It is Christmas time, so lots of people are waiting in line, and talking about it. Naturally, this leads to some OR related articles. Ivars Peterson in his Math Trek column at Science News talks about parking strategies: do you park and walk, or cycle through hoping for a better spot (seems the better is generally the better strategy). And the Birminham News is the latest to talk about Dick Larson’s work on queue behavior.
OR and Suicide Bombs
This past weekend, the NY Times Magazine in its annual Ideas issue reports on Ed Kaplan’s work with Moshe Kress on damage done by suicide bombers. Here is an excerpt:
Even if you manage to detect a suicide bomber, what do you do next? This question was taken up by Edward H. Kaplan, a professor of public health at Yale, in a paper he published in July, written with Moshe Kress of the Naval Postgraduate School in Monterey, Calif. Kaplan and Kress investigated the physics of a belt-bomb blast and reached some unexpected conclusions. It turns out that very few people are killed by the concussive force of a suicide explosion; the deadly weapon is in fact the shrapnel – the ball bearings, nails or pieces of metal that the attacker attaches to the outside of his bomb. The explosions, though, are usually not powerful enough to send these projectiles all the way through a human body, which means that if your view of a suicide bomber is entirely obscured by other people at the moment of detonation, you are much more likely to escape serious injury. Because of the geometry of crowds, Kaplan found, a belt bomb set off in a heavily populated room will actually yield fewer casualties than one set off in a more sparsely populated area; the unlucky few nearest to the bomb will absorb all of its force.
The authors used these calculations to question some assumptions about what authorities should do if they detect a bomber. The International Association of Chiefs of Police issued guidelines this year suggesting that police officers who find a bomber in a crowd should fire shots into the air to cause people near the bomber to scatter or hit the deck. But Kaplan’s calculations demonstrate that in many cases, this would make things worse – as a packed crowd ran away from a bomber or dropped to the ground, the circle of potential victims around him would get wider and thus more populous, and more lives could be lost.
Amazing what a little OR will teach you!
Book blurb
Springer has just published a new book containing introductory tutorials on a range of optimization subjects: Search Methodologies: Introductory Tutorials in Optimization and Decision Support Techniques. Of course, I have a vested interest: Bob Bosch and I wrote the Integer Programming chapter, which I think came out quite nice!
Wikipedia and Operations Research
Wikipedia is in the news, due to some inaccurate/obnoxious/insulting entries regarding a journalist and alleged involvement in the Kennedy assassinations.
Wikipedia is an interesting effort to harness the knowledge and energy of hundreds of thousands of people to form a new type of encyclopedia. The key aspect is the ability to freely enter information (or misinformation) and edit what is there.
The entry for operations research is pretty good, covering both the general (what is OR) and the specific (some military and other examples). The pointers are well selected. The “history” gives all the past edits, and it is striking how many edits went into this entry alone. That is one advantage of multiple people working on it: 100 edits don’t seem bad if 50 people are doing them.
But what is to stop someone from putting “Operations Research is an upsidedown cake with cherries” into the entry? As this Wired article describes, with enough interested volunteers, such vandalism can get corrected within minutes. Of course, that assumes the entry is in an area with knowledgeable and active volunteers. I wonder how much misinformation is stored in the backwaters of wiki.
Despite the doubts, I think a wiki based Encyclopedia of OR would be a tremendous asset for the field. While some expert-based encyclopedias are fine (Gass and Harris have a good one, at an astounding $620 new), a community based wiki would be more up-to-date and be able to cover far more.
More on Scientific Publishing
Wired magazine pointed me to www.badscience.net, which is fascinating! There was a posting relevant to issues of publishing science which seems very relevant to OR:
Science is done by scientists, who write it up. Then a press release is written by a non-scientist, who runs it by their non-scientist boss, who then sends it to journalists without a science education who try to convey difficult new ideas to an audience of either lay people, or more likely – since they’ll be the ones interested in reading the stuff – people who know their way around a t-test a lot better than any of these intermediaries. Finally, it’s edited by a whole team of people who don’t understand it.
Lots of good stuff there.
OR in the media and Virginia Postrel
Operations Research has a hard time getting into the press. Partially it is our fault: OR people as a whole are pretty modest and are great at seeing two sides to every issue (after all, it is this dynamic that makes for the best models: start small, and add to address issues with the resulting solutions). I am certainly guilty of this: I turned down more requests to talk about my work with Major League Baseball than I can count, due primarily to worries about client relations and being accurately portrayed.
Partially it is the fault of much of the media, who are unwilling to assume a very high “lowest common denominator” in their readership. OR continues to be a weird black box, and our core values of analytical thinking, using data, and working with models are rarely portrayed.
One person who does get it is Virginia Postrel, who wrote a very fine article for the Boston Globe last year (and I don’t say that just because I am quoted!). Here is what she had to say about publishing about operations research (in the context of the rise in productivity, which OR certainly contributes to):
In today’s Boston Globe Ideas section, I look at one piece of that very big story [the rapid rise in productivity]: the spreading use of operations research techniques once confined to theory. (What’s operations research? The story explains that too, or tries to without using any math, graphs, or jargon about optimizing subject to constraints or finding interior solutions. For more on the field, see the INFORMS site.)
Of course, very few general-interest publications would let a writer spend nearly 2,000 words writing about operations research–or, for that matter, rising productivity.
I stumbled across her website and blog, and she has a number of interesting posts on productivity (generally her 2004 posts, like the one I quoted that mentions her OR article). Her recent work is on glamor and aesethics, which doesn’t appeal to me as much but might to others. Lots of interesting things to read throughout her website.
Initial Plans for INFORMS 2006
Now that INFORMS 2005 New Orleans/San Francisco is over, it is the Pittsburgh crowd’s turn to put together INFORMS 2006. We are well on the way planning, with a list of tutorials and invited sessions that I think will be very good.
The theme of the conference in OR Renaissance, and we hope to highlight some of the exciting new directions and applications for OR/MS. We’ll have some new things in store for 2006 including
1) A new Tuesday reception, planned for PNC Park, the site of the 2006 MLB All-Star game.
2) “Renaissance Sessions”, highlighting the best of where OR will be in 5 years.
3) Plans for a revamped awards ceremony.
The Monday General Reception will be at the wonderful Carnegie Museums: the Museum of Art and the Museum of Natural History (they are connected).
The conference will be November 5-8, 2006. If you have some thoughts on how INFORMS conferences could be made better, I’d love to hear them. Comment here or mail to me.
Rothkopf’s Rankings of University Contributions to the Practice Literature
Mike Rothkopf has just published his sixth ranking of universities in publishing in the practice of operations research in the journal Interfaces (subscription required to access full paper). The definition of “practice” is naturally a complicated thing: most OR people (myself included) claim relevance to practice on pretty slim connections. For this ranking, “practice” means publishing either in Interfaces or in the OR Practice area of Operations Research. My own school, Carnegie Mellon, came out on top with 10 such publications (1998-2004). The next part of the rankings are
2. University of Pennsylvania
3 (tie). Georgia Institute of Technology and Naval Postgraduate School
5. Cornell, Stanford, University of Texas at Austin, and University of Virginia
Non-US schools are ranked separately. Top is Erasmus University with 7 papers (same as Cornell, etc. on the US list).
John Muth
John Muth passed away in October. He was an early faculty member at the school I am at (the Tepper School at Carnegie Mellon) and a key researcher in the economic area of rational expections. What is the OR content? Check out his obituary and note the work in operations management and forecasting along with his work in economics and finance. Not many people about these days that would try to span those areas!
INFORMS Awards
I just got back from the INFORMS Award ceremonies. Some of the highlights:
1) John von Neumann Theory Prize went to Robert Aumann, who also won the Nobel Prize in Economics this year. The committee picked Aumann before the Nobel Prize was announced. Pretty good year for Robert!
2) The Lanchester Prize (best publication in English) went to Garrett van Ryzin and Kalyan Talluri for their Revenue Management book.
3) Peter Bell of the University of Western Ontario won the INFORMS Prize for the Teaching of OR/MS Practice. Peter has worked a lot on cases in OR/MS.
There were a number of other prizes. I am sure the INFORMS Prize web page will have these shortly.