San Francisco 2014
 
Blog:
blog home Home
 
       Follow:
RSS
 
Nov 11 14

Morse Lecture

by Harrison Schramm

Just came from the Phillip McCord Morse lecture. Dimitri Bertsimas showed us a differnent way of thinking about statistics — though an optimization lens. Status: Mind. Blown.

Quote: “The advances in computing power over the past 20 years, resulting in a 200 Billion fold increase in efficiency – should lead us to rethink our definition of ‘tracitble'”

Nov 11 14

Keeping a horse in the race, 2014 edition

by Marc-Andre Carle
This year’s conference is a huge success. More than 5000 participants, so many events, and so many talks! A huge proportion of those talks propose new or improved solution methods for mixed-integer programs. And the most popular way to demonstrate the interest of these methods is to benchmark against a generic solver. Last year, I heard John Siirola called this keeping a horse in the race. I very much liked the expression.

 

Over the last year, MIP solvers have improved. Often to the point that many researchers are close to being beaten by the horse they put in the race. For example, I attended some talks in the stochastic programming track. The authors have implemented a state-of-the-art decomposition algorithm (including acceleration techniques) for a multistage stochastic programming problem. Yet, the MIP solver they compare against still beats them for about 20-30% of instances. Solvers are not only catching up with applications, in some fields they are pretty much catching up with tailored algorithms from academia.

 

There are many possible reasons for this. First is that many ideas that get into solvers these days do not come from research papers anymore. Availability of multi-core processors is also a factor, and I assume the typical code from academia doesn’t match the level of sophistication that is found in code-optimized commercial solvers. They also have access to a huge model base to test their new ideas which is considerably richer than the typical 5 to 30 instances found in a research paper. With all the new features presented at this Conference, I imagine next year’s horses will be even more challenging to beat!
Nov 11 14

INFORMS: Perspectives – Banafsheh Behzad on Women in OR/MS

by David Morrison

It’s the elephant in the room that people don’t really like to talk about: the lack of diversity in STEM fields. It’s an important issue, though, because a group of people with different backgrounds and perspectives will always be stronger than a group of people who all have similar stories to tell. So for the past year, I’ve been a student liason with WORMS — the Women in Operations Research and Management Science forum at INFORMS. I believe these are important issues that need to be discussed openly, not only by our female colleagues, but by everyone. I’m not a student any more, so I can’t keep being a student liason, but I’m planning to continue my involvement with WORMS in the coming years. And today, for my entry in INFORMS: Perspectives, I interviewed one of my friends and colleagues, and a prominent voice in WORMS, Banafsheh Behzad.


banafsheh

Banafsheh is an Assistant Professor at the California State University at Long Beach, and in addition to her work with WORMS, she’s had a strong research career. She’s interested in applying game-theoretic techniques to real-world problems, particularly in areas of public health, safety, policy, and social networks. Her dissertation work was at the University of Illinois, Urbana-Champaign with Sheldon Jacobson (another INFORMS fellow!), where she performed a game-theoretic analysis of the public sector pediatric vaccine pricing market in the United States. The results that Banafsheh found are quite interesting: it turns out that pediatric vaccines are sold in the United States at higher prices than what the equilibrium prices from her models suggest. Banafsheh has some connections with researchers in the Centers for Disease Control and Prevention (CDC), and her hope is that her research can be used by the CDC to negotiate better vaccine prices in the public sector, and thereby increase the availability of pediatric vaccines in the United States to improve public health.

Banafsheh has quite a few other research projects that she’s worked on in addition to her dissertation work; her graduate advisor refers to these side projects as “bedtime research”—the problems that you’re thinking about as you fall asleep. Banafsheh’s bedtime research has revolved around the obesity problem in the United States; in particular, she’s looked at correlations between obesity and seat belt usage, and demonstrated correlations between obesity and seat belt usage in states with primary and secondary seat belt laws—that is, states where failing to wear a seat belt is a ticketable offense by itself, and states where failure to wear a seat belt can only be penalized in conjunction with another traffic violation.

Banafsheh told me that these research projects have been quite rewarding because of the opportunities that she’s had to directly impact and improve public health and safety. The challenges she’s faced? Mainly they revolve around getting data. She didn’t always have access to the data she needed, and had to make approximations, perform sensitivity analysis, and constantly search for better data sources. But she says it’s important to not let the lack of data inhibit the research progress. “Find a problem that you’re interested in, and try to find the best methodology for that problem. Don’t limit yourself to only the methodologies that you’re familiar with,” she says.

I then turned the conversation to her involvement with WORMS, a topic about which she is equally passionate. Banafsheh has been attending INFORMS for five years, and went to the WORMS business meeting at her very first INFORMS. At that time, she was the only student at the business meeting, and very few people knew what WORMS was or why it existed. She wasn’t even sure if students were allowed to participate in WORMS! Of course, they can, and it’s partly through her efforts that WORMS now has much more visibility and much more student involvement. Over the past year, she’s been a student liason for WORMS, helping to raise awareness of the forum and promote discussion about diversity in STEM. But there’s still much work to be done.

I asked Banafsheh about the most significant barriers to success for women and minorities in STEM fields, and she told me this: “As an undergraduate student, if all of your professors are male, it is hard to imagine that you can succeed in this field.” Banafsheh told me a story of a student of hers who had never considered a career in STEM, because all of the professors in the area were male. But after taking Banafsheh’s class, her student realized that gender shouldn’t be a barrier to success in STEM, and has been talking with Banafsheh about pursuing an advanced degree.

Stories like this are encouraging, but we need more people telling them. “It is the responsibility of all of us (not just women, but also male faculty) to encourage and support women in OR/MS,” Banafsheh says. “If you have more female colleagues, more female students, it will be a healthier environment in your class, in your department. It doesn’t make sense to say that men are more capable—it’s just not true!” she concluded.

WORMS is one way to get more people telling these stories, and despite its increased visibility over the last five years, Banafsheh still meets women faculty and students at INFORMS who have no idea what WORMS is and why it exists. So she wants to encourage people to get the word out: come to the WORMS luncheon at INFORMS (sold out this year with 190 people!), come to the WORMS business meeting, find out what issues women and minorities in our field face, and take responsibility for correcting them. And particularly if you’re a student and want to get involved, there’s no better time! WORMS is looking for more student liasons to help with the forum, so if you’re interested in being a student liason next year, let Banafsheh know!

Nov 11 14

COIN-OR at INFORMS 2014

by Matthew Saltzman

OK I guess I’d better blog or they might take away my ribbon, which would threaten my clear frontrunner status in the max-cardinality badge ribbon competition.

2014-11-10 06.27.39

Of course, the biggest COIN-OR news here is that, on Sunday, a group of us were privileged to accept the Impact Prize on behalf of the COIN-OR community.  We are honored to be so recognized, but we in turn recognize that the impact comes not only from our efforts but from those of contributors, adopters, and believers throughout the OR field, and we thank all of you for your support.  This is a huge next step in our campaign for world domination!

On Monday, the tenth COIN-OR Cup (the most coveted award in computational OR) was awarded to Mehdi Towhidi and Dominique Orban for their work on CyLP, a flexible Python interface to COIN-OR’s linear and mixed-integer programming solvers (CLP, CBC, and CGL).  Unfortunately, the winners were unable to join in the festivities, but they will receive their prize (as usual, a tacky souvenir cup from the conference location filled with COIN-OR chocolate coin swag) in the mail.

In addition, the first ever COIN-OR Distinguished Contribution Award.was presented to Robin Lougee, for all she has done in support of the project over the years.

The reception was held at the Bluestem Brasserie. It was well attended, the food was excellent, and it appeared that a good time was had by all.  Professional optimizers couldn’t have planned better, as we hit the minimum bill amount just as the last couple of tables were finishing up.  Unfortunately, I’m still a bit of a social media tyro, so there are no pictures and we didn’t pull off the human pyramid.  (If anyone does have pictures, please let me know.)

The COIN-OR exhibit booth is still open and we still have enough chocolate coins to implement  QE4, so drop by and nosh.

Nov 11 14

Three questions to answer when the conference is over

by Walt DeGrange

It is only Tuesday but folks are starting to depart. By Wednesday evening almost everyone will be in some mode of transportation back to their place of origination. Here are there questions to answer on the plane, train, or automobile on the way home.

1) What just happened?

Do the past few days seem like a blur? Good, now start freeform writing about your experiences. You could start when you arrived and go forward or go in reverse (my personal favorite). Just start writing about presentations, conversations and observations. This is not a trip report, just a vehicle to trigger memories.

2) Who did you meet?

Meeting new people is fun and enriching. Make a simple list of folks by going through business cards, downloading the list from the INFORMS conference app or kick it old school, from your memory.

3) What do I do now?

Did you promise to send your presentation to someone? Perhaps mentioned that you would call to catchup with an old colleague? Do you want to become more involved in a section, society or other INFORMS group? Time to schedule actions in the near future. You don’t want to find yourself at the 2015 Philadelphia conference wondering how a year passed so quickly without taking action.

Answering these three simple questions can greatly enhance your investment of time and resources to attend the INFORMS Annual Meeting.

Nov 11 14

Remembering the Dantzig Century and a Conversation with His Daughter

by Anna Nagurney

What a thrill it was to listen to Professor Richard “Dick” Cottle of Stanford deliver his plenary talk “Remembering the Dantzig Century” to a standing room only overflow crowd. In a voice that was clear, steady, and filled with clear great respect for his adviser and mentor, Professor Cottle took us on a journey of Dantzig’s life and contributions to our field. Through a historical timeline of milestones and achievements illustrated with photos we all felt as though George was back with us and truly he has never left us.

From the fact as to how he received his PhD with Neyman as his supervisor, which I relate to every OR/MS class that I teach, to his work at the Pentagon and Rand, to is work at Berkeley and Stanford we were mesmerized. Dantzig emphasized the following order: a real-world problem, then the mathematical modeling, and then the theory and algorithms.

A lady came during the talk and stood next to me and there was a spirit about her that I immediately noticed. When Cottle showed a photo of Dantzig’s office after the earthquake in Loma Linda, she proceeded to fill me in on more facts, such as that George had been working in his office when it struck. I left the plenary realizing that I wanted to have a conversation with this interesting female.

And, as for serendipity, which was the theme of my previous blogpost, it happened again! I spoke with the INFORMS Executive Director Melissa Moore about what a great plenary it had been and what great memories I had of George from so many conferences at which I enjoyed his kindness and warmth, including the first conference I ever spoke at as a fresh PhD at which he came to my presentation (such a memory one never forgets). She presented me with a request and an offer to meet with Dantzig’s daughter, which, of course, I accepted.

When his daughter walked in to our meeting area yesterday afternoon, it was precisely the lady that stood next to me during Cottle’s plenary! We had a conversation for almost an hour (I had to chair a session but could have continued to talk for days). I now have the business card of Dantzig’s daughter, Jessica Dantzig Klass, through which his generosity, energy, and wisdom live on!

Nov 11 14

Captain’s Blog

by Shiva

Captain’s blog star-date 2014-11-10 at planet #informs2014. Significant levels of #orms activity, previously unknown to mankind, was detected at the upper levels of the planet. Upon investigation, an entire new floor full of missing ORMS sessions was discovered. Local search is useful, but global search can be helpful too. Attended a Daniel Wagner Practice Prize candidate presentation and learned a lot.

Nov 11 14

Chapters/Fora breakfast

by Burcu B. Keskin

Great leadership and great awards! Congratulations to Laura McLay for earning Moving Spirit Award with her involvement in Women in Operations Research and Management Science. image

We also had two great student leaders who got the Judith Liebman Award:  Michelle McGaha Alvarado and Kimia Ghobadi, and Ruixie Guo.

Other student leaders are exchanging ideas among each other to understand how to reach a wider audience (undergrads, industry), how to improve the financial stability of student chapters, and how to increase activity level. Great discussion led by Informs VP-Chapters/Fora David Hunt!! If you have questions/concerns or answers for these issues, contact David or Courtney Biefeld (Courtney.Biefeld@informs.org).

 

 

 

Nov 11 14

My Way of Annual Meeting

by Sertalp B Çay

I need to say this first: This annual meeting is way crazier than last year’s. It is very likely to see someone you know in 15 minutes if you are traveling in the wild corridors of Hilton.

Planning my schedule among 83 parallel tracks is an optimization problem itself. Someone should write a paper about it! But I can say that meeting is going great so far. Although I couldn’t make it to many talks as I had originally planned, I benefit from the annual meeting in many different ways.

Career Fair on Sunday was a great way to meet with people to talk about internship opportunities. I talked with representatives from Google, Amazon, FedEx, Apple and SAS. Wow, that was cool. I am really impressed by their attitude towards PhD students. The fair was more beneficial for soon-to-be graduates, but I am happy that I was able to learn more about career opportunities overall.

I had chance to attend Optimization Society Business Meeting and IOL & Social Networking Reception. I wish such events will be in different time slots in the following annual meetings. Both were great. Meeting finally with Thiago Serra and Marco Lübbecke, seeing Marc-Andre (@ORNinja), Paul Rubin, Phil Kim and Mary again made my day. We had a long discussion with Marc-Andre and Thiago about pronouncing names. As a person who is used to hear his name to be pronounced incorrectly, I quite enjoyed the conversation.

To be honest, I am still having difficulties to plan my remaining days, but preparing my presentation is taking all my time and energy. I am quite nervous about my upcoming presentation and would be happy to hear your suggestions!

Nov 11 14

Recognizing the impact of operations research

by Polly Mitchell-Guthrie

Today I had a dilemma due to scheduling conflicts – I could go watch my boss Radhika Kulkarni be officially welcomed as an INFORMS Fellow or watch my colleague Ivan Oliveira give his Wagner talk. I figured Radhika had already won but Ivan still needed my support, so I went with the latter.

Jim Wilson (NC State University), Radhika Kulkarni (SAS), and Vidyadhar Kulkarni (University of North Carolina)

Jim Wilson (NC State University), Radhika Kulkarni (SAS), and Vidyadhar Kulkarni (University of North Carolina)

Fortunately, colleague Brad Klenz sent a photo of he took of Radhika at the luncheon, posing with her nominator Jim Wilson and husband Vidyadhar Kulkarni. The 2014 INFORMS Fellows are an impressive bunch. Other 2014 winners include Terry Harrison, with whom I’ve had the pleasure of working on the Analytics Certification Board, and I’ve been quite impressed with Russ Labe through various practice activities and hearing him speak multiple times. I was particularly glad to see Radhika’s hard work recognized. She is such a passionate believer in operations research that she has been an incredible advocate for it internally at my company SAS and externally to the companies we work with. Radhika has grown our investment in optimization and simulation and extended its impact to more organizations through the development of more full-featured optimization solutions that solve complex problems like marketing optimization or revenue management and price optimization. As INFORMS member and Gartner analyst Lisa Kart reported at the 2013 INFORMS Executive Forum, in one survey they did only 3% were using prescriptive analytics (a term some have adopted for optimization), vs. 70% reporting usage of descriptive analytics and 16% using predictive analytics. Clearly a lot of work needs to be done to extend the huge potential impact of operations research to more places.

Ivan Oliveira (SAS)

Ivan Oliveira (SAS)

Speaking of the impact of optimization, the session I did attend instead was a one where two of the six finalists for the Daniel H. Wagner Prize for Excellence in Operations Research Practice presented. This prize emphasizes good writing, strong analytical content, and verifiable practice. I went to hear my colleague Ivan Oliveira present, along with Kevin Norwood of Procter and Gamble, on “Statistical and Optimization Techniques for Laundry Portfolio Optimization at P&G.” They represented a broader contingent from each company who had worked together on this problem. P&G makes well-known laundry detergents like Tide, Gain, Share and more, and the impact of this project was about reducing the production costs involved in making the detergents within certain constraints. Each detergent is comprised of a number of raw ingredients that can be combined in a variety of ways, as long as the outcomes consumers desire remain consistent. When you wash your clothes, you care about what they call SRI, or stain removal index, wanting to make sure that it gets your clothes clean. And blood stains differently than grass, which is different than wine, etc., all of which produce different demands on the raw ingredients to work effectively. The many constraints led to a highly convex mixed integer linear programming problem that the team had not seen addressed in the literature. Their ultimate solution saved P&G $5-20 million (depending on the combinations), helped them balance cost with the complexity of the ingredient mix, and allowed them to evaluate more combinations due to the speed of the solution.

The other talk in this session was “Gerrymandering for Justice: Redistricting U.S. Liver Allocation,” given by Sommer Gentry from the US Naval Academy and Dorry Segev from the Johns Hopkins School of Medicine. The focus was on reducing the geographic disparity related to liver transplants, since your likelihood of getting a liver varies hugely depending on where you live. Their model, if implemented, is projected to save lives as well as money. We can all hope that the transplant authorities responsible for approval put it into action, since it would certainly increase the public immeasurably.

It is exciting to see the impact of operations research on so many projects and lives in so many diverse ways. But behind this impact are individuals who make incredible contributions to the profession and to society through their hard work. I feel fortunate to work with so many hardworking, intelligent people at SAS and through INFORMS and to have seen their work recognized today.