Search 
 
 
 


Past Operations and Logistics Division and Centre for Operations Excellence Seminars

Past OPLOG Division Seminars September 2017 - August 2018
Past OPLOG Division Seminars - September 2016 - July 2017
Past OPLOG Division and COE Seminars - September 2015 - August 2016

Past OPLOG Seminars January 2015 -  August 2015
Past OPLOG Seminars - August 2014 - December 2014
Past OPLOG Seminars September 2013 - May 2014

Past OPLOG Seminars - September 2012 - May 2013

OPLOG Division Seminars September 2018 - Onwards


Date: Monday, October 22nd, 2018

Speaker: Henry Wolkowicz, University of Waterloo
Title:  "Facial Reduction in Cone Optimization with Applications to Matrix Completions"
Time: 2:30 PM - 3:30 PM
Place: HA 968

Abstract
: Strict feasibility is at the heart of convex optimization. This is needed for optimality conditions, stability, and algorithmic development. New optimization modelling techniques and convex relaxations for hard nonconvex problems have shown that the loss of strict feasibility is a much more pronounced phenomenon than previously realized. These new developments suggest a reappraisal. We describe the various reasons for the loss of strict feasibility, whether due to poor modelling choices or (more interestingly) rich underlying structure, and describe ways to cope with it and, in particular, "take advantage of it".


 

Date: Monday, October 29th, 2018

Speaker: Jussi Keppo, National University of Singapore Business School
Title: "Investment Decisions and Falling Cost of Data Analytics"
Time: 2:30-3:30 PM
Place: Henry Angus 968

Abstract: We model a risk-averse decision maker who optimizes the size of an investment, leverage used in the investment, and the level of information on the investment through costly data analytics. We show that borrowing-constrained or highly risk-averse investors have low demand for data analytics. We also show that the demand of data analytics is highest for investment opportunities with high expected returns, and the demand of data analytics is lumpy for opportunities with low expected returns even without analytics fixed cost. Furthermore, the falling cost of data analytics raises investors' leverage, which leads to higher losses during crises.


Date: Friday, November 9th, 2018

Speaker: Miguel Anjos, Polytechnique Montreal
Title:  "Mathematical optimization approaches for facility layout problems: The state-of-the-art and future research directions"
Time: Noon - 1:00 PM
Place: Henry Angus 333

Abstract
: Facility layout problems are an important class of operations research problems that has been studied for several decades. Most variants of facility layout are NP-hard, therefore global optimal solutions are difficult or impossible to compute in reasonable time. Mathematical optimization approaches that guarantee global optimality of solutions or tight bounds on the global optimal value have nevertheless been successfully applied to several variants of facility layout. This review covers three classes of layout problems, namely row layout, unequal-areas layout, and multifloor layout.  We summarize the main contributions to the area made using mathematical optimization, mostly mixed integer linear optimization and conic optimization. For each class of problems, we also briefly discuss directions that remain open for future research.


Date:  Monday, November 19th, 2018

Speaker: Martin Puterman, UBC Sauder School of Business
Title: "Points Gained: a New Metric for Assessing National Football League (NFL) Team Performance"
Time: 2:30 PM - 3:30 PM
Place: HA 968

 Abstract: In this talk I describe our revisit to Carter and Machol’s 1971 paper “Operations Research on Football” with new data, a theoretical foundation and new ways to use the results. Our research uses rather obscure Markov Decision Process theory to develop a rigorous framework for deriving league value functions. It shows that the obvious value function equation has a unique solution that equals the bias of the underlying Markov reward process thus providing a precise interpretation for the value function.  Our research proceeds by extending Brodie’s recent work on “shots gained” in golf, by showing how to use the Bellman equation to derive a “points gained” metric for football.  Data analysis from the 2013-2016 NFL seasons shows how the points gained metric identifies and provides new insights into specific factors that distinguish team performance.

This talk is based on joint work with Tim Chan and Craig Fernandes of the University of Toronto.  


Date: Monday, December 3rd, 2018

Speaker: Karla Hoffman, George Mason University
Title:  "The use of hybrid optimization algorithms for problems that are either unsolvable by state-of-the-art software packages or for problems where such codes are too slow for the real-time application"
Time: 2:30 PM - 3:30 PM
Place: Henry Angus 968

Abstract
:  Optimization algorithms have increasingly played a role in improving the operations of many standard corporate activities such as the supply chain, real-time scheduling and routing, and determining the allocation and pricing of goods and services through auctions. In this talk, we examine problems that arose in a high-profile government auction where the Federal Communications Commission (FCC) bought back spectrum from TV stations, packed the remaining broadcasters into a smaller swath of spectrum and sold the acquired spectrum to the wireless industry. Optimization is used before, during and after this auction to assure that multiple governmental goals are met.

Our second example considers how the military can use similar optimization strategies to allocate limited spectrum during combat situations when spectrum is scarce and communication vital. In both instances, the problems require the solution to extremely difficult optimization problems.  In each case, our approach is to use a combination of heuristics, decompositions, and constraint programming to create an overall algorithm that is capable of solving to global or near global optimality problems with millions of variables and hundreds of thousands of constraints.

We finally present an example of a real-time routing and scheduling problem where one needs near-optimal solutions in less than a second.  In this case, we ask the question:  Under what conditions are the following algorithms the best choice:  enumeration, constraint programming, heuristics, decision diagrams or global optimization techniques best?

In all applications, we use realistic data sets and make the data sets available to the research community.




Date:  Tuesday, June 19th, 2018 Speaker: David Stanford, The University of Western Ontario Title:  "An updated approach to Emergency Department prioritization in light of empirical data" Time: Noon - 1:00 PM Place: Henry Angus 968
Date:  Tuesday, June 19th, 2018 Speaker: David Stanford, The University of Western Ontario Title:  "An updated approach to Emergency Department prioritization in light of empirical data" Time: Noon - 1:00 PM Place: Henry Angus 968
Date:  Tuesday, June 19th, 2018 Speaker: David Stanford, The University of Western Ontario Title:  "An updated approach to Emergency Department prioritization in light of empirical data" Time: Noon - 1:00 PM Place: Henry Angus 968
Date:  Tuesday, June 19th, 2018 Speaker: David Stanford, The University of Western Ontario Title:  "An updated approach to Emergency Department prioritization in light of empirical data" Time: Noon - 1:00 PM Place: Henry Angus 968