bit.ly/mh101webinar.

OVERVIEW
During the webinar, you will be introduced to a unique, step-by-step, systematic approach to making material handling decisions. This method is an ideal way for someone who doesn’t have any formal material handling training to get up to speed.

Join Lee Hales as he touches upon the below issues:

This is your opportunity to get introduced to the basics!

Attendees will receive a discount* to the course, “Material Handling 101: Fundamentals, Analysis and Selection,” April 8-9, 2015 in Atlanta. (*Webinar discount cannot be combined with any other discount.) 


ABOUT THE PRESENTER

H. Lee Hales is President of Richard Muther & Associates and instructor for SCL's Material Handling 101 course. He is author and co-author of several books on industrial planning, including: Systematic Planning of Industrial Facilities (with Richard Muther) and Computer-Aided Facilities Planning. As a materials and distribution manager, Mr. Hales planned, set up and managed a network of distribution facilities in the oil field equipment industry and as a consultant, Mr. Hales has assisted on a wide range of projects in the US, Asia, Europe, and South America.

]]> Andy Haleblian 1 1425320589 2015-03-02 18:23:09 1492118393 2017-04-13 21:19:53 0 0 event Attend this free webinar and be introduced to a unique, step-by-step, systematic approach to making material handling decisions.

]]>
2015-03-12T14:30:00-04:00 2015-03-12T15:30:00-04:00 2015-03-12T15:30:00-04:00 2015-03-12 18:30:00 2015-03-12 19:30:00 2015-03-12 19:30:00 2015-03-12T14:30:00-04:00 2015-03-12T15:30:00-04:00 America/New_York America/New_York datetime 2015-03-12 02:30:00 2015-03-12 03:30:00 America/New_York America/New_York datetime <![CDATA[]]> webinar@scl.gatech.edu]]> 383521 383521 image <![CDATA[FREE Material Handling 101 webinar, Thursday, March 12 1:30 - 2:30 pm EDT!]]> image/jpeg 1449246246 2015-12-04 16:24:06 1475894393 2016-10-08 02:39:53 <![CDATA[Register Online to Attend]]> <![CDATA[Material Handling 101: Fundamentals, Analysis and Selection (MODEX)]]>
<![CDATA[PhD Thesis Defense - Rodrigue Ngueyep]]> 27187 TITLE: Model Selection and Estimation in High Dimensional Settings

STUDENT:  Rodrigue Ngueyep Tzoumpe

Several statistical problems can be described as estimation problem, where the goal is to learn a set of parameters, from some data, by maximizing a criterion. These type of problems are typically encountered in a supervised learning setting, where we want to relate an output (or many outputs) to multiple inputs. The relationship between these outputs and these inputs can be complex, and this complexity can be attributed to the high dimensionality of the space containing the inputs and the outputs; the existence of a structural prior knowledge within the inputs or the outputs that if ignored may lead to inefficient estimates of the parameters; and the presence of a non-trivial noise structure in the data.

In Chapter 2, we study one of the most commonly used multivariate time series model, the Vector Autoregressive Model (VAR). VAR is generally used to identify lead, lag and contemporaneous relationships describing Granger causality within and between time series. In this chapter, we investigate VAR methodology for analyzing data consisting of multi-layer time series which are spatially interdependent. When modeling VAR relationships for such data, the dependence between time series is both a curse and a blessing. The former because it requires modeling the between time series correlation and the contemporaneous relationships which may be challenging when using likelihood-based methods. The latter because the spatial correlation structure can be used to specify the lead-lag relationships within and between time series, within and between layers. To address these challenges, we propose a L1\L2 regularized likelihood estimation method. The lead, lag and contemporaneous relationsh ips are estimated using a new coordinate descent algorithm that exploits sparsity in the VAR structure, accounts for the spatial dependence and models the error dependence. We assess the performance of the proposed VAR model and compare it with existing methods within a simulation study. We also apply the proposed methodology to a large number of state-level US economic time series.

In the third chapter, we propose a new methodology to tackle the problem of high-dimensional nonparametric learning in the multi-responses or multitask learning setting. We impose sparsity constraints that allow the recovery of the additive functions that are the most influential accross tasks and responses. The methodology applies a functional L1\L2 norm to each group of additive functions. Each group contains all the additive functions associated with a specific predictor. We derive a novel thresholding condition for the union support recovery in the nonparametric setting. A new functional block coordinate descent algorithm is developed to solve for all the additive functions. By applying the methodology to a benchmark cancer data set, we are able to perfectly classify 83 cancer patients to 4 distinct cancer categories by using only 12 out of 2308 genes. The method is also used to uncover the determinants of health that drive the county level cost of care in the state of No rth Carolina from 2005 to 2009.

Motivated by the analysis of a Positron Emission Tomography (PET) imaging data considered in Bowen et al. (2012), we introduce in chapter 4, a semiparametric topographical mixture model able to capture the characteristics of dichotomous shifted response-type experiments. We propose a pointwise estimation procedure of the proportion and location functions involved in our model. Our estimation procedure is only based on the symmetry of the local noise and does not require any finite moments on the errors (e.g. Cauchy-type errors). We establish under mild conditions minimax properties and asymptotic normality of our estimators. Moreover, Monte Carlo simulations are conducted to examine their finite sample performance. Finally a statistical analysis of the PET imaging data in Bowen et al. (2012) is illustrated for the proposed method.

]]> Anita Race 1 1425567897 2015-03-05 15:04:57 1492118391 2017-04-13 21:19:51 0 0 event 2015-03-20T15:00:00-04:00 2015-03-20T15:00:00-04:00 2015-03-20T15:00:00-04:00 2015-03-20 19:00:00 2015-03-20 19:00:00 2015-03-20 19:00:00 2015-03-20T15:00:00-04:00 2015-03-20T15:00:00-04:00 America/New_York America/New_York datetime 2015-03-20 03:00:00 2015-03-20 03:00:00 America/New_York America/New_York datetime <![CDATA[]]>
<![CDATA[Statistics Seminar - Le Song]]> 27187 TITLE: Scalable Kernel Methods via Doubly Stochastic Gradients

SPEAKER:  Le Song

ABSTRACT:

Kernel methods are widely used in many machine learning models, such as the support vector machines and Gaussian processes. However, the general perception is that kernel methods are not scalable, and neural networks are the methods of choice for large scale nonlinear learning problems. Or we simply have not tried hard enough for kernel methods? Here we propose an approach that scales up kernel methods using a novel concept called "doubly stochastic functional gradients". Our approach relies on the fact that many kernel methods can be expressed as convex optimization problems, and we solve the problems by making two unbiased stochastic approximations to the functional gradient, one using random training points and another using random functions associated with a positive definite kernel function, and then descending using this noisy functional gradient. We show that a function estimated by this procedure after t iterations converges to the optimal function in rate O(t^-1), and achieves a generalization guarantee of O(t^-1/2). This doubly stochasticity also allows us to avoid keeping the support vectors and to implement the algorithm in a small memory footprint, which is linear in number of iterations and independent of data dimension. Our approach can readily scale kernel methods up to the regimes which are dominated by neural networks. We show that our method can achieve competitive performance to neural networks in problems such as classifying 8 million handwritten digits from MNIST, regressing 2.3 million energy materials from MolecularSpace, and categorizing 1 million photos from ImageNet.

Short Bio: 
Le Song is an assistant professor in the Department of Computational Science and Engineering, College of Computing, Georgia Institute of Technology. He received his Ph.D. in Computer Science from University of Sydney and NICTA in 2008, and then conducted his post-doctoral research in the School of Computer Science, Carnegie Mellon University, between 2008 and 2011. Before he joined Georgia Institute of Technology, he worked briefly as a research scientist at Google. His principal research interests lie in nonparametric and kernel methods, probabilistic graphical models, spatial/temporal dynamics of networked processes, and the applications of machine learning to interdisciplinary problems. He is the recipient of NSF CAREER Award 2014, IPDPS'15 Best Paper Award, NIPS’13 Outstanding Paper Award and ICML’10 Best Paper Award.

]]> Anita Race 1 1425568026 2015-03-05 15:07:06 1492118391 2017-04-13 21:19:51 0 0 event 2015-03-10T13:00:00-04:00 2015-03-10T13:00:00-04:00 2015-03-10T13:00:00-04:00 2015-03-10 17:00:00 2015-03-10 17:00:00 2015-03-10 17:00:00 2015-03-10T13:00:00-04:00 2015-03-10T13:00:00-04:00 America/New_York America/New_York datetime 2015-03-10 01:00:00 2015-03-10 01:00:00 America/New_York America/New_York datetime <![CDATA[]]>
<![CDATA[PhD Thesis Defense - Matthew Plumlee]]> 27187 TITLE: Fast methods for identifying high dimensional systems using observations

ABSTRACT:

Computational modeling is a popular tool to understand a diverse set of complex systems. The output from a computational model depends on a set of parameters which are unknown to the designer, but a modeler can estimate them by collecting physical data. In the second chapter of this thesis, we study the action potential of ventricular myocytes and our parameter of interest is a function as opposed to a scalar or a set of scalars. We develop a new modeling strategy to nonparametrically study the functional parameter using Bayesian inference with Gaussian process priors. We also devise a new Markov chain Monte Carlo sampling scheme to address this unique problem. In the more general case, computational simulation is expensive. Emulators avoid the repeated use of a stochastic simulation by performing a designed experiment on the computer simulation and developing a predictive distribution.  Random field models are considered the standard in analysis of computer experiments, but the current framework fails in high dimensional scenarios because of the cost of inference. The third chapter of this thesis shows by using a class of experimental designs, the computational cost of inference from random fields scales significantly better in high dimensions. Exact prediction and likelihood evaluation with close to half a million design points is possible in seconds using only a laptop computer. Compared to the more common space-filling designs, the proposed designs are shown to be competitive in terms of prediction accuracy through simulation and analytic results. The fourth chapter of this thesis proposes a method to construct an emulator for a stochastic simulation. Existing emulators have focused on estimation of the mean of the simulation output, but this work presents an emulator for the distribution of the output in a nonparametric setting. This construction provides both an explicit distribution and a fast sampling scheme. Beyond describing the emulator, this work demonstrates that the emulator's convergence rate is asymptotically rate optimal among all possible emulators using the same sample size.  Lastly, the fifth chapter of this work investigates the use of a modified version of the above method to study patterns of defects on products. We achieve efficient inference on the defect patterns by developing a novel estimate of an inhomogeneous point process that is both computationally tractable and asymptotically appealing. ]]> Anita Race 1 1425886422 2015-03-09 07:33:42 1492118389 2017-04-13 21:19:49 0 0 event 2015-03-26T11:30:00-04:00 2015-03-26T11:30:00-04:00 2015-03-26T11:30:00-04:00 2015-03-26 15:30:00 2015-03-26 15:30:00 2015-03-26 15:30:00 2015-03-26T11:30:00-04:00 2015-03-26T11:30:00-04:00 America/New_York America/New_York datetime 2015-03-26 11:30:00 2015-03-26 11:30:00 America/New_York America/New_York datetime <![CDATA[]]>
<![CDATA[DOS Seminar - Murat Yildirim]]> 27187 TITLE: Sensor Driven Condition Based Maintenance Models for Generator Fleets

ABSTRACT:

We provide a framework that links low-level performance and condition monitoring data with high-level operational and maintenance decisions for generators. The operational decisions identify the optimal commitment and dispatch to satisfy demand and transmission constraints. Maintenance decisions focus on arriving at an optimal condition based maintenance (CBM) schedule that accounts for optimal asset-specific CBM schedules driven by the condition monitoring data. We propose new mixed-integer optimization models and efficient algorithms that exploit the special structure of the proposed formulation. We present extensive computational experiment results to show proposed models achieve significant improvements in cost and reliability. This is a joint work with Andy Sun and Nagi Gebraeel.
]]> Anita Race 1 1425886555 2015-03-09 07:35:55 1492118389 2017-04-13 21:19:49 0 0 event 2015-03-11T13:00:00-04:00 2015-03-11T13:00:00-04:00 2015-03-11T13:00:00-04:00 2015-03-11 17:00:00 2015-03-11 17:00:00 2015-03-11 17:00:00 2015-03-11T13:00:00-04:00 2015-03-11T13:00:00-04:00 America/New_York America/New_York datetime 2015-03-11 01:00:00 2015-03-11 01:00:00 America/New_York America/New_York datetime <![CDATA[]]>
<![CDATA[PhD Thesis Defense - Liujia Hu]]> 27187 TITLE: Convergent Algorithms in Simulation Optimization

ABSTRACT:

It is frequently the case that deterministic optimization models could be made more practical by explicitly incorporating uncertainty. The resulting stochastic optimization problems are in general more difficult to solve than their deterministic counterparts, because the objective function cannot be evaluated exactly and/or because there is no explicit relation between the objective function and the corresponding decision variables. This thesis develops random search algorithms for solving optimization problems with continuous decision variables when the objective function values can be estimated with some noise via simulation. Our algorithms will maintain a set of sampled solutions, and use simulation results at these solutions to guide the search for better solutions.

In the first part of the thesis, we propose an Adaptive Search with Resampling and Discarding (ASRD) approach for solving continuous stochastic optimization problems. Our ASRD approach is a framework for designing provably convergent algorithms that are adaptive both in seeking new solutions and in keeping or discarding already sampled solutions. The framework is an improvement over the Adaptive Search with Resampling (ASR) method of Andradottir and Prudius in that it spends less effort on inferior solutions (the ASR method does not discard already sampled solutions). We present conditions under which the ASRD method is convergent almost surely and carry out numerical studies aimed at comparing the algorithms. Moreover, we show that whether it is beneficial to resample or not depends on the problem, and analyze when resampling is desirable. Our numerical results show that the ASRD approach makes substantial improvements on ASR, especially for difficult problems with large numbers of local optima.

In traditional simulation optimization problems, noise is only involved in the objective functions. However, many real world problems involve stochastic constraints. Such problems are more difficult to solve because of the added uncertainty about feasibility. The second part of the thesis presents an Adaptive Search with Discarding and Penalization (ASDP) method for solving continuous simulation optimization problems involving stochastic constraints. Rather than addressing feasibility separately, ASDP utilizes the penalty function method from deterministic optimization to convert the original problem into a series of simulation optimization problems without stochastic constraints. We present conditions under which the ASDP algorithm converges almost surely from inside feasible region, and under which it converges to the optimal solution but without feasibility guarantee. We also conduct numerical studies aimed at assessing the efficiency and tradeoff under the two different convergence modes.

Finally, in the third part of the thesis, we propose a random search method named Gaussian Search with Resampling and Discarding (GSRD) for solving simulation optimization problems with continuous decision spaces. The method combines the ASRD framework with a sampling distribution based on a Gaussian process that not only utilizes the current best estimate of the optimal solution but also learns from past sampled solutions and their objective function observations. We prove that our GSRD algorithm converges almost surely, and carry out numerical studies aimed at studying the effects of utilizing the Gaussian sampling strategy. Our numerical results show that the GSRD framework performs well when the underlying objective function is multi-modal. However, it takes much longer to sample solutions, especially in higher dimensions.

]]> Anita Race 1 1427100360 2015-03-23 08:46:00 1492118382 2017-04-13 21:19:42 0 0 event 2015-03-27T14:00:00-04:00 2015-03-27T16:00:00-04:00 2015-03-27T16:00:00-04:00 2015-03-27 18:00:00 2015-03-27 20:00:00 2015-03-27 20:00:00 2015-03-27T14:00:00-04:00 2015-03-27T16:00:00-04:00 America/New_York America/New_York datetime 2015-03-27 02:00:00 2015-03-27 04:00:00 America/New_York America/New_York datetime <![CDATA[]]>
<![CDATA[PhD Thesis Defense - Cristobal Guzman]]> 27187 TITLE: Information, Complexity and Structure in Convex Optimization

ABSTRACT:

This thesis is focused on the limits of performance of large-scale convex optimization algorithms. Classical theory of oracle complexity, first proposed by Nemirovski and Yudin in 1983, successfully established the worst-case behavior of methods based on local oracles (a generalization of first-order oracle for smooth functions) for nonsmooth convex minimization, both in the large-scale and low-scale regimes; and the complexity of approximately solving linear systems of equations (equivalent to convex quadratic minimization) over Euclidean balls, under a matrix-vector multiplication oracle.

Our work extends the applicability of lower bounds in two directions:

Worst-Case Complexity of Large-Scale Smooth Convex Optimization: We generalize lower bounds on the complexity of first-order methods for convex optimization, considering classes of convex functions with Holder continuous gradients. Our technique relies on the existence of a smoothing kernel, which defines a smooth approximation for any convex function via infimal convolution. As a consequence, we derive lower bounds for ell_p/ell_q-setups, where 1 <= p,q <= \infty, and extend to its matrix analogue: Smooth (w.r.t. Schatten q-norm) convex minimization over matrices with bounded Schatten p-norm.

The major consequences of this result are the near-optimality of the Conditional Gradient method over box-type domains (p=q=\infty), and the near-optimality of Nesterov's accelerated method over the cross-polytope (p=q=1).

The thesis is available for public inspection in the School of
Mathematics lounge (Skiles 236), the ARC lounge (Klaus 2222), the ISyE
PhD student lounge, and at the URL

        http://www.aco.gatech.edu/dissert/guzman.html

]]> Anita Race 1 1427210532 2015-03-24 15:22:12 1492118381 2017-04-13 21:19:41 0 0 event 2015-03-30T10:00:00-04:00 2015-03-30T10:00:00-04:00 2015-03-30T10:00:00-04:00 2015-03-30 14:00:00 2015-03-30 14:00:00 2015-03-30 14:00:00 2015-03-30T10:00:00-04:00 2015-03-30T10:00:00-04:00 America/New_York America/New_York datetime 2015-03-30 10:00:00 2015-03-30 10:00:00 America/New_York America/New_York datetime <![CDATA[]]>
<![CDATA[ISyE Seminar - Huseyin Topaloglu]]> 27187 TITLE: Revenue Management Under Markov Chain Choice Model

ABSTRACT:

A recent choice model is based on modeling the customer choice process through a Markov chain. In this choice model, a customer arrives into the system with the intention of purchasing a particular product. If this product is available for purchase, then the customer purchases it and leaves the system. Otherwise, the customer transitions to another product according to a Markov chain and considers purchasing the other product. In this way, the customer transitions between the products until she reaches a product that is available for purchase or she decides to leave the system without purchasing anything. In this talk, we consider revenue management problems when customers choose according to the Markov chain choice model. For single-leg revenue management, we study the dynamic programming formulation of the problem. We show that the efficient offer sets are nested and the optimal policy can be characterized by nested protection levels. For network revenue management, we study a deterministic linear program that offers each subset of products with a certain probability. In the deterministic linear program, there is one decision variable for each subset of products. Thus, the number of decision variables grows exponentially fast with the number of products, and it is common to solve the deterministic linear program through column generation. We show that if the customers choose according to the Markov chain choice model, then the deterministic linear program can immediately be reduced to an equivalent one whose numbers of decision variables and constraints grow only linearly with the number of products. (This work is joint with Jacob Feldman.)

 

]]> Anita Race 1 1427210686 2015-03-24 15:24:46 1492118381 2017-04-13 21:19:41 0 0 event 2015-03-31T12:00:00-04:00 2015-03-31T13:00:00-04:00 2015-03-31T13:00:00-04:00 2015-03-31 16:00:00 2015-03-31 17:00:00 2015-03-31 17:00:00 2015-03-31T12:00:00-04:00 2015-03-31T13:00:00-04:00 America/New_York America/New_York datetime 2015-03-31 12:00:00 2015-03-31 01:00:00 America/New_York America/New_York datetime <![CDATA[]]> Alejandro Toriello

atoriello@isye.gatech.edu

]]>
<![CDATA[ISyE Seminar - Tara Javidi]]> 27187 TITLE: Size-dependent Noisy Search as a Problem of Information Acquisition

ABSTRACT:

Information acquisition problems form a class of stochastic decision problems in which a decision maker, by carefully controlling a sequence of actions with uncertain outcomes, dynamically refines the belief about a time-varying (Markov) parameter of interest. Examples arise in patient care, computer vision, spectrum utilization, and joint source--channel coding. In the first part of the talk, we consider this generalization of hidden Markov models (HMMs), the corresponding dynamic program, and provide some structural results. 

 In the second part of the talk, as a special case of information acquisition, we consider the problem of noisy search with size-dependent noise. We connect De Groot's "information utility" framework with the Shannon theoretic concept of "uncertainty reduction" to introduce a symmetrized divergence measure: Extrinsic Jensen-Shannon (EJS) divergence.  We use this divergence to provide (tight) lower and upper bounds on the optimal performance and  strengthen Chernoff's analysis to account for the resolution of the search. These bounds, as a corollary, provide the (asymptotic) performance gain of adaptive search strategies over the non-adaptive (open loop) and non-sequential ones.  This is joint work with Anusha Lalitha, Mohammad Naghshvar, Yonatan Kaspi, and Ofer Shayevitz. Bio: Tara Javidi studied electrical engineering at Sharif University of Technology, Tehran, Iran from 1992 to 1996. She received her MS degrees in electrical engineering (systems), and in applied mathematics (stochastics) from the University of Michigan, Ann Arbor, in 1998 and 1999, respectively. She received her Ph.D. in electrical engineering and computer science from the University of Michigan, Ann Arbor, in 2002. From 2002 to 2004, she was an assistant professor at the Electrical Engineering Department, University of Washington, Seattle. She joined University of California, San Diego, in 2005, where she is currently an associate professor of electrical and computer engineering.]]> Anita Race 1 1427284475 2015-03-25 11:54:35 1492118379 2017-04-13 21:19:39 0 0 event 2015-03-26T12:00:00-04:00 2015-03-26T12:00:00-04:00 2015-03-26T12:00:00-04:00 2015-03-26 16:00:00 2015-03-26 16:00:00 2015-03-26 16:00:00 2015-03-26T12:00:00-04:00 2015-03-26T12:00:00-04:00 America/New_York America/New_York datetime 2015-03-26 12:00:00 2015-03-26 12:00:00 America/New_York America/New_York datetime <![CDATA[]]>
<![CDATA[Statistics Seminar - Jun Zhang]]> 27187 TITLE:  Bayesian Change-Detection with Identification: Sequential Algorithm and Neuronal Evidence

ABSTRACT:

How do animal and people combine prior knowledge with evidence gathered on a moment-to-moment evidence for decision-making has been well-studied in psychology using choice-reaction time paradigm. Here, we consider the problem of change detection along with identification in multi-hypotheses setting, assuming a known prior. A Bayesian sequential 
updating algorithm is proposed, along with the threshold-crossing stopping rule common in sequential analysis. The value of an absorbing boundary is shown to exactly equal the hit rate of a decision-maker conditioned on that response. Computer simulation reveals that the 
algorithm shares many similarities with human performance in stimulus detection/identification experiments. Given recent evidence from neuroscience in support of sequential analysis algorithm, we report a re-analysis of  the neuronal data recorded by Roitman and Shadlen (2002) in the monkey’s lateral intraparietal cortex (LIP). Results show that neuronal activity accumulates during each trial up until monkey’s behavioral response, that the accumulation (“buildup”) rate is monotonically related to the strength of the stimulus, and that buildup activities encode intended movement rather than sensory information.

Short Bio:
Dr. Jun Zhang is a Professor of Psychology and Professor of Mathematics at the University of Michigan, Ann Arbor. He received the B.Sc. degree in Theoretical Physics from Fudan University in 1985, and Ph.D. degree in Neurobiology from the University of California, Berkeley in 1992. He has also held visiting positions at the University of Melbourne, the 
University of Waterloo, and RIKEN Brain Science Institute. During 2007-2010, he worked as the Program Manager for the U.S. Air Force Office of Scientific Research (AFOSR) in charge of the basic research portfolio for Cognition and Decision in the Directorate of Mathematics, 
Information, and Life Sciences. Dr. Zhang served as the President for the Society for Mathematical Psychology (SMP) and serves on the Federation of Associations in Brain and Behavioral Sciences (FABBS). He is Associate Editor for the Journal of Mathematical Psychology, and a Fellow of the Association for Psychology Sciences (APS). Dr. Zhang’s 
publications span the fields of vision, mathematical psychology, cognitive psychology, cognitive neuroscience, game theory, machine learning, information geometry, etc. His research has been funded by the National Science Foundation (NSF), Air Force Office for Scientific Research (AFOSR), and Army Research Office (ARO).

]]> Anita Race 1 1427365883 2015-03-26 10:31:23 1492118379 2017-04-13 21:19:39 0 0 event 2015-04-02T12:00:00-04:00 2015-04-02T12:00:00-04:00 2015-04-02T12:00:00-04:00 2015-04-02 16:00:00 2015-04-02 16:00:00 2015-04-02 16:00:00 2015-04-02T12:00:00-04:00 2015-04-02T12:00:00-04:00 America/New_York America/New_York datetime 2015-04-02 12:00:00 2015-04-02 12:00:00 America/New_York America/New_York datetime <![CDATA[]]>
<![CDATA[PhD Thesis Defense - Sue Reynolds]]> 27187 TITLE:  Statistical Estimation and Changepoint Detection Methods in Public Health Surveillance

ABSTRACT:

This thesis focuses on assessing and improving statistical methods implemented in two areas of public health research.  The first topic involves estimation of national influenza-associated mortality rates via mathematical modeling.  The second topic involves the timely detection of infectious disease outbreaks using statistical process control monitoring.

For over fifty years, the Centers for Disease Control and Prevention has been estimating annual rates of U.S. deaths attributable to influenza.  These estimates have been used to determine costs and benefits associated with influenza prevention and control strategies.  Quantifying the effect of influenza on mortality, however, can be challenging since influenza infections typically are not confirmed virologically nor specified on death certificates.  Consequently, a wide range of ecologically based, mathematical modeling approaches have been applied to specify the association between influenza and mortality.  To date, all influenza-associated death estimates have been based on mortality data first aggregated at the national level and then modeled.  Unfortunately, there are a number of local-level seasonal factors that may confound the association between influenza and mortality thus suggesting that data be modeled at the local level and then pooled to make national estimates of death. 

The first component of the thesis topic involving mortality estimation addresses this issue by introducing and implementing a two-stage hierarchical Bayesian modeling approach.  In the first stage, city-level data with varying trends in mortality and weather were modeled using semi-parametric, generalized additive models.  In the second stage, the log-relative risk estimates calculated for each city in stage 1 represented the “outcome” variable, and were modeled two ways: (1) assuming spatial independence across cities using a Bayesian generalized linear model, and (2) assuming correlation among cities using a Bayesian spatial correlation model.  Results from these models were compared to those from a more-conventional approach.

The second component of this topic examines the extent to which seasonal confounding and collinearity affect the relationship between influenza and mortality at the local (city) level.  Disentangling the effects of temperature, humidity, and other seasonal confounders on the association between influenza and mortality is challenging since these covariates are often temporally collinear with influenza activity.  Three modeling strategies with varying representations of background seasonality were compared.  Seasonal covariates entered into the model may have been measured (e.g., ambient temperature) or unmeasured (e.g., time-based smoothing splines or Fourier terms).  An advantage of modeling background seasonality via time splines is that the amount of seasonal curvature can be controlled by the number of degrees of freedom specified for the spline.  A comparison of the effects of influenza activity on mortality based on these varying representations of seasonal confounding is assessed. 

The third component of this topic explores the relationship between mortality rates and influenza activity using a flexible, natural cubic spline function to model the influenza term.  The conventional approach of fitting influenza-activity terms linearly in regression was found to be too constraining.  Results show that the association is best represented nonlinearly.

The second area of focus in this thesis involves infectious disease outbreak detection.  A fundamental goal of public health surveillance, particularly syndromic surveillance, is the timely detection of increases in the rate of unusual events.  In syndromic surveillance, a significant increase in the incidence of monitored disease outcomes would trigger an alert, possibly prompting the implementation of an intervention strategy.  Public health surveillance generally monitors count data (e.g., counts of influenza-like illness, sales of over-the-counter remedies, and number of visits to outpatient clinics).  Statistical process control charts, designed for quality control monitoring in industry, have been widely adapted for use in disease and syndromic surveillance.  The behavior of these detection methods on discrete distributions, however, has not been explored in detail.  

For this component of the thesis, a simulation study was conducted to compare the CuSum and EWMA methods for detection of increases in negative binomial rates with varying amounts of dispersion.  The goal of each method is to detect an increase in the mean number of cases as soon as possible after an upward rate shift has occurred.  The performance of the CuSum and EWMA detection methods is evaluated using the conditional expected delay criterion, which is a measure of the detection delay, i.e., the time between the occurrence of a shift and when that shift is detected.  Detection capabilities were explored under varying shift sizes and times at which the shifts occurred.

]]> Anita Race 1 1427380711 2015-03-26 14:38:31 1492118378 2017-04-13 21:19:38 0 0 event 2015-04-01T10:00:00-04:00 2015-04-01T10:00:00-04:00 2015-04-01T10:00:00-04:00 2015-04-01 14:00:00 2015-04-01 14:00:00 2015-04-01 14:00:00 2015-04-01T10:00:00-04:00 2015-04-01T10:00:00-04:00 America/New_York America/New_York datetime 2015-04-01 10:00:00 2015-04-01 10:00:00 America/New_York America/New_York datetime <![CDATA[]]>
<![CDATA[Statistics Seminar - Douglas M. Ray]]> 27187 TITLE: Applications of Sensitivity Testing Methods to US Army Technologies

ABSTRACT:

This presentation will discuss some applications of sensitivity testing to US Army products and systems, analytical approaches, and some common issues encountered in this unique approach to testing.  

The US Army Armament Research, Development and Engineering Center is located at Picatinny Arsenal, in northern NJ.  The ARDEC strives to support the Army's efforts to ensure Soldier survivability and enhance platform and area protection by providing engineering, design and development support. This support is essential to the rapid delivery of critical technologies to our Warfighters.  ARDEC's Statistics Group's applied statisticians provide project consultation and support on test design and statistical analysis to ARDEC's nearly 3,000 engineers and scientists.  Much of this support focuses on systematically designed experiments (DOE).  

Army technologies deal with a lot of destructive testing with binary response data, where efficient test methods are required.  There are a variety of test approaches tailored to these test problems and their unique test requirements.  One of the test approaches often used by the Statistics Group is fully adaptive sensitivity test methods, including the 3pod (three phase optimal design) recently developed by C. F. Jeff Wu.  

This presentation will focus on the unique needs inherent to destructive sensitivity test using a number of ARDEC technology case-studies as a backdrop.  We will discuss approaches being used in test design, statistical analysis of the test data, and resulting reliability, safety, and performance predictions.  We will also highlight the need for future research in the area of sensitivity test methods to support DoD test activities.

]]> Anita Race 1 1427453826 2015-03-27 10:57:06 1492118378 2017-04-13 21:19:38 0 0 event 2015-03-31T12:00:00-04:00 2015-03-31T12:00:00-04:00 2015-03-31T12:00:00-04:00 2015-03-31 16:00:00 2015-03-31 16:00:00 2015-03-31 16:00:00 2015-03-31T12:00:00-04:00 2015-03-31T12:00:00-04:00 America/New_York America/New_York datetime 2015-03-31 12:00:00 2015-03-31 12:00:00 America/New_York America/New_York datetime <![CDATA[]]>
<![CDATA[The Physical Internet - GT Savannah Learning Series]]> 27233 Sustainability Challenges and Solutions for Port Regions

Building Momentum toward a Smart Interconnected Era of Efficient and Sustainable Logistics, Supply Chains and Transportation

A lecture and discussion seeking to convey an understanding of Port Regions’ challenges with sustainability and other supply chain components in the face of continued growth, inadequate infrastructure and unsustainable inefficiencies. The Physical Internet concept is introduced as an emerging long-term solution to achieving a more sustainable global supply chain. The Physical Internet is an open global-logistics system founded on physical, digital and operational interconnectivity, through encapsulation, interfaces and protocols. The Physical Internet is intended to replace current logistical models.

Who Should Attend

How You Will Benefit

The Instructor

Dr. Benoit Montreuil, Professor and Coca-Cola Material Handling & Distribution Chair

About the PI Initiative

The Physical Internet Initiative aims at transforming the way physical objects
are moved, stored, realized, supplied and used; enabling economic, environmental and societal efficiency and sustainability breakthroughs by applying the Internet metaphor to the real world.

]]> Andy Haleblian 1 1427688771 2015-03-30 04:12:51 1492118378 2017-04-13 21:19:38 0 0 event A lecture and discussion seeking to convey an understanding of Port Regions’ challenges with sustainability and other supply chain components in the face of continued growth, inadequate infrastructure and unsustainable inefficiencies.

]]>
2015-06-04T09:00:00-04:00 2015-06-04T10:30:00-04:00 2015-06-04T10:30:00-04:00 2015-06-04 13:00:00 2015-06-04 14:30:00 2015-06-04 14:30:00 2015-06-04T09:00:00-04:00 2015-06-04T10:30:00-04:00 America/New_York America/New_York datetime 2015-06-04 09:00:00 2015-06-04 10:30:00 America/New_York America/New_York datetime <![CDATA[]]> info@scl.gatech.edu

]]>
391381 391381 image <![CDATA[The Physical Internet - GT Savannah Learning Series]]> image/jpeg 1449246312 2015-12-04 16:25:12 1475894403 2016-10-08 02:40:03 <![CDATA[Register Online to Attend]]>
<![CDATA[ISyE Seminar - Sanjay Mehrotra]]> 27187 TITLE: Optimizing Healthcare and Using Healthcare to Motivate Optimization Methods

ABSTRACT:

Healthcare, particularly in US, is a large and complex system.  Policies are determined based on legislated priorities, and decisions are often made based on suboptimal algorithms.   There is a growing interest in optimal resource utilization, while preserving the ethical equipoise between equity, justice and utility in healthcare.  Solutions require a trans-disciplinary collaborative approach, where industrial and systems engineers, operations researchers, and management scientists can make significant contributions by developing realistic data-driven and model based approaches to promote evidence based decision making and informing policy changes.  The need to bring greater realism to the decision models also motivates new methodological developments that can then benefit application in areas other than health.  The central consideration in developing innovative strategies to improve the health system is to save patients’ lives and to improve their quality of life. This must be balanced against risks and cost to individuals and society.   This leads to problems with multiple objectives, and input from multiple experts weighing in on these objectives. The parameters of the functions modeling the objectives and constraints are uncertain as model recommendations have implications on an unknown future. 

In this presentation we will focus on a few specific examples from our research illustrating how work in healthcare has provided insights towards developing concepts of robust Pareto optimality in multi-objective decision making; generation of parametric inequalities to significantly reduce the complexity of two-stage stochastic programs; and the need to consider out-of-sample error minimization towards developing a distributionally robust support vector machine.

Brief Biography:
Sanjay Mehrotra is the director of Center for Engineering at Health at Northwestern University.  He received his PhD in Operations Research from Columbia University.  Mehrotra is widely known for his methodology research in optimization that has spanned from linear, convex, mixed integer, stochastic, multi-objective, distributionally robust, and risk adjusted optimization.  His healthcare research includes topics in predictive modeling, budgeting, hospital operations, and policy modeling using modern operations research tools.  He is the immediate past chair of the INFORMS Optimization Society.  He has also been a INFORMS vice-president of (Chapter/Fora) and the chair of the INFORMS subdivision council.  He is the current Health Department editor of Institute for Industrial Engineering journal IIE-Transactions, and also held the role of Optimization Department editor for the same journal.

]]> Anita Race 1 1427814630 2015-03-31 15:10:30 1492118377 2017-04-13 21:19:37 0 0 event 2015-04-09T12:00:00-04:00 2015-04-09T12:00:00-04:00 2015-04-09T12:00:00-04:00 2015-04-09 16:00:00 2015-04-09 16:00:00 2015-04-09 16:00:00 2015-04-09T12:00:00-04:00 2015-04-09T12:00:00-04:00 America/New_York America/New_York datetime 2015-04-09 12:00:00 2015-04-09 12:00:00 America/New_York America/New_York datetime <![CDATA[]]>
<![CDATA[Seminar - Dr. Larry Burns]]> 27187 TITLE:  Making Value: The Future of Manufacturing, Technology and Work

ABSTRACT:

The scope of manufacturing has expanded beyond the factory floor.  In addition to “making things,” leading companies are focused on “making value.” This entails providing compelling and consistently positive customer experiences by integrating innovation, design, engineering and production throughout the entire value chain. Such experiences have proven to be key to building brand equity and realizing superior financial returns.

This presentation frames the future of manufacturing in terms of “making value.”  It first views manufacturing as an integrated system that turns resources into compelling customer experiences.   It then looks at how emerging technology (e.g., math-based design and engineering, intelligent machine-to-machine systems, additive manufacturing, nano-technology, the materials “genome”, advanced robotics, and business analytics) is combining with new business models (e.g., selling integrated products and services) to transform how companies compete to make value. 

The presentation concludes by summarizing the findings and recommendations of a recently released National Academy of Engineering report focused on “Making Value for America.”  This report emphasizes that

The best way to help people being left behind is to advance their skills and create an environment for innovation in the U.S. that continually attracts and creates skilled jobs

]]> Anita Race 1 1427815095 2015-03-31 15:18:15 1492118377 2017-04-13 21:19:37 0 0 event 2015-04-07T16:00:00-04:00 2015-04-07T16:00:00-04:00 2015-04-07T16:00:00-04:00 2015-04-07 20:00:00 2015-04-07 20:00:00 2015-04-07 20:00:00 2015-04-07T16:00:00-04:00 2015-04-07T16:00:00-04:00 America/New_York America/New_York datetime 2015-04-07 04:00:00 2015-04-07 04:00:00 America/New_York America/New_York datetime <![CDATA[]]>
<![CDATA[PhD Thesis Defense - Linwei Xin]]> 27187 TITLE: New approaches to inventory control: algorithms, asymptotics and robustness

ABSTRACT:

The fundamental problem of managing an inventory over time in the presence of stochastic demand is one of the core problems of operations research. This thesis is motivated by the need (in both government and industry) to understand such complex inventory systems used to model many of society's most important problems. In particular, we investigate simple, efficient and robust inventory policies for several fundamental models commonly used in the study of stochastic inventory systems. Some of these policies have been already implemented in practice and we provide theoretic support for their practical utilization in industry. Furthermore, the results on the performance of these policies often yield a rule-of-thumb that is applicable in a variety of settings.

There are four main chapters in this thesis. In the second chapter, we study lost sales inventory model with positive lead time. Although such a model has been studied for over fifty years, the structure of the optimal policy remains poorly understood. Furthermore, it is notoriously difficult to solve the dynamic program especially when the lead time is large due to the curse of dimensionality. Recently, \cite{Goldberg12} proved that a simple constant-order policy (proposed earlier by \cite{Reiman04}) is asymptotically optimal as the lead time grows large. However, the bound proven there is impractical. Numerical experiments of \cite{Zipkin08b} suggested the good performance of constant-order policies even for small lead times, and there remains a huge gap between the existing theoretical bound and the actual performance of constant-order policies. In this chapter, we significantly improve the bound. In particular, we prove that the same constant-order policy actually conver ges exponentially fast to optimality as the lead time grows. In addition, our bound is simple and explicit, demonstrating good performance of constant-order policies for realistic lead time values. Our results provide theoretical justification for the good performance of such simple policies, and open the window to making the results and methodology practical.

In the third chapter, we investigate dual-sourcing inventory systems, which arise naturally in many real-world supply chains (cf. \cite{Allon10}). These systems are notoriously difficult to optimize due to the complex structure of the optimal solution and the curse of dimensionality. Recently, so-called Tailored Base-Surge (TBS) policies have been proposed as a heuristic for the dual-sourcing problem, and analyzed in \cite{Allon10} and \cite{Janakiraman14a}. Under such a policy, a constant order is placed at the regular source in each period, while the order placed at the express source follows a simple order-up-to rule. Numerical experiments by several authors have suggested that such policies perform well as the lead time difference between the two sources grows large, which is exactly the setting in which the curse of dimensionality leads to the problem becoming intractable. However, providing a theoretical foundation for this phenomenon has remained a major open probl em. In this chapter, we provide such a theoretical foundation by proving that a simple TBS policy is indeed asymptotically optimal as the lead time of the regular source grows large, with the lead time of the express source held fixed. Interestingly, the simple TBS policy performs nearly optimally exactly when standard DP-based methodologies become intractable due to the aforementioned ``curse of dimensionality". Furthermore, as the ``best" TBS policy can be computed by solving a convex program that does not depend on the lead time of the regular source, our results lead directly to very efficient algorithms (with complexity independent of the lead time of the regular source) with asymptotically optimal performance guarantees. Perhaps most importantly, since many companies are already implementing such TBS policies (cf. \cite{Allon10}), our results provide strong theoretical support for the widespread use of TBS policies in practice.

In the fourth chapter, we explore the concept of time consistency in the context of distributionally robust inventory models with second moment constraints. Recently, several communities have observed that a subtle phenomena known as time inconsistency, which never happen in the classic (non-robust) setting, can arise in the framework of distributionally robust optimization. In particular, there have been two fundamentally different approaches taken in the literature regarding the specification of uncertainty in the underlying joint distribution, depending on whether the underlying optimization model is static or dynamic in nature. In a multistage static formulation, one specifies a family of joint distributions for demand over time, typically by fixing means, covariances, and supports (or some generalization thereof), and then solves an associated global minimax optimization problem. Such static formulations generally cannot be decomposed and solved by dynamic programming an d are generally referred to as time-inconsistent in the literature. Alternatively, in a multistage dynamic formulation, the underlying family of potential joint distributions must implicitly satisfy certain conditional independence properties, and thus allow for a resolution by dynamic programming. The existence of such a decomposition is generally referred to as the rectangularity property (cf. \cite{Iyengar05}, \cite{Nilim05}). In this chapter, we provide several illustrative examples showing that here the question of time consistency can be quite subtle and complement these observations by providing simple sufficient conditions for time consistency. We also prove that, although the multistage-dynamic formulation always has an optimal policy of base-stock form, there may be no such optimal policy for the multistage-static formulation. Interestingly, our results show that time consistency may hold even when rectangularity does not.

In the fifth chapter, we study distributionally robust inventory control with martingale demand. Many inventory models assume perfect knowledge of the demand distribution. However, exact knowledge of such distributions is rarely available in practice, and there has been a growing interest in developing inventory control policies which are robust to model misspecification. In the context of distributionally robust inventory control which was initiated in \cite{Scarf58}, one assumes that the joint distribution of future demand belongs to some set of joint distributions, and solves the min-max problem of computing the control policy which is optimal against a worst-case distribution belonging to this set. Although such models have been analyzed previously, the cost and policy implications of positing different dependency structures remains poorly understood (cf. \cite{Agrawal12}). In this chapter, we combine the framework of distributionally robust optimization with the theory o f martingales, and study a novel distributionally robust model in which the sequence of future demands is assumed to belong to a family of martingales. We explicitly compute the optimal policy and shed light on the interplay between the optimal policy and worst-case distribution. In particular, we show that at optimality, in each period the adversary always selects a demand distribution with two-point support, with one of these points equal to zero. Combined with the martingale property, this implies that at optimality, the worst-case demand distribution corresponds to the setting in which demand may become obsolete at a random time, a scenario of practical interest which has been studied previously in the literature (cf. \cite{Song96}). We also compare to the analogous setting in which demand is independent across periods (previously studied in \cite{Shapiro12}). Interestingly, we find the ratio of the optimal cost under the martingale and independent models to be exactly 1 /2 in the perfectly symmetric case. Our results shed light on several intriguing phenomena regarding the impact of correlations on distributionally robust models, and provide a first step towards developing a conditional-expectation based theory of dynamic distributionally robust forecasting.

]]> Anita Race 1 1427814364 2015-03-31 15:06:04 1492118377 2017-04-13 21:19:37 0 0 event 2015-04-13T13:00:00-04:00 2015-04-13T13:00:00-04:00 2015-04-13T13:00:00-04:00 2015-04-13 17:00:00 2015-04-13 17:00:00 2015-04-13 17:00:00 2015-04-13T13:00:00-04:00 2015-04-13T13:00:00-04:00 America/New_York America/New_York datetime 2015-04-13 01:00:00 2015-04-13 01:00:00 America/New_York America/New_York datetime <![CDATA[]]>
<![CDATA[DOS Seminar - Burak Kocuk]]> 27187 TITLE: Strong SOCP Relaxations for Optimal Power Flow Problem

ABSTRACT:

Optimal Power Flow is a fundamental optimization problem in electrical power systems analysis. Although local optimal solution methods are generally successful, they do not provide guarantees on the quality. Recently, much research has focused on Semidefinite Programming (SDP) relaxations to obtain strong lower bounds. In this work, we instead utilize Second Order Cone Programming (SOCP) relaxations due to their superior computational power. However, since SOCPs are weaker than their SDP counterparts, we propose three improvements to strengthen SOCP relaxation and show that two of them are incomparable to SDP relaxation. Finally, we present extensive computational experiments with standard benchmark instances from literature to demonstrate the accuracy and efficiency of SOCP-based methods.

This is joint work with Santanu S. Dey and X. Andy Sun

]]> Anita Race 1 1428330235 2015-04-06 14:23:55 1492118373 2017-04-13 21:19:33 0 0 event 2015-04-08T13:00:00-04:00 2015-04-08T13:00:00-04:00 2015-04-08T13:00:00-04:00 2015-04-08 17:00:00 2015-04-08 17:00:00 2015-04-08 17:00:00 2015-04-08T13:00:00-04:00 2015-04-08T13:00:00-04:00 America/New_York America/New_York datetime 2015-04-08 01:00:00 2015-04-08 01:00:00 America/New_York America/New_York datetime <![CDATA[]]>
<![CDATA[Statistics Seminar - Vadim Zipunnikov]]> 27187 TITLE:  Makings Sense of Sensors

ABSTRACT:

The crashing wave of activity tracking “wearables” opens up an opportunity to unveil previously hidden but pivotal signatures of disability and disease. To achieve this promise, the understanding, interpretation and analysis of complex multimodal data produced by such devices becomes crucial. The first part of my talk will provide an overview of the instruments that are available for real-time measurement of physical activity as well as a quick review of the strengths and limitations of current methods for measuring physical activity. In the second part, I will talk about analysis of data collected on 700+ subjects wearing an Actiheart device that collects minute-by-minute activity counts and heart rate for one week as a part of the Baltimore Longitudinal Study of Aging. I will discuss recent multilevel functional data approaches to separate and quantify the systematic and random circadian patterns of physical activity as functions of time of day, age, and gender in this population.


Brief Bio:

Vadim Zipunnikov earned his PhD in Statistics from Cornell University. After three years as a postdoctoral fellow at the Department of Biostatistics at Johns Hopkins University, he joined the department as a faculty. He is deeply involved in analyzing and modeling accelerometry measured physical activity, heart rate, and ecological momentary assessment (EMA) data in large-scale epidemiological studies such as Baltimore Longitudinal Study of Aging (BLSA), NIMH Family Study of Health and Behavior, and National Health and Nutrition Examination Survey (NHANES).]]> Anita Race 1 1428566457 2015-04-09 08:00:57 1492118371 2017-04-13 21:19:31 0 0 event 2015-04-13T15:00:00-04:00 2015-04-13T15:00:00-04:00 2015-04-13T15:00:00-04:00 2015-04-13 19:00:00 2015-04-13 19:00:00 2015-04-13 19:00:00 2015-04-13T15:00:00-04:00 2015-04-13T15:00:00-04:00 America/New_York America/New_York datetime 2015-04-13 03:00:00 2015-04-13 03:00:00 America/New_York America/New_York datetime <![CDATA[]]>
<![CDATA[Statistics Seminar - Ciprian Crainiceanu]]> 27187 TITLE: Not everybody, but some people move like you: Wearable computing in public health

ABSTRACT:

Accelerometers are now used extensively in health studies, where they increasingly replace self-report questionnaires. The sudden success of accelerometers in these studies is due to the fact that they are cheap, easy to wear, collect millions of data points at high frequency (10-100Hz or more), store months worth of data, and can be paired with other devices, such as heart, gps, or skin temperature sensors. I will discuss the multi-resolution structure of the data and will introduce methods for movement recognition both for in-the-lab and in-the-wild data using second- and sub-second level data. I will introduce movelets, a powerful dictionary learning approach, designed for quick identification of movement patterns. At the minute level I will describe activity intensity measures (activity counts, vector magnitude, and activity intensity) and introduce functional data approaches for characterizing the circadian rhythm of activity and its association with health. The natural data structure induced by such observational studies is that of multilevel functional data (activity intensity measured at every minute for multiple days observed within each subject.) I will introduce fast functional data analysis approaches that can deal with the data complexity, describe its structure and its association with health outcomes. In particular, I will discuss results from a motivating study of the association between age, body mass index (BMI) and the circadian rhythm of activity. I will also explain why you should not trust your calorie counter.


Brief Bio:

Ciprian obtained his PhD at Cornell University in 2003 and moved to Johns Hopkins University, where he is now Professor of Biostatistics. He is the co-founder and co-leader of the SMART (www.smart-stats.org) research group, which does extensive research in biosignals, wearable biosensors, and clinical brain imaging. He is Fellow of ASA, has published over 100 papers both in Statistics and Scientific journals with emphasis on studies of Multiple Sclerosis, Stroke, Aging, Alzheimer, Acute Lung Injury, and child growth. His methodological work is concerned with smoothing, functional data analysis, and brain imaging including fMRI, sMRI, DCE-MRI, and CT. His most important accomplishments are his twin almost-to-be-seven-year-old twin daughters.
]]> Anita Race 1 1428566590 2015-04-09 08:03:10 1492118371 2017-04-13 21:19:31 0 0 event 2015-04-14T15:00:00-04:00 2015-04-14T16:00:00-04:00 2015-04-14T16:00:00-04:00 2015-04-14 19:00:00 2015-04-14 20:00:00 2015-04-14 20:00:00 2015-04-14T15:00:00-04:00 2015-04-14T16:00:00-04:00 America/New_York America/New_York datetime 2015-04-14 03:00:00 2015-04-14 04:00:00 America/New_York America/New_York datetime <![CDATA[]]>
<![CDATA[PhD Thesis Defense - Fan Ye]]> 27187 TITLE: Information Relaxation in Stochastic Optimal Control

ABSTRACT:

Dynamic programming is a principal method for analyzing stochastic optimal control problems. However, the exact computation of dynamic programming can be intractable in large-scale problems due to the "curse of dimensionality". Various approximation methods have been proposed to address this issue and often generate suboptimal policies, though it is generally difficult to tell how far these policies are from optimal. This thesis concerns with studying the stochastic control problems from a duality perspective and generating upper bounds on maximal expected rewards, which complements the lower bounds associated with suboptimal policies. If the gap between the lower and upper bounds is small, it implies that a suboptimal policy must be close to optimal. The approach considered in this thesis is called "information relaxation", that is, it relaxes the non-anticipativity constraint that requires the decisions to depend only on the information available to the decision maker and imposes a penalty that punishes such a violation.

In the first part of the thesis, we study the interaction of Lagrangian relaxation and information relaxation in weakly coupled dynamic programs. A commonly studied approach builds on the property that this high-dimensional problem can be decoupled by dualizing the resource constraints via Lagrangian relaxation. We generalize the information relaxation approach to improve upon the Lagrangian bound in an infinite-horizon setting. We also develop a computational method to tackle large-scale problems and provide insightful interpretation and performance guarantee.

In the second part, we formulate the information relaxation-based duality in an important class of continuous-time decision-making models --- controlled Markov diffusion, which is widely used in portfolio optimization and risk management. We find that this continuous-time model admits an optimal penalty in compact form --- an Ito stochastic integral, which enables us to construct approximate penalties in simple forms and achieve tight dual bounds. We demonstrate its use in a dynamic portfolio choice problem subject to position and consumption constraints.

In the third part, we consider the problem of optimal stopping of discrete-time continuous-state partially observable Markov processes. We develop a filtering-based dual approach, which relies on the martingale duality formulation of the optimal stopping problem and the particle filtering technique. We carry out error analysis and illustrate the effectiveness of our method in an example of pricing American options under partial observation of stochastic volatility.

]]> Anita Race 1 1428657409 2015-04-10 09:16:49 1492118370 2017-04-13 21:19:30 0 0 event 2015-04-22T14:00:00-04:00 2015-04-22T14:00:00-04:00 2015-04-22T14:00:00-04:00 2015-04-22 18:00:00 2015-04-22 18:00:00 2015-04-22 18:00:00 2015-04-22T14:00:00-04:00 2015-04-22T14:00:00-04:00 America/New_York America/New_York datetime 2015-04-22 02:00:00 2015-04-22 02:00:00 America/New_York America/New_York datetime <![CDATA[]]>
<![CDATA[PhD Thesis Defense - Tonya Woods]]> 27187 TITLE:  Extracting Meaningful Statistics for the Characterization and Classification of Biological, Medical, and Financial Data

ABSTRACT:

This thesis is focused on extracting meaningful statistics for the characterization and classification of biological, medical, and financial data and contains four chapters. The first chapter contains theoretical background on scaling and wavelets, which supports the work in chapters two and three.

In the second chapter, we outline a methodology for representing sequences of DNA nucleotides as numeric matrices in order to analytically investigate important structural characteristics of DNA. This methodology involves assigning unit vectors to nucleotides, placing the vectors into columns of a matrix, and accumulating across the rows of this matrix. Transcribing the DNA in this way allows us to compute the 2-D wavelet transformation and assess regularity characteristics of the sequence via the slope of the wavelet spectra. In addition to computing a global slope measure for a sequence, we can apply our methodology for overlapping sections of nucleotides to obtain an evolutionary slope.

In the third chapter, we describe various ways wavelet-based scaling may be used for cancer diagnostics. There were nearly half of a million new cases of ovarian, breast, and lung cancer in the United States last year. Breast and lung cancer have highest prevalence, while ovarian cancer has the lowest survival rate of the three. Early detection is critical for all of these diseases, but substantial obstacles to early detection exist in each case. In this work, we use wavelet-based scaling on metabolic data and radiography images in order to produce meaningful features to be used in classifying cases and controls. Computer-aided detection (CAD) algorithms for detecting lung and breast cancer often focus on select features in an image and make a priori assumptions about the nature of a nodule or a mass. In contrast, our approach to analyzing breast and lung images captures information contained in the background tissue of images as well as information about specific featu res and makes no such a priori assumptions.

In the fourth chapter, we investigate the value of social media data in building commercial default and activity credit models. We use random forest modeling, which has been shown in many instances to achieve better predictive accuracy than logistic regression in modeling credit data. This result is of interest, as some entities are beginning to build credit scores based on this type of publicly available online data alone. Our work has shown that the addition of social media data does not provide any improvement in model accuracy over the bureau only models. However, the social media data on its own does have some limited predictive power.

]]> Anita Race 1 1429005611 2015-04-14 10:00:11 1492118370 2017-04-13 21:19:30 0 0 event 2015-05-01T11:00:00-04:00 2015-05-01T11:00:00-04:00 2015-05-01T11:00:00-04:00 2015-05-01 15:00:00 2015-05-01 15:00:00 2015-05-01 15:00:00 2015-05-01T11:00:00-04:00 2015-05-01T11:00:00-04:00 America/New_York America/New_York datetime 2015-05-01 11:00:00 2015-05-01 11:00:00 America/New_York America/New_York datetime <![CDATA[]]>
<![CDATA[FREE Supply and Demand Planning webinar, Thursday, April 30 1:30 - 2:30 pm EDT!]]> 27233 Attend our free The Three Cornerstones for Effective Supply and Demand Planning webinar and receive a *discount code towards any of the courses that are part of the Supply and Demand Planning (SDP) series taking place the week of June 1, 2015 in Atlanta (*please note that the discount cannot be combined with any other discount).

OVERVIEW

Companies are facing challenges with globalization, market volatility and the intense pressure to stay competitive. To address these issues, progressive supply chain leaders are taking advantage of the best-of-industry practices and advanced technologies for better end-to-end visibility and improved Supply and Demand Planning.

There are three cornerstones for effective Supply and Demand Planning. Each serves a role in tackling the challenges of today and the potential risks of tomorrow with cross-functional collaboration and a unified objective to increase financial growth.

This next generation of Supply and Demand Planning is only now possible with advances in methodologies and technologies for each of the following Supply and Demand planning cornerstones:

The webinar will provide an overview of these advances and how they enable companies to address today’s realities by making smarter decisions that drive significant competitive advantage.

ABOUT THE PRESENTERS

Richard Sharpe is the CEO of Competitive Insights, a professional Software as a Service (SaaS) company focused on driving business value through the adoption of process and technology innovations. He is also instructor for our Integrated Business Planning and Supply Chain Risk courses. Richard is a frequent speaker at national gatherings of The Council of Supply Chain Management Professionals (CSCMP) and other industry related forums. He actively promotes awareness of supply chain management, challenging executives to consider its role in gaining competitive advantage.

Bryan Garland has worked in the supply chain planning area for the past four years, focusing on process industry projects on various planning layers. He developed expertise in S&OP through several customer projects at organizations including Monsanto, DSM Pharmaceuticals, Patheon and Axalta Coatings.

]]> Andy Haleblian 1 1429094565 2015-04-15 10:42:45 1492118367 2017-04-13 21:19:27 0 0 event Attend our free The Three Cornerstones for Effective Supply and Demand Planning webinar to learn about the three cornerstones for effective Supply and Demand Planning and how to approach dealing with today's realities and risks through making smarter decisions that drive significant competitive advantage.

]]>
2015-04-30T14:30:00-04:00 2015-04-30T15:30:00-04:00 2015-04-30T15:30:00-04:00 2015-04-30 18:30:00 2015-04-30 19:30:00 2015-04-30 19:30:00 2015-04-30T14:30:00-04:00 2015-04-30T15:30:00-04:00 America/New_York America/New_York datetime 2015-04-30 02:30:00 2015-04-30 03:30:00 America/New_York America/New_York datetime <![CDATA[]]> webinar@scl.gatech.edu]]> 396511 396511 image <![CDATA[The Three Cornerstones for Effective Supply and Demand Planning]]> image/jpeg 1449246361 2015-12-04 16:26:01 1475895112 2016-10-08 02:51:52 <![CDATA[Register Online to Attend]]> <![CDATA[Supply & Demand Planning brochure (PDF)]]> <![CDATA[Supply Chain Risk Management course]]> <![CDATA[World Class Sales & Operations Planning course]]> <![CDATA[Integrated Business Planning course]]>
<![CDATA[ISyE Colloquium Seminar - Sebastien Bubeck]]> 27187 TITLE:  The entropic barrier: a simple and optimal universal self-concordant barrier

ABSTRACT:

A fundamental result in the theory of Interior Point Methods is Nesterov and Nemirovski's construction of a universal self-concordant barrier. In this talk I will introduce the entropic barrier, a new (and in some sense optimal) universal self-concordant barrier. The entropic barrier connects many topics of interest in Machine Learning: exponential families, convex duality, log-concave distributions, Mirror Descent, and exponential weights.

]]> Anita Race 1 1429516540 2015-04-20 07:55:40 1492118366 2017-04-13 21:19:26 0 0 event 2015-04-27T12:00:00-04:00 2015-04-27T12:00:00-04:00 2015-04-27T12:00:00-04:00 2015-04-27 16:00:00 2015-04-27 16:00:00 2015-04-27 16:00:00 2015-04-27T12:00:00-04:00 2015-04-27T12:00:00-04:00 America/New_York America/New_York datetime 2015-04-27 12:00:00 2015-04-27 12:00:00 America/New_York America/New_York datetime <![CDATA[]]>
<![CDATA[ISyE Seminar - Tim Huh]]> 27187 TITLE: Pricing under the Nested Attraction Model with a Multi-stage Choice Structure

ABSTRACT:

We develop a solution approach to the centralized pricing problem of a nested attraction model with a multi-stage tree structure. We identify conditions under which the optimal solution can be uniquely determined and characterize the optimal solution as a fixed-point of a single variable. In the special case of a multi-stage nested logit model, we show the impact of asymmetry in price sensitivity and adjustment index (also known as the dissimilarity index) and we derive a closed-form solution when the tree structure is symmetric. We show that the equal mark-up property which holds for the single-stage nested attraction models is not valid in the multi-stage nested choice structure even when price sensitivities are the same for all products.

Joint work with Hongmin Li (ASU).

Bio: Woonghee Tim Huh is an associate professor in the Sauder School of Business at the University of British Columbia. His current research interests include supply chain management, inventory control and dynamic pricing. He received a B.A. in sociology, B.Math in computer science and M.Math in combinatorics & optimization from the University of Waterloo, and holds an M.Sc. and a Ph.D. in Operations Research from Cornell University. He is an associate editor for Management Science, Naval Research Logistics and Operations Research Letters, and a senior editor for Production and Operations Management. _______________________________________________

]]> Anita Race 1 1429516726 2015-04-20 07:58:46 1492118364 2017-04-13 21:19:24 0 0 event 2015-04-23T12:00:00-04:00 2015-04-23T12:00:00-04:00 2015-04-23T12:00:00-04:00 2015-04-23 16:00:00 2015-04-23 16:00:00 2015-04-23 16:00:00 2015-04-23T12:00:00-04:00 2015-04-23T12:00:00-04:00 America/New_York America/New_York datetime 2015-04-23 12:00:00 2015-04-23 12:00:00 America/New_York America/New_York datetime <![CDATA[]]>
<![CDATA[DOS Seminar - Arefin Hug]]> 27187 TITLE: The matching problem has no small symmetric SDP

ABSTRACT:

Yannakakis showed that the matching problem does not have a small symmetric linear program. Rothvoß recently proved that any (not necessarily symmetric) linear program also has exponential size. It is natural to ask whether the matching problem can be expressed compactly in a framework such as semidefinite programming (SDP) that is more powerful than linear programming but still allows efficient optimization. We answer this question negatively for symmetric SDPs: any symmetric SDP for the matching problem has exponential size. We also show that an O(k)-round Lasserre SDP relaxation for the metric traveling salesperson problem (TSP) yields at least as good an approximation as any symmetric SDP relaxation of size n^k. The key technical ingredient underlying both these results is an upper bound on the degree needed to derive polynomial identities that hold over the space of matchings or traveling salesperson tours.

This is joint work with Jonah Brown-Cohen, Prasad Raghavendra and Benjamin Weitz from Berkeley, and Gabor Braun, Sebastian Pokutta, Aurko Roy and Daniel Zink at Georgia Tech.

Speaker Bio:

Arefin Huq is a PhD student in the Theory group of the College of Computing at Georgia Tech, currently focusing on the connection between computational complexity and combinatorial optimization. He has previously done work on Kolmogorov complexity, machine learning, and automated emotion recognition in music.




]]> Anita Race 1 1429516899 2015-04-20 08:01:39 1492118364 2017-04-13 21:19:24 0 0 event 2015-04-22T13:00:00-04:00 2015-04-22T13:00:00-04:00 2015-04-22T13:00:00-04:00 2015-04-22 17:00:00 2015-04-22 17:00:00 2015-04-22 17:00:00 2015-04-22T13:00:00-04:00 2015-04-22T13:00:00-04:00 America/New_York America/New_York datetime 2015-04-22 01:00:00 2015-04-22 01:00:00 America/New_York America/New_York datetime <![CDATA[]]>
<![CDATA[ISyE Seminar - Barry Nelson]]> 27187 TITLE: Discrete Optimzation via Simulation using Gaussian Markov Random Fields

ABSTRACT:

We consider maximizing or minimizing the expected value of a stochastic performance measure that can be observed by running a dynamic, discrete-event simulation when the feasible solutions are defined by integer-ordered decision variables. Inventory sizing, call center staffing and manufacturing system design are common applications. Standard approaches are ranking and selection, which takes no advantage of spatial structure, and adaptive random search, which exploits it but in a heuristic way (“good solutions tend to be clustered”). Instead, we construct an optimization procedure built on discrete Gaussian Markov random fields (GMRFs). This enables computation of the expected improvement (EI) that could be obtained by running the simulation for any feasible solution. This computation can be numerically challenging; however, GMRFs are defined by their precision matrices which can be constructed to be sparse. Thus, we can use sparse matrix techniques to calculate expressions that involve the precision matrix. We also introduce a new EI criterion that incorporates the uncertainty in stochastic simulation by treating the value at the current optimal solution as a random variable.

]]> Anita Race 1 1429782977 2015-04-23 09:56:17 1492118364 2017-04-13 21:19:24 0 0 event 2015-04-28T12:00:00-04:00 2015-04-28T12:00:00-04:00 2015-04-28T12:00:00-04:00 2015-04-28 16:00:00 2015-04-28 16:00:00 2015-04-28 16:00:00 2015-04-28T12:00:00-04:00 2015-04-28T12:00:00-04:00 America/New_York America/New_York datetime 2015-04-28 12:00:00 2015-04-28 12:00:00 America/New_York America/New_York datetime <![CDATA[]]>
<![CDATA[ISyE Seminar - David D. Yao]]> 27187 TITLE:  A Stochastic-Network Approach to Modeling Contagion Dynamics

ABSTRACT:

Stochastic network has been a useful and effective tool in studying diverse engineering and service systems such as semiconductor wafer fab’s, supply chains, call centers and the Internet.  Here we present two novel applications of stochastic network in (a) resiliency studies of critical urban infrastructures (such as the water distribution system and the power grid), and (b) systemic risk analysis of financial networks. In both applications, the interdependence and interactions within the system are characterized by a dynamic complementarity problem, also known as Skorohod problem; and the cascading effect of a “shock” or extreme event can be analytically (or computationally) captured by the solution to the Skorohod problem.

 

Bio:

David Yao is the Piyasombatkul Family Professor of Industrial Engineering and Operations Research at Columbia University. He is the founding chair of the Financial and Business Analytics Center at Columbia Data Science Institute. Author/co-author of some 200 scientific publications, he is a principal investigator of over thirty grants and contracts from government agencies and industrial sources, and a holder of eight U.S. patents.  His honors and awards include the Presidential Young Investigator Award from the National Science Foundation, Guggenheim Fellowship from the John Simon Guggenheim Foundation, Franz Edelman Award from the Institute for Operations Research and Management Sciences, SIAM Outstanding Paper Prize from the Society for Industrial and Applied Mathematics, Outstanding Technical Achievement Award from IBM Research, Great Teacher Award from the Society of Columbia Graduates, and the IBM Faculty Award. He is an IEEE Fellow, an INFORMS Fellow, and a member of the National Academy of Engineering.

]]> Anita Race 1 1429865639 2015-04-24 08:53:59 1492118364 2017-04-13 21:19:24 0 0 event 2015-04-30T12:00:00-04:00 2015-04-30T12:00:00-04:00 2015-04-30T12:00:00-04:00 2015-04-30 16:00:00 2015-04-30 16:00:00 2015-04-30 16:00:00 2015-04-30T12:00:00-04:00 2015-04-30T12:00:00-04:00 America/New_York America/New_York datetime 2015-04-30 12:00:00 2015-04-30 12:00:00 America/New_York America/New_York datetime <![CDATA[]]>
<![CDATA[FREE Defining and Implementing Effective Sourcing Strategies webinar, Tuesday, May 12 1:30 - 2:30 pm EDT!]]> 27233 Attend our free Defining and Implementing Effective Sourcing Strategies webinar to learn how to go about formulating an integrated supply strategy while taking into consideration both local geographical and global issues. As a thank you for attending, you will receive a *discount code towards towards the course in Brunswick, GA being held June 23-25 (*please note that the discount cannot be combined with any other discount).

OVERVIEW

Leading companies throughout the world are looking to formulate integrated supply strategies on both a local geographical and a global basis. This includes strategically sourcing materials and components worldwide and selecting global locations for key supply and distribution centers. The growth in global trade will continue to have a major impact on supply chains and requires firms to practice professional strategic supply management in order to help ensure continuity of supply and to contain and reduce costs. 

Strategic sourcing enhances value, ultimately impacting the profitability of an entire organization. In this essential course, you’ll learn how to develop and implement a sourcing strategy that aligns with overall competitive strategy. The course and the associated case studies, activities and discussions provide the context and a framework for making effective sourcing decisions including a comprehensive approach to strategic sourcing. 

This webinar will provide an overview of the LIVE course planned through the Supply Chain and Logistics Institute at Georgia Tech which will take place in Brunswick, Georgia, near St Simons island June 23-25, 2015. 

ABOUT THE PRESENTER

Lew specializes in the fields of Business Performance Improvement and Supply Chain Management. His consulting and management development career over the last twenty five years has included working with Ryder, UPS-SCS, Caliber Logistics, DHL, Menlo Worldwide, BAX Global, Owens Corning, Coca-Cola, Georgia-Pacific, Baxter, De Beers, Mercedes Benz, Nissan and many other major firms worldwide. He has studied under Dr. Richard Schonberger, who is recognized as one of the worlds leading authorities in the fields of JIT and Business Performance Improvement.

]]> Andy Haleblian 1 1430158781 2015-04-27 18:19:41 1492118363 2017-04-13 21:19:23 0 0 event Attend our free Defining and Implementing Effective Sourcing Strategies webinar to learn how to go about formulating an integrated supply strategy while taking into consideration both local geographical and global issues.

]]>
2015-05-12T14:30:00-04:00 2015-05-12T15:30:00-04:00 2015-05-12T15:30:00-04:00 2015-05-12 18:30:00 2015-05-12 19:30:00 2015-05-12 19:30:00 2015-05-12T14:30:00-04:00 2015-05-12T15:30:00-04:00 America/New_York America/New_York datetime 2015-05-12 02:30:00 2015-05-12 03:30:00 America/New_York America/New_York datetime <![CDATA[]]> webinar@scl.gatech.edu]]> 400001 400001 image <![CDATA[FREE Defining and Implementing Effective Sourcing Strategies webinar, Tuesday, May 12 1:30 - 2:30 pm EDT!]]> image/jpeg 1449246388 2015-12-04 16:26:28 1475895117 2016-10-08 02:51:57 <![CDATA[Register Online to Attend]]> <![CDATA[Defining and Implementing Effective Sourcing Strategies Course]]> <![CDATA[Link to Course Flyer]]>
<![CDATA[PhD Thesis Defense - Xinchang Wang]]> 27187 TITLE: Revenue Management with Customer Choice and Sellers Competition

ABSTRACT:

Revenue management is concerned with managing demands of customers and has been found successful in broad areas such as airline, hotel and retailing industries. In revenue management, decisions of sellers such as designing a product portfolio or choosing prices of products are often made based on customer choice. It is thus important to understand customer choice behavior and analyze how it affects sellers' decisions, especially when customers' choice exhibits specific behavioral phenomena that deviate from axioms of rational choice (e.g., Luce's axiom of choice) and sellers compete.

My thesis is focused on revenue management problems, with particular emphasis on customer choice behavior, and it consists of three essential chapters.

 

In the first chapter, we build a variety of customer booking choice models for a major airline that operates in a very competitive origin-destination market, including the multinomial logit (MNL) models, nested logit (NL) models, mixed-logit (ML) models and latent logit class (LCL) models. The latter three types of models are aimed at incorporating unobserved heterogeneous customer preferences for different departure times of flights and identifying latent customer types. More interestingly, we incorporate in all our models the context effect that the attractiveness of a fare class is influenced by the other fare classes offered in the same assortment, which is not standard in the literature of discrete choice modeling. The estimation results show that including these factors into choice models dramatically affects price sensitivity estimates, and therefore matters.

 

Previously available algorithms are inefficient for estimating choice models from large sets of data (observations), especially for estimating advanced choice models that usually involve high-dimensional integrals, such as the ML-type models. In the second chapter, we present a stochastic trust region algorithm for ML-type model estimations. The algorithm embeds two sampling processes: (i) a data sampling process and (ii) a Monte Carlo sampling process. The second process is employed to compute the sample average approximation of a high-dimensional integral. The algorithm dynamically controls the sample sizes based on the magnitude of the errors incurred due to the two sampling processes. First, the algorithm controls the size of Monte Carlo samples for each observation in the dataset to minimize the total sample size subject to a constraint on the variance of the objective estimate. Second, the algorithm controls sampling from the dataset according to the magnitude of data sampling error relative to the Monte Carlo sampling error. The first-order convergence (w.p. 1) is proved based on generalized uniform law of large numbers theories for both the objective function and its gradient. The efficiency of the algorithm is tested with data and compared with other algorithms.

 

In the third chapter, we study how a specific behavioral phenomenon, called the decoy effect, affects the decisions of sellers in product assortment competition in a duopoly. We propose a discrete choice model to capture decoy effects, and we use the model to provide a complete characterization of the Nash equilibria and their dependence on choice model parameters. For the cases in which there are multiple equilibria, we consider dynamical systems models of the sellers responding to their competitors using Cournot adjustment or fictitious play to study the evolution of the assortment competition and the stability of the equilibria. Our results show that all pure-strategy Nash equilibria can provide reliable forecasts of the outcome of the competition in the sense that they have large domains of attraction. In contrast, mixed-strategy Nash equilibria have negligible domains of attraction, except for a special case, and thus we conclude that mixed-strategy Nash equilibria do not provide reliable forecasts of the outcome of the competition. Our results also provide a simple geometric characterization of the dynamics of fictitious play for general $2 \times 2$ games that is more complete than previous characterizations.

]]> Anita Race 1 1430214447 2015-04-28 09:47:27 1492118363 2017-04-13 21:19:23 0 0 event 2015-05-07T13:00:00-04:00 2015-05-07T13:00:00-04:00 2015-05-07T13:00:00-04:00 2015-05-07 17:00:00 2015-05-07 17:00:00 2015-05-07 17:00:00 2015-05-07T13:00:00-04:00 2015-05-07T13:00:00-04:00 America/New_York America/New_York datetime 2015-05-07 01:00:00 2015-05-07 01:00:00 America/New_York America/New_York datetime <![CDATA[]]>
<![CDATA[PhD Thesis Defense - Yi Xiao]]> 27187 TITLE:  Some results in High-dimensional Statistics

ABSTRACT:

High-dimensional statistics is one of the most active research topics in modern statistics. The complexity of data both in size and structure brings new challenges to statisticians to extract useful information apart from the noises in an efficient and accurate manner. The purpose of this thesis is to narrow the gap between theory and practice in high-dimensional statistics by studying some of the widely adopted assumptions in the literature and introducing new testing procedures. To be more specific, it consists of two parts covering $l_1$-regularized estimation for time series and testing for sparse Gaussian graphical model.

 

The first chapter studied the applications of $l_1$-regularized regression methods for Gaussian vector autoregressive processes. We decomposed the classical regression model into smaller submodels and obtained sparse solutions by applying $l_1$-penalties. We showed that under mild conditions the design matrices corresponding to the submodels are actually generated from some $\alpha$-mixing processes. Therefore, a more general problem is how good is an $l_1$-regularized estimate for a linear model with a random design matrix that is generated by an $\alpha$-mixing Gaussian process with exponential decay rate. Our main result verified the restricted eigenvalue assumption for the mixing random design based on the generic chaining technique and derived the $l_p$ error bound for Lasso and Dantzig selector. We also studied the sufficient conditions for a VAR(p) model to guarantee tight error bound of the solutions. Finally, we illustrated the variable selection and estimation performance of Lasso by several sets of simulation.

 

In the second chapter, we proposed a new statistic to test the decomposable structure of a Gaussian graphical model in the high-dimensional setting. It is based on the quadratic forms of the sample covariance matrix eigenvalues. In the case when the null hypothesis corresponds to a group independence structure, we derived the asymptotic distribution of the proposed statistic and showed that it is invariant under non-singular linear transformations within each group. When testing an arbitrary decomposable structure, a simple asymptotic distribution of the statistic is not available. We suggested a simulation-based method to approximate the null distribution and calculate the corresponding $p$ value. We also gave some numerical results including both simulation and an empirical example to study the proposed testing procedure in different scenarios.

]]> Anita Race 1 1430317480 2015-04-29 14:24:40 1492118363 2017-04-13 21:19:23 0 0 event 2015-05-11T15:00:00-04:00 2015-05-11T15:00:00-04:00 2015-05-11T15:00:00-04:00 2015-05-11 19:00:00 2015-05-11 19:00:00 2015-05-11 19:00:00 2015-05-11T15:00:00-04:00 2015-05-11T15:00:00-04:00 America/New_York America/New_York datetime 2015-05-11 03:00:00 2015-05-11 03:00:00 America/New_York America/New_York datetime <![CDATA[]]>
<![CDATA[4th India's Supply Chain: Markets and Opportunities Conference (ISCMOC)]]> 27233 OVERVIEW

An executive forum designed for business decision-makers, strategic business planners, consultants and senior academics interested in the business opportunities and profit potential afforded by the rapid expansion and globalization of India’s Supply Chain

PRESENTING PARTNERS
UIBS, Georgia Ports Authority and Georgia Tech CIBER

SUPPORTING PARTNERS
Emirates Airlines, Taj Group, The U.S. Commercial Service (USCS), The Georgia Department of Economic Development (GDEcD), The Metro Atlanta Chamber (MAC), World Trade Center, Savannah, Confederation of Indian Industry (CII), the Georgia Tech Supply Chain and Logistics Institute (GTSCL) and USIBRC.

PRESENTATIONS AND SESSIONS

Fireside Chat with CEOs

Awards Reception and Networking

]]> Andy Haleblian 1 1431434225 2015-05-12 12:37:05 1492118360 2017-04-13 21:19:20 0 0 event The one day event will include keynote remarks and presentations by Chris Logan, Sr. Director of  Trade Development for the Georgia Ports Authority and Scott Fenwick, Senior Director at Manhattan Associates, Inc. Over a dozen other distinguished speakers will include high ranking officials from U.S. Commercial Service, Government of India, and subject matter experts from US and Indian companies and organizations.

]]>
2015-06-03T14:00:00-04:00 2015-06-03T18:00:00-04:00 2015-06-03T18:00:00-04:00 2015-06-03 18:00:00 2015-06-03 22:00:00 2015-06-03 22:00:00 2015-06-03T14:00:00-04:00 2015-06-03T18:00:00-04:00 America/New_York America/New_York datetime 2015-06-03 02:00:00 2015-06-03 06:00:00 America/New_York America/New_York datetime <![CDATA[]]> Ani Agnihotri, Program Chair – ISCMOC (Ani@usaindiabusinesssummit.com)

]]>
404391 404391 image <![CDATA[4th India's Supply Chain: Markets and Opportunities Conference (ISCMOC)]]> image/jpeg 1449254135 2015-12-04 18:35:35 1475895127 2016-10-08 02:52:07 <![CDATA[Download the event program]]> <![CDATA[Conference Registration Link]]> <![CDATA[Download the event flyer]]>
<![CDATA[FREE Developing a Demand-Driven Supply Chain Strategy webinar, Tuesday, July 14 1:30 - 2:30 pm EDT!]]> 27233 Attend our free Developing a Demand-Driven Supply Chain Strategy webinar to learn about the essential elements of demand-driven supply chains and the keys to developing a demand-driven supply chain strategy. As a thank you for attending, you will receive a *discount code towards towards the August 18-21 course in being held in Atlanta, GA (*please note that the discount cannot be combined with any other discount).

OVERVIEW

As supply chain executives become more instrumental in supporting long-term strategic objectives of their firms, they need to complement traditional supply chain operational knowledge with a more strategic view of their role in delivering aligned results to the business. This includes assessing their organization's current supply chain strategy and formulating a new one as well as understanding the key issues needed to implement a demand-driven supply chain strategy and how to continually review and align it. 

Our webinar will discuss the essential elements of demand-driven supply chains and the keys to developing a demand-driven supply chain strategy. The webinar will also provide a preview of the upcoming course in August where participants will have the opportunity to work with an extended simulation game of a fictional company to see the impact of supply chain strategic decisions in real time, while monitoring their ability to manage uncertainty and deliver financial results for the firm. 

ABOUT THE PRESENTER

Lew Roberts specializes in the fields of Business Performance Improvement and Supply Chain Management. His consulting and management development career over the last twenty five years has included working with Ryder, UPS-SCS, Caliber Logistics, DHL, Menlo Worldwide, BAX Global, Owens Corning, Coca-Cola, Georgia-Pacific, Baxter, De Beers, Mercedes Benz, Nissan and many other major firms worldwide. He has studied under Dr. Richard Schonberger, who is recognized as one of the worlds leading authorities in the fields of JIT and Business Performance Improvement.

]]> Andy Haleblian 1 1432648404 2015-05-26 13:53:24 1492118357 2017-04-13 21:19:17 0 0 event Attend our free Developing a Demand-Driven Supply Chain Strategy webinar to learn about the essential elements of demand-driven supply chains and the keys to developing a demand-driven supply chain strategy.

]]>
2015-07-14T14:30:00-04:00 2015-07-14T15:30:00-04:00 2015-07-14T15:30:00-04:00 2015-07-14 18:30:00 2015-07-14 19:30:00 2015-07-14 19:30:00 2015-07-14T14:30:00-04:00 2015-07-14T15:30:00-04:00 America/New_York America/New_York datetime 2015-07-14 02:30:00 2015-07-14 03:30:00 America/New_York America/New_York datetime <![CDATA[]]> webinar@scl.gatech.edu]]> 407371 407371 image <![CDATA[FREE Developing a Demand-Driven Supply Chain Strategy webinar]]> image/jpeg 1449254168 2015-12-04 18:36:08 1475895132 2016-10-08 02:52:12 <![CDATA[Register Online to Attend]]> <![CDATA[Link to Course Flyer]]> <![CDATA[Course webpage within the SCL website]]>
<![CDATA[FREE Optimizing Packaging’s Impact in the Supply Chain webinar, Wed July 29 1:30 - 2:30 pm EDT!]]> 27233 Attend our free Optimizing Packaging’s Impact in the Supply Chain webinar to learn how to take a holistic supply chain view toward packaging decisions, helping to achieve financial benefits as well as direct and positive sustainability impacts that can be measured in emissions and waste reduction. As a thank you for attending, you will receive a *discount code towards towards the September 1-2 course in being held in Atlanta, GA (*please note that the discount cannot be combined with any other discount).

OVERVIEW

Our interactive webinar will provide an overview of our two-day Professional Education course titled “Optimizing Packaging’s impact in the Supply Chain” as well as give attendees an opportunity to ask supply chain packaging-related questions. 

Packaging’s effect on the supply chain is often underestimated and can create inefficiencies that drive increased costs and wastes in logistics. Damage in shipping is detrimental to customer satisfaction, drives increased costs and negates sustainability efforts. The webinar will highlight the 7 Hazards in Distribution, a holistic view of packaging,optimized container and system utilization and strategies, and tactics to tackle packaging situations and turn them into savings. Designed to be interactive, the webinar is an opportunity for listeners and course candidates to ask course-specific questions as well as questions related to other supply chain packaging-related issues. We look forward to having you join us for a glimpse of the September event as well as relevant content for anyone looking to improve packaging to gain supply chain and sustainability benefits. 

ABOUT THE PRESENTER

Tom Blanck leads the Packaging Optimization Practice at Chainalytics as Principal where he manages the delivery of supply chain packaging optimization and complex package engineering services as well as packaging value improvement programs. Working with leading organizations, Tom has enabled packaging improvement across multiple industries, including CPG, Manufacturing, Medical, and Technology.

]]> Andy Haleblian 1 1432659436 2015-05-26 16:57:16 1492118356 2017-04-13 21:19:16 0 0 event Attend our free Optimizing Packaging’s Impact in the Supply Chain webinar to learn how to take a holistic supply chain view toward packaging decisions, helping to achieve financial benefits as well as direct and positive sustainability impacts that can be measured in emissions and waste reduction.

]]>
2015-07-29T14:30:00-04:00 2015-07-29T15:30:00-04:00 2015-07-29T15:30:00-04:00 2015-07-29 18:30:00 2015-07-29 19:30:00 2015-07-29 19:30:00 2015-07-29T14:30:00-04:00 2015-07-29T15:30:00-04:00 America/New_York America/New_York datetime 2015-07-29 02:30:00 2015-07-29 03:30:00 America/New_York America/New_York datetime <![CDATA[]]> webinar@scl.gatech.edu]]> 407431 407431 image <![CDATA[Optimizing Packaging’s Impact in the Supply Chain]]> image/jpeg 1449254168 2015-12-04 18:36:08 1475895132 2016-10-08 02:52:12 <![CDATA[Register Online to Attend]]> <![CDATA[Optimizing Packaging's Impact in the Supply Chain Course]]> <![CDATA[Link to Course Flyer]]>
<![CDATA["Mobile Supply Chain" presented by TAG Mobility & Supply Chain and Logistics]]> 27233 Please join TAG for a panel discussion moderated by Aaron's Inc. and HD Supply, with panelists from nuVizz, Inc., Smart Gladiator LLC, and TrackBlue LLC.

Event Details

Where: Heritage Sandy Springs (6110 Blue Stone Road, Sandy Springs, GA 30328)

When: Tuesday, Aug 18, 2015, 7:30 - 9AM

Overview
Mobile applications are rapidly changing the way our supply chains operate. In fact, we're seeing dramatic transformations in the areas of shipment visibility, order and inventory management, and warehouse management. Hear industry experts discuss challenges and share best practices for leveraging the mobile supply chain.

Description
Most consumers are intimately familiar with the use of mobile applications to enhance their retail and services experiences. B2C mobile applications are pervasive and relatively simple to create. But mobile applications also have a place in the B2B solution landscape, especially in the area of logistics and supply chain management. Shipment visibility, order and inventory management, and warehouse management are all applications. How are mobile applications changing the way our supply chains operate? How will supply chains evolve as mobility technology become less reliant on people? In this discussion, we will share challenges of and ideas for creating the mobile supply chain.

Register Today!

TAG members: FREE; Non-members $20

For more information and to register online, please visit http://bit.ly/TAG-mobilesc.

]]> Andy Haleblian 1 1435768980 2015-07-01 16:43:00 1492118349 2017-04-13 21:19:09 0 0 event Mobile applications are rapidly changing the way our supply chains operate. In fact, we're seeing dramatic transformations in the areas of shipment visibility, order and inventory management, and warehouse management. Hear industry experts discuss challenges and share best practices for leveraging the mobile supply chain.

]]>
2015-08-18T08:30:00-04:00 2015-08-18T10:00:00-04:00 2015-08-18T10:00:00-04:00 2015-08-18 12:30:00 2015-08-18 14:00:00 2015-08-18 14:00:00 2015-08-18T08:30:00-04:00 2015-08-18T10:00:00-04:00 America/New_York America/New_York datetime 2015-08-18 08:30:00 2015-08-18 10:00:00 America/New_York America/New_York datetime <![CDATA[]]> Please email membership@tagonline.org with any questions.

]]>
420041 420041 image <![CDATA[TAG Mobility & Supply Chain and Logistics present "Mobile Supply Chain"]]> image/png 1449254269 2015-12-04 18:37:49 1475895157 2016-10-08 02:52:37 <![CDATA[Register Online]]> <![CDATA[TAG Supply Chain & Logistics Society]]>
<![CDATA[PhD Thesis Defense - Tugce Isik]]> 27187 TITLE: Optimal Control of Queueing Systems with Non-Collaborating Servers

ABSTRACT:

In this thesis, we focus on effective management of cross-trained workforce in manufacturing systems. In particular, we analyze non-collaborative queueing networks where cross-trained (flexible) servers are not allowed to work at a station together. Our contributions to the understanding of systems with non-collaborative flexible servers can be summarized in two parts. (i)  we characterize the structure of optimal server assignment policies and draw insights to improve decision making in these systems. (ii) we  develop easy-to-implement policies with near-optimal performance for systems where the optimal policy is difficult to implement or is not analytically tractable.   In the first study, our goal is to identify the server assignment policy that maximizes the long-run average throughput in tandem networks with finite buffers and non-collaborative flexible servers. For Markovian systems with two stations and two servers, we characterize the optimal server assignment policy and demonstrate that the structure of the optimal policy is insensitive to the service requirement distributions. For larger tandem networks, we propose server assignment heuristics. Our numerical results suggest that in these systems, near-optimal throughputs can be achieved even if the server allocation decisions are made myopically. We also examine how lack of collaboration affects the performance of queueing systems with flexible servers. We show that the improvement that can be gained through collaboration is dependent on similarity of the tasks in the system, as well as the buffer sizes.   The second part focuses on tandem queueing networks with finite buffers and flexible servers where server reassignments  result in setups. For systems of arbitrary size with general service requirement distributions, we show that the policy that maximizes the long-run average profit becomes dedicated as the setup costs increase. We also characterize the profit-optimal server assignment policy for Markovian tandem lines with two stations, homogeneous tasks, and constant setup costs. Our results demonstrate that the structure of the optimal policy depends both on the magnitude of the setup costs and the buffer size. For systems with non-homogeneous tasks and/or non-constant setup costs, we provide near-optimal server assignment heuristics. Our computational results suggest that the relative performances of dynamic and dedicated server assignment policies are also dependent on the structure of the tasks.  Finally, we extend our analysis to queueing networks with general topology and routing. For non-collaborative networks with infinite buffers, we formulate a linear program that yields an upper bound on the long-run average throughput. We also introduce a processor sharing scheme for general queueing networks, and identify the optimal processor sharing policy for tandem lines with homogeneous tasks. For Markovian systems with two stations, finite buffers, and homogeneous tasks, we prove that processor sharing achieves the non-collaborative optimal throughput as the buffer size grows. For systems where processor sharing is not implementable, we propose a class of round-robin server assignment policies and show that they approximate processor sharing in systems with two stations. We evaluate the performance of the proposed class of policies in systems with various topologies and finite buffers, and demonstrate that they are near-optimal.]]> Anita Race 1 1436341007 2015-07-08 07:36:47 1492118347 2017-04-13 21:19:07 0 0 event 2015-07-21T14:00:00-04:00 2015-07-21T17:00:00-04:00 2015-07-21T17:00:00-04:00 2015-07-21 18:00:00 2015-07-21 21:00:00 2015-07-21 21:00:00 2015-07-21T14:00:00-04:00 2015-07-21T17:00:00-04:00 America/New_York America/New_York datetime 2015-07-21 02:00:00 2015-07-21 05:00:00 America/New_York America/New_York datetime <![CDATA[]]>
<![CDATA[PhD Thesis Defense - Javad Feizollahi]]> 27187 TITLE:  Large-Scale Unit Commitment: Decentralized Mixed Integer Programming Approaches

ABSTRACT:

In this dissertation, we investigate theory and application of decentralized optimization for mixed integer programming (MIP) problems. Our focus is on loosely coupled MIPs where different blocks of the problem have mixed integer linear feasible sets and a small number of linear constraints couple these blocks together. We develop decentralized optimization approaches based on Lagrangian and augmented Lagrangian duals for such MIPs. The contributions of this dissertation are as follows: a) proof of exactness of augmented Lagrangian dual (ALD) for MIPs, b) decentralized exact and heuristic algorithms for MIPs, and c) application to decentralized unit commitment (UC).

First, we prove that ALD is able to close the duality gap for MIPs. In particular, we show that with non-negative level bounded augmenting functions, ALD is able to asymptotically achieve zero duality gap for MIPs, when the penalty coefficient is allowed to go to infinity. We further show that, under some mild conditions, using any norm as the augmenting function ALD is able to close the duality gap of a MIP with a finite penalty coefficient.  

Nonlinear objective functions in ALD destroy the decomposability which exists in classical Lagrangian dual for a loosely coupled MIP. A key challenge is that, because of the non-convex nature of MIPs, classical distributed and decentralized optimization approaches such as alternating direction method of multipliers (ADMM) cannot be applied directly to find their optimal solutions. We propose three exact and one heuristic decentralized algorithms, which are based on extensions of ADMM and dual decomposition techniques.

 Finally, we apply the developed algorithms to solve decentralized UC. We present mathematical formulations for the UC problem which are appropriate for the proposed decentralized algorithms. Privacy concerns of the participants in UC are taken into account in these formulations. We propose a solution approach for decentralized UC, which exploits the structure of UC in our decentralized algorithms. We present extensive computational experiments for solving UC instances with different decentralized approaches. We illustrate the challenges arising from nonconvexity of UC problem and show how the proposed algorithms overcome these challenges. We demonstrate remarkable performance of parallel implementation of the heuristic decentralized algorithm to solve large scale UC instances on power systems of more than 3,000 buses. We also show that for small UC instances, the proposed exact algorithms are able to find global optimal solutions.

 

 

]]> Anita Race 1 1436784435 2015-07-13 10:47:15 1492118346 2017-04-13 21:19:06 0 0 event 2015-07-30T14:00:00-04:00 2015-07-30T16:00:00-04:00 2015-07-30T16:00:00-04:00 2015-07-30 18:00:00 2015-07-30 20:00:00 2015-07-30 20:00:00 2015-07-30T14:00:00-04:00 2015-07-30T16:00:00-04:00 America/New_York America/New_York datetime 2015-07-30 02:00:00 2015-07-30 04:00:00 America/New_York America/New_York datetime <![CDATA[]]>
<![CDATA[Using Lean Supply Chain Concepts to Maximize Customer Value, Employee Engagement and Improve Your Balance Sheets webinar, Wednesday, August 12 1:30 - 2:30 pm EDT!]]> 27233 Attend our free Using Lean Supply Chain Concepts to Maximize Customer Value, Employee Engagement and Improve Your Balance Sheets webinar for a live Q&A session and hear how an Atlanta-based organization has utilized lean supply chain education to maximize customer value, employee engagement, and overall balance sheet improvements. As a thank you for attending, you will receive a *discount code towards towards the September 15-17 course in being held in Atlanta, GA (*please note that the discount cannot be combined with any other discount).

OVERVIEW

How do we maximize employee engagement toward C-level goals? How does the work inside of an organization connect to the C-level targets? How can we maximize value at the lowest cost with respect for employees?

Join our webinar for a live Q&A session with a Lean Supply Chain Certificate 2011 graduate and Supply Chain Director and hear how this Atlanta-based organization has utilized lean supply chain education to maximize customer value, employee engagement, and overall balance sheet improvements.

The purpose of the Lean Supply Chain Professional certificate program is to transform the way supply chain professionals think, act and lead. Upon completion of the program, participants will be able to develop, lead and implement strategic and tactical elements of lean principles in the supply chain. We are committed to building individuals into serious, results-based Lean Supply Chain Professionals.

ABOUT THE PRESENTERS
Brad Bossence is instructor of the Georgia Tech Building the Lean Supply Chain Problem Solver course and Vice President of Leancor Supply Chain Group. Brad will be joined by Chris Jenkins, Director or Supply Chain for Mueller Co.

]]> Andy Haleblian 1 1437423402 2015-07-20 20:16:42 1492118344 2017-04-13 21:19:04 0 0 event Attend our free Using Lean Supply Chain Concepts to Maximize Customer Value, Employee Engagement and Improve Your Balance Sheets webinar for a live Q&A session with a Lean Supply Chain Certificate 2011 graduate and Supply Chain Director and hear how an Atlanta-based organization has utilized lean supply chain education to maximize customer value, employee engagement, and overall balance sheet improvements.

]]>
2015-08-12T14:30:00-04:00 2015-08-12T15:30:00-04:00 2015-08-12T15:30:00-04:00 2015-08-12 18:30:00 2015-08-12 19:30:00 2015-08-12 19:30:00 2015-08-12T14:30:00-04:00 2015-08-12T15:30:00-04:00 America/New_York America/New_York datetime 2015-08-12 02:30:00 2015-08-12 03:30:00 America/New_York America/New_York datetime <![CDATA[]]> webinar@scl.gatech.edu]]> 426771 426771 image <![CDATA[Using Lean Supply Chain Concepts to Maximize Customer Value, Employee Engagement and Improve Your Balance Sheets]]> image/jpeg 1449254342 2015-12-04 18:39:02 1475895165 2016-10-08 02:52:45 <![CDATA[Register Online to Attend]]> <![CDATA[Link to Series Flyer]]> <![CDATA[Course webpage within the SCL website]]>
<![CDATA[Savannah Logistics Lunch]]> 27233 The Center of Innovation for Logistics and HunterMaclean have developed a new event focused specifically on the coastal Georgia region: the Savannah Logistics Lunch. The 1st annual event will take place on Tuesday, Aug. 18, Embassy Suites, Savannah Airport. At this inaugural event, the centerpiece of the lunch is an impressive list of panelists who will address the topic of Cold Chain and Perishables. You won't want to miss it!

Learn more about the event at the Business in Savannah website.

]]> Andy Haleblian 1 1437840110 2015-07-25 16:01:50 1492118341 2017-04-13 21:19:01 0 0 event Continuing the conversation from the Georgia Logistics Summit. Hear about local updates on growth in persihables and cold chain infrastructure.

]]>
2015-08-18T12:45:00-04:00 2015-08-18T14:30:00-04:00 2015-08-18T14:30:00-04:00 2015-08-18 16:45:00 2015-08-18 18:30:00 2015-08-18 18:30:00 2015-08-18T12:45:00-04:00 2015-08-18T14:30:00-04:00 America/New_York America/New_York datetime 2015-08-18 12:45:00 2015-08-18 02:30:00 America/New_York America/New_York datetime <![CDATA[]]> bharmon@huntermaclean.com,cmcrae@huntermaclean.com, or 912-236-0261

]]>
428421 428421 image <![CDATA[Savannah Logistics Lunch]]> image/jpeg 1449254358 2015-12-04 18:39:18 1475895167 2016-10-08 02:52:47 <![CDATA[Article in Business in Savannah website]]> <![CDATA[Registration and more info within Center of Innovation for Logistics website]]>
<![CDATA[PhD Thesis Defense - Zhi Han]]> 27187 TITLE: Applications of Stochastic Control and Statistical Inference in Macroeconomics and High-Dimensional Data

ABTSRACT:

This thesis is focused on the optimality of stochastic control in macroeconomics and the fast algorithm of statistical inference. The first topic involves the proof and the calculation of the optimal drift control policy in foreign exchange reserve management. The second topic involves the fast computing algorithm of partial distance covariance statistics with its application in feature screening in high dimensional data.

In the first part of the dissertation, we study the problem of optimally controlling the level of foreign exchange reserve held by a country. When a reserve authority accumulates foreign exchange reserves to meet changing economic conditions, he faces the challenge of finding the right balance between the holding costs and the operational costs involved in adjusting the reserve size. We consider a foreign exchange reserve whose inventory fluctuation is modeled by a Brownian motion with drift, and at any moment the reserve manager can adjust the inventory level by varying the drift rate at which the reserve accumulates or depletes, but incurs a cost satisfies triangle inequality. When the reserve is accumulating or depleting, it also incurs a maintaining cost related with the current drift rate. The inventory level must be nonnegative at all times and continuously incurs a linear holding cost. The manager's problem is to decide when and how to change the drift rate so that the long run expected discounted cost of maintaining the foreign exchange reserve is minimized. We show that, under certain conditions, the control band policies are optimal for the discounted cost drift control problem and explicitly calculate the parameters of the optimal control band policy. In the two drift case, this form of policy is described by two parameters $\{L, U\}$, $0 < L < U$. When the inventory falls to $L$ (rises to $U$), the controller switch the drift rate to depletion (accumulation). We also extend the result to the multiple drift case and develop an algorithm to calculate the optimal thresholds of the optimal control band policy.

In the second part of the dissertation we study the problem of fast computing algorithm of partial distance covariance. If the computation of partial distance covariance is implemented directly accordingly to its definition then its computational complexity is $O(n2)$ which may hinder the application of an algorithm. To illustrate it, if $n$ is equal to $10^6$, an $O(n2)$ algorithm will need $10^{12}$ numerical operations, which is impossible even for modern computers. In comparison, an $O(n \log n)$ algorithm will only require around $106$ numerical operations, which is doable. In this part of the thesis, we show that an $O(n \log n)$ algorithm for a version of the partial distance covariance exits. The derivation of the fast algorithm involves significant reformulation from the original version of partial distance covariance. We also demonstrate its application in feature screening in high dimensional data in the following part of the thesis.

In the final part of the thesis we further study the feature screening problem in high dimensional data. We propose an iterative feature screening procedure based on the partial distance covariance. This procedure can simultaneously address the two issue when using sure independence screening (SIS) procedure. First, an important predictor that is marginally uncorrelated but jointly correlated with the response cannot be picked by SIS and thus will not enter the estimation model. Second, SIS works only for linear models, and performance is very unstable in other nonlinear models. To the best of our knowledge, this is the first time that the ``new metric'' -- partial distance covariance -- is used for feature screening, and the idea of conditional screening is formally developed.

]]> Anita Race 1 1438158403 2015-07-29 08:26:43 1492118341 2017-04-13 21:19:01 0 0 event 2015-08-07T14:00:00-04:00 2015-08-07T14:00:00-04:00 2015-08-07T14:00:00-04:00 2015-08-07 18:00:00 2015-08-07 18:00:00 2015-08-07 18:00:00 2015-08-07T14:00:00-04:00 2015-08-07T14:00:00-04:00 America/New_York America/New_York datetime 2015-08-07 02:00:00 2015-08-07 02:00:00 America/New_York America/New_York datetime <![CDATA[]]>
<![CDATA[PhD Thesis Defense - Josh McDonald]]> 27187 TITLE:  Topics in the Statistical Aspects of Simulation

ABSTRACT:

In the first part of the thesis, we apply various variance reduction techniques to the estimation of Asian averages and options and propose an easy-to-use quasi-Monte Carlo method that can provide significant variance reductions with minimal increases in computational time. We have also extended these techniques to estimate higher moments of the Asians averages.


In the second part, we then use these estimated moments to efficiently implement Gram-Charlier based estimators for probability distribution functions of Asian averages and options.


Finally, in the third part, we investigate a ranking and selection application that uses post hoc analysis to determine how the circumstances of procedure termination affect the probability of correct selection.

 

]]> Anita Race 1 1438599270 2015-08-03 10:54:30 1492118339 2017-04-13 21:18:59 0 0 event 2015-08-11T10:00:00-04:00 2015-08-11T10:00:00-04:00 2015-08-11T10:00:00-04:00 2015-08-11 14:00:00 2015-08-11 14:00:00 2015-08-11 14:00:00 2015-08-11T10:00:00-04:00 2015-08-11T10:00:00-04:00 America/New_York America/New_York datetime 2015-08-11 10:00:00 2015-08-11 10:00:00 America/New_York America/New_York datetime <![CDATA[]]>
<![CDATA[PhD Thesis Defense - Isil Alev]]> 27187 TITLE:  Operational Perspectives on Extended Producer Responsibility for Durable and Consumable Products

ABSTRACT:

Growing post-consumer waste and associated environmental and public health concerns have resulted in more regulated waste management. In this context, Extended Producer Responsibility (EPR) has emerged as an environmental policy concept that focuses on the “polluter pays” principle. This principle shifts the economic burden of waste management on producers by imposing collection and recycling obligations. Over the last two decades, EPR has gained momentum all around the world for several product categories including batteries, carpet, leftover paint, pharmaceuticals, and electronics. This thesis consists of three essays that contribute to the understanding of the economic implications of EPR from an operational perspective by analyzing how EPR affects the markets for certain durable (such as electronics) and consumable (such as pharmaceuticals) products.

In the first essay “Extended Producer Responsibility and Secondary Markets”, we investigate the effect of EPR-based policy on a durable good producer's secondary market strategy. We consider incentives of durable good producers to recover used products from the secondary markets and discard them. We base our analysis on a discrete-time, sequential, producer-consumer game over an infinite time horizon, where the producer offers a buy-back program that pays consumers a fair market value for the return of their used products (e.g. Dell, HP, Fujitsu, Apple).  We completely characterize the secondary market strategy of the producer at Markov-perfect, stationary equilibrium. We find that secondary market interference through buying back may deteriorate environmental outcomes by increasing new production and reducing reuse levels. We provide insights into how to set EPR obligations to avoid these adverse outcomes. Furthermore, we validate our results by calibrating with real-life data and considering a number of extensions that represent different operational environments. Our analysis collectively uncovers possible strategic approach of durable good producers to EPR obligations and suggests that EPR obligations may result in unintended outcomes in a durable setting.

In the second essay “A Market-Based Extended Producer Responsibility Implementation - The Case of Minnesota Electronics Recycling Act”, we investigate the operational implementation details of EPR-based policy on the ground. In the implementation of EPR-based policies, ``market-based" approach has recently become the mostly advocated approach. Its main premises are to promote cost efficiency and to achieve better environmental outcomes by adopting free market principle and setting desirable targets for collection with broad flexibilities. In this essay, we analyze whether these premises hold by focusing on the Minnesota Electronics Recycling Act, a prevailing example of marked-based EPR policy implementation in the US. Based on publicly available reports and our interviews with the stakeholders, we explore its implementation rules, stakeholder perspectives, and resulting outcomes together with underlying dynamics. We find that the Minnesota Act appears to achieve the premises of the market-based approach, but this possibly occurs at the expense of several environmental disadvantages. Our analysis suggests that these disadvantages arise from market dynamics at the implementation stage and associated stakeholder interactions.

In the third essay “Extended Producer Responsibility for Pharmaceuticals”, we focus on EPR-based policies for unused pharmaceuticals, a category of products with a consumable nature. EPR-based policies have recently gained popularity for pharmaceuticals to address their recently recognized environmental and public health externalities (e.g. EU, British Columbia, Alameda and King Counties in the US). However, little is known regarding the effectiveness of these policies for pharmaceuticals and little guidance can be obtained from EPR implementations for durable products, because product characteristics demand structures and market dynamics are very different. Motivated by this, we analyze how the EPR concept can be effectively operationalized for pharmaceuticals by analyzing the interactions in the pharmaceutical chain as they relate to EPR. We build a game-theoretical model that focuses on major stakeholders (pharmaceutical producers, doctors, patients, the environment and public health) and their unique and complex interactions as well as moderating factors for these interactions (pharmaceutical promotions, mediated demand structure due to doctor-patient interaction). With this framework, we investigate the effectiveness of EPR-based policies and demonstrate that the preferred policy from the welfare perspective depends on the healthcare and externality characteristics of the medicine together with collection-related requirements in place. This shows that experiences and well-established premises learned from EPR implementations for durable products do not necessarily hold for consumables such as pharmaceuticals.

]]> Anita Race 1 1439542393 2015-08-14 08:53:13 1492118331 2017-04-13 21:18:51 0 0 event 2015-08-28T12:00:00-04:00 2015-08-28T12:00:00-04:00 2015-08-28T12:00:00-04:00 2015-08-28 16:00:00 2015-08-28 16:00:00 2015-08-28 16:00:00 2015-08-28T12:00:00-04:00 2015-08-28T12:00:00-04:00 America/New_York America/New_York datetime 2015-08-28 12:00:00 2015-08-28 12:00:00 America/New_York America/New_York datetime <![CDATA[]]>
<![CDATA[ISyE Seminar Series - Paul Glasserman]]> 27187 TITLE:  Robust Monte Carlo

ABSTRACT:

Simulation methodology has traditionally focused on measuring and reducing sampling error in simulating well-specified models; it has given less attention to quantifying the effect of model error or model uncertainty.  But simulation actually lends itself well to bounding this sort of model risk. In particular, if the set of alternative models consists of all models within a certain “distance” of a baseline model, then the potential effect of model risk can be estimated at low cost within a simulation of the baseline model. I will illustrate this approach to making Monte Carlo robust with examples from finance, where concerns about model risk have received heightened attention. The problem of bounding “wrong-way risk” in  counterparty risk presents a related question in which model uncertainty is limited to the nature of the dependence between two otherwise certain marginal models for market and credit risk.  The effect of uncertain dependence can be bounded through convenient combinations of simulation with linear programming and/or convex optimization. This talk is based on work with Xingbo Xu and Linan Yang.

 

]]> Anita Race 1 1439826440 2015-08-17 15:47:20 1492118328 2017-04-13 21:18:48 0 0 event 2015-09-09T16:00:00-04:00 2015-09-09T17:00:00-04:00 2015-09-09T17:00:00-04:00 2015-09-09 20:00:00 2015-09-09 21:00:00 2015-09-09 21:00:00 2015-09-09T16:00:00-04:00 2015-09-09T17:00:00-04:00 America/New_York America/New_York datetime 2015-09-09 04:00:00 2015-09-09 05:00:00 America/New_York America/New_York datetime <![CDATA[]]>
<![CDATA[ISyE Seminar Series - Erhan Bayraktar]]> 27187 TITLE:  On Hedging American Options (or hedging with them) under Model Uncertainty

ABSTRACT:

We consider as given a discrete time financial market with a risky asset and options written on that asset, and we determine both the sub- and superhedging prices of an American option in the model independent framework. We do this first when some (path dependent) European options are available for static hedging. We then analyze what happens when some American options are added to the mix of options that are available for hedging.

 Bio:

 Erhan Bayraktar is a full professor of Mathematics at the University of Michigan, where he has been since 2004. He is the holder of the Susan M. Smith Chair since 2010 and also is the co-director of the Quantitative Finance and Risk Management Masters Program. Professor Bayraktar’s research is in stochastic analysis, control, applied probability and mathematical finance. He has over 90 publications in prestigious journals in these areas. He is in the editorial boards of Mathematics of Operations Research, Mathematical Finance and the SIAM Journal on Control and Optimization. His research has been continually funded by the National Science Foundation. In particular, he received a CAREER grant in 2010. He received the inaugural junior scientist prize of the SIAM Activity Group on Financial Mathematics and Engineering in 2010.

 Professor Bayraktar received his Bachelor’s degree (a double major in Electrical Engineering and Mathematics) from Middle East Technical University in Ankara in 2000. He received his Ph.D. degree from Princeton in 2004.

]]> Anita Race 1 1439826522 2015-08-17 15:48:42 1492118328 2017-04-13 21:18:48 0 0 event 2015-09-16T16:00:00-04:00 2015-09-16T17:00:00-04:00 2015-09-16T17:00:00-04:00 2015-09-16 20:00:00 2015-09-16 21:00:00 2015-09-16 21:00:00 2015-09-16T16:00:00-04:00 2015-09-16T17:00:00-04:00 America/New_York America/New_York datetime 2015-09-16 04:00:00 2015-09-16 05:00:00 America/New_York America/New_York datetime <![CDATA[]]>
<![CDATA[ISyE Seminar Series - Jiawei Zhang]]> 27187 TITLE: Recent Results on Process Flexibility

ABSTRACT:

Process flexibility has been widely applied in many industries as a competitive strategy to improve responsiveness to demand uncertainty. In their seminal work, Jordan and Graves proposed an important flexibility concept, called the long chain design, and studied its empirical performance. In this talk, we summarize recent theoretical developments in understanding the effectiveness of the long chain design as well as other sparse flexibility structures. 

Bio:

Jiawei Zhang is Professor of Information, Operations, and Management Sciences and Harold MacDowell Faculty Fellow at the Stern School of Business, New York University. He holds a PhD from Stanford University and a B.S. degree and an M.S. degree from Tsinghua University.  Professor Zhang's research interests include business analytics and optimization. His work has appeared in Management ScienceManufacturing and Service OperationsMathematical Programming, Mathematics of Operations ResearchOperations ResearchSIAM Journal on Computing, etc. Professor Zhang was a recipient of the INFORMS Optimization Prize for Young Researchers in 2004. He currently serves as Associate Editor for several journals including Operations Research (2006-), Mathematics of Operations Research (2009-), and Management Science (2014-).

 

]]> Anita Race 1 1439826616 2015-08-17 15:50:16 1492118328 2017-04-13 21:18:48 0 0 event 2015-10-14T16:00:00-04:00 2015-10-14T17:00:00-04:00 2015-10-14T17:00:00-04:00 2015-10-14 20:00:00 2015-10-14 21:00:00 2015-10-14 21:00:00 2015-10-14T16:00:00-04:00 2015-10-14T17:00:00-04:00 America/New_York America/New_York datetime 2015-10-14 04:00:00 2015-10-14 05:00:00 America/New_York America/New_York datetime <![CDATA[]]>
<![CDATA[ISyE Seminar Series - Retsef Levi]]> 27187 TITLE:  Exploration vs. Exploitation: Reducing Uncertainty in Operational Problems

ABSTRACT:

Motivated by several core operational applications, we introduce a new class of multistage stochastic optimization models that capture a fundamental tradeoff between performing work and making decisions under uncertainty (exploitation) and investing capacity (and time) to reduce the uncertainty in the decision making (exploration/testing). Unlike existing models, in which the exploration-exploitation tradeoffs typically relate to learning the underlying distributions, the models we introduce assume a known probabilistic characterization of the uncertainty, and focus on the tradeoff of learning (or partially learning) the exact realizations.

Focusing on core scheduling models, we derive insightful structural results on the optimal policies that lead to: (i) Low dimensional dynamic programming formulations; (ii) quantification of the value of learning; (iii) surprising results on the optimality of local (myopic) decision rules for when it is optimal to explore (learn). We then generalize some of the results to a general class of stochastic combinatorial optimization models defined over contra-polymatroids.

The talk is based on several papers that are joint work with Chen Atias, Robi Krauthgamer, Tom Magnanti and Yaron Shaposhnik. 

 Bio:

Retsef Levi is the J. Spencer Standish (1945) Professor of Operations Management at the MIT Sloan School of Management. He is a member of the Operations Management Group at MIT Sloan and affiliated with the Operations Research Center and the Computational for Design and Optimization Program. Prior to joining MIT in 2006, he spent a year as a Goldstine Postdoctoral Fellow at the IBM T.J. Watson Research Center.  He received a Bachelor's degree in Mathematics from Tel-Aviv University in 2001, and a PhD in Operations Research from Cornell University in 2005. He spent more than 11 years in the Israeli Defense Forces as an Office in the Intelligence Wing. Levi's current research is focused on the design of analytical data-driven decision support models and tools addressing complex business and system design decisions under uncertainty in areas, such as health and healthcare management, supply chain, procurement and inventory management, revenue management, pricing optimization and logistics. He is interested in the theory underlying these models and algorithms, as well as their computational and organizational applicability in practical settings. Levi is leading several industry-based collaborative research efforts with some of the major academic hospitals in the Boston area, such as Mass General Hospital (MGH), Beth Israel Deaconess Medical Center (BIDMC), Children’s Hospital, and across the US (e.g., Memorial Sloan Kettering Cancer Center, NYC Prebyterian Hospital System and the American Association of Medical Colleges). Levi is the lead PI on an MIT contract with the Federal Drug Administration (FDA) to develop systematic risk management approach to address risk related to economically motivated adulterations of food and drug products manufactured in China. Levi received the NSF Faculty Early Career Development award, the 2008 INFORMS Optimization Prize for Young Researchers and the 2013 Daniel H. Wagner Prize.

]]> Anita Race 1 1439826734 2015-08-17 15:52:14 1492118328 2017-04-13 21:18:48 0 0 event 2015-10-21T16:00:00-04:00 2015-10-21T17:00:00-04:00 2015-10-21T17:00:00-04:00 2015-10-21 20:00:00 2015-10-21 21:00:00 2015-10-21 21:00:00 2015-10-21T16:00:00-04:00 2015-10-21T17:00:00-04:00 America/New_York America/New_York datetime 2015-10-21 04:00:00 2015-10-21 05:00:00 America/New_York America/New_York datetime <![CDATA[]]>
<![CDATA[ISyE Seminar Series - Mariel Lavieri]]> 27187 TITLE:  Optimal Coinsurance Rates for a Heterogeneous Patient Population under Constraints on Inequality and Resources

ABSTRACT:

While operations research has contributed heavily to the derivation of optimal treatment guidelines for chronic disease, patient adherence to treatment plans is low and variable. It is estimated that half of all patients at risk of cardiovascular disease are not fully adherent to their prescribed medications. One mechanism for improving patient adherence to guidelines is to tailor coinsurance rates for prescription medications to patient characteristics. For patients insured by Medicare, we seek to find coinsurance rates which maximize the welfare of the heterogeneous patient population at risk for cardiovascular disease. We analyze the problem as a bilevel optimization model where the lower optimization problem has the structure of a Markov decision process which determines the optimal hypertension treatment plan for each patient class. The upper optimization problem is a nonlinear resource allocation problem with constraints on total expenditures and coinsurance inequality. The models are parameterized using data from the National Health and Nutrition Examination Survey (NHANES). We find that optimizing coinsurance rates can be a cost-effective intervention for improving patient adherence and health outcomes, particularly for those patients at high risk for cardiovascular disease. This research is done in collaboration with Greggory J. Schell (former PhD student) as well as University of Michigan and U.S. Department of Veteran Affairs clinicians: Rodney A. Hayward, and Jeremy B. Sussman.

BIO:

Dr. Mariel Lavieri is an Assistant Professor in the Department of Industrial and Operations Engineering at the University of Michigan.  She has bachelor's degrees in Industrial and Systems Engineering and Statistics and a minor in String Bass Performance from the University of Florida. She holds a Masters and PhD in Management Science from the University of British Columbia. In her work, she applies operations research to healthcare topics. Her most recent research develops dynamic programming, stochastic control, and continuous, partially observable state space models to guide screening, monitoring and treatment decisions of chronic disease patients. She has also developed models for health workforce planning which take into account training requirements, workforce attrition, capacity planning, promotion rules and learning. Dr. Lavieri is the recipient of the Bonder Scholarship, and an honorary mention in the George B. Dantzig Dissertation award. She received the 2009 Pierskalla Award for the best paper presented in the Health Applications Society at INFORMS and mentored students who won the 2012 Doing Good with Good OR, the 2013 Society for Medical Decision Making Lee Lusted Award for Quantitative Methods and Theoretical Developments and the 2015 IBM Research Service Science Best Student Paper Award (hosted by the INFORMS Service Science Section). Dr. Lavieri was named the 2013 Young Participant with Most Practical Impact by the International Conference on Operations Research.

]]> Anita Race 1 1439826839 2015-08-17 15:53:59 1492118328 2017-04-13 21:18:48 0 0 event 2015-11-11T15:00:00-05:00 2015-11-11T16:00:00-05:00 2015-11-11T16:00:00-05:00 2015-11-11 20:00:00 2015-11-11 21:00:00 2015-11-11 21:00:00 2015-11-11T15:00:00-05:00 2015-11-11T16:00:00-05:00 America/New_York America/New_York datetime 2015-11-11 03:00:00 2015-11-11 04:00:00 America/New_York America/New_York datetime <![CDATA[]]>
<![CDATA[ISyE Seminar Series - Dan Bienstock]]> 27187 TITLE: Recent results on polynomial optimization problems

ABSTRACT:

Polynomial optimization problems, as the name suggests, are optimization problem where the objective function as well as the constraints are described by polynomials.  Such problems have acquired increased interest to some degree because of applications in engineering and science, where constraints arise because of physics, and also because of increased theoretical understanding.  In this talk I will focus on two topics where I am working, the CDT (Celis Dennis Tapia) problem, which concerns the solution of a system of quadratic inequalities over R^n, and mixed-integer polynomial optimization problems over graphs with structural sparsity, i.e. low treewidth.  We will describe our results, but also we will discuss how these problems relate to classical problems in various branches of mathematics.

 

Bio:

Daniel Bienstock is a professor at the departments of Industrial Engineering and Operations Research, and Applied Physics and Applied Mathematics, Columbia University, where he has been since 1989.   He received the PhD from MIT in Operations Research. His research focuses on discrete and nonconvex optimization, from both a theoretical and a computational standpoint, and applications, such as the mathematics of power grids.  He was plenary speaker at the 2005 SIAM Conference on Optimization, semi-plenary speaker at the 2006 ISMP meeting, became an INFORMS fellow in 2013, and is editor-in-chief of Mathematical Programming C.

]]> Anita Race 1 1439826940 2015-08-17 15:55:40 1492118328 2017-04-13 21:18:48 0 0 event 2015-11-18T15:00:00-05:00 2015-11-18T16:00:00-05:00 2015-11-18T16:00:00-05:00 2015-11-18 20:00:00 2015-11-18 21:00:00 2015-11-18 21:00:00 2015-11-18T15:00:00-05:00 2015-11-18T16:00:00-05:00 America/New_York America/New_York datetime 2015-11-18 03:00:00 2015-11-18 04:00:00 America/New_York America/New_York datetime <![CDATA[]]>
<![CDATA[ISyE Seminar Series - Alfred Hero]]> 27187 TITLE:  Graph continuum limits in data science

ABSTRACT:

Many problems in data science fields including data mining, computer vision, and machine learning involve combinatorial optimization over a graphs, e.g., minimal spanning trees, traveling salesman tours, or k-point minimal graphs over a feature space.  Certain properties of minimal graphs like their length, minimal paths, or span have continuum limits  as the number of nodes approaches infinity. These include problems  arising in spectral clustering, statistical classification, multi-objective learning, and anomaly detection. In some cases these continuum limits lead to analytical approximations that can  break the combinatorial bottleneck.  In this talk, I will present an overview of some of the remarkable theory of graph continuum limits and illustrate with data science applications.

Bio:
Alfred O. Hero III received the B.S. (summa cum laude) from Boston University (1980) and the Ph.D from Princeton University (1984), both in Electrical Engineering. Since 1984 he has been with the University of Michigan, Ann Arbor, where he is the R. Jamison and Betty Williams Professor of Engineering and co-director of the Michigan Institute for Data Science (MIDAS) . His primary appointment is in the Department of Electrical Engineering and Computer Science and he also has appointments, by courtesy, in the Department of Biomedical Engineering and the Department of Statistics. From 2008 to 2013 he held the Digiteo Chaire d'Excellence, sponsored by Digiteo Research Park in Paris, located at the Ecole Superieure d'Electricite, Gif-sur-Yvette, France. He has held other visiting positions at LIDS Massachusetts Institute of Technology (2006), Boston University (2006), I3S University of Nice, Sophia-Antipolis, France (2001), Ecole Normale Sup\'erieure de Lyon (1999), Ecole Nationale Sup\'erieure des T\'el\'ecommunications, Paris (1999), Lucent Bell Laboratories (1999), Scientific Research Labs of the Ford Motor Company, Dearborn, Michigan (1993), Ecole Nationale Superieure des Techniques Avancees (ENSTA), Ecole Superieure d'Electricite, Paris (1990), and M.I.T. Lincoln Laboratory (1987 - 1989).

Alfred Hero is a Fellow of the Institute of Electrical and Electronics Engineers (IEEE). He received the University of Michigan Distinguished Faculty Achievement Award (2011). He has been plenary and keynote speaker at several workshops and conferences. He has received several best paper awards including: an IEEE Signal Processing Society Best Paper Award (1998), a Best Original Paper Award from the Journal of Flow Cytometry (2008), a Best Magazine Paper Award from the IEEE Signal Processing Society (2010), a SPIE Best Student Paper Award (2011), an IEEE ICASSP Best Student Paper Award (2011), an AISTATS Notable Paper Award (2013), and an IEEE ICIP Best Paper Award (2013). He received an IEEE Signal Processing Society Meritorious Service Award (1998), an IEEE Third Millenium Medal (2000), an IEEE Signal Processing Society Distinguished Lecturership (2002), and a IEEE Signal Processing Society Technical Achievement Award (2014).  In 2015 he received the Society Award, which is the highest award bestowed by
the IEEE Signal Processing Society.

Al Hero was President of the IEEE Signal Processing Society (2006-2007). He was a member of the IEEE TAB Society Review Committee (2009), the IEEE Awards Committee (2010-2011), and served on the Board of Directors of the IEEE (2009-2011) as Director of Division IX (Signals and Applications). He served on the IEEE TAB Nominations and Appointments Committee (2012-2014). Alfred Hero is currently a member of the Big Data Special Interest Group (SIG) of the IEEE Signal Processing Society. Since 2011 he has been a member of the Committee on Applied and Theoretical Statistics (CATS) of the US National Academies of Science.

Alfred Hero's recent research interests are in the data science of high dimensional spatio-temporal data, statistical signal processing, and machine learning. Of particular interest are applications to networks, including social networks, multi-modal sensing and tracking, database indexing and retrieval, imaging, biomedical signal processing, and biomolecular signal processing.

]]> Anita Race 1 1439827052 2015-08-17 15:57:32 1492118328 2017-04-13 21:18:48 0 0 event 2015-12-09T15:00:00-05:00 2015-12-09T16:00:00-05:00 2015-12-09T16:00:00-05:00 2015-12-09 20:00:00 2015-12-09 21:00:00 2015-12-09 21:00:00 2015-12-09T15:00:00-05:00 2015-12-09T16:00:00-05:00 America/New_York America/New_York datetime 2015-12-09 03:00:00 2015-12-09 04:00:00 America/New_York America/New_York datetime <![CDATA[]]>
<![CDATA[FREE Challenges and Opportunities in Inventory Management webinar, Tues August 25 1:30pm EDT!]]> 27233 Attend our free Challenges and Opportunities in Inventory Management webinar to learn about some of the key challenges of inventory management as well as opportunities to help improve overall supply chain performance through better practice. As a thank you for attending, you will receive a *discount code towards towards the September 22-24 course in being held in Atlanta, GA (*please note that the discount cannot be combined with any other discount).

OVERVIEW

Our interactive webinar will provide an overview of our three-day Professional Education course titled “Inventory Planning and Management” as well as give attendees an opportunity to ask their inventory-related questions. 

Traditional inventory models focus on the trade-off between customer service and the costs of overstocking. The inappropriate use of these models, however, has led many organizations to do poorly on both measures. During our webinar, we will discuss some of the key challenges of inventory management as well as opportunities to help improve overall supply chain performance through better practice. 

ABOUT THE PRESENTER

Paul Griffin is the Joseph C. Mello Chair in the Stewart School of Industrial & Systems Engineering at Georgia Tech and research director for healthcare delivery in the Center for Health & Humanitarian Systems at Georgia Tech. Dr. Griffin is also one of the instructors for the Inventory Planning and Management course.

]]> Andy Haleblian 1 1439830834 2015-08-17 17:00:34 1492118326 2017-04-13 21:18:46 0 0 event Attend our free Challenges and Opportunities in Inventory Management webinar to learn about some of the key challenges of inventory management as well as opportunities to help improve overall supply chain performance through better practice.

]]>
2015-08-25T14:30:00-04:00 2015-08-25T15:30:00-04:00 2015-08-25T15:30:00-04:00 2015-08-25 18:30:00 2015-08-25 19:30:00 2015-08-25 19:30:00 2015-08-25T14:30:00-04:00 2015-08-25T15:30:00-04:00 America/New_York America/New_York datetime 2015-08-25 02:30:00 2015-08-25 03:30:00 America/New_York America/New_York datetime <![CDATA[]]> webinar@scl.gatech.edu]]> 435431 435431 image <![CDATA[Challenges and Opportunities in Inventory Management webinar]]> image/jpeg 1449256162 2015-12-04 19:09:22 1475895174 2016-10-08 02:52:54 <![CDATA[Register Online to Attend]]> <![CDATA[Inventory Planning and Management course]]> <![CDATA[Link to Course Flyer]]>
<![CDATA[ISyE/SCL September 2015 Supply Chain Day]]> 27233 ISyE students, please join us for our first fall Supply Chain Day! The 3-hour session will host supply chain representatives from UPS, CaterpillarChainalyticsLeanCor, Alexander Proudfoot and West Rock who will be on campus to educate ISyE students about their organizations and available employment opportunities. Plus, enjoy a free pizza lunch!

EVENT DETAILS

Where: ISyE Main Bldg, 2nd Floor Atrium

When: Wednesday, September 9, 10AM-1PM

What: The session will include:

Please plan on staying for the duration of the event and bring copies of your resume and business cards. Dress is business casual.

REGISTER ONLINE by September 4! Seating is limited!

EVENT SPONSOR

The event is sponsored through the generosity and support of APICS - Atlanta Chapter. APICS is a non-profit educational organization addressing operations management and supply chain management issues, and providing professional development opportunities to our members. Visit http://www.apicsatlanta.org/ to learn more and make sure to stop by the APICS table at the event.

]]> Andy Haleblian 1 1440150338 2015-08-21 09:45:38 1492118319 2017-04-13 21:18:39 0 0 event ISyE students, please join us for our first fall Supply Chain Day! The 3-hour session will host supply chain representatives from UPS, CaterpillarChainalyticsLeanCor, Alexander Proudfoot and West Rock who will be on campus to educate ISyE students about their organizations and available employment opportunities. Plus, enjoy a free pizza lunch!

]]>
2015-09-09T11:00:00-04:00 2015-09-09T14:00:00-04:00 2015-09-09T14:00:00-04:00 2015-09-09 15:00:00 2015-09-09 18:00:00 2015-09-09 18:00:00 2015-09-09T11:00:00-04:00 2015-09-09T14:00:00-04:00 America/New_York America/New_York datetime 2015-09-09 11:00:00 2015-09-09 02:00:00 America/New_York America/New_York datetime <![CDATA[]]> event@scl.gatech.edu

]]>
438471 438471 image <![CDATA[Supply Chain Day - September 9, 2015]]> image/png 1449256175 2015-12-04 19:09:35 1475895176 2016-10-08 02:52:56 <![CDATA[Register online to attend (for ISyE students)]]>
<![CDATA[Statistics Seminar - Ray-Bing Chen]]> 27187 TITLE:  Bayesian Sparse Group Selection

ABSTRACT:

A Bayesian approach is proposed for the sparse group selection problem in the regression model. In this problem, the variables are partitioned into different disjoint groups. It is assumed that only a small number of groups are active for explaining the response variable, and it is further assumed that within each active group only a small number of variables are active. We adopt a Bayesian hierarchical formulation, where each candidate group is associated with a binary variable indicating whether the group is active or not. Within each group, each candidate variable is also associated with a binary indicator, too. Thus the sparse group selection problem can be solved by sampling from the posterior distribution of the two layers of indicator variables. We adopt a group-wise Gibbs sampler for posterior sampling. We demonstrate the proposed method by simulation studies as well as real examples. The simulation results show that the proposed method performs better than the sparse group Lasso in terms of selecting the active groups as well as identifying the active variables within the selected groups.

]]> Anita Race 1 1440687362 2015-08-27 14:56:02 1492118312 2017-04-13 21:18:32 0 0 event 2015-09-03T12:00:00-04:00 2015-09-03T12:00:00-04:00 2015-09-03T12:00:00-04:00 2015-09-03 16:00:00 2015-09-03 16:00:00 2015-09-03 16:00:00 2015-09-03T12:00:00-04:00 2015-09-03T12:00:00-04:00 America/New_York America/New_York datetime 2015-09-03 12:00:00 2015-09-03 12:00:00 America/New_York America/New_York datetime <![CDATA[]]>
<![CDATA[FREE Effectively Managing Global Supply and Risk webinar, Wed Sept 2 1:30pm EDT!]]> 27233 Attend our free Effectively Managing Global Supply and Risk webinar to understand the risks and issues associated with sourcing globally, and also learn how to approach and develop mitigation strategies for these associated risks. As a thank you for attending, you will receive a *discount code towards towards the October 6-8 course being held in Atlanta, GA (*please note that the discount cannot be combined with any other discount).

OVERVIEW

Today's supply managers are finding that they need to do much work in terms of conceptualizing, designing, and implementing initiatives that may be effective globally. In addition, supply managers need to understand the risks inherent in sourcing globally and be able to develop mitigation strategies for these risks. Our interactive webinar will provide an overview of our three-day Professional Education course titled “Effectively Managing Global Supply and Risk in an Increasingly Complex World” as well as give attendees an opportunity to ask questions about the topics presented.

ABOUT THE PRESENTER

Lew Roberts is an SCL professional education instructor and President and founder of L. Roberts & Associates Inc., a USA based firm which provides a wide range of professional consulting and management development services aimed at improving business performance, with an emphasis in the field of supply chain management.

]]> Andy Haleblian 1 1440689210 2015-08-27 15:26:50 1492118312 2017-04-13 21:18:32 0 0 event Attend our free Effectively Managing Global Supply and Risk webinar to understand the risks and issues associated with sourcing globally.

]]>
2015-09-02T14:30:00-04:00 2015-09-02T15:30:00-04:00 2015-09-02T15:30:00-04:00 2015-09-02 18:30:00 2015-09-02 19:30:00 2015-09-02 19:30:00 2015-09-02T14:30:00-04:00 2015-09-02T15:30:00-04:00 America/New_York America/New_York datetime 2015-09-02 02:30:00 2015-09-02 03:30:00 America/New_York America/New_York datetime <![CDATA[]]> webinar@scl.gatech.edu]]> 441061 441061 image <![CDATA[Effectively Managing Global Supply and Risk]]> image/jpeg 1449256190 2015-12-04 19:09:50 1475895179 2016-10-08 02:52:59 <![CDATA[Register Online to Attend]]> <![CDATA[Effectively Managing Global Supply and Risk in an Increasingly Complex World (Course)]]> <![CDATA[Link to Course Flyer]]>
<![CDATA[FREE Lean Warehousing webinar, Monday September 14 1:30pm EDT!]]> 27233 Attend our free Lean Warehousing webinar for an introduction to using proven principles of lean management to cut waste and improve warehouse efficiencies. As a thank you for attending, you will receive a *discount code towards towards the October 20-22 course being held in Atlanta, GA (*please note that the discount cannot be combined with any other discount).

OVERVIEW

Our interactive webinar will provide an overview of our three-day Professional Education course titled “Lean Warehousing” as well as give attendees an opportunity to ask their LEAN and warehousing related questions. 

Our host will also discuss:

ABOUT THE PRESENTER

Brad Bossence is Vice President at LeanCor and a professional education instructor for the Georgia Tech Supply Chain & Logistics Institute. Brad is the lead instructor for the Lean Warehousing course as well as courses that are part of our Lean Supply Chain Professional certificate series.

]]> Andy Haleblian 1 1441108114 2015-09-01 11:48:34 1492118308 2017-04-13 21:18:28 0 0 event Attend our free Lean Warehousing webinar for an introduction to using proven principles of lean management to cut waste and improve warehouse efficiencies.

]]>
2015-09-14T14:30:00-04:00 2015-09-14T15:30:00-04:00 2015-09-14T15:30:00-04:00 2015-09-14 18:30:00 2015-09-14 19:30:00 2015-09-14 19:30:00 2015-09-14T14:30:00-04:00 2015-09-14T15:30:00-04:00 America/New_York America/New_York datetime 2015-09-14 02:30:00 2015-09-14 03:30:00 America/New_York America/New_York datetime <![CDATA[]]> webinar@scl.gatech.edu]]> 443111 443111 image <![CDATA[Lean Warehousing]]> image/jpeg 1449256205 2015-12-04 19:10:05 1475895182 2016-10-08 02:53:02 <![CDATA[Register Online to Attend]]> <![CDATA[Course webpage within the SCL website]]> <![CDATA[Link to Course Flyer]]>
<![CDATA[FREE Introduction to International Logistics and Compliance webinar, Wednesday September 16 1:30pm EDT!]]> 27233 Attend our free Introduction to International Logistics and Compliance webinar for an overview relating to the topics covered as part of our 3-day course offering. As a thank you for attending, you will receive a *discount code towards towards the October 21-23 course being held in Savannah, GA (*please note that the discount cannot be combined with any other discount).

OVERVIEW

Our interactive webinar will provide an overview of our three-day Professional Education course titled Introduction to International Logistics and Compliance as well as give attendees an opportunity to ask their international logistics-related questions. 

Our host will also discuss:

ABOUT THE PRESENTER

Dan Gardner is an SCL professional education instructor and President of Trade Facilitators, Inc., a supply chain consulting and training firm.

]]> Andy Haleblian 1 1441380300 2015-09-04 15:25:00 1492118303 2017-04-13 21:18:23 0 0 event Attend our free Introduction to International Logistics and Compliance webinar for an overview relating to the topics covered as part of our 3-day course offering.

]]>
2015-09-16T14:30:00-04:00 2015-09-16T15:30:00-04:00 2015-09-16T15:30:00-04:00 2015-09-16 18:30:00 2015-09-16 19:30:00 2015-09-16 19:30:00 2015-09-16T14:30:00-04:00 2015-09-16T15:30:00-04:00 America/New_York America/New_York datetime 2015-09-16 02:30:00 2015-09-16 03:30:00 America/New_York America/New_York datetime <![CDATA[]]> webinar@scl.gatech.edu]]> 445001 445001 image <![CDATA[Introduction to International Logistics and Compliance]]> image/jpeg 1449256205 2015-12-04 19:10:05 1475895184 2016-10-08 02:53:04 <![CDATA[Register Online to Attend]]> <![CDATA[Introduction to International Logistics and Compliance]]> <![CDATA[Link to Course Flyer]]>
<![CDATA[FREE Measuring and Managing Performance in Supply Chain and Logistics Operations webinar, Wednesday October 7 1:30pm EDT!]]> 27233 Attend our free Measuring and Managing Performance in Supply Chain and Logistics  webinar for an overview relating to the topics covered as part of our 3-day course offering. As a thank you for attending, you will receive a *discount code towards towards the November 4-6 course being held in Atlanta, GA (*please note that the discount cannot be combined with any other discount).

OVERVIEW

Are you having difficulties aligning your supply chain metrics with your company's financial goals? This difficulty grows exponentially with ever more complex global operations and the use of external business partners. During this free one-hour webinar, learn how to tailor your metrics and measurement processes to focus on the most important aspects of operations that impact overall corporate goals.

ABOUT THE PRESENTER

Paula Ferguson is the instructor for our "Measuring and Managing Performance in Supply Chain and Logistics Operations" course and a renowned expert in the field of supply chain performance management.

]]> Andy Haleblian 1 1441897123 2015-09-10 14:58:43 1492118301 2017-04-13 21:18:21 0 0 event Attend our free Measuring and Managing Performance in Supply Chain and Logistics Operations webinar for an overview relating to the topics covered as part of our 3-day course offering.

]]>
2015-10-07T14:30:00-04:00 2015-10-07T15:30:00-04:00 2015-10-07T15:30:00-04:00 2015-10-07 18:30:00 2015-10-07 19:30:00 2015-10-07 19:30:00 2015-10-07T14:30:00-04:00 2015-10-07T15:30:00-04:00 America/New_York America/New_York datetime 2015-10-07 02:30:00 2015-10-07 03:30:00 America/New_York America/New_York datetime <![CDATA[]]> webinar@scl.gatech.edu]]> 446551 446551 image <![CDATA[Measuring and Managing Performance in Supply Chain and Logistics Operations]]> image/jpeg 1449256217 2015-12-04 19:10:17 1475895187 2016-10-08 02:53:07 <![CDATA[Register Online to Attend]]> <![CDATA[Course webpage within the SCL website]]> <![CDATA[Link to Course Flyer]]>
<![CDATA[TAG Panel | Towards a Physical Internet: Meeting the Global Logistics Sustainability Challenge]]> 27233 Please join TAG Supply Chain & Logistics for a panel discussion with members of the Georgia Tech Stewart School of Industrial and Systems Engineering, Fortna, and Kurt Salmon.

Event Description

The global logistics system is not sustainable from an economic, environmental or social perspective. Re-thinking logistics is a grand challenge as big as the information technology grand challenge that resulted in the (digital) Internet. It is proposed that in order to meet the logistics sustainability grand challenge, our logistics systems will evolve towards a Physical Internet. The merits and challenges of implementing the vision will be presented from three viewpoints and illustrated through demonstrations in the Physical Internet Center.

Panelists: 

Event Details

Where: Georgia Tech Stewart School of Industrial & Systems Engineering/Supply Chain & Logistics Institute (ISyE, 755 Ferst Drive, N.W. Atlanta, GA 30332-0205 | 404-894-2343 Google Maps). Please arrive at least 30 minutes early to give yourself ample time to find parking and to walk to the venue. We suggest using Visitor Lot/Area 3: Ferst Drive & Regents or Visitor Lot/Area 2: Student Center (directions and parking map with areas). If prior visitor lots are full, Visitor Lot/Area 4: State Street & Ferst Drive and Visitor Lot/Area 8: Tech Square, Georgia Tech Hotel & Conference Center are options (ride the Tech Trolley to the Campus Recreation Center and exit, ISyE is across the street from CRC).

When: Wednesday, November 11, 2015, 11:30 a.m. – 1:30 p.m.

Register Today!

TAG members: $10; Non-members $30

For more information and to register online, please visit http://bit.ly/1Il3GDr.

 Sponsors

Gigabyte: Manhattan Associates

Megabyte: Software AG SuiteOutSystems, Blackstone+Cullen

Hosted by the Georgia Tech Supply Chain & Logistics Institute

]]> Andy Haleblian 1 1442317732 2015-09-15 11:48:52 1492118297 2017-04-13 21:18:17 0 0 event Please join TAG Supply Chain & Logistics for a panel discussion with members of the Georgia Tech Stewart School of Industrial and Systems Engineering, Fortna, and Kurt Salmon.

]]>
2015-11-11T11:30:00-05:00 2015-11-11T13:30:00-05:00 2015-11-11T13:30:00-05:00 2015-11-11 16:30:00 2015-11-11 18:30:00 2015-11-11 18:30:00 2015-11-11T11:30:00-05:00 2015-11-11T13:30:00-05:00 America/New_York America/New_York datetime 2015-11-11 11:30:00 2015-11-11 01:30:00 America/New_York America/New_York datetime <![CDATA[]]> Please email lindsay@tagonline.org with any questions.

]]>
448001 448001 image <![CDATA[Towards a Physical Internet]]> image/jpeg 1449256246 2015-12-04 19:10:46 1475895189 2016-10-08 02:53:09 <![CDATA[Register Online]]> <![CDATA[The Supply Chain and Logistics Institute at Georgia Tech]]> <![CDATA[TAG Supply Chain & Logistics Society]]> <![CDATA[Event Flyer, Map and Directions]]>
<![CDATA[ISyE & IIE Homecoming Kickoff Activities]]> 27233 The H. Milton Stewart School of Industrial Engineering (ISyE), Supply Chain & Logistics Institute (SCL), Georgia Tech IIE, IIE Atlanta Chapter, and GT Supply Chain Network (GTSCN) will be co-hosting an Open House and Homecoming Meet & Greet Tailgate.

OPEN HOUSE - Friday, October 23, 2015

To be held at SCL's suite in the ISyE complex (Groseclose Building, 2nd floor, Suite 228 http://www.scl.gatech.edu/about/visiting). All interested parties are invited to stop by, mix and mingle, hear the latest on what is going on in research, professional education, degree programs, forums, etc. in Supply Chain at Georgia Tech. A tour of the campus will be offered.

TAILGATE MEET & GREET - Saturday, October 24, 2015

We will host tents in the Tech Green Tailgating Area (http://tailgateguys.com/georgia-tech), between Skiles and the Student Center. We anticipate being set up at 4pm, kickoff is at 7pm. Snacks and drinks will be available.

If you are able to join us for either event, please RSVP online at http://bit.ly/isye-hc2015.

]]> Andy Haleblian 1 1442603720 2015-09-18 19:15:20 1492118293 2017-04-13 21:18:13 0 0 event The H. Milton Stewart School of Industrial Engineering (ISyE), Supply Chain & Logistics Institute (SCL), Georgia Tech IIE, IIE Atlanta Chapter, and GT Supply Chain Network (GTSCN) will be co-hosting an Open House and Homecoming Meet & Greet Tailgate

]]>
2015-10-23T11:00:00-04:00 2015-10-24T18:00:00-04:00 2015-10-24T18:00:00-04:00 2015-10-23 15:00:00 2015-10-24 22:00:00 2015-10-24 22:00:00 2015-10-23T11:00:00-04:00 2015-10-24T18:00:00-04:00 America/New_York America/New_York datetime 2015-10-23 11:00:00 2015-10-24 06:00:00 America/New_York America/New_York datetime <![CDATA[]]> event@scl.gatech.edu

]]>
449851 449851 image <![CDATA[ISyE & IIE Homecoming Kickoff Activities]]> image/jpeg 1449256264 2015-12-04 19:11:04 1475895192 2016-10-08 02:53:12 <![CDATA[More information and to RSVP online]]>
<![CDATA[DOS Seminar - Merve Bodur]]> 27187 TITLE:  Cutting Planes from Extended LP Formulations

ABSTRACT:

For mixed-integer sets, we study extended formulations of their LP relaxations and study the effect of adding cutting planes in the extended space. In terms of optimization, extended LP formulations do not lead to better bounds as their projection onto the original space is precisely the original LP relaxation. However, we show that applying split cuts to such extended formulations can be more effective than applying split cuts to the original formulation. For any 0-1 mixed-integer set with n integer and k continuous variables, we construct an extended formulation with 2n+k-1 variables whose elementary split closure is integral. We extend this idea to general mixed-integer sets and construct the best extended formulation with respect to split cuts. 
This is joint work with Sanjeeb Dash and Oktay Gunluk.]]> Anita Race 1 1442821166 2015-09-21 07:39:26 1492118293 2017-04-13 21:18:13 0 0 event 2015-09-23T14:00:00-04:00 2015-09-23T14:00:00-04:00 2015-09-23T14:00:00-04:00 2015-09-23 18:00:00 2015-09-23 18:00:00 2015-09-23 18:00:00 2015-09-23T14:00:00-04:00 2015-09-23T14:00:00-04:00 America/New_York America/New_York datetime 2015-09-23 02:00:00 2015-09-23 02:00:00 America/New_York America/New_York datetime <![CDATA[]]>
<![CDATA[ISyE Seminar Series - Jeff Wu]]> 27187 TITLE:  From real world problems to esoteric research:  examples and personal experience

ABSTRACT:

Young (and some not-so-young) researchers often wonder how to extract good research ideas and develop useful methodologies from solving real world problems. The path is rarely straightforward and its success depends on the circumstances, tenacity and luck. I will use three examples to illustrate how I trod the path. The first involved an attempt to find optimal growth conditions for nano structures. It led to the development of a new method “sequential minimum energy design (smed)”, which exploits an analogy to potential energy of charged particles. After a few years of frustrated efforts and relentless pursuit, we realized that smed is more suitable for generating samples adaptively to mimic an arbitrary distribution rather than for optimization. The main objective of the second example was to build an efficient statistical emulator based on finite element simulation results with two mesh densities in cast foundry operations. It eventually led to the development of a class of nonstationary Gaussian process models that can be used to connect simulation data of different precisions and speeds. The third example is about sequential design that works well for small samples in sensitivity testing. I will describe three major papers in a span of 30 years and how each paper had one new idea that inspired the next paper. In each example, the developed methodology has broader applications beyond the original problem. I will explain the thought process in each example.  Finally, I will share some secrets about a “path to innovation”.

]]> Anita Race 1 1442821353 2015-09-21 07:42:33 1492118293 2017-04-13 21:18:13 0 0 event 2015-09-23T16:00:00-04:00 2015-09-23T16:00:00-04:00 2015-09-23T16:00:00-04:00 2015-09-23 20:00:00 2015-09-23 20:00:00 2015-09-23 20:00:00 2015-09-23T16:00:00-04:00 2015-09-23T16:00:00-04:00 America/New_York America/New_York datetime 2015-09-23 04:00:00 2015-09-23 04:00:00 America/New_York America/New_York datetime <![CDATA[]]>
<![CDATA[PhD Thesis Defense - Ross Hilton]]> 27187 TITLE: Model-Based Data Mining Methods for Identifying Patterns in Medical and Health Data

ABSTRACT:

Every day humans and machines are responsible for the creation of massive amounts of data. Alongside the growth of these data banks, a new field of study, data science, has emerged. The central role of data science is to infer knowledge on the data in the form of models and estimates employing methods at the intersection of computer science, data mining, mathematics, and statistics. In this thesis we provide statistical and model-based data mining methods for pattern detection with applications to biomedical and healthcare data sets. In particular, we examine applications in costly acute or chronic disease management. Health data are extremely varied: at the macro-level, one can examine the healthcare utilization of millions of patients in the insurance systems like Medicare and Medicaid, while at the micro-level, a single snapshot from a medical imaging device may be used to diagnose cancerous cells in the body. In all, statisticians can contribute methods that extract structure from large, noisy data.

In Chapter II, we consider NMR experiments in which we seek to locate and de-mix smooth, yet highly localized components in a noisy two-dimensional signal. By using wavelet-based methods we are able to separate components from the noisy background, as well as from other neighboring components. In Chapter III, we pilot methods for identifying profiles of patient utilization of the healthcare system from large, highly-sensitive, patient-level data. We combine model-based data mining methods with clustering analysis in order to extract longitudinal utilization profiles. We transform these profiles into simple visual displays that can inform policy decisions and quantify the potential cost savings of interventions that improve adherence to recommended care guidelines. In Chapter IV, we propose new methods integrating survival analysis models and clustering analysis to profile patient-level utilization behaviors while controlling for variations in the population’s demographic and healthcare characteristics and explaining variations in utilization due to different state-based Medicaid programs, as well as access and urbanicity measures.

]]> Anita Race 1 1442821544 2015-09-21 07:45:44 1492118293 2017-04-13 21:18:13 0 0 event 2015-10-07T11:15:00-04:00 2015-10-07T11:15:00-04:00 2015-10-07T11:15:00-04:00 2015-10-07 15:15:00 2015-10-07 15:15:00 2015-10-07 15:15:00 2015-10-07T11:15:00-04:00 2015-10-07T11:15:00-04:00 America/New_York America/New_York datetime 2015-10-07 11:15:00 2015-10-07 11:15:00 America/New_York America/New_York datetime <![CDATA[]]>
<![CDATA[Statistics Seminar - Richard Peng]]> 27187 TITLE:  L_p Row Sampling by Lewis Weights

ABSTRACT:

We give an algorithm that efficiently samples the rows of a matrix while preserving the L_1-norm of its product with vectors. Given an n-by-d matrix A, we find with high probability and in input sparsity time A' consisting of about dlogd rescaled rows of A such that |Ax|_1
is close to |A’x|_1 for all vectors x. We also show similar results giving nearly optimal sample bounds for all L_p-norms.

Our results are based on sampling by ``Lewis weights'', which can be viewed as generalizations of statistical leverage scores to non-linear settings. We also give an elementary proof of an L_1 matrix concentration bound that governs the convergence of this sampling
process.
Joint work with Michael Cohen

]]> Anita Race 1 1443017926 2015-09-23 14:18:46 1492118290 2017-04-13 21:18:10 0 0 event 2015-09-29T12:00:00-04:00 2015-09-29T12:00:00-04:00 2015-09-29T12:00:00-04:00 2015-09-29 16:00:00 2015-09-29 16:00:00 2015-09-29 16:00:00 2015-09-29T12:00:00-04:00 2015-09-29T12:00:00-04:00 America/New_York America/New_York datetime 2015-09-29 12:00:00 2015-09-29 12:00:00 America/New_York America/New_York datetime <![CDATA[]]>
<![CDATA[ISyE Seminar Series - Henry Lam]]> 27187 TITLE:  Model Uncertainty and Robust Stochastic Modeling

ABSTRACT:

Virtually any performance analysis in stochastic modeling relies on input model assumptions that, to some extent, deviate from the truth. This talk will investigate a robust framework to quantify these model errors, by positing worst-case optimizations over the input probability distributions subject to the modeler's partial, nonparametric information. We illustrate these optimization formulations in several stylized contexts in stochastic modeling, describe their computational challenges, and present some simulation-based machinery in approximating their solutions. We also illustrate their statistical connections with conventional input modeling in the stochastic simulation literature.

]]> Anita Race 1 1443425844 2015-09-28 07:37:24 1492118288 2017-04-13 21:18:08 0 0 event 2015-09-30T16:00:00-04:00 2015-09-30T16:00:00-04:00 2015-09-30T16:00:00-04:00 2015-09-30 20:00:00 2015-09-30 20:00:00 2015-09-30 20:00:00 2015-09-30T16:00:00-04:00 2015-09-30T16:00:00-04:00 America/New_York America/New_York datetime 2015-09-30 04:00:00 2015-09-30 04:00:00 America/New_York America/New_York datetime <![CDATA[]]>
<![CDATA[DOS Seminar - Bistra Dilkina]]> 27187 TITLE:  Learning to Branch in Mixed Integer Programming

ABSTRACT:

Choosing good variables to branch on often leads to a dramatic reduction in terms of the number of nodes needed to solve an instance. The design of strategies for branching in Mixed Integer Programming (MIP) is guided by cycles of parameter tuning and offline experimentation on an extremely heterogeneous testbed, using the average performance. Once devised, these strategies (and the values of their parameters) are essentially input-agnostic. To address these issues, we develop a novel framework for data-driven, on-the-fly design of variable selection strategies. By leveraging recent advances in supervised ranking, we aim to produce strategies that gather the best of all properties: 1) using a small number of search nodes approaching the good performance of SB, 2) maintaining a low computation footprint as in PC, and 3) selecting variables adaptively based on the properties of the given instance. Based on a limited number of observed Strong Branching decisions at the start of the search and a set of dynamic features computed for each candidate variable at a search node, we learn an easy-to-evaluate surrogate function that mimics the SB strategy by solving a ``learning-to-rank'' problem, common in ML. We will show an instantiation of this framework using CPLEX, a state-of-the-art MIP solver, and evaluate performance the MIPLIB benchmark.

 

This is joint work with Elias Khalil, Le Song, Pierre Le Bodic and George Nemhauser.

]]> Anita Race 1 1443425983 2015-09-28 07:39:43 1492118288 2017-04-13 21:18:08 0 0 event 2015-09-30T14:00:00-04:00 2015-09-30T14:00:00-04:00 2015-09-30T14:00:00-04:00 2015-09-30 18:00:00 2015-09-30 18:00:00 2015-09-30 18:00:00 2015-09-30T14:00:00-04:00 2015-09-30T14:00:00-04:00 America/New_York America/New_York datetime 2015-09-30 02:00:00 2015-09-30 02:00:00 America/New_York America/New_York datetime <![CDATA[]]>
<![CDATA[DOS Seminar - Avinash Bhardwaj]]> 27187 TITLE:  Submodular Knapsacks - A Polyhedral Discussion

ABSTRACT:

Submodular and supermodular knapsack sets arise naturally when modeling utilities, risk and probabilistic constraints on discrete variables. Although submodular functions have been explored plentiful with respect to set function optimization, understanding of level sets of these set functions is still underdeveloped. The results with respect to submodular optimization cannot be translated as such when optimizing over the level sets of submodular functions. In particular, even though the convex hull of convex lower envelope of a general submodular function f can be completely characterized, the convex hull description of the lower level set of f is unknown even in the special case when f is monotone. In this talk, we will study the facial structure of the convex hull of the level sets of a given (not necessarily monotone) submodular set function f : {0, 1}^n

]]> Anita Race 1 1443541230 2015-09-29 15:40:30 1492118286 2017-04-13 21:18:06 0 0 event 2015-10-02T13:00:00-04:00 2015-10-02T13:00:00-04:00 2015-10-02T13:00:00-04:00 2015-10-02 17:00:00 2015-10-02 17:00:00 2015-10-02 17:00:00 2015-10-02T13:00:00-04:00 2015-10-02T13:00:00-04:00 America/New_York America/New_York datetime 2015-10-02 01:00:00 2015-10-02 01:00:00 America/New_York America/New_York datetime <![CDATA[]]>
<![CDATA[PhD Thesis Defense - Hyunwoo Park]]> 27187 TITLE:  Computational Analysis of Technological Innovation in Complex Enterprise Systems

ABSTRACT:

Technological innovation in complex enterprise systems requires coordinated interplay between a heterogeneous set of industrial players. The complexity in how firms form relationship with each other perplexes the decision-making processes for individual players when they explore the technological search space in order to achieve breakthrough innovation. Building upon the extant literature on business ecosystem, interfirm alliance, new product development, and technology management, this dissertation explores the interplay between technological innovation and interfirm relationship as well as the alliance formation patterns in the business ecosystem context. In Chapter 2, We begin with providing a macroscopic perspective on the information and communication technology ecosystem, followed by in-depth empirical investigations in the mobile handset industry. We employ network visualization, sequence clustering, and organizational simulation methodologies for the macroscopic analysis. Our microscopic analyses presented in Chapters 3 through 5 borrow methodologies from econometrics including regression analysis, difference-in-differences estimation, and event study. Our results propose an effective way to visualize the whole industrial ecosystem and show how the enterprise system has transformed over time. The results from the microscopic analysis show how interfirm relationship shapes technological innovation and how technological innovation is materialized in firm value. Lastly in Chapter 6, we present an integrated computation framework to infer alliance formation strategy. We contribute to the literature by providing generative methods and empirical evidences that accommodate a more complete view on the innovation process in the business ecosystem setting. The dissertation ends with suggesting directions for future research and highlighting implications for research and practice in the area of technological innovation in business ecosystem.

]]> Anita Race 1 1444031993 2015-10-05 07:59:53 1492118283 2017-04-13 21:18:03 0 0 event 2015-10-19T14:00:00-04:00 2015-10-19T14:00:00-04:00 2015-10-19T14:00:00-04:00 2015-10-19 18:00:00 2015-10-19 18:00:00 2015-10-19 18:00:00 2015-10-19T14:00:00-04:00 2015-10-19T14:00:00-04:00 America/New_York America/New_York datetime 2015-10-19 02:00:00 2015-10-19 02:00:00 America/New_York America/New_York datetime <![CDATA[]]>
<![CDATA[SCL Course: Building the Lean Supply Chain Professional]]> 27233 Course Description

One of the most misunderstood elements of Lean thinking is that Lean is primarily about effective supply chain management. Therefore, we need to be able to answer the question "How does Lean apply to the supply chain?" Connecting Lean to supply chain management is the core purpose of Module 2. Students will build upon their knowledge in module one and learn how to apply Lean principles and problem solving to supply chain functions.

The overarching theme for module 2 is "systems thinking," where we understand how "pull and one piece flow" will lead to reductions to "total cost" of the supply chain. Students will once again be challenged to questions mental models such as economies of scale and replace them with mental models such as economies of time. Having completed module 2, students will not only be Lean problem solvers, they will understand how to connect lean and waste reduction to supply chain functions.

Who Should Attend

Supply chain professionals, logistics professionals, material managers, production control managers, transportation managers, warehousing managers and purchasing managers

How You Will Benefit

What Is Covered

Day 1 - Setting the Corporate Stage
Day 2- Tactical Implementation
Day 3 - Learning from Case Studies
]]> Andy Haleblian 1 1444126454 2015-10-06 10:14:14 1492118282 2017-04-13 21:18:02 0 0 event One of the most misunderstood elements of lean thinking is that lean is primarily about effective supply chain management. Therefore, we need to be able to answer the question, "How does lean apply to the supply chain?" Connecting lean to supply chain management is the core purpose of this second course in a 3-course series on becoming a Lean Supply Chain professional. Participants will build upon their knowledge from the first course and learn how to apply lean principles and problem solving to supply chain functions.

]]>
2015-10-13T09:00:00-04:00 2015-10-15T18:00:00-04:00 2015-10-15T18:00:00-04:00 2015-10-13 13:00:00 2015-10-15 22:00:00 2015-10-15 22:00:00 2015-10-13T09:00:00-04:00 2015-10-15T18:00:00-04:00 America/New_York America/New_York datetime 2015-10-13 09:00:00 2015-10-15 06:00:00 America/New_York America/New_York datetime <![CDATA[]]> EMAIL: info@scl.gatech.edu or CALL: (404) 385-3501 between 9:00a.m. and 4:00p.m., Eastern time.

]]>
<![CDATA[2015-16 LEAN Supply Chain Brochure (PDF)]]> <![CDATA[Course webpage within the SCL website]]> <![CDATA[Register Online via the GT Professional Education website]]>
<![CDATA[PhD Thesis Defense - Yanling Chang]]> 27187 TITLE: A Leader-Follower Partially Observed Markov Game

ABSTRACT:

Models of sequential decision-making under uncertainty provide a rich normative framework for one or more intelligent decision-makers to improve, e.g., optimize, the operation of a system subject to control over a horizon containing a sequence of decision epochs. The solutions of such models can provide guidance as to how decision-makers should select actions, based on currently available data, in order to achieve their objectives. This dissertation models and analyzes a sequential stochastic game involving two intelligent and adaptive decision-makers.  Each of these decision-makers partially observes the other decision-maker's state at each decision epoch. 

 Chapter II presents a model of and analyzes a leader-follower, multi-objective partially observed infinite horizon Markov game, where it is assumed that the follower selects its policy with complete knowledge of the policy selected by the leader. We show how the results of this POMG can be used to support decision-making involving a leader having multiple objectives.

Chapter III considers the single objective version of the problem considered in Chapter II and investigates the impact of how accurately the leader observes the follower's state on the performance of the leader, thus representing an analysis of the value of information for this class of POMGs.

Chapter IV applies the results of the first two chapters in order to quantify the risk of a food production facility to an intelligent and adaptive adversary intent on delivering a chemical or biological toxin to the general population through use of the food supply chain. The goal of this chapter is to develop a new model of dynamic risk analysis that can explicitly describe the strategic interaction between two intelligent and adaptive agents with different objectives, and to provide decision support to the defender as to when and what action should be taken in order to achieve the defender's (possibly multiple) objectives.

]]> Anita Race 1 1444130148 2015-10-06 11:15:48 1492118282 2017-04-13 21:18:02 0 0 event 2015-10-22T10:30:00-04:00 2015-10-22T10:30:00-04:00 2015-10-22T10:30:00-04:00 2015-10-22 14:30:00 2015-10-22 14:30:00 2015-10-22 14:30:00 2015-10-22T10:30:00-04:00 2015-10-22T10:30:00-04:00 America/New_York America/New_York datetime 2015-10-22 10:30:00 2015-10-22 10:30:00 America/New_York America/New_York datetime <![CDATA[]]>
<![CDATA[DOS Seminar - Sheldon Jacobson]]> 27187 TITLE:  Balance Optimization Subset Selection (BOSS): An Alternative Approach for Casual Inference with Observational Data

ABSTRACT:

Researchers in medicine and the social sciences attempt to identify and document causal relationships. Those not fortunate enough to be able to design and implement randomized control trials must resort to observational studies.  To preserve the ability to make causal inferences outside the experimental realm, researchers attempt to post-process observational data to draw meaningful insights and conclusions.  Finding the subset of data that most closely resembles experimental data is a challenging, complex problem. However, the rise in computational power and discrete optimization algorithmic advances suggests an operations research solution as an alternative to methods currently being employed.

Joint work with Jason J. Sauppe (University of Wisconsin - Lacrosse)

Biography:

Sheldon H. Jacobson is a Professor and Director of the Simulation and Optimization Laboratory at the University of Illinois.  He has a broad set of basic and applied research interests, including problems related to optimal decision-making, national security, and public health. His research has been disseminated in numerous archival journals, including Operations Research, Mathematical Programming, and SIAM Journal on Control and Optimization, among others.  His research has been supported by grants from the National Science Foundation and the Air Force Office of Scientific Research.  He is a Fellow of INFORMS and IIE.

]]> Anita Race 1 1444130341 2015-10-06 11:19:01 1492118282 2017-04-13 21:18:02 0 0 event 2015-10-08T12:00:00-04:00 2015-10-08T12:00:00-04:00 2015-10-08T12:00:00-04:00 2015-10-08 16:00:00 2015-10-08 16:00:00 2015-10-08 16:00:00 2015-10-08T12:00:00-04:00 2015-10-08T12:00:00-04:00 America/New_York America/New_York datetime 2015-10-08 12:00:00 2015-10-08 12:00:00 America/New_York America/New_York datetime <![CDATA[]]>
<![CDATA[Statistics Seminar - Lifeng Lai]]> 27187 TITLE: Distributed Statistical Inference with Compressed Data

ABSTRACT:

In the classic statistical inference problems, all data is available at a centralized location. With the explosion of the size of data set, it is increasingly common that data is stored in multiple terminals connected by links with a limited communication capacity. In this scenario, terminals have to exchange compressed data and perform statistical inference using the compressed data. In this talk, we will discuss our recent work that addressed the following two questions: 1) Suppose we would like to achieve the same optimal inference as that of the centralized case, how much data compression can be performed?; and 2) Suppose we compress the data extremely (zero-rate compression), what is the optimal inference performance?

Bio: 

Lifeng Lai received the B.E. and M. E. degrees from Zhejiang University, Hangzhou, China in 2001 and 2004 respectively, and the PhD degree from The Ohio State University at Columbus, OH, in 2007. He was a postdoctoral research associate at Princeton University from 2007 to 2009. He is now an assistant professor at Worcester Polytechnic Institute. Dr. Lai’s research interests include information theory, stochastic signal processing and their applications in wireless communications, security and other related areas.

Dr. Lai is a co-recipient of the Best Paper Award from IEEE Global Communications Conference (Globecom) in 2008, the Best Paper Award from IEEE Conference on Communications (ICC) in 2011 and the Best Paper Award from IEEE Smart Grid Communications (SmartGridComm) in 2012. He received the National Science Foundation CAREER Award in 2011, and Northrop Young Researcher Award in 2012. He served as a Guest Editor for IEEE Journal on Selected Areas in Communications, Special Issue on Signal Processing Techniques for Wireless Physical Layer Security. He is currently serving as an Editor for IEEE Transactions on Wireless Communications, and an Associate Editor for IEEE Transactions on Information Forensics and Security.

 

]]> Anita Race 1 1444315253 2015-10-08 14:40:53 1492118279 2017-04-13 21:17:59 0 0 event 2015-10-15T12:00:00-04:00 2015-10-15T12:00:00-04:00 2015-10-15T12:00:00-04:00 2015-10-15 16:00:00 2015-10-15 16:00:00 2015-10-15 16:00:00 2015-10-15T12:00:00-04:00 2015-10-15T12:00:00-04:00 America/New_York America/New_York datetime 2015-10-15 12:00:00 2015-10-15 12:00:00 America/New_York America/New_York datetime <![CDATA[]]>
<![CDATA[DOS Seminar - Liron Yedidsion]]> 27187 TITLE: A Polynomial Time Approximation Scheme (PTSA) for the bi-scenario sum of completion times problem

ABSTRACT:

An influential aspect of any scheduling problem is the processing time of a tasks (job), which typically can be deterministic, stochastic or even uncertain. Scheduling according to unique and known processing times (a.k.a. Nominal) may be naïve, since real production systems are usually subject to inherent uncertainty. Moreover, typically, there are several objectives that the decision-maker seeks to satisfy. We offer a novel approach in the context of deterministic scheduling, borrowed from scenario-based optimization. The new approach copes with uncertainty by simultaneously optimizing a the sum of completion times criterion under two different instances of the processing times. We term the new problem a bi-scenario trade-off problem. We develop a PTAS that approximates the Pareto-optimal set of solutions and show that it is tight. Then we introduce a Branch-and-Bound based algorithm that solves the problem optimally.

]]> Anita Race 1 1444644542 2015-10-12 10:09:02 1492118278 2017-04-13 21:17:58 0 0 event 2015-10-15T12:00:00-04:00 2015-10-15T12:00:00-04:00 2015-10-15T12:00:00-04:00 2015-10-15 16:00:00 2015-10-15 16:00:00 2015-10-15 16:00:00 2015-10-15T12:00:00-04:00 2015-10-15T12:00:00-04:00 America/New_York America/New_York datetime 2015-10-15 12:00:00 2015-10-15 12:00:00 America/New_York America/New_York datetime <![CDATA[]]>
<![CDATA[DOS Seminar - Juan Pablo Vielma]]> 27187 TITLE:  Embedding Formulations and Complexity for Unions of Polyhedra

ABSTRACT:

We consider a systematic procedure to construct small but strong Mixed Integer Programming (MIP) formulations for unions of polyhedra. A key of the procedure is the flexible use of 0-1 variables that signal the selection among the polyhedra. This flexibility is achieved by considering several possible embeddings of the polyhedra in a higher dimensional space. An important characteristic of these formulations is that they do not use auxiliary variables other than the 0-1 variables strictly needed to construct a formulation (in contrast to "extended" formulations, which are allowed to use any number of auxiliary variables). Nonetheless, the formulations obtained through this embedding procedure can be smaller that the smallest known alternative formulation (extended or not). Furthermore, the Linear Programming (LP) relaxation of these formulations often have extreme points that naturally satisfy the appropriate integrality constraints (such formulations are usually denoted "ideal"). These characteristics allow some versions of these formulations to provide a significant computational advantage over alternative formulations. 

The embedding formulation approach leads to two notions of complexity for unions of polyhedra: 1) the size of the smallest non-extended formulation, and 2) the size of the smallest ideal non-extended formulations. We give upper and lower bounds for these complexity measures for several classes of polyhedra. We also study how the embedding selection affects the sizes of the associated formulations. Finally, we compare these measures to other complexity notions such as the size of the convex hull of the union of the polyhedra and the extension complexity of this convex hull.

]]> Anita Race 1 1444644764 2015-10-12 10:12:44 1492118278 2017-04-13 21:17:58 0 0 event 2015-10-16T13:00:00-04:00 2015-10-16T13:00:00-04:00 2015-10-16T13:00:00-04:00 2015-10-16 17:00:00 2015-10-16 17:00:00 2015-10-16 17:00:00 2015-10-16T13:00:00-04:00 2015-10-16T13:00:00-04:00 America/New_York America/New_York datetime 2015-10-16 01:00:00 2015-10-16 01:00:00 America/New_York America/New_York datetime <![CDATA[]]>
<![CDATA[ISyE/SCL November 2015 Supply Chain Day]]> 27233 ISyE students, please join us for our second fall Supply Chain Day! The 3-hour session will host supply chain representatives from AmazonAmericoldCiscoDeloitteThe Home Depot and Sears Holdings Corporation who will be on campus to educate ISyE students about their organizations and available employment opportunities. Plus, enjoy a free pizza lunch!

EVENT DETAILS

Where: ISyE Main Bldg, 2nd Floor Atrium

When: Wednesday, Nov 4, 10AM-1PM

What: The session will include:

Please plan on staying for the duration of the event and bring copies of your resume and business cards. Dress is business casual.

REGISTER ONLINE by October 30! Seating is limited!

EVENT SPONSOR

The event is sponsored through the generosity and support of APICS - Atlanta Chapter. APICS is a non-profit educational organization addressing operations management and supply chain management issues, and providing professional development opportunities to our members. Visit http://www.apicsatlanta.org/ to learn more and make sure to stop by the APICS table at the event.

]]> Andy Haleblian 1 1444651046 2015-10-12 11:57:26 1492118278 2017-04-13 21:17:58 0 0 event ISyE students, please join us for our second fall Supply Chain Day! The 3-hour session will host supply chain representatives from Amazon, Americold, Cisco, Deloitte, The Home Depot and Sears Holdings Corporation who will be on campus to educate ISyE students about their organizations and available employment opportunities. Plus, enjoy a free pizza lunch!

]]>
2015-11-04T10:00:00-05:00 2015-11-04T13:00:00-05:00 2015-11-04T13:00:00-05:00 2015-11-04 15:00:00 2015-11-04 18:00:00 2015-11-04 18:00:00 2015-11-04T10:00:00-05:00 2015-11-04T13:00:00-05:00 America/New_York America/New_York datetime 2015-11-04 10:00:00 2015-11-04 01:00:00 America/New_York America/New_York datetime <![CDATA[]]> event@scl.gatech.edu

]]>
458151 458151 image <![CDATA[Supply Chain Day - November 4, 2015]]> image/png 1449256347 2015-12-04 19:12:27 1475895202 2016-10-08 02:53:22 <![CDATA[Register online to attend (for ISyE students)]]>
<![CDATA[PhD Thesis Defense - Niao He]]> 27187 TITLE: Saddle Point Techniques in Convex Composite and Error-in-Measurement Optimization

ABSTRACT:

This dissertation aims to develop efficient algorithms with improved scalability and stability properties for large-scale optimization and optimization under uncertainty, and to bridge some of the gaps between modern optimization theories and recent applications emerging in the Big Data environment. To this end, the dissertation is dedicated to two important subjects -- i) Large-scale Convex Composite Optimization and ii) Error-in-Measurement Optimization. In spite of the different natures of these two topics, the common denominator, to be presented, lies in their accommodation for systematic use of saddle point techniques for mathematical modeling and numerical processing. The main body can be split into three parts.

 In the first part, we consider a broad class of variational inequalities with composite structures, allowing to cover the saddle point/variational analogies of the classical convex composite minimization (i.e. summation of a smooth convex function and a simple nonsmooth convex function). We develop novel composite versions of the state-of-the-art Mirror Descent and Mirror Prox algorithms aimed at solving such type of problems. We demonstrate that the algorithms inherit the favorable efficiency estimate of their prototypes when solving structured variational inequalities. Moreover, we develop several variants of the composite Mirror Prox algorithm along with their corresponding complexity bounds, allowing the algorithm to handle the case of imprecise prox mapping as well as the case when the operator is represented by an unbiased stochastic oracle.

 In the second part, we investigate four general types of large-scale convex composite optimization problems, including (a) multi-term composite minimization, (b) linearly constrained composite minimization, (c) norm-regularized nonsmooth minimization, and (d) maximum likelihood Poisson imaging. We demonstrate that the composite Mirror Prox, when integrated with saddle point techniques and other algorithmic tools, can solve all these optimization problems with the best known so far rates of convergences. Our main related contributions are as follows. Firstly, regards to problems of type (a), we develop an optimal algorithm by integrating the composite Mirror Prox with a saddle point reformulation based on exact penalty. Secondly, regards to problems of type (b), we develop a novel algorithm reducing the problem to solving a ``small series'' of saddle point subproblems and achieving an optimal, up to log factors, complexity bound. Thirdly, regards to problems of type (c), we develop a Semi-Proximal Mirror-Prox algorithm by leveraging the saddle point representation and linear minimization over problems' domain and attain optimality both in the numbers of calls to the first order oracle representing the objective and calls to the linear minimization oracle representing problem's domain. Lastly, regards to problem (d), we show that the composite Mirror Prox when applied to the saddle point reformulation circumvents the difficulty with non-Lipschitz continuity of the objective and exhibits better convergence rate than the typical rate for nonsmooth optimization. We conduct extensive numerical experiments and illustrate the practical potential of our algorithms in a wide spectrum of applications in machine learning and image processing.

 In the third part, we examine error-in-measurement optimization, referring to decision-making problems with data subject to measurement errors; such problems arise naturally in a number of important applications, such as privacy learning, signal processing, and portfolio selection. Due to the postulated observation scheme and specific structure of the problem, straightforward application of standard stochastic optimization techniques such as Stochastic Approximation (SA) and Sample Average Approximation (SAA) are out of question. Our goal is to develop computationally efficient and, hopefully, not too conservative data-driven techniques applicable to a broad scope of problems and allowing for theoretical performance guarantees. We present two such approaches -- one depending on a fully algorithmic calculus of saddle point representations of convex-concave functions and the other depending on a general approximation scheme of convex stochastic programming. Both approaches allow us to convert the problem of interests to a form amenable for SA or SAA. The latter developments are primarily focused on two important applications -- affine signal processing and indirect support vector machines.

]]> Anita Race 1 1444986649 2015-10-16 09:10:49 1492118275 2017-04-13 21:17:55 0 0 event 2015-10-27T10:30:00-04:00 2015-10-27T10:30:00-04:00 2015-10-27T10:30:00-04:00 2015-10-27 14:30:00 2015-10-27 14:30:00 2015-10-27 14:30:00 2015-10-27T10:30:00-04:00 2015-10-27T10:30:00-04:00 America/New_York America/New_York datetime 2015-10-27 10:30:00 2015-10-27 10:30:00 America/New_York America/New_York datetime <![CDATA[]]>
<![CDATA[Statistics Seminar - Youngdeok Hwang]]> 27187 TITLE:  Large Scale Multi-level Model: Solar Energy Prediction Combining Information from Multiple Sources

ABSTRACT:

We propose a hierarchical modeling approach appropriate for a data where the measurements and associated information are taken repeatedly over a large monitoring network. The proposed method is to make improved inferences by dividing a large scale data into manageable sizes and combining them. Our approach also provides a natural and flexible framework for situations where the data are available in different resolution. The proposed method is applied to the solar energy prediction problem for U.S. Department of Energy's SunShot initiative.

]]> Anita Race 1 1445247508 2015-10-19 09:38:28 1492118275 2017-04-13 21:17:55 0 0 event 2015-10-22T12:00:00-04:00 2015-10-22T12:00:00-04:00 2015-10-22T12:00:00-04:00 2015-10-22 16:00:00 2015-10-22 16:00:00 2015-10-22 16:00:00 2015-10-22T12:00:00-04:00 2015-10-22T12:00:00-04:00 America/New_York America/New_York datetime 2015-10-22 12:00:00 2015-10-22 12:00:00 America/New_York America/New_York datetime <![CDATA[]]>
<![CDATA[2015 Health & Humanitarian Logistics Conference]]> 27233 Now in its seventh year, the conference will be hosted in Africa for the first time by the University of Pretoria Gordon Institute of Business Science and co-organized by Georgia Tech, INSEAD, MIT and Northeastern University. The mission of the conference is to stimulate innovation and build capacity to manage global health and humanitarian supply chains around the world. It brings together high level speakers from across the health and humanitarian sectors, including non-governmental organizations (NGOs), industry, government, etc. Discussions will focus on the role of logistics in areas such as disaster response, health systems and food security as well as highlight the unique logistical challenges for humanitarian response and long-term development in Africa. You are welcome to email your ideas for panel themes, workshop topics, and potential speakers/facilitators to humlogconf@gatech.eduFor a listing of panel topics and associated workshops, view the conference agenda.

The conference is intended to be highly interactive, where participants will have ample time to discuss different view points rather than simply listening to speakers. In many conferences, even if there is time for participants to ask questions, only a handful speak up. Therefore we have adapted the conference agenda to allow for various kinds of interaction. For instance, panel speakers will give brief presentations, allowing plenty of time for participants to discuss the topics amongst those at their tables. A representative from each table will then direct a summary statement back to the panel with one or two resulting questions or challenges.

There will be a number of break-out workshops of small groups of participants led by experts in a given topic. These sessions will focus on discussion and exchange, allowing for different participants to share their views and make connections with each other. Following the workshops, a final plenary discussion will summarize the conclusions from the various workshops and the panel will discuss outcomes and potential action steps to move the agenda forward following the conference.


]]> Andy Haleblian 1 1445949847 2015-10-27 12:44:07 1492118269 2017-04-13 21:17:49 0 0 event Join us November 18-20, 2015 in South Africa for the 7th annual conference on Health and Humanitarian Logistics. The conference will be hosted in Africa for the first time by the University of Pretoria Gordon Institute of Business Science and co-organized by Georgia Tech, INSEAD, MIT and Northeastern University.

]]>
2015-11-18T13:00:00-05:00 2015-11-20T23:00:00-05:00 2015-11-20T23:00:00-05:00 2015-11-18 18:00:00 2015-11-21 04:00:00 2015-11-21 04:00:00 2015-11-18T13:00:00-05:00 2015-11-20T23:00:00-05:00 America/New_York America/New_York datetime 2015-11-18 01:00:00 2015-11-20 11:00:00 America/New_York America/New_York datetime <![CDATA[]]> https://www.scl.gatech.edu/humlog2015/contact/

]]>
462961 462961 image <![CDATA[2015 Health & Humanitarian Logistics Conference]]> image/jpeg 1449256385 2015-12-04 19:13:05 1475895209 2016-10-08 02:53:29 <![CDATA[Visit the Conference website]]>
<![CDATA[Statistics Seminar - Robert Gramacy]]> 27187 TITLE: Local Gaussian process approximation for large computer experiments

ABSTRACT:

We provide a new approach to approximate emulation of large computer experiments. By focusing expressly on desirable properties of the predictive equations, we derive a family of local sequential design schemes that dynamically define the support of a Gaussian process predictor based on a local subset of the data. We further derive expressions for fast sequential updating of all needed quantities as the local designs are built-up iteratively. Then we show how independent application of our local design strategy across the elements of a vast predictive grid facilitates a trivially parallel implementation. The end result is a global predictor able to take advantage of modern multicore architectures, GPUs, and cluster computing, while at the same time allowing for a non stationary modeling feature as a bonus. We demonstrate our method on examples utilizing designs sized in the tens of thousands to over a million data points.  Comparisons are made to the method of compactly supported covariances, and we present applications to computer model calibration of a radiative shock and the calculation of satellite drag.

]]> Anita Race 1 1446191911 2015-10-30 07:58:31 1492118267 2017-04-13 21:17:47 0 0 event 2015-11-05T11:00:00-05:00 2015-11-05T11:00:00-05:00 2015-11-05T11:00:00-05:00 2015-11-05 16:00:00 2015-11-05 16:00:00 2015-11-05 16:00:00 2015-11-05T11:00:00-05:00 2015-11-05T11:00:00-05:00 America/New_York America/New_York datetime 2015-11-05 11:00:00 2015-11-05 11:00:00 America/New_York America/New_York datetime <![CDATA[]]>
<![CDATA[Health Services Seminar - Eric Finkelstein]]> 27187 TITLE:  Use of Stated Preference Surveys to Quantify Preferences and Willingness to Pay for End of Life Care

ABSTRACT:

The goal of this presentation is to provide an understanding of decision-making around high cost medical care at the end of life. We use  a series of discrete choice (conjoint) analyses to quantify preferences and willingness to pay (WTP) for various EOL treatment options of community dwelling older adults (CDOAs), advanced cancer patients, caregivers, and physicians. We show that community dwelling older adults systematically underestimate their willingness to pay for high cost medical care, and that caregivers often want to provide greater levels of care than patients prefer for themselves. There is significant heterogeneity in treatment recommendations for similar patients by physicians. These results have implications for how best to finance and deliver medical care to patients with life limiting illnesses.

Bio:

Dr Eric A. Finkelstein, Ph.D., M.H.A. is the Director of Lien Centre for Palliative Care and Professor of the Signature Research Program in Health Services and Systems Research at the Duke-National University of Singapore Graduate Medical School and Research Professor at Duke University Global Health Institute. He received his BA in Mathematics/Economics from the University of Michigan, and a Ph.D. in economics and Masters in Health Administration from the University of Washington. Over the past ten years Professor Finkelstein has established himself as a leading international health economist doing research in the economics of health behaviours. His research focuses on economic incentives, behavioural economics, the economics of obesity, discrete choice analysis, economic evaluation, burden of illness analysis and cost effectiveness analyses. He has published over 70 peer-reviewed manuscripts, 2 books, and several book chapters in these areas. He also has experience as a Principal or Co-Investigator on research projects funded by the U.S. National Institutes of Health, the U.S. Centers for Disease Control and Prevention, and the Robert Wood Johnson Foundation. His research has been showcased in the Economist, Wall Street Journal, New York Times and other television, print, and media outlets throughout the world.

]]> Anita Race 1 1447324397 2015-11-12 10:33:17 1492118259 2017-04-13 21:17:39 0 0 event 2015-11-17T16:00:00-05:00 2015-11-17T16:00:00-05:00 2015-11-17T16:00:00-05:00 2015-11-17 21:00:00 2015-11-17 21:00:00 2015-11-17 21:00:00 2015-11-17T16:00:00-05:00 2015-11-17T16:00:00-05:00 America/New_York America/New_York datetime 2015-11-17 04:00:00 2015-11-17 04:00:00 America/New_York America/New_York datetime <![CDATA[]]>
<![CDATA[Statistics Seminar - Gary Lorden]]> 27187 TITLE: Confidence Intervals for a Proportion - the Good, the Bad, and the Ugly

ABSTRACT:

Introductory statistics books teach how to calculate confidence intervals for a proportion as an easy application of formulas for the standard error in sampling from binomial or hypergeometric distributions.   Assurances that “the approximation is good if the sample size n is at least 30” are typical.  But even for much larger n the performance of the usual confidence intervals can best be described as “ugly”.   Using better approximate methods only improves the results to “bad”.  Analyzing how to get “good” performance turns out to be interesting and forces us to address questions about what we want confidence intervals to do.

 

]]> Anita Race 1 1447415759 2015-11-13 11:55:59 1492118259 2017-04-13 21:17:39 0 0 event 2015-11-19T16:00:00-05:00 2015-11-19T16:00:00-05:00 2015-11-19T16:00:00-05:00 2015-11-19 21:00:00 2015-11-19 21:00:00 2015-11-19 21:00:00 2015-11-19T16:00:00-05:00 2015-11-19T16:00:00-05:00 America/New_York America/New_York datetime 2015-11-19 04:00:00 2015-11-19 04:00:00 America/New_York America/New_York datetime <![CDATA[]]>
<![CDATA[PhD Thesis Defense - Timothy Sprock]]> 27187 TITLE:  A Metamodel of Operational Control for Discrete Event Logistics Systems

ABSTRACT:

Discrete event logistics systems (DELS) are a class of dynamic systems that are defined by the transformation of discrete flows through a network of interconnected subsystems. The DELS domain includes systems such as supply chains, manufacturing systems, transportation networks, warehouses, and health care delivery systems.

In the future, smart operational control mechanisms must not only integrate real-time data from system operations, but also formulate and solve a wide variety of optimal-control analyses quickly and efficiently and then translate the results into executable commands.

However in contemporary DELS practice, these optimal-control analyses, and analyses in general, are often purpose-built to answer specific questions, with an implicit system model and many possible analysis implementations depending on the question, the instance data, and the solver. Automated and cost-effective access to multiple analyses from a single conceptual model of the target system would broaden the usage and implementation of analysis-based decision support and system optimization.

The fundamental contribution of this dissertation is concerned with interoperability and bridging the gap between operations research analysis models and practical applications of the results. This dissertation closes this gap by constructing a standard domain-specific language, standard problem definitions, and a standard analysis methodology to answer the control questions and execute the prescribed control actions.

The domain specific language meets a broader requirement for facilitating interoperability for DELS, including system integration, plug-and-play analysis methods and tools, and system design methodologies. The domain-specific language formalizes a recurring product, process, resource, and facility description of the DELS domain. It provides a common language to discuss our systems, including the questions that we want to ask about our systems, the problems that we need to solve in order to answer those questions, and the mechanisms to deploy the solution.

A canonical set of control questions defines the comprehensive functional specification of all the decision-making mechanisms that a controller needs to be able to provide; i.e. a model of analysis models or a metamodel of operational control. These questions refine the interoperability mechanism between system and analysis models by mapping classes of control analysis models to implementation and execution mechanisms in the system model.

A standard representation of each class of control problems is only a partial solution to fully addressing operational control. The final contribution of this dissertation constructs a round-trip analysis methodology that completes the bridge between operations research analysis models and deployable control mechanisms. This contribution formalizes an analysis pathway, from formulating an analysis model to executing a control action, that is grounded in a more fundamental insight into how analysis methods are executed to support operational control decision-making.

]]> Anita Race 1 1447843889 2015-11-18 10:51:29 1492118257 2017-04-13 21:17:37 0 0 event 2015-12-01T16:00:00-05:00 2015-12-01T16:00:00-05:00 2015-12-01T16:00:00-05:00 2015-12-01 21:00:00 2015-12-01 21:00:00 2015-12-01 21:00:00 2015-12-01T16:00:00-05:00 2015-12-01T16:00:00-05:00 America/New_York America/New_York datetime 2015-12-01 04:00:00 2015-12-01 04:00:00 America/New_York America/New_York datetime <![CDATA[]]>
<![CDATA[School Seminar - Peng Shi]]> 27187 TITLE:  Prediction and Optimzation in School Choice

ABSTRACT:

In public school choice, students are not assigned to a designated school based on home location, but submit preference rankings for a given set of schools to the school board, which takes into account everyone’s choices to compute the assignment. Such systems exist in Boston, Chicago, Denver, Miami, Minneapolis, New York City, New Orleans, and San Francisco. An important policy lever is what choice options to offer to each neighborhood, and how to prioritize between students. A key tradeoff is between giving students equitable chances to go to the schools they want and controlling the city’s school busing costs.

We study the optimization problem of choosing the choice menus and priorities for each neighborhood in order to maximize the sum of utilitarian and max-min welfare, subject to capacity and transportation constraints. The optimization is built on-top of a predictive model of how students will choose given new choice menus, which we validate using both out-of-sample testing and a field experiment. We show that under a fluid approximation, the optimization reduces to an assortment planning problem in which the objective is social-welfare rather than revenue. We show how to efficiently solve this sub-problem under MNL, Nested-Logit and Markov Chain choice models, and use this to produce better menus and priorities for Boston, which we evaluate by discrete simulations while taking into account possible errors in parameters.

]]> Anita Race 1 1448873824 2015-11-30 08:57:04 1492118252 2017-04-13 21:17:32 0 0 event 2015-12-01T16:00:00-05:00 2015-12-01T16:00:00-05:00 2015-12-01T16:00:00-05:00 2015-12-01 21:00:00 2015-12-01 21:00:00 2015-12-01 21:00:00 2015-12-01T16:00:00-05:00 2015-12-01T16:00:00-05:00 America/New_York America/New_York datetime 2015-12-01 04:00:00 2015-12-01 04:00:00 America/New_York America/New_York datetime <![CDATA[]]>
<![CDATA[IE Seminar - John Carlsson]]> 27187 TITLE:  Some new results in the continuous approximation theory of transportation

ABSTRACT:

One of the fundamental concerns in the analysis of logistical systems is the trade-off between localized, independent provision of goods and services versus provision along a centralized infrastructure such as a backbone network.  In this talk, we use the "continuous approximation paradigm" to study two instances of this trade-off:  in the first instance, we calculate the improvements in efficiency that arise when one uses unmanned aerial vehicles, or "drones", to deliver packages to customers, and in the second instance, we determine the social benefit that arises when cities introduce delivery services for groceries.

 

Bio:  John Gunnar Carlsson is an assistant professor in the Department of Industrial and Systems Engineering at the University of Southern California.  He received a Ph.D. in computational mathematics from ICME in Stanford University in 2009 and an A.B. in mathematics and music from Harvard College in 2005.  He is the recipient of the 2013 INFORMS Computing Society (ICS) Prize, the 2014 Air Force Young Investigator Prize, the 2012 DARPA Young Faculty Award, and the 2010 INFORMS Interactive Session Prize.  His research is supported by DARPA, the Office of Naval Research, the Air Force Office of Scientific Research, the National Science Foundation, the US Department of Transportation (MnDOT), and the Boeing Company, and has appeared in Operations Research,  Scientific Reports, Transportation Science, the INFORMS Journal on Computing, and the ACM Transactions on Algorithms.  He is an associate editor of Operations Research and Management Science.

]]> Anita Race 1 1448961535 2015-12-01 09:18:55 1492118252 2017-04-13 21:17:32 0 0 event 2015-12-02T20:00:00-05:00 2015-12-02T20:00:00-05:00 2015-12-02T20:00:00-05:00 2015-12-03 01:00:00 2015-12-03 01:00:00 2015-12-03 01:00:00 2015-12-02T20:00:00-05:00 2015-12-02T20:00:00-05:00 America/New_York America/New_York datetime 2015-12-02 08:00:00 2015-12-02 08:00:00 America/New_York America/New_York datetime <![CDATA[]]>
<![CDATA[IE Seminar - He Wang]]> 27187 TITLE:  Dynamic Learning and Optimization for Online Revenue Management

ABSTRACT:

In a dynamic pricing problem where the demand function is unknown a priori, price experimentation can be used for demand learning. In practice, however, sellers are faced with business constraints when learning demand, such as the inability to conduct extensive experimentation, short sales window and limited inventory. In this talk I will discuss models and algorithms that combine price optimization with demand learning, and report implementation results at a large e-commerce marketplace for daily deals.

 

Biography:

He Wang is a PhD candidate in the Operations Research Center at MIT, advised by David Simchi-Levi.  He received master's degree in Transportation from MIT in 2013, and dual bachelor's degree in Industrial Engineering and Mathematics from Tsinghua University in 2011. He is a recipient of Edward Linde (1962) MIT Presidential Graduate Fellowship, a Finalist in IBM Service Science Best Student Paper Award, and second place in CSAMSE Best Student Paper Award.  His current research focuses on data-driven methods in revenue management and supply chains.

]]> Anita Race 1 1449047103 2015-12-02 09:05:03 1492118251 2017-04-13 21:17:31 0 0 event 2015-12-03T16:00:00-05:00 2015-12-03T16:00:00-05:00 2015-12-03T16:00:00-05:00 2015-12-03 21:00:00 2015-12-03 21:00:00 2015-12-03 21:00:00 2015-12-03T16:00:00-05:00 2015-12-03T16:00:00-05:00 America/New_York America/New_York datetime 2015-12-03 04:00:00 2015-12-03 04:00:00 America/New_York America/New_York datetime <![CDATA[]]>
<![CDATA[School Seminar - Xuan Wang]]> 27187 TITLE:  Performance guarantees of the long chain design in resource allocation

ABSTRACT:

We consider a class of resource allocation problems in which there are n capacitated resources and n demand types. The resources are flexible, where resource j can be used to fulfill both demand type j and j+1. This is known as the long chain design proposed by Jordan and Graves (1995), which has been an important concept in the design of sparse flexible processes. In this talk, we discuss the theoretical performance of the long chain in two different settings.

In the first setting, the resource allocation decisions are made after all the demand has realized. We obtain a distribution-free bound on the ratio of the expected unit sales of the long chain relative to that of full flexibility. In a special case with i.i.d. demand and uniform capacity, we are able to derive the bound in closed form. Our bound depends only on the mean and standard deviation of the random demand, but compares very well with the ratio that uses complete information of the demand distribution.

In the second setting, the demand arrives sequentially and reveals its type upon arrival, and the allocation decisions must be made in real time. We show that the long chain is still very effective even under simple myopic online allocation policies. In particular, we show that the expected total number of lost sales only depends on the number of resources n, and is independent of how large the market size is.

Bio

Xuan Wang is a fifth year doctoral candidate in the Operations Management group at Stern School of Business, New York University. Xuan’s primary research interest lies in the field of supply chain management, optimization and business analytics. Prior to joining Stern, Xuan received her Bachelor's degree in industrial engineering and operations research from Tsinghua University in 2011. During her junior year, Xuan also spent one semester in the H. Milton Stewart School of Industrial & Systems Engineering at Georgia Institute of Technology as an exchange student.

]]> Anita Race 1 1449475033 2015-12-07 07:57:13 1492118250 2017-04-13 21:17:30 0 0 event 2015-12-08T16:00:00-05:00 2015-12-08T16:00:00-05:00 2015-12-08T16:00:00-05:00 2015-12-08 21:00:00 2015-12-08 21:00:00 2015-12-08 21:00:00 2015-12-08T16:00:00-05:00 2015-12-08T16:00:00-05:00 America/New_York America/New_York datetime 2015-12-08 04:00:00 2015-12-08 04:00:00 America/New_York America/New_York datetime <![CDATA[]]>
<![CDATA[School Seminar - Xuan Wang]]> 27187 TITLE:  Performance guarantees of the long chain design in resource allocation

ABSTRACT:

We consider a class of resource allocation problems in which there are n capacitated resources and n demand types. The resources are flexible, where resource j can be used to fulfill both demand type j and j+1. This is known as the long chain design proposed by Jordan and Graves (1995), which has been an important concept in the design of sparse flexible processes. In this talk, we discuss the theoretical performance of the long chain in two different settings.

In the first setting, the resource allocation decisions are made after all the demand has realized. We obtain a distribution-free bound on the ratio of the expected unit sales of the long chain relative to that of full flexibility. In a special case with i.i.d. demand and uniform capacity, we are able to derive the bound in closed form. Our bound depends only on the mean and standard deviation of the random demand, but compares very well with the ratio that uses complete information of the demand distribution.

In the second setting, the demand arrives sequentially and reveals its type upon arrival, and the allocation decisions must be made in real time. We show that the long chain is still very effective even under simple myopic online allocation policies. In particular, we show that the expected total number of lost sales only depends on the number of resources n, and is independent of how large the market size is.

Bio

Xuan Wang is a fifth year doctoral candidate in the Operations Management group at Stern School of Business, New York University. Xuan’s primary research interest lies in the field of supply chain management, optimization and business analytics. Prior to joining Stern, Xuan received her Bachelor's degree in industrial engineering and operations research from Tsinghua University in 2011. During her junior year, Xuan also spent one semester in the H. Milton Stewart School of Industrial & Systems Engineering at Georgia Institute of Technology as an exchange student.

]]> Anita Race 1 1449475040 2015-12-07 07:57:20 1492118250 2017-04-13 21:17:30 0 0 event 2015-12-08T16:00:00-05:00 2015-12-08T16:00:00-05:00 2015-12-08T16:00:00-05:00 2015-12-08 21:00:00 2015-12-08 21:00:00 2015-12-08 21:00:00 2015-12-08T16:00:00-05:00 2015-12-08T16:00:00-05:00 America/New_York America/New_York datetime 2015-12-08 04:00:00 2015-12-08 04:00:00 America/New_York America/New_York datetime <![CDATA[]]>
<![CDATA[School Seminar - Kimon Drakopoulos]]> 27187 TITLE:  Control of contagion processes on networks

ABSTRACT:

We consider the propagation of a contagion process (“epidemic”) on a network and study the problem of dynamically allocating a fixed curing budget to the nodes of the graph, at each time instant. We provide a dynamic policy for the rapid containment of a contagion process modeled as an SIS epidemic on a bounded degree undirected graph with n nodes. We show that if the budget r of curing resources available at each time is Ω(W), where W is the CutWidth of the graph, and also of order Ω(logn), then the expected time until the extinction of the epidemic is of order O(n/r), which is within a constant factor from optimal, as well as sublinear in the number of nodes. Furthermore, if the CutWidth increases only sublinearly with n, a sublinear expected time to extinction is possible with a sublinearly increasing budget r.

In contrast, for bounded degree graphs, we provide a lower bound on the expected time to extinction under any such dynamic allocation policy, in terms of a combinatorial quantity that we call the resistance of the set of initially infected nodes, the available budget, and the number of nodes n. Specifically, we consider the case of bounded degree graphs, with the resistance growing linearly in n. We show that if the curing budget is less than a certain multiple of the resistance, then the expected time to extinction grows exponentially with n. As a corollary, if all nodes are initially infected and the CutWidth of the graph grows linearly, while the curing budget is less than a certain multiple of the CutWidth, then the expected time to extinction grows exponentially in n.

The combination of these two results establishes a fairly sharp phase transition on the expected time to extinction (sublinear versus exponential) based on the relation between the CutWidth and the curing budget.

Bio

Kimon Drakopoulos   (M’13) received the diploma in electrical and computer engineering from the National Technical University of Athens, Athens, Greece, in 2009 and the M.Sc. degree in electrical engineering and computer science from the Massachusetts Institute of Technology, Cambridge, MA, in 2011.  From 2011 to present, he is a Ph.D. candidate at the Laboratory for Information and Decision Systems, Massachusetts Institute of Technology, Cambridge,
MA. His current research interests include social network analysis, network science,
applied probability, game theory and network economics.

]]> Anita Race 1 1449653011 2015-12-09 09:23:31 1492118248 2017-04-13 21:17:28 0 0 event 2015-12-10T16:00:00-05:00 2015-12-10T16:00:00-05:00 2015-12-10T16:00:00-05:00 2015-12-10 21:00:00 2015-12-10 21:00:00 2015-12-10 21:00:00 2015-12-10T16:00:00-05:00 2015-12-10T16:00:00-05:00 America/New_York America/New_York datetime 2015-12-10 04:00:00 2015-12-10 04:00:00 America/New_York America/New_York datetime <![CDATA[]]>
<![CDATA[SCL Course: Effectively Managing Global Supply and Risk in an Increasingly Complex World]]> 27233 COURSE DESCRIPTION

In recent years, increasing numbers of companies have become aware that the marketplace encompasses the world, not just the country in which they do business. For example, many firms have found that evaluating offshore sourcing alternatives is essential to a well-run, cost effective organization. Alternatively, by developing export markets, firms have highlighted the need for effective supply networks throughout the world. Conversely, companies located in other countries have also broadened their sourcing and marketing considerations geographically-they look toward global supply strategies and operations to provide competitive advantage through efficiency, effectiveness and differentiation.

Today’s supply managers are finding that they need to do much work in terms of conceptualizing, designing, and implementing initiatives that may be effective globally. In addition, supply managers need to understand the risks inherent in sourcing globally and be able to develop mitigation strategies for these risks. This course and the associated case studies, activities and discussions address key issues and topics that are essential to the global aspects of supply and risk management.

WHO SHOULD ATTEND

Sourcing and procurement managers, supply chain managers, logistics and supply chain planners, leaders/individuals involved in supply chain cost reduction initiatives, individuals needing to expand sourcing knowledge

HOW YOU WILL BENEFIT

Upon completion of this course, you will be able to:

WHAT IS COVERED

The Changing Global Landscape
Understanding International Sourcing
Sourcing in Emerging and Low Cost Countries
Identifying, Preventing and Managing Global Supply Risk
Pursuing Global Supply Management Excellence
Developing Global Supply Strategies
Creating a World Class Global Organization

COURSE MATERIALS

Required

Provided

COURSE PREREQUISITES

None.

CERTIFICATE INFORMATION

This course is part of the Strategic Sourcing and Supply Management (SSSM) Certificate.

PROGRAM TIMES

On the first day, please check in at least 30 minutes before the class start time.

]]> Andy Haleblian 1 1442399423 2015-09-16 10:30:23 1475892830 2016-10-08 02:13:50 0 0 event Today’s supply managers are finding that they need to do much work in terms of conceptualizing, designing, and implementing initiatives that may be effective globally. In addition, supply managers need to understand the risks inherent in sourcing globally and be able to develop mitigation strategies for these risks. This course and the associated case studies, activities and discussions address key issues and topics that are essential to the global aspects of supply and risk management.

]]>
2015-10-06T09:00:00-04:00 2015-10-08T18:00:00-04:00 2015-10-08T18:00:00-04:00 2015-10-06 13:00:00 2015-10-08 22:00:00 2015-10-08 22:00:00 2015-10-06T09:00:00-04:00 2015-10-08T18:00:00-04:00 America/New_York America/New_York datetime 2015-10-06 09:00:00 2015-10-08 06:00:00 America/New_York America/New_York datetime <![CDATA[]]> info@scl.gatech.edu

]]>
<![CDATA[Course registration page]]> <![CDATA[Course webpage within the SCL website]]>
<![CDATA[SCL Course: Lean Warehousing]]> 27233 COURSE DESCRIPTION

This course will demonstrate how warehouse operations are a key enabler to a successful supply chain implementation and the starting point for a transformation. It is critical to manage safety, quality and efficiency. Learn to leverage the lean supply chain modifications to improve customer responsiveness and reduce operating costs and in doing so contributing to a supply chain that creates a competitive advantage for a company. To accomplish this goal, we must bring lean principles into the warehouse and distribution center.

WHO SHOULD ATTEND

Supply chain professionals, logistics professionals, material managers, production control managers, transportation managers, warehousing managers and purchasing managers

HOW YOU WILL BENEFIT

Upon completion of this course, you will be able to:

Benefits:

WHAT IS COVERED

COURSE MATERIALS

Required

Provided

COURSE PREREQUISITES

None.

CERTIFICATE INFORMATION

This course is part of the Distribution Operations Analysis and Design (DOAD) Certificate.

PROGRAM TIMES

On the first day, please check in at least 30 minutes before the class start time.

]]> Andy Haleblian 1 1442400521 2015-09-16 10:48:41 1475892830 2016-10-08 02:13:50 0 0 event This course will demonstrate how warehouse operations are a key enabler to a successful supply chain implementation and the starting point for a transformation. It is critical to manage safety, quality and efficiency. Learn to leverage the lean supply chain modifications to improve customer responsiveness and reduce operating costs and in doing so contributing to a supply chain that creates a competitive advantage for a company. To accomplish this goal, we must bring lean principles into the warehouse and distribution center.

]]>
2015-10-20T09:00:00-04:00 2015-10-22T18:00:00-04:00 2015-10-22T18:00:00-04:00 2015-10-20 13:00:00 2015-10-22 22:00:00 2015-10-22 22:00:00 2015-10-20T09:00:00-04:00 2015-10-22T18:00:00-04:00 America/New_York America/New_York datetime 2015-10-20 09:00:00 2015-10-22 06:00:00 America/New_York America/New_York datetime <![CDATA[]]> info@scl.gatech.edu

]]>
<![CDATA[Course registration page]]> <![CDATA[Course webpage within the SCL website]]>
<![CDATA[SCL Course: Introduction to International Logistics and Compliance (Savannah, GA)]]> 27233 COURSE DESCRIPTION

This course is designed to provide students with an understanding of the complexities of global trade, its impact on logistics, and key areas of concern for international logistics managers. Key topics are investigated such as: Incoterms, global trade compliance, harmonized tariff schedules, US import and export regulations, US Free Trade Agreements, and supply chain security.

WHO SHOULD ATTEND

HOW YOU WILL BENEFIT

Upon completion of this course, you will be able to:

WHAT IS COVERED

COURSE MATERIALS

Participants receive a course notebook.

COURSE PREREQUISITES

None.

CERTIFICATE INFORMATION

This course is part of the Supply Chain Management (SCM) Certificate.

PROGRAM TIMES

On the first day, please check in at least 30 minutes before the class start time.

]]> Andy Haleblian 1 1442401546 2015-09-16 11:05:46 1475892830 2016-10-08 02:13:50 0 0 event This course is designed to provide students with an understanding of the complexities of global trade, its impact on logistics, and key areas of concern for international logistics managers. Key topics are investigated such as: Incoterms, global trade compliance, harmonized tariff schedules, US import and export regulations, US Free Trade Agreements, and supply chain security.

]]>
2015-10-21T09:00:00-04:00 2015-10-23T18:00:00-04:00 2015-10-23T18:00:00-04:00 2015-10-21 13:00:00 2015-10-23 22:00:00 2015-10-23 22:00:00 2015-10-21T09:00:00-04:00 2015-10-23T18:00:00-04:00 America/New_York America/New_York datetime 2015-10-21 09:00:00 2015-10-23 06:00:00 America/New_York America/New_York datetime <![CDATA[]]> info@scl.gatech.edu

]]>
<![CDATA[Course registration page]]> <![CDATA[Course webpage within the SCL website]]>
<![CDATA[SCL Course: Warehouse/Distribution Center Layout]]> 27233 COURSE DESCRIPTION

Do you work with problems involving the use of material handling equipment in plants, warehouses and other commercial enterprises? Focus on material handling and distribution problems from the source of raw material through manufacturing and distribution systems to the final consumer. All techniques presented are field-proven and derived from successful implementation. Case exercises are adapted from real situations and projects.

WHO SHOULD ATTEND

Industrial engineers and systems analysts, warehouse supervisors and team leaders, warehouse/distribution center managers, logistics and supply chain planners, planning teams for new or expanded facilities and leaders of supply chain and lean initiatives

HOW YOU WILL BENEFIT

Upon completion of this course, you will be able to:

WHAT IS COVERED

COURSE MATERIALS

Participants receive a course notebook.

COURSE PREREQUISITES

None.

CERTIFICATE INFORMATION

This course is part of the Distribution Operations Analysis and Design (DOAD) Certificate.

PROGRAM TIMES

On the first day, please check in at least 30 minutes before the class start time.

]]> Andy Haleblian 1 1442404417 2015-09-16 11:53:37 1475892830 2016-10-08 02:13:50 0 0 event Do you work with problems involving the use of material handling equipment in plants, warehouses and other commercial enterprises? Focus on material handling and distribution problems from the source of raw material through manufacturing and distribution systems to the final consumer. All techniques presented are field-proven and derived from successful implementation. Case exercises are adapted from real situations and projects.

]]>
2015-10-27T09:00:00-04:00 2015-10-29T18:00:00-04:00 2015-10-29T18:00:00-04:00 2015-10-27 13:00:00 2015-10-29 22:00:00 2015-10-29 22:00:00 2015-10-27T09:00:00-04:00 2015-10-29T18:00:00-04:00 America/New_York America/New_York datetime 2015-10-27 09:00:00 2015-10-29 06:00:00 America/New_York America/New_York datetime <![CDATA[]]> info@scl.gatech.edu

]]>
<![CDATA[Course registration page]]> <![CDATA[Course webpage within the SCL website]]>
<![CDATA[SCL Course: Measuring and Managing Performance in Supply Chain and Logistics Operations]]> 27233 COURSE DESCRIPTION

Develop effective supply chain performance measurement processes to drive alignment across the corporation. Corporations often struggle with a lack of alignment between financial goals and operational metrics. Additionally, today’s information technology often overwhelms management with data and metrics. In this 3-day course you will learn how to develop metrics that synchronize supply chain and logistics metrics with key company financial metrics and goals. Through coursework and hands-on exercises, students learn to tailor metrics and measurement processes to focus on the most important aspects of the operations and overall corporate goals. As part of the course, performance dashboards from participating student companies and best-in-class organizations will be critiqued. The methods learned in the course are intended to be immediately put to use in the corporate environment.

WHO SHOULD ATTEND

Chief supply chain officers, executive vice presidents of supply chain, executive vice presidents of procurement, company owners, logistics service providers, consultants, vice presidents of sales operations, vice presidents/directors of process improvement, executive/senior/vice presidents/directors of supply chain, executive/senior/vice presidents/directors of logistics, executive/senior/vice presidents/directors of procurement, executive/senior/vice presidents/directors of manufacturing, executive/senior/vice presidents/directors of distribution

HOW YOU WILL BENEFIT

Upon completion of this course, you will be able to:

WHAT IS COVERED

COURSE MATERIALS

Participants receive a course notebook.

COURSE PREREQUISITES

None.

CERTIFICATE INFORMATION

This course is part of the Supply Chain Management (SCM) Certificate.

PROGRAM TIMES

On the first day, please check in at least 30 minutes before the class start time.

]]> Andy Haleblian 1 1442406022 2015-09-16 12:20:22 1475892830 2016-10-08 02:13:50 0 0 event Develop effective supply chain performance measurement processes to drive alignment across the corporation. Corporations often struggle with a lack of alignment between financial goals and operational metrics. Additionally, today’s information technology often overwhelms management with data and metrics. In this 3-day course you will learn how to develop metrics that synchronize supply chain and logistics metrics with key company financial metrics and goals. Through coursework and hands-on exercises, students learn to tailor metrics and measurement processes to focus on the most important aspects of the operations and overall corporate goals. As part of the course, performance dashboards from participating student companies and best-in-class organizations will be critiqued. The methods learned in the course are intended to be immediately put to use in the corporate environment.

]]>
2015-11-04T08:00:00-05:00 2015-11-06T17:00:00-05:00 2015-11-06T17:00:00-05:00 2015-11-04 13:00:00 2015-11-06 22:00:00 2015-11-06 22:00:00 2015-11-04T08:00:00-05:00 2015-11-06T17:00:00-05:00 America/New_York America/New_York datetime 2015-11-04 08:00:00 2015-11-06 05:00:00 America/New_York America/New_York datetime <![CDATA[]]> info@scl.gatech.edu

]]>
<![CDATA[Course registration page]]> <![CDATA[Course webpage within the SCL website]]>
<![CDATA[Center of Innovation for Logistics with Georgia Tech Hosts Logistics Open House]]> 27233 Join the Georgia Center of Innovation for Logistics the morning of Wednesday, September 2nd for an open house to discover more about the organization, learn what it offers logistics providers and suppliers in the state, and to make connections with fellow logistics industry professionals in the Savannah region. 

On hand with be Dr. Chip White, professor and Schneider National Chair in Transportation and Logistics in the Stewart School of Industrial & Systems Engineering at Georgia Tech. Dr. White will kickoff the event with a brief presentation at 8am, with networking and light breakfast to follow.

Drop in as you like! Doors open at 7:30am and the center staff will be available to mix-and-mingle until about 10am. Please RSVP by Friday, August 28th by emailing coilogistics@gmail.com.

 

]]> Andy Haleblian 1 1440065654 2015-08-20 10:14:14 1475892790 2016-10-08 02:13:10 0 0 event Join the Georgia Center of Innovation for Logistics the morning of September 2nd for an open house on to discover more about the organization and what it offers logistics providers and suppliers in the state.

]]>
2015-09-02T08:30:00-04:00 2015-09-02T11:00:00-04:00 2015-09-02T11:00:00-04:00 2015-09-02 12:30:00 2015-09-02 15:00:00 2015-09-02 15:00:00 2015-09-02T08:30:00-04:00 2015-09-02T11:00:00-04:00 America/New_York America/New_York datetime 2015-09-02 08:30:00 2015-09-02 11:00:00 America/New_York America/New_York datetime <![CDATA[]]> Please RSVP by Friday, August 28th by emailing coilogistics@gmail.com

]]>
437651 437651 image <![CDATA[Georgia Center of Innovation for Logistics]]> image/jpeg 1449256162 2015-12-04 19:09:22 1475895176 2016-10-08 02:52:56
<![CDATA[SCL Course: Supply Chain Project Management Fundamentals]]> 27233 COURSE DESCRIPTION

Supply Chain Management projects can span a wide range of project types including supply chain strategy, network analysis, facility design build, supply chain technology selection and implementation, and continuous process improvement initiatives. This course provides an overview of project management methodologies as applied in the supply chain environment. Class discussion and projects provide an understanding of how fundamental project management approaches and industry best practices can be used to effectively manage the complexities. Supply chain projects typically require managing resources, stakeholder alignment, risk management, customer impact, and effective communication across many internal and external business partners.

HOW YOU WILL BENEFIT

Upon completion of this course, you will be able to:

WHAT IS COVERED

ON-CAMPUS COURSE MATERIALS

COURSE PREREQUISITES

None.

CERTIFICATE INFORMATION

For those interested in earning the Supply Chain Project Management Certificate, this course is the first of the four-course certificate program. To earn the certificate, participants must register and complete the following courses in the below sequence within four years, plus one elective.

  1. Supply Chain Project Management: Fundamentals
  2. Supply Chain Project Management: Vendor Selection & Management
  3. Supply Chain Project Management: Effectively Managing Transformation Projects

For a list of courses that can be used as electives towards this certificate, please visit the Georgia Tech Professional Education website.

]]> Andy Haleblian 1 1439976593 2015-08-19 09:29:53 1475892786 2016-10-08 02:13:06 0 0 event Supply Chain Management projects can span a wide range of project types including supply chain strategy, network analysis, facility design build, supply chain technology selection and implementation, and continuous process improvement initiatives. This course provides an overview of project management methodologies as applied in the supply chain environment.

]]>
2015-09-08T09:00:00-04:00 2015-09-10T18:00:00-04:00 2015-09-10T18:00:00-04:00 2015-09-08 13:00:00 2015-09-10 22:00:00 2015-09-10 22:00:00 2015-09-08T09:00:00-04:00 2015-09-10T18:00:00-04:00 America/New_York America/New_York datetime 2015-09-08 09:00:00 2015-09-10 06:00:00 America/New_York America/New_York datetime <![CDATA[]]> info@scl.gatech.edu

]]>
<![CDATA[Course registration page]]> <![CDATA[Course webpage within the SCL website]]>
<![CDATA[SCL Course: Building the Lean Supply Chain Leader]]> 27233 Course Description

Transforming an organization from traditional thinking to lean thinking requires leadership. This is the third course in a 3-course series on becoming a Lean Supply Chain professional. While the previous courses focus on strategic and tactical implementation of the lean supply chain, this final course "builds" the individual into a lean leader. This transformation is critical to navigate through the waters of change management that is required to successfully execute and sustain the lean supply chain journey.

In the third course, participants will complete a deep dive into the concept of the House of Lean and explore the main aspects of lean leadership. Among other important topics, participants will learn "go see" management, "A3 thinking" and "leader-as-teacher" concepts.

Who Should Attend

Supply chain professionals, logistics professionals, material managers, production control managers, transportation managers, warehousing managers and purchasing managers.

How You Will Benefit

 

What Is Covered

Day 1 - The Lean Culture
Day 2 - Lean Principles Deep Dive
Day 3 - The Lean Leader
  ]]> Andy Haleblian 1 1444129877 2015-10-06 11:11:17 1475892768 2016-10-08 02:12:48 0 0 event In this course, students will complete a deep dive of the House of Lean and explore the main aspects of lean leadership. Among other important topics, students will learn "go see" management, A3 thinking and leader as teacher concepts.

]]>
2015-11-10T08:00:00-05:00 2015-11-12T17:00:00-05:00 2015-11-12T17:00:00-05:00 2015-11-10 13:00:00 2015-11-12 22:00:00 2015-11-12 22:00:00 2015-11-10T08:00:00-05:00 2015-11-12T17:00:00-05:00 America/New_York America/New_York datetime 2015-11-10 08:00:00 2015-11-12 05:00:00 America/New_York America/New_York datetime <![CDATA[]]> EMAIL: info@scl.gatech.edu or CALL: (404) 385-3501 between 9:00a.m. and 4:00p.m., Eastern time.

]]>
<![CDATA[2015-16 LEAN Supply Chain Brochure (PDF)]]> <![CDATA[Register Online via the GT Professional Education website]]> <![CDATA[Course webpage within the SCL website]]>
<![CDATA[SCL Course: Demand-Driven Supply Chain Strategy]]> 27233 COURSE DESCRIPTION

As supply chain executives become more instrumental in supporting long-term strategic objectives, they need to complement traditional supply chain operational knowledge with a more strategic view of their role in delivering aligned results to the business. Learn about different lenses of strategic planning applied to supply chain management and the specific implications on supply chain operations.

During the course, attendees will have several opportunities to assess their current supply chain strategy, formulate a new one, discuss about keys to implement a demand driven supply chain strategy, and how to review and align one.

For the duration of the course, participants will have the opportunity to work with an extended simulation game of a fictional company, and see the impact of supply chain strategic decisions in real time, while monitoring their ability to manage uncertainty and deliver financial results for the firm.

WHO SHOULD ATTEND

Those responsible for determining the future position of supply chain strategy, executing specific supply chain processes that must support specific business initiatives, innovating supply chain strategy to better align with operational goals, understanding the strategic impact of supply chain decisions in the firm and the extended enterprise network and advising clients about their specific supply chain strategic positioning.

HOW YOU WILL BENEFIT

WHAT IS COVERED

ON-CAMPUS COURSE MATERIALS

Participants receive a course notebook.

COURSE PREREQUISITES

None.

REQUIRED MATERIAL

Laptop computer and calculator to be provided by the student.

CERTIFICATE INFORMATION

This course is part of the Supply Chain Management (SCM) Certificate and can also be used towards the Strategic Sourcing and Supply Management (SSSM) Certificate.

PROGRAM TIMES

On the first day, please check in at least 30 minutes before the class start time.

]]> Andy Haleblian 1 1436359764 2015-07-08 12:49:24 1475892746 2016-10-08 02:12:26 0 0 event As supply chain executives become more instrumental in supporting long-term strategic objectives, they need to complement traditional supply chain operational knowledge with a more strategic view of their role in delivering aligned results to the business. Learn about different lenses of strategic planning applied to supply chain management and the specific implications on supply chain operations.

]]>
2015-08-18T09:00:00-04:00 2015-08-21T13:00:00-04:00 2015-08-21T13:00:00-04:00 2015-08-18 13:00:00 2015-08-21 17:00:00 2015-08-21 17:00:00 2015-08-18T09:00:00-04:00 2015-08-21T13:00:00-04:00 America/New_York America/New_York datetime 2015-08-18 09:00:00 2015-08-21 01:00:00 America/New_York America/New_York datetime <![CDATA[]]> info@scl.gatech.edu

]]>
<![CDATA[Course webpage within the SCL website]]> <![CDATA[Course registration page]]>
<![CDATA[SCL Course: Building the Lean Supply Chain Problem Solver]]> 27233 COURSE DESCRIPTION

To become a lean supply chain professional, you need to first become a lean thinker and lean problem solver. This is the first course in a 3-course series on becoming a Lean Supply Chain professional. Students will be introduced to lean thinking and critical lean concepts. In addition, students will become proficient problem solvers through gained skills of waste identification and use of fundamental problem solving tools to eliminate waste at the root cause.

This first course is a pivot point in the educational process, and this is where current mental models and business paradigms will be challenged. Participants will learn to see operations from a new vantage point. Upon arrival back to the workplace, participants will be able to "talk" the talk of lean and also have a keen eye for operational waste. However, the real test of the first course is when the participant can "walk" the lean walk and solve business problems at the root cause, completely eliminating the problem for the organization.

WHO SHOULD ATTEND

Supply chain professionals, logistics professionals, material managers, production control managers, transportation managers, warehousing managers and purchasing managers

HOW YOU WILL BENEFIT

WHAT IS COVERED

Day 1 - Lean Thinking

Day 2 - Fundamental Problem Solving

Day 3 - Sustaining Improvements

Post Classroom Training

ON-CAMPUS COURSE MATERIALS

Participants receive a course notebook.

COURSE PREREQUISITES

None.

CERTIFICATE INFORMATION

This is a required course for the Lean Supply Chain Professional (LSCP) Certificate and can also be used as an elective for the Supply Chain Management (SCM) Certificate.

PROGRAM TIMES

On the first day, please check in at least 30 minutes before the class start time.

]]> Andy Haleblian 1 1436360236 2015-07-08 12:57:16 1475892746 2016-10-08 02:12:26 0 0 event To become a lean supply chain professional, you need to first become a lean thinker and lean problem solver. This is the first course in a 3-course series on becoming a Lean Supply Chain professional. Students will be introduced to lean thinking and critical lean concepts. In addition, students will become proficient problem solvers through gained skills of waste identification and use of fundamental problem solving tools to eliminate waste at the root cause.

]]>
2015-09-15T09:00:00-04:00 2015-09-17T18:00:00-04:00 2015-09-17T18:00:00-04:00 2015-09-15 13:00:00 2015-09-17 22:00:00 2015-09-17 22:00:00 2015-09-15T09:00:00-04:00 2015-09-17T18:00:00-04:00 America/New_York America/New_York datetime 2015-09-15 09:00:00 2015-09-17 06:00:00 America/New_York America/New_York datetime <![CDATA[]]> info@scl.gatech.edu

]]>
<![CDATA[Course registration page]]> <![CDATA[Course webpage within the SCL website]]>
<![CDATA[SCL Course: World Class Sales and Operations Planning]]> 27233 COURSE DESCRIPTION

This course focuses on defining, executing, and improving the S&OP process. Participants will be introduced to the appropriate stakeholders of S&OP, the importance of S&OP to corporate performance, S&OP cadence, and the use of visionary technology to bring S&OP to the next level. Business cases will be used to show concrete examples of companies where S&OP is effectively applied.

WHO SHOULD ATTEND

HOW YOU WILL BENEFIT

Upon completion of this course, you will be able to:

LEARNING OBJECTIVES

WHAT IS COVERED

ON-CAMPUS COURSE MATERIALS

Participants receive a course notebook, an in-class software demonstration, and hand-out materials.

COURSE PREREQUISITES

None.

REQUIRED MATERIAL

Laptop computer by the student.

CERTIFICATE INFORMATION

This course is part of the Supply and Demand Planning (SDP) Certificate.

PROGRAM TIMES

On the first day, please check in at least 30 minutes before the class start time.

]]> Andy Haleblian 1 1432284835 2015-05-22 08:53:55 1475892726 2016-10-08 02:12:06 0 0 event This course focuses on defining, executing, and improving the S&OP process. Participants will be introduced to the appropriate stakeholders of S&OP, the importance of S&OP to corporate performance, S&OP cadence, and the use of visionary technology to bring S&OP to the next level. Business cases will be used to show concrete examples of companies where S&OP is effectively applied.

]]>
2015-06-01T09:00:00-04:00 2015-06-02T13:00:00-04:00 2015-06-02T13:00:00-04:00 2015-06-01 13:00:00 2015-06-02 17:00:00 2015-06-02 17:00:00 2015-06-01T09:00:00-04:00 2015-06-02T13:00:00-04:00 America/New_York America/New_York datetime 2015-06-01 09:00:00 2015-06-02 01:00:00 America/New_York America/New_York datetime <![CDATA[]]> info@scl.gatech.edu

]]>
<![CDATA[Course registration page]]> <![CDATA[World Class Sales & Operations Planning course]]>
<![CDATA[SCL Course: Integrated Business Planning]]> 27233 COURSE DESCRIPTION

This course provides a holistic view toward corporate profitability and supports effective complexity management. Participants will learn about the challenges of today's operating environment with "big data," cross-functional consensus and strategies that impact profitability. Integrated Business Planning (IBP) building blocks will be provided that solve these challenges. Break-out sessions will allow participants to apply these IBP concepts with an interactive tool.

WHO SHOULD ATTEND

HOW YOU WILL BENEFIT

Upon completion of this course, you will be able to:

LEARNING OBJECTIVES

WHAT IS COVERED

ON-CAMPUS COURSE MATERIALS

Participants receive a course notebook and reading materials.

COURSE PREREQUISITES

None.

REQUIRED MATERIAL

Laptop computer by the student.

CERTIFICATE INFORMATION

This course is part of the Supply and Demand Planning (SDP) Certificate.

PROGRAM TIMES

On the first day, please check in at least 30 minutes before the class start time.

]]> Andy Haleblian 1 1432300035 2015-05-22 13:07:15 1475892726 2016-10-08 02:12:06 0 0 event This course provides a holistic view toward corporate profitability and supports effective complexity management. Participants will learn about the challenges of today's operating environment with "big data," cross-functional consensus and strategies that impact profitability. Integrated Business Planning (IBP) building blocks will be provided that solve these challenges. Break-out sessions will allow participants to apply these IBP concepts with an interactive tool.

]]>
2015-06-02T14:00:00-04:00 2015-06-03T18:00:00-04:00 2015-06-03T18:00:00-04:00 2015-06-02 18:00:00 2015-06-03 22:00:00 2015-06-03 22:00:00 2015-06-02T14:00:00-04:00 2015-06-03T18:00:00-04:00 America/New_York America/New_York datetime 2015-06-02 02:00:00 2015-06-03 06:00:00 America/New_York America/New_York datetime <![CDATA[]]> info@scl.gatech.edu

]]>
<![CDATA[Course registration page]]> <![CDATA[Integrated Business Planning course]]>
<![CDATA[SCL Course: Supply Chain Risk Management]]> 27233 COURSE DESCRIPTION

In today's global economy, operating risks are increasingly on the minds of executives. The specific context of operating risk can range from general areas of business continuity to the effects of natural disasters. In this course participants will gain a solid understanding of Supply Chain Risk Management principals including effective ways to identify, mitigate and measure the impact of potential supply chain disruptions.

WHO SHOULD ATTEND

HOW YOU WILL BENEFIT

Upon completion of this course, you will be able to:


LEARNING OBJECTIVES

WHAT IS COVERED

ON-CAMPUS COURSE MATERIALS

Participants receive a course notebook and reading materials.

COURSE PREREQUISITES

None.

REQUIRED MATERIAL

Laptop computer by the student.

CERTIFICATE INFORMATION

This course is part of the Supply and Demand Planning (SDP) Certificate.

PROGRAM TIMES

On the first day, please check in at least 30 minutes before the class start time.

]]> Andy Haleblian 1 1432300374 2015-05-22 13:12:54 1475892726 2016-10-08 02:12:06 0 0 event In today's global economy, operating risks are increasingly on the minds of executives. The specific context of operating risk can range from general areas of business continuity to the effects of natural disasters. In this course participants will gain a solid understanding of Supply Chain Risk Management principals including effective ways to identify, mitigate and measure the impact of potential supply chain disruptions.

]]>
2015-06-04T09:00:00-04:00 2015-06-05T13:00:00-04:00 2015-06-05T13:00:00-04:00 2015-06-04 13:00:00 2015-06-05 17:00:00 2015-06-05 17:00:00 2015-06-04T09:00:00-04:00 2015-06-05T13:00:00-04:00 America/New_York America/New_York datetime 2015-06-04 09:00:00 2015-06-05 01:00:00 America/New_York America/New_York datetime <![CDATA[]]> info@scl.gatech.edu

]]>
<![CDATA[Course registration page]]> <![CDATA[Supply Chain Risk Management course]]>
<![CDATA[SCL Course: Defining and Implementing Effective Sourcing Strategies]]> 27233 COURSE DESCRIPTION

Leading companies throughout the world are looking to formulate integrated supply strategies on both a local geographical and a global basis. This includes strategically sourcing materials and components worldwide and selecting global locations for key supply and distribution centers. The growth in global trade will continue to have a major impact on supply chains and requires firms to practice professional strategic supply management in order to help ensure continuity of supply and to contain and reduce costs.

Strategic sourcing enhances value, ultimately impacting the profitability of an entire organization. In this essential course, you’ll learn how to develop and implement a sourcing strategy that aligns with overall competitive strategy. The course and the associated case studies, activities and discussions provide the context and a framework for making effective sourcing decisions including a comprehensive approach to strategic sourcing.

WHO SHOULD ATTEND

HOW YOU WILL BENEFIT

Upon completion of this course, you will be able to:

 

WHAT IS COVERED

The Role and Nature of Procurement and Supply Management in a Supply Chain Context

Supply Management and Strategy Development

Types of Supply Strategies

Procurement Analysis and Classification

Assessing Supply Markets

Developing a Sourcing Strategy

Implementing the Strategy

Institutionalizing the Strategy

Trends & Future Directions in Sourcing

ON-CAMPUS COURSE MATERIALS

Participants receive a course notebook and the book "Strategic Supply Management" by Dr. Robert Trent.

COURSE PREREQUISITES

None.

REQUIRED MATERIAL

Laptop computer and calculator to be provided by the student.

CERTIFICATE INFORMATION

This course is part of the Supply and Demand Planning (SDP) Certificate.

PROGRAM TIMES

On the first day, please check in at least 30 minutes before the class start time.

]]> Andy Haleblian 1 1432300819 2015-05-22 13:20:19 1475892726 2016-10-08 02:12:06 0 0 event Leading companies throughout the world are looking to formulate integrated supply strategies on both a local geographical and a global basis. This includes strategically sourcing materials and components worldwide and selecting global locations for key supply and distribution centers. The growth in global trade will continue to have a major impact on supply chains and requires firms to practice professional strategic supply management in order to help ensure continuity of supply and to contain and reduce costs.

Strategic sourcing enhances value, ultimately impacting the profitability of an entire organization. In this essential course, you’ll learn how to develop and implement a sourcing strategy that aligns with overall competitive strategy. The course and the associated case studies, activities and discussions provide the context and a framework for making effective sourcing decisions including a comprehensive approach to strategic sourcing.

]]>
2015-06-23T09:00:00-04:00 2015-06-25T18:00:00-04:00 2015-06-25T18:00:00-04:00 2015-06-23 13:00:00 2015-06-25 22:00:00 2015-06-25 22:00:00 2015-06-23T09:00:00-04:00 2015-06-25T18:00:00-04:00 America/New_York America/New_York datetime 2015-06-23 09:00:00 2015-06-25 06:00:00 America/New_York America/New_York datetime <![CDATA[]]> info@scl.gatech.edu

]]>
<![CDATA[Course registration page]]> <![CDATA[Defining and Implementing Effective Sourcing Strategies Course]]>
<![CDATA[Faculty Candidate Seminar - Pierre Nyquist]]> 27187 TITLE: Large deviations and accelerated Monte Carlo methods

ABSTRACT:

Monte Carlo methods have emerged as a set of indispensable tools in the applied sciences and engineering. In situations where the underlying stochastic model is too complex for analytical calculations to be tractable they offer a convenient way to obtain numerical approximations. However, the problem of rare-event sampling can often be a hindrance to the use such methods. In order to overcome this problem one must use some type of accelerated Monte Carlo method, in which a control mechanism is used to guide the particles in the simulation into the relevant parts of the state space. Earlier results in the area have shown that intuition can be misleading in the design of such controls and a proper theoretical analysis of the simulation method of choice is often needed.

The aim of this talk is to discuss the connection between Monte Carlo methods, and the rare-event sampling problem, and large deviations. Large deviation theory is the branch of probability theory that deals with rare events. In addition to providing estimates to the probabilities of such events, the theory also gives insight into how the events will occur. This is precisely the kind of insight needed to develop efficient Monte Carlo methods. After a brief overview of these two topics I will focus on the method known as importance sampling and how one can analyze and design efficient algorithms by means of large deviation theory. In particular, I will discuss connections to Hamilton-Jacobi equations and a recent results of ours on representations of solutions to such PDE’s and its applications to rare-event simulation.

]]> Anita Race 1 1428650028 2015-04-10 07:13:48 1475892699 2016-10-08 02:11:39 0 0 event 2015-04-14T12:00:00-04:00 2015-04-14T12:00:00-04:00 2015-04-14T12:00:00-04:00 2015-04-14 16:00:00 2015-04-14 16:00:00 2015-04-14 16:00:00 2015-04-14T12:00:00-04:00 2015-04-14T12:00:00-04:00 America/New_York America/New_York datetime 2015-04-14 12:00:00 2015-04-14 12:00:00 America/New_York America/New_York datetime <![CDATA[]]>
<![CDATA[Dean May's ISyE Office Hour]]> 27187 Office Hour with the Dean

The visit will have no specific agenda and Dean May will not make a presentation nor have any prepared remarks.  This is an opportunity for the faculty and staff to ask questions, share concerns, or offer suggestions to the Dean.  You may come in one at a time or as a small group but the intent is to have people coming and going throughout the hour.

]]> Anita Race 1 1421835738 2015-01-21 10:22:18 1475892683 2016-10-08 02:11:23 0 0 event 2015-02-23T11:00:00-05:00 2015-02-23T12:00:00-05:00 2015-02-23T12:00:00-05:00 2015-02-23 16:00:00 2015-02-23 17:00:00 2015-02-23 17:00:00 2015-02-23T11:00:00-05:00 2015-02-23T12:00:00-05:00 America/New_York America/New_York datetime 2015-02-23 11:00:00 2015-02-23 12:00:00 America/New_York America/New_York datetime <![CDATA[]]>
<![CDATA[Visiting Speaker Seminar]]> 27187 TITLE: Lattice-Dependent Base-Stock and Rationing Policies for Assemble-to-Order Systems

SPEAKER: Mustafa Akan

ABSTRACT:

We consider an assemble-to-order generalized M-system with multiple components and products, batch ordering of components, random lead times, and lost sales. We model the system as an infinite-horizon Markov decision process and seek an optimal control policy, which specifies when a batch of components should be produced and whether an arriving demand for each product should be satisfied. To facilitate our analysis, we introduce new functional characterizations for convexity and submodularity with respect to certain non-unitary directions. These help us characterize optimal inventory replenishment and allocation policies under a mild condition on component batch sizes: lattice-dependent base-stock and lattice-dependent rationing (LBLR). We conduct numerical experiments to evaluate the use of an LBLR policy for general ATO systems (which violate the M-system product structure) as a heuristic, comparing it to two other heuristics from the literature: a state-dependent base-stock and state-dependent rationing (SBSR) policy, and a fixed base-stock and fixed rationing (FBFR) policy. Remarkably, LBLR yields the globally optimal cost in every experiment. LBLR and SBSR perform significantly better than FBFR when replenishment batch sizes imperfectly match the component requirements of the most valuable or most highly demanded product. In addition, LBLR substantially outperforms SBSR if a significant amount of inventory must be held for rationing. Finally, we approximate the optimal cost function by reducing the state space of the original problem through a novel hard aggregation technique. We establish that LBLR is the optimal policy obtained by solving the aggregate problem. We derive error bounds for this approximation and present preliminary computational results.

]]> Anita Race 1 1420732755 2015-01-08 15:59:15 1475892664 2016-10-08 02:11:04 0 0 event 2015-01-20T11:00:00-05:00 2015-01-20T12:00:00-05:00 2015-01-20T12:00:00-05:00 2015-01-20 16:00:00 2015-01-20 17:00:00 2015-01-20 17:00:00 2015-01-20T11:00:00-05:00 2015-01-20T12:00:00-05:00 America/New_York America/New_York datetime 2015-01-20 11:00:00 2015-01-20 12:00:00 America/New_York America/New_York datetime <![CDATA[]]>
<![CDATA[Visiting Speaker Seminar]]> 27187 TITLE: "Sparse" computation of gradients for optimization with large data sets

SPEAKER: Dr. Guanghui Lan

ABSTRACT:

The last few years have seen an increasing interest in utilizing optimization for large-scale data analysis. However, optimization problems arising from these applications often involve, in addition to expensive smooth components for data fitting, nonsmooth and nonseparable regularization terms/constraints to enforce certain structural properties for the generated solutions (e.g, low rank or group sparsity). It is well-known that such nonsmooth components can significantly slow down the convergence of existing first-order optimization algorithms, leading to a large number of traverses through the data sets. To address this issue, we present a new class of optimization techniques, referred as to gradient sliding and conditional gradient sliding methods, which can skip the computation of gradients from time to time while still maintaining the overall optimal rate of convergence. In particular, the number of gradient evaluations required for these algorithms will be the same as if the aforementioned nonsmooth and nonseparable components do not exist. When applied to data analysis problems, these algorithms can reduce the number of scans through the data sets by orders of magnitude. Numerical experiments have been conducted to illustrate the effectiveness of these techniques.

 Short-bio: Guanghui (George) Lan obtained his Ph.D. degree in Industrial and Systems Engineering from Georgia Institute of Technology in August, 2009. He then joined the Department of Industrial and Systems Engineering at the University of Florida as an assistant professor thereafter. His main research interests lie in stochastic optimization, nonlinear programming, simulation-based optimization, and their applications in data sciences. His research has been supported by the National Science Foundation and Office of Naval Research. The academic honors that he received include the INFORMS Computing Society Student Paper Competition First Place (2008), INFORMS George Nicholson Paper Competition Second Place (2008), Mathematical Optimization Society Tucker Prize Finalist (2012), INFORMS Junior Faculty Interest Group (JFIG) Paper Competition First Place (2012) and the National Science Foundation CAREER Award (2013).]]> Anita Race 1 1420033267 2014-12-31 13:41:07 1475892654 2016-10-08 02:10:54 0 0 event 2015-01-15T11:00:00-05:00 2015-01-15T12:00:00-05:00 2015-01-15T12:00:00-05:00 2015-01-15 16:00:00 2015-01-15 17:00:00 2015-01-15 17:00:00 2015-01-15T11:00:00-05:00 2015-01-15T12:00:00-05:00 America/New_York America/New_York datetime 2015-01-15 11:00:00 2015-01-15 12:00:00 America/New_York America/New_York datetime <![CDATA[]]>
<![CDATA[Faculty Meeting]]> 27187 Faculty Meeting

]]> Anita Race 1 1420450933 2015-01-05 09:42:13 1475892654 2016-10-08 02:10:54 0 0 event 2015-01-22T11:00:00-05:00 2015-01-22T11:00:00-05:00 2015-01-22T11:00:00-05:00 2015-01-22 16:00:00 2015-01-22 16:00:00 2015-01-22 16:00:00 2015-01-22T11:00:00-05:00 2015-01-22T11:00:00-05:00 America/New_York America/New_York datetime 2015-01-22 11:00:00 2015-01-22 11:00:00 America/New_York America/New_York datetime <![CDATA[]]>
<![CDATA[Faculty Meeting]]> 27187 Faculty Meeting

]]> Anita Race 1 1420451023 2015-01-05 09:43:43 1475892654 2016-10-08 02:10:54 0 0 event 2015-02-12T11:00:00-05:00 2015-02-12T11:00:00-05:00 2015-02-12T11:00:00-05:00 2015-02-12 16:00:00 2015-02-12 16:00:00 2015-02-12 16:00:00 2015-02-12T11:00:00-05:00 2015-02-12T11:00:00-05:00 America/New_York America/New_York datetime 2015-02-12 11:00:00 2015-02-12 11:00:00 America/New_York America/New_York datetime <![CDATA[]]>
<![CDATA[Faculty Meeting]]> 27187 Faculty Meeting

]]> Anita Race 1 1420451089 2015-01-05 09:44:49 1475892654 2016-10-08 02:10:54 0 0 event 2015-03-12T12:00:00-04:00 2015-03-12T12:00:00-04:00 2015-03-12T12:00:00-04:00 2015-03-12 16:00:00 2015-03-12 16:00:00 2015-03-12 16:00:00 2015-03-12T12:00:00-04:00 2015-03-12T12:00:00-04:00 America/New_York America/New_York datetime 2015-03-12 12:00:00 2015-03-12 12:00:00 America/New_York America/New_York datetime <![CDATA[]]>
<![CDATA[Faculty Meeting]]> 27187 Faculty Meeting

]]> Anita Race 1 1420451152 2015-01-05 09:45:52 1475892654 2016-10-08 02:10:54 0 0 event 2015-04-16T12:00:00-04:00 2015-04-16T12:00:00-04:00 2015-04-16T12:00:00-04:00 2015-04-16 16:00:00 2015-04-16 16:00:00 2015-04-16 16:00:00 2015-04-16T12:00:00-04:00 2015-04-16T12:00:00-04:00 America/New_York America/New_York datetime 2015-04-16 12:00:00 2015-04-16 12:00:00 America/New_York America/New_York datetime <![CDATA[]]>
<![CDATA[Faculty Candidate Seminar]]> 27187 TITLE:  Learning to optimize via efficient experimentation

SPEAKER:  Daniel Russo

ABSTRACT:

The information revolution is spawning systems that require very frequent decisions and provide high volumes of data concerning past outcomes. Fueling the design of algorithms used in such systems is a vibrant research area at the intersection of sequential decision-making and machine learning that addresses how to balance between exploration and exploitation and learn over time to make increasingly effective decisions.  In this talk, I will formulate a broad family of such problems that greatly extends the classical multi-armed bandit problem by allowing samples of one action to inform the decision-maker's assessment of other actions. I'll describe the rising importance of this problem class, and then discuss two recent methodological advances. One advance is Thompson sampling, a simple and tractable approach that is provably efficient for many relevant problem classes. The other is information-directed sampling, a new algorithm we propose that is inspired by an information-theoretic perspective and can offer greatly superior statistical efficiently. We provide new insight into both algorithms and establish general theoretical guarantees. ]]> Anita Race 1 1419852341 2014-12-29 11:25:41 1475892651 2016-10-08 02:10:51 0 0 event 2015-01-05T11:00:00-05:00 2015-01-05T12:00:00-05:00 2015-01-05T12:00:00-05:00 2015-01-05 16:00:00 2015-01-05 17:00:00 2015-01-05 17:00:00 2015-01-05T11:00:00-05:00 2015-01-05T12:00:00-05:00 America/New_York America/New_York datetime 2015-01-05 11:00:00 2015-01-05 12:00:00 America/New_York America/New_York datetime <![CDATA[]]>
<![CDATA[Faculty and Staff Honors Luncheon]]> 27187 Anita Race 1 1406538459 2014-07-28 09:07:39 1475892503 2016-10-08 02:08:23 0 0 event 2015-04-17T13:00:00-04:00 2015-04-17T14:30:00-04:00 2015-04-17T14:30:00-04:00 2015-04-17 17:00:00 2015-04-17 18:30:00 2015-04-17 18:30:00 2015-04-17T13:00:00-04:00 2015-04-17T14:30:00-04:00 America/New_York America/New_York datetime 2015-04-17 01:00:00 2015-04-17 02:30:00 America/New_York America/New_York datetime <![CDATA[]]> <![CDATA[Faculty Candidate Seminar]]> 27187 TITLE: Distributed estimation and inference for sparse regression

SPEAKER:  Yuekai Sun

ABSTRACT:

We address two outstanding challenges in sparse regression: (i) computationally efficient estimation in distributed settings (ii) valid inference for the selected coefficients. The main computational challenge in a distributed setting is harnessing the computational capabilities of all the machines while keeping communication costs low. We devise an approach that requires only a single round of communication among the machines. We show the approach recovers the convergence rate of the (centralized) lasso as long as each machine has access to an adequate number of samples. Turning to the second challenge, we devise an approach to post-selection inference by conditioning on the selected model. In a nutshell, our approach gives inferences with the same frequency interpretation as those given by data/sample splitting, but it is more broadly applicable and more powerful. The validity of our approach also does not depend on the correctness of the selected model; i.e. it gives valid inferences even when the selected model is incorrect.

Joint work with Jason Lee, Qiang Liu, Dennis Sun, Jonathan Taylor


]]> Anita Race 1 1418745648 2014-12-16 16:00:48 1475891368 2016-10-08 01:49:28 0 0 event 2015-01-08T11:00:00-05:00 2015-01-08T12:00:00-05:00 2015-01-08T12:00:00-05:00 2015-01-08 16:00:00 2015-01-08 17:00:00 2015-01-08 17:00:00 2015-01-08T11:00:00-05:00 2015-01-08T12:00:00-05:00 America/New_York America/New_York datetime 2015-01-08 11:00:00 2015-01-08 12:00:00 America/New_York America/New_York datetime <![CDATA[]]>
<![CDATA[Get Your Voice on the Table: Leadership Strategies for Female Faculty]]> 27187 TITLE: Get Your Voice on the Table: Leadership Strategies for Female Faculty

SPEAKER:  Nancy Houfek

Many highly successful women in academia can easily identify what changes need to happen for their organization to succeed, but find it difficult to lead others to confront the challenges facing their group. They often find themselves not heard in the way that they had hoped. The heart of this workshop is to help women use a strategic approach to getting their voices on the table: differentiating types of challenges, understanding factions and what's at stake, using alliances, getting on the balcony, and preparing for meetings in a new way. This workshop also will offer vocal skills designed to help participants be more effective as leaders. The goal is to be heard and understood so that we can successfully influence decision-making in both our professional and our daily lives. Participants will leave with new strategies — theoretical, psychological, and physical — for successfully leading change in their institution or organization.

View Agenda

View Biography

http://www.isye.gatech.edu/apps/RSVP/ADVANCE-event/

 

]]> Anita Race 1 1425481358 2015-03-04 15:02:38 1475891312 2016-10-08 01:48:32 0 0 event 2015-03-30T10:30:00-04:00 2015-03-30T15:30:00-04:00 2015-03-30T15:30:00-04:00 2015-03-30 14:30:00 2015-03-30 19:30:00 2015-03-30 19:30:00 2015-03-30T10:30:00-04:00 2015-03-30T15:30:00-04:00 America/New_York America/New_York datetime 2015-03-30 10:30:00 2015-03-30 03:30:00 America/New_York America/New_York datetime <![CDATA[]]>
<![CDATA[Visiting Speaker Seminar - Guodong (Gordon) Pang]]> 27187 TITLE: Large-Scale Fork-Join Networks with Non-Exchangeable Synchronization

SPEAKER: Dr. Guodong (Gordon) Pang

ABSTRACT:

Fork-join networks consist of a set of service stations that serve job requests simultaneously and sequentially according to pre-designated deterministic precedence constraints. We consider non-exchangeable synchronization (NES), that is, each job can be synchronized only after all of its tasks are completed. Such networks have many applications in manufacturing, telecommunications, patient flow analysis in healthcare and parallel computing. When each station has multiple servers and operates under the FCFS service discipline, the main mathematical challenge to study fork-join networks with NES is the resequencing of tasks’ arrival orders after service completion. We develop a new framework to solve the resequencing problem in the many-server heavy-traffic regimes where the arrival rates and the numbers of servers in each station get large appropriately.

In this talk, we focus on a fundamental fork-join network model with a single class of jobs and NES. Service times of the parallel tasks of each job can be correlated. Upon service completion, each parallel task will join a buffer associated with its service station and wait for synchronization. The goal is to understand the waiting buffer dynamics for synchronization as well as the service dynamics. We show functional central limit theorems for the number of tasks in each waiting buffer for synchronization jointly with the number of tasks in each parallel service station and the number of synchronized jobs, in the many-server asymptotic regimes. All the limiting processes are functionals of two independent processes: the arrival limit process and the generalized multiparameter Kiefer process driven by the service vectors for the parallel tasks of each job. We characterize the transient and stationary distributions of these limiting processes. We also discuss generalizations of the framework to study more complex fork-join networks with NES constraints.

 

 

]]> Anita Race 1 1424956281 2015-02-26 13:11:21 1475891229 2016-10-08 01:47:09 0 0 event 2015-03-03T11:00:00-05:00 2015-03-03T11:00:00-05:00 2015-03-03T11:00:00-05:00 2015-03-03 16:00:00 2015-03-03 16:00:00 2015-03-03 16:00:00 2015-03-03T11:00:00-05:00 2015-03-03T11:00:00-05:00 America/New_York America/New_York datetime 2015-03-03 11:00:00 2015-03-03 11:00:00 America/New_York America/New_York datetime <![CDATA[]]>
<![CDATA[ISyE Seminar - Viswanath Nagarajan]]> 27187 TITLE: Approximation Algorithms for Stochastic Orienteering

SPEAKER: Viswanath Nagarajan

ABSTRACT:

Adaptive optimization deals with optimization problems having stochastic input, where a solution involves multiple stages of decisions, each of which reveals additional information about the realized input. We introduce and study a stochastic version of the orienteering problem in this setting. The deterministic orienteering problem is an extensively studied vehicle routing problem with many applications. Given a set of locations with associated rewards, distance bound B and a designated depot, the objective in orienteering is to find a path of length at most B originating from the depot that maximizes the total reward. In the stochastic orienteering problem, each location also has a random processing time which is realized only when that location is visited. The objective here is to find a non-anticipatory policy to visit locations that maximizes the expected reward subject to the total distance plus processing time being at most B. Due to its adaptive nature, even storing a feasible policy might require exponential space. We focus on a simpler class of “non-adaptive” policies that are just specified by a permutation of the locations, and obtain positive and negative results on how well such policies approximate the optimal adaptive policy. This talk is based on joint works with Nikhil Bansal, Anupam Gupta, Ravishankar Krishnaswamy and R.Ravi. BIO:Viswanath Nagarajan is an Assistant Professor in the Department of Industrial and Operations Engineering, University of Michigan. From 2009-14 he was a Research Staff Member at the IBM T.J. Watson Research Center. He has a Ph.D. in Algorithms, Combinatorics and Optimization from Carnegie Mellon University (2009) and a B.Tech. in Computer Science from IIT Bombay (2003). Dr. Nagarajan's research interests are in combinatorial optimization and approximation algorithms, especially as applied to routing, location and scheduling. He received the Gerald L. Thompson dissertation award at CMU (2009), a best paper award at the European Symposium on Algorithms (2010), and two Outstanding Technical Achievement awards at IBM (2012 and 2014).

 

]]> Anita Race 1 1425389799 2015-03-03 13:36:39 1475891208 2016-10-08 01:46:48 0 0 event 2015-03-10T12:00:00-04:00 2015-03-10T12:00:00-04:00 2015-03-10T12:00:00-04:00 2015-03-10 16:00:00 2015-03-10 16:00:00 2015-03-10 16:00:00 2015-03-10T12:00:00-04:00 2015-03-10T12:00:00-04:00 America/New_York America/New_York datetime 2015-03-10 12:00:00 2015-03-10 12:00:00 America/New_York America/New_York datetime <![CDATA[]]>