ORC IAP Seminar 2020

1/27/20 | 9:30am-3:30pm | E51-345

Operations Research: From Theory to Practice 

If you have any questions, please contact us via email: orc_iapcoordinators@mit.edu.

Date: Monday, January 27, 2020
Time: 9:30am-3:30pm 





Vivek Farias

Robert N. Noyce Career Development Professor, MIT 

The Omni-Channel Fulfillment Problem: From Concept to Practice
The prevalence of omni-channel purchase behavior motivates a slew of new supply chain problems. Here, we describe one such new problem: the 'omni-channel fulfillment problem'. We formulate this problem as an online optimization problem. We propose an algorithm for this problem based on the primal-dual schema. We provide a performance analysis and an upper bound on achievable competitive ratios establishing that our algorithm is optimal in the face of adversarial demand. Our algorithm has been implemented at multiple large retailers. We describe one such large-scale implementation. This implementation processes as many as hundreds of thousand of orders on peak demand days. We discuss the savings achieved through optimal order-fulfillment decisions that simultaneously increase turn and lower shipping costs for this implementation.

Vivek is interested in the development of new methodologies and applications for large scale dynamic optimization. He received his Ph.D. in Electrical Engineering from Stanford University in 2007 and is the Patrick J. McGovern (1959) Professor at MIT. Vivek is a recipient of an INFORMS MSOM Student Paper Prize (2006), an INFORMS JFIG paper prize (2009, 2011), the NSF CAREER award (2011),  MIT Sloan’s Outstanding Teacher award (2013), the INFORMS Simulation Society Best Publication Award (2014), the INFORMS Pricing and Revenue Management Best Publication Award (2015), and the INFORMS MSOM Best Publication award in Management Science (2016). His practice based work has been judged a finalist for the Pierskalla Award (2011), the Gary L. Lilien ISMS-MSI Practice Prize (2016) and the Wagner Prize (2018). Outside of academia, Vivek was most recently co-founder/CTO at Celect (acquired by Nike), and serves on several technology startup advisory boards.



Emil Iantchev

Head of Portfolio Construction Research, Fidelity Investments

Prospect Theory and Robust Life-Cycle Investing

This talk will examine the problem of investing over the life-cycle in order to provide funds in retirement. A solution method is presented which uses robust control to design an investment strategy that accounts for the salient characteristics of investor preferences as postulated by Kahneman & Tversky’s prospect theory, namely loss aversion and overweighting of tail events.


Emil Iantchev is the Head of Portfolio Construction Research for Fidelity’s Global Asset Allocation division (GAA) and Strategic Advisers, Inc. (SAI). With over $460 billion in assets under management, Fidelity’s Global Asset Allocation division (GAA) is a leading provider of asset allocation solutions for retail and institutional clients. Strategic Advisers, Inc. (SAI) is a fast growing fund of funds managed account business currently with $250 billion in assets under management. In his current role, Emil specializes in designing systematic investment processes for multi-asset class mandates including strategic asset allocation over the life-cycle as well as dynamic asset allocation over shorter horizons. Prior to joining Fidelity in 2013, Emil spent time in academia teaching at Boston University and Syracuse University. Emil holds a PhD in Economics from The University of Chicago and a BA in Economics and Mathematics from Eckerd College in St. Petersburg, FL.



LUNCH BREAK (not provided)


Ross Anderson

Software Engineer, Google 

Optimizing Over Trained Neural Networks
 We show how to solve optimization problems containing trained neural networks in the objective or constraints. These problems arise (1) in the “predict and optimize” framework, (2) when verifying the robustness of a neural network model, and (3) as a subroutine in sequential optimization methods, e.g., approximate dynamic programming and Bayesian optimization. To solve these problems, we develop strong mixed integer programming (MIP) formulations, good primal heuristics, and efficient bound tightening methods specifically for these problems. We demonstrate our method by verifying robustness of an image classifier, control of a robotics simulator, a DNA sequence design problem, and in ongoing work on both a vehicle routing problem and neural architecture search.

Ross Anderson is a software engineer in Google's Operations Research team in Cambridge MA. His recent work has been on problems in the intersection of optimization and machine learning. He obtained his PhD from the MIT Operations Research Center in 2014.



Martin Copenhaver

Operations Researcher, Massachusetts General Hospital

Lessons and Opportunities for Operations Research in Hospitals

In this talk, I will focus on two examples of operations research projects that were implemented in hospitals. Throughout I will discuss the full life cycle of an operations research project, from initial methodological work to final implementation and monitoring. In doing so, I will highlight some of the challenges of translating operations research into practice as well as the significant opportunities, both methodological and applied in nature, for ongoing research in this domain.

Martin Copenhaver is a research scientist at Massachusetts General Hospital and a lecturer in operations research and statistics at MIT Sloan. His research interests lie in the design and implementation of new health care systems with the use of tools from optimization and statistics. Martin received his Ph.D. from MIT's Operations Research Center where he was advised by Dimitris Bertsimas.

A PDF of the schedule can be found here.

Event Time: 

2020 - 08:30