Cybersecurity Workforce Optimization

This afternoon we had the great honor and pleasure of hosting INFORMS Fellow Dr. Les Servi of the MITRE Corporation who spoke in the UMass Amherst  INFORMS Speaker Series, which is organized by our magnificent UMass Amherst INFORMS Student Chapter.

Dr. Servi has two degrees in Applied Math from Brown University and a PhD in Engineering from Harvard. He has had a terrific career, having worked at Bell Labs, GTE (now Verizon), MIT Lincoln Labs, and now MITRE.  He is a champion of Operations Research and our professional society INFORMS and he is a good friend. Les is the Group Leader of the Decision Analytics Group at MITRE. (A disclaimer - he hosted me for my talk at MITRE on March 14).

Joining Les were: Brian Kulig, who is a UMass Amherst alum, and also Ken Kato, both of MITRE.

I hosted the lunch in his honor at the University Club, where the service was great and the food very good. We arrived at 11:30AM to make sure that we would have enough time to eat before his talk at the Isenberg School at 2PM. Joining us were Associate Dean of Engineering Tilman Wolf, Professor Eric Sommers of Math/Stats, and Professor Weibo Gong, of ECE, who shared the same advisor at Harvard with Les - Dr. Ho, who is still publishing at age 82.

The Title of Les's talk was: A Two-Stage Stochastic Shift Scheduling Model for Cybersecurity Workforce Optimization with On Call Options.

We took a photo when we arrived at the Isenberg School for his talk.

His presentation was very interesting and clear. He explained both stochastic programming and column generation in a very accessible way. I was very impressed and intrigued by the possibilities of Operations Research and Optimization in workforce optimization in this challenging sector. The model was for scheduling over a two week period three shifts of workers with quite interesting constraints and uncertain demand with the option of calling in more workers. The three classes of cyber workers included an expert group (which someone caught did not want to work the night shift).  He even compared the results of the stochastic model with a deterministic model and presented quite extensive computational results.

I was very pleased that, although it was a holiday weekend, the audience had undergraduate students, PhD students from Engineering, Math, and the Isenberg School, and even MBA students. Two of the latter have internships at MITRE in Bedford this summer, so the talk was perfectly timed.

And, as always happens after a great presentation, the audience members enjoy milling around, sampling refreshments, networking, and chatting.

Afterwards, I was very pleased that Professor Brian Levine of our Computer Science Department, who is the Director of the new Cybersecurity Institute at UMass Amherst, and the PI on the $4.2 million NSF Scholars for Service (SFS) grant, could meet with Les and MITRE colleagues in my office.

Les has spoken in our series multiple times and he always impresses, whether he is speaking on tracking pirates, extracting information and emotions from social media, including Twitter, or on cybersecurity workforce optimization!

In April, he will be heading u to Montana to investigate research on rural healthcare and then will be off right afterwards to the INFORMS Analytics conference in Orlando!

Many thanks to Dr. Les Servi for coming out to UMass Amherst and speaking in our series today. Special thanks also to MITRE colleagues, Brian Kulig and Ken Kato, for joining in and making the visit very rewarding for students and faculty alike!

Supply Chain Networks with Global Outsourcing and Quick-Response Production -- Operations Research and Fast Fashion

Last Sunday's New York Times magazine had a terrific article by Suzy Hansen, "How Zara Grew Into the World's Largest Fashion Retailer,"  which I very much enjoyed reading. I do admit that,  when I live in Gothenburg, Sweden, I enjoy visiting the Zara, H&M, and TopShop stores there since I conduct research on a variety of network systems, including supply chain networks.  Along with my former doctoral student, who is now an Assistant Professor, Min Yu,  we have published two papers on what is known as fast fashion.

The first paper on the topic that we wrote is: Fashion Supply Chain Management Through Cost and Time Minimization from a Network Perspective, Anna Nagurney and Min Yu, in Fashion Supply Chain Management: Industry and Business Analysis (2011), T.M. Choi, Editor, IGI Global, Hershey, PA, pp 1-20.

The second is: Sustainable Fashion Supply Chain Management Under Oligopolistic Competition and Brand Differentiation, Anna Nagurney and Min Yu, International Journal of Production Economics, Special Section on Green Manufacturing and Distribution in the Fashion and Apparel Industries 135: (2012) pp 532-540.

In The New York Times article,  Hansen writes about the company Inditex, which owns the Zara brand and series of stores. She writes:  More than half of Inditex’s manufacturing takes place either in the factories it owns or within proximity to company headquarters, which is to say in Europe or Northern Africa. Inditex owns factories in Spain and outsources production to factories in Portugal, Morocco and Turkey — considered costly labor markets, typically. The rest of its clothes are produced in China, Bangladesh, Vietnam and Brazil, among other countries. The trendiest items are made closest to home, however, so that the production process, from start to finish, takes only two to three weeks. Inditex’s higher labor costs are offset by greater flexibility — no extra inventory lying around — and on faster turnaround speed. 

The article also has some very nice quotes from Professor Fraiman of Columbia University that got me excited because of the language and terminology. Fraiman, in commenting on Inditex's  ideas on expanding in China stated that: “Their factories in La Coruña have a finite capacity to respond quickly. You open more and more stores, and you don’t have flexibility of the last-minute response. Once they have a big thrust in China, then what happens is that they will have to take the whole model” — the processing of customer reactions, the quick-turnaround design teams, the logistics platform — “and replicate it in China.” But the bigger Inditex gets, he says, the more it will lose control over quality and efficiency. 

In a recent paper of ours, Supply Chain Networks with Global Outsourcing and Quick-Response Production Under Demand and Cost Uncertainty, Zugang Liu and Anna Nagurney, which is in press in a special issue of the Annals of Operations Research, we use Zara, as well as toy production, as some of the motivating examples. The paper is also interesting from a methodological perspective since we integrate stochastic programming, game theory, and variational inequality theory, as well as real options from finance.

Specifically, in our paper, which my co-author presented at the INFORMS conference in Phoenix and will also be presenting at the DSI conference in San Francisco next week, we developed a modeling and computational framework for supply chain networks with global outsourcing and quick-response production under demand and cost uncertainty. The  model considers multiple off-shore suppliers, multiple manufacturers, and multiple demand markets. Using variational inequality theory, we were able to formulate the governing equilibrium conditions of the competing decision-makers (the manufacturers) who are faced with two-stage stochastic programming problems but who also have to cooperate with the other decision makers (the off-shore suppliers). Our theoretical and analytical results shed light on the value of outsourcing from novel real option perspectives. In addition, our simulation studies reveal important managerial insights regarding how demand and cost uncertainty affects the profits, the risks, as well as the global outsourcing and quick-production decisions of supply chain firms under competition.

 The paper will appear in a special issue dedicated to the memory of Professor Cyrus Derman, who passed away last April at age 85. According to Columbia University, Professor Emeritus Cyrus Derman, was considered the driving force behind the success of Columbia Engineering's Department of Industrial Engineering and Operations Research (IEOR). He had also served as the doctoral dissertation advisor of well-known colleagues in Operations Research, including Professor Michael Katehakis, one of the co-editors of the special memorial volume, Peter Kolesar, and Art Veinott, Jr.

The theme of the special issue is Optimization under Uncertainty Costs, Risks and Revenues.

Supply Chain Networks, Global Outsourcing, and Quick Response Production -- Integrating Methodologies for Theory and Practice

During the SAMSI Workshop on Engineering and Renewable Energy, which was one of the events organized under this year's theme of Uncertainty Quantification, I was asked, after my presentation on Sustainability, as to the uncertainty underlying supply chains.

In response, I noted both cost and demand uncertainty, issues that we have been researching in applications from the design of supply chain networks for critical needs products to assessing the performance of supply chains under cost and demand disruptions. Of course, in global supply chains, exchange rate risk is also a major issue that we have investigated.

I had also mentioned, in a discussion with a workshop participant, that another one of our papers on supply chain uncertainty was under review for a journal and, yesterday, we received the official acceptance.

Our paper, Supply Chain Networks with Global Outsourcing and Quick-Response Production Under Demand and Cost Uncertainty, by Zugang Liu and Anna Nagurney, has been accepted for publication in the Annals of Operations Research. It will appear in a special issue dedicated to the memory of Professor Cyrus Derman, who passed away last April at age 85. According to Columbia University, Professor Emeritus Cyrus Derman, was considered the driving force behind the success of Columbia Engineering's Department of Industrial Engineering and Operations Research (IEOR). He had also served as the doctoral dissertation advisor of well-known colleagues in Operations Research, including Professor Michael Katehakis, one of the co-editors of the special memorial volume, Peter Kolesar, and Art Veinott, Jr.

The theme of the special issue is Optimization under Uncertainty Costs, Risks and Revenues.

The issue was originally formulated to celebrate Professor Derman's 86th birthday and I had been told that he was very excited about it but then passed away -- hence, the issue is now a memorial one.

I am very, if I may say, "fond" of the Annals of Operations Research, which was edited for many years by Professor Peter Hammer, who died tragically in a car accident in December 2006. I had last seen him at the European Conference on Operational Research in Iceland in 2006 and enjoyed my conversation with him and his wife very much. This journal is now edited by Professor Endre Boros, who actually informed many of us of the passing of Professor Hammer. I had edited a special issue of the Annals back in 1993, entitled, Advances in Equilibrium Modeling, Analysis and Computation, vol. 44, and have also published in the journal since.

In our paper, Supply Chain Networks with Global Outsourcing and Quick-Response Production Under Demand and Cost Uncertainty, we integrate the methodologies of stochastic programming and variational inequalities in order to construct a modeling and computational framework for supply chain networks with
global outsourcing and quick-response production under demand and cost uncertainty. Our model handles multiple off-shore suppliers, multiple manufacturers, and multiple demand markets.

Using variational inequality theory, we formulated the governing equilibrium conditions of the competing manufacturers, who are faced with two-stage stochastic programming problems but who also have to cooperate with the off-shore suppliers. Our theoretical and analytical results shed light on the value of outsourcing from novel real option perspectives as in finance. In addition, our simulation studies revealed important managerial insights regarding how demand and cost uncertainty affects the profits, the risks, as well as the global outsourcing and quick-production decisions of supply chain firms under competition.

The modeling and computational framework is relevant to many industries in which quick response production and enhanced supply chain flexibility and responsiveness may provide a competitive advantage.

We dedicated the paper to the memory of Professor Cyrus Derman.

A Terrific Operations Research Lecture on Multistage Optimization

Today we were treated to a brilliant lecture by Professor Dimitris Bertsimas of the Sloan School at MIT, who spoke on Advances in Multistage Optimization in our Spring 2010 Speaker Series in Operations Research / Management Science at the Isenberg School at UMass Amherst. Professor Bertsimas is my "academic brother-in-law" since he is married to Professor Georgia Perakis, also a Professor at the Sloan School, whose doctoral dissertation advisor at Brown University was my dissertation advisor, Professor Stella Dafermos. Both Georgia and I have PhDs in Applied Math from Brown (with concentrations in Operations Research (O.R.)), so we are "academic sisters."

Professor Bertsimas began his lecture by recalling his conversations with Professor George Dantzig (now deceased), of Stanford University, when he visited Stanford 16 years ago. He said how much Dantzig (who is considered the Father of our field) had valued and had worked on planning under uncertainty. Above I have a photo from Bertsimas' lecture today that displays a photo of Dantzig. Professor Bertsimas discussed the challenges of stochastic programming (and probability theory) from a practical, computational perspective. He then moved on to adaptive optimization and discussed the advantages of robust optimization and some bounds that he had obtained which were not only powerful, but also surprising. He discussed applications to electric power generation, finance, and inventory management. Robust optimization "takes a deterministic view to protect yourself." The problem class in robust optimization that he focused on in his presentation is tractable and practically solvable.

The audience included faculty from Management Science, Finance, Computer Science, Economics, as well as students from many departments and guests who came even from Boston. We had a splendid lunch afterwards at the University Club with conversations that related wonderful stories about academic adventures, travels, research, and various universities (always fun to compare notes on such topics).

It was an honor to be able to host Professor Bertsimas in our Speaker Series!