Use OR for your NCAA Picks

Joel Sokol and his LRMC method have made their picks for this year’s NCAA College Basketball Tournament. Joel’s method is based on work he did with Paul Kvam on a “Logistic Regression/Markov Chain Model for NCAA Basketball” published in Naval Research Logistics. This approach agrees in part with the NCAA selectors in that the predicted final four are the four number one seeds. It disagrees strongly with some of the rankings though. It would not have put UNC as a number one seed at all, let alone the top number one seed, preferring Duke over UNC. The number six seeds, who the NCAA ranks, presumably, 24-27 or so, are ranked 16 (Marquette), 24, 26, and 34 (Oklahoma). My city’s Pitt Panthers, an NCAA 4, seed are only ranked 23rd by Sokol’s method. Perhaps the biggest difference is 11th seed Kansas State, who is 19th (4th seed) under LRMC.  Kansas State over USC is also the big first round upset.

More do OR and make money!

Francisco Marco-Serrano, author of the fmwaves blog pointed out another contest to determine a recommender system.  Unlike the Netfix contest, which might not end up with a winner, the MyStrands contest appears to guarantee $100,000, and all that is needed is an idea.  Again, this seems like a good opportunity for those in OR to use our approaches to attack these sorts of problems (it seems that CS “owns” recommender systems, for no obvious reason).

Hmmmm…. blog recommendations?  systems that allow faculty to rate students?  sabbatical destination recommendations?   What do I need recommended, and how can OR help?

Poker and Operations Research

I just attended a fantastic talk by Michael Bowling of the University of Alberta entitled “AI After Dark: Computers Playing Poker” where he described work by the U of A Computer Poker Research Group. Like most artificial intelligence research (particularly in games), there is no veneer of trying to mimic the human brain in this work: it is all about algorithms to compute optimal strategies, which puts it solidly within operations research by my view.

Much of his talk revolved around a recent “Man-Machine Poker Championship” where a program they have developed played against two professional poker players: Phil “The Unabomber” Laak and Ali Eslami. Laak is known from TV appearances; I haven’t seen Eslami, but he has a computer science background and understands the work the U of A researchers do, so that might make him an even more formidable opponent. The results were, at one level, inconclusive. The humans (playing in “duplicate” poker) won two of the four 500 deal matches, lost one, and one was tied. The overall amount of money “lost” by the computer was very small. I have no doubt that most humans facing professionals would have lost a lot more. So having a competitive program is already a big step. Like most who lose at poker, Michael claimed “bad luck” but he has the technical skills to correct for the luck, and was pretty convincing that the computer outplayed the humans.

One interesting aspect is that the program does no “opponent analysis” yet, though that is an extremely active research area (75% of the U of A’s efforts, Michael said). Given a couple more years, I am pretty confident that these programs will start giving the pros a run for their money. Michael said that one goal of their work could be stated: they want to make Phil Helmuth cry. That seems a little less likely.

On the technical side, the presentation concentrated on some new ways for systems to learn to solve huge (1000000000000 state) extensive form games. They have a neat system for having systems learn by playing against themselves. It takes a month of serious computation to tune the poker player, but the method may have other applications in economics. Check out Michael’s publications for more information.

Definitely one of the best talks I have heard in a long time!

Get your registration in for the INFORMS Practice Meeting

I really like the INFORMS Practice Meeting.  It is much different than the regular INFORMS conference.  The key difference is that not everyone speaks.  At regular INFORMS (or EURO or IFORS), practically everyone there will give a 20-25 minute talk on their own research.  At the Practice Meeting, speakers are carefully selected in order to present the best practical work, along with the most important methodological advances (generally in the form of tutorials).  As an example of the talks, here is the “Supply Chain” track:

  • Procter & Gamble – William Tarlton, Supply Chain R&D Manager, Personal Beauty Care Products, on implementing inventory optimization at P&G.
  • Pepsi Bottling Group – Arzum Akkas, Senior Project Manager, Supply Chain Technology, on retail out-of-stock reduction in a direct store delivery environment .
  • Pennsylvania State University –Terry Harrison, Professor of Business, Professor of Supply Chain & Information Systems, and Thomas Robbins, Instructor in Supply Chain & Information Systems, on services supply chain.
  • Xilinx – Alex Brown, Principal Engineer and Supply Chain Architect, on collecting and using demand information from customers and distributors to improve forecasts and supply chain performance.
  • IBM  –Markus Ettl, Manager of Supply Chain Analytics & Architecture, IBM Research, and Blair Binney, Manager of Demand/Supply Planning Process Transformation, IBM Integrated Supply Chain, on how analytics and collaborative processes improve distributor and IBM performance in the supply chain.
  • University of North Carolina – Brian Tomlin, Assistant Professor of Operations, Technology and Innovation Management, on supply chain risk management.

This is a great mix of academics and business executives, and I guarantee that the speakers will have spent significant time honing their presentations.

Of course, the conference is pretty pricey, but if you are doing (or considering doing) OR in practice, or if you teach courses on the practical use of OR, this is a must see conference.  The hotel deadline is in a couple of days.

I’ll be there, and perhaps do a bit of live blogging along the way.

In Praise of a … Yankee?

Since I live in Pittsburgh and attend many Pittsburgh Pirate baseball games, it is tough to be very positive about the Yankees. For European readers, it is kinda like a Barça fan saying something nice about Real Madrid. But I have a new favorite player: Russ Ohlendorf, relief pitcher for the Yankees!

Why is he my favorite player? It has to do with operations research, of course. Russ received his undergraduate degree from Princeton (he is the third Princeton graduate to play in the majors) in operations research. From the New York Times article on him:

Ohlendorf graduated with a 3.75 grade-point average and a degree in operations research and financial engineering. His curiosity extends widely. Someday, Ohlendorf said, he may want to try investment banking or entrepreneurship. He has also thought about pursuing a front-office job in baseball.

“He seems to have a real interest in people from all walks of life,” Curtis Ohlendorf said. “He’s always been pretty active. He needs to be involved in something.”

For his senior thesis, Ohlendorf studied the value of draft picks for major league teams. His conclusion — which the Yankees now embrace — was that teams generally double their investment in the draft based on the production of players in their first few major league seasons.

I can’t find the thesis on the web: I would love to see the approaches used. Statistics and baseball go way back, of course, but real economic analysis or operations research modeling is a lot rarer.

Gene Woolsey and Consulting in Operations Research

Gene Woolsey is one of my favorite people in OR, though I have met him just a couple of times. Gene has very strong views on how OR should be taught, and he implements them at the Colorado School of Mines. A key aspect of his approach is that OR is about doing and solving problems. And no problem can be solved without spending significant amounts of time with the people currently doing the job. So if you think you have a stocking problem at a grocery store, most of us OR people would ask for data, create models, solve them, and send back the results, without ever setting foot in a store. Gene (or, more likely now, his students) would spend days working with the stock people in the grocery store, and gain a hands-on understanding of the real situation. More often than not, this allows Gene to understand what the real problem is, and to avoid wasting time on unneeded analysis.  I will confess I like this approach much more in the abstract than in practice. I worked for a while on US Postal System reorganization without spending much time at all in a postal sorting facility. That is not the way Gene would do it!

Gene writes a regular column in Interfaces, my favorite OR journal (make sure your organization or university subscribes if you are at all serious about OR) entitled “The Fifth Column”. In general, the column is about doing OR, particularly doing OR as a consultant. In the November, 2007 issue, he wrote on “How to Consult and Not Be Paid”. It is a great column that hits a bit close to home (most of my consulting seems to be of the unpaid nature). In short, Gene gets a call from a possible client. He comes up immediately with an insightful, clever, and very easy to implement solution. And then he blurts it out. The prospective client thanks him profusely, hangs up, and that is all there is. Learning not to blurt out solutions is a good lesson!

OR at P&G

Let me be late in the OR blogging game and note that there is a great article on operations research at Proctor and Gamble on bnet. It is wonderful advertising for our field, including phrases like “P&G’s Killer Apps in OR”. INFORMS was strongly involved in the article, with quotes from Past-President Brenda Dietrich and Executive Director Mark Doherty:

P&G, GE, Merrill Lynch, UPS — the list of Fortune 500 companies getting into the OR game is expanding, says Mark Doherty, executive director of the Hanover, MD-based Institute for Operations Research and Management Sciences (INFORMS), an OR think tank. “In the private sector, OR is the secret weapon that helps companies tackle complex problems in manufacturing, supply chain management, health care, and transportation,” he says. “In government, OR helps the military create and evaluate strategies. It also helps the Department of Homeland Security develop models of terrorist threats. That’s why OR is increasingly referred to as the ‘science of better.’”

Having sat in on a few too many board meetings, I think calling INFORMS an “OR think tank” is going a little far. But the article does project the very best vision of our field.

Check out another take on the article at Punk Rock Operations Research (and I thought I saw it on another OR blog, but it escapes me at the moment).

Soo-Haeng Cho and the Influenza Vaccine

I’ve been back in the US for about six weeks now, and am getting used to being back in my academic life. A sign of the slowness of this transition, however, is that our operations management group here at the Tepper School is hiring a junior faculty person, and I didn’t notice, so I have missed most of the job talks. I feel bad about that: the best part of hiring is seeing the best new research from around the world.

Lasts week, Soo-Haeng Cho interviewed here (and did very well by all reports). Soo-Haeng works in a number of areas of operations management, including the use of OM methods in medical decisions. He has a very nice paper on choosing the correct flu vaccine each year. This issue has been in the news recently (including CNN ) because the current vaccine is missing quite a few of the flu bugs. Cho’s paper talks about many of the issues that go into the choice of vaccine, not all of which are reasonably covered by the popular press. In particular, I hadn’t realized the strong advantage of doing the same as last year in terms of getting reliable vaccines out to people.  From his paper:

The production yields of strains are variable and unknown owing to its biological characteristic (Matthews 2006). Moreover, yield uncertainty is increased significantly when a vaccine strain is changed. The magnitude of this challenge is illustrated in the following quote from an industry representative (Committee
2003):
“certainly the best way to ensure this predictability of supply is not to recommend any [strain] changes, … a second best way is to minimize the number of strain changes. Each new strain can yield anywhere from 50 to 120 percent of the average strain.”
Thus, even if a new virus strain is predominant, a change is made only when the benefit from improved efficacy outweighs the risk associated with making the change in production. For instance, although new A/Fujian-like virus strains were widely spreading during the 2002-3 season, the Committee did not select
that strain because it was not suitable for large-scale manufacturing.

It is clear that understanding the medical decision making requires the understanding of manufacturing operations, which I think is a great theme in the upcoming years for our field.

Summary of Quantum Computing

Scott Aaronson, whose writings I both admire and am jealous of, has an article in the month’s Scientific American on the limits of quantum computing.  He has posted a preliminary version of the paper on his site.  I found this extremely useful in trying to make sense of what quantum computing can and can’t do.  It is a shame the writers at SlashDot don’t read the paper before making the comments showing their confusions!