Month: September 2011

Grizzlies, Pandas, and Optimal Ecological Structures

Last year, a group of students in one of my classes did a project on designing a grizzly bear habitat, inspired by the work at Cornell’s wonderful Institute for Computational Sustainability. In that project, the goal was to pick out a collection of geographic areas that formed a contiguous zone that the grizzly’s could move freely through. As the ICS description says:

Land development often results in a reduction and fragmentation of natural habitat, which makes wildlife populations more vulnerable to local extinction. One method for alleviating the negative impact of land fragmentation is the creation of conservation corridors, which are continuous areas of protected land that link zones of biological significance.

My colleague, Willem van Hoeve, had worked on variants of this problem and had some nice data for the students to work with. The models were interesting in their own right, with the “contiguity” constraints causing the most challenge to the students. The results of the project were corridors that were much cheaper (by a factor of 10) than the estimates of the cost necessary to support the wildlife. The students did a great job (as Tepper MBA students generally do!) using AIMMS to model and find solutions (are there other MBA students who come out knowing AIMMS? Not many I would bet!). But I was left with a big worry. The goal here was to find corridors linking “safe” regions for the grizzlies. But what keeps the grizzlies in the corridors? If you check out the diagram (not from the student project but from a research paper by van Hoeve and his coauthors), you will see the safe areas in green, connected by thin brown lines representing the corridors.    It does seem that any self-respecting grizzly would say:  “Hmmm…. I could walk 300 miles along this trail they have made for me, or go cross country and save a few miles.”  The fact that the cross country trip goes straight through, say, Bozeman Montana, would be unfortunate for the grizzly (and perhaps the Bozemanians).  But perhaps the corridors could be made appealing enough for the grizzlies to keep them off the interstates.

I thought of this problem as I was planning a trip to China (which I am taking at the end of November).  After seeing a picture of a ridiculously cute panda cub (not this one, but similarly cute), my seven-year-old declared that we had to see the pandas.  And, while there are pandas in the Beijing zoo, it was necessary to see the pandas in the picture, who turned out to be from Chengdu.  So my son was set on going to Chengdu.  OK, fine with me:  Shanghai is pretty expensive so taking a few days elsewhere works for me.

As I explored the panda site, I found some of the research projects they are exploring.  And one was perfect for me!

Construction and Optimization of the Chengdu Research Base of Giant Panda Breeding Ecological System

I was made for this project!  First, I have the experience of watching my students work on grizzly ecosystems (hey, I am a professor:  seeing a student do something is practically as good as doing it myself).  Second, and more importantly, I have extensive experience in bamboo, which is, of course, the main food of the panda.  My wife and I planted a “non-creeping” bamboo plant three years ago, and I have spent two years trying to exterminate it from our backyard without resorting to napalm.  I was deep into negotiations to import a panda into Pittsburgh to eat the cursed plant before I finally seemed to gain the upper hand on the bamboo.  But I fully expect the plant to reappear every time we go away for a weekend.

Between my operations research knowledge and my battle-scars in the Great Bamboo Battle, I can’t think of anyone better to design Panda Ecological Systems.  So, watch out “Chengdu Research Base on Giant Panda Breeding”:  I am headed your way and I am ready to optimize your environment.  And if I sneak away with a panda, rest assured that I can feed it well in Pittsburgh.

This is part of the INFORMS September Blog Challenge on operations research and the environment.

The Importance of Accurate Data

I have been spending the last couple of weeks assigning faculty to courses and helping staff think about scheduling issues. I wish I could say that I have been using operations research techniques to do this sort of work. After all, most of my work has been in some form of timetabling optimization. But that has not been the case: for the most part I have simply done the work manually. Partially this is because I inherited a schedule that was 90% done, so I was really in a “rework” phase. But the main reason is that I am new at this job, so I don’t really understand the constraints (though I think I have a pretty good idea of the objective and variables). Gene Woolsey of the Colorado School of Mines had the philosophy that his students had to go out and do a job before they could do any modeling or optimization. So students worked production lines or helped drivers deliver packages first. Only after spending a few weeks on the job, could they think about how operations research could improve things. If I was Gene’s student, I would definitely pick an application in sports or entertainment rather than, say, high-rise steelwork.  For now, I am emulating that approach by first handling the courses manually then thinking about optimization.

Doing the course assignment and scheduling has been eyeopening, and a little worrisome. Just as I worry at the beginning of the season for every sports league I schedule (“Why are there three teams in Cleveland this weekend?”), I worried over the beginning of the fall term as the first of my assignments rolled out. Would all the faculty show up? Would exactly one faculty member show up for each course? Oh, except for our three co-taught courses. And …. etc. etc.

It turns out there is one issue I hadn’t thought of, though fortunately it didn’t affect me. From the University of Pennsylvania (AP coverage based on the Under the Button blog entry):

PHILADELPHIA (AP) — University of Pennsylvania students who were puzzled by a no-show professor later found out why he missed the first day of class: He died months ago.

The students were waiting for Henry Teune (TOO’-nee) to teach a political science class at the Ivy League school in Philadelphia on Sept. 13.

University officials say that about an hour after the class’s start time, an administrator notified students by email that Teune had died. The email apologized for not having canceled the class.

I hadn’t thought to check on the life status of the faculty.  I guess I will add “Read obituaries” to my to-do list.

Operations Research: The Sort of Decisions That Will Get You Fired

I just saw an ad for “Moneyball”, a new movie based on the book by Michael Lewis. A baseball manager (Billy Beane of the Oakland Athletics) used analytics (“Sabremetrics” in the baseball world) to choose players who were undervalued by the rest of the baseball world.  Beane had a constrained optimization problem:  he had to get as many wins as possible with a highly binding budget constraint.  His solution to that problem was to concentrate on statistics that seemed to be undervalued in the market, notably “on base percentage” (if you don’t know baseball, this gets a bit opaque, but getting on base is not as “sexy” as hitting home runs:  home run hitters are expensive; players that just get on base were cheap at the time).

There is a great line in the ad.  A colleague (the “stats guy”) of Beane says:

This is the type of decision that will get you fired!

Brad Pitt, playing Beane,  looks worried, but perseveres.  See the ad at about 25 seconds the official ad at about 18 seconds.

[Unofficial ad deleted.]

I love that line, since it really does sum up what operations research (and make no mistake: “Moneyball” is an operations research film) is all about. When you do operations research, you create models of reality. You do not create models of decisions. The decisions come from the models. And sometimes, the decisions don’t look at all like what you expected. And that is when it gets interesting.

Sometimes these unexpected decisions are due to modeling failures: you have forgotten a constraint, or a key modeling assumption turns out to not only be incorrect (assumptions almost always have some level of incorrectness) but critically incorrect. Optimization is really good at putting solutions right where the models are the weakest. And so you modify the model, not in order to change the decision, but in order to better represent reality. And you get new decisions. And you iterate between modeling and decisions until you reach a model that you believe represents reality. At that point, the decisions are of two types. They might tell you to do what you are doing, but do it better. And that is comforting and probably improves the decision making in the organization.

Or they tell you to do something completely different. And that is when you get to “Decisions that might get you fired.” That is when you need to decide whether you believe in your model and believe in the decisions it has generated. It would certainly be easy to change the model, not to reflect reality, but to force the decisions you believe are right. But if you really believe the model, then you need to avoid that easy path. You really need to decide whether you believe in the model and the resulting decisions.

I worked with a company a few years ago on their supply chain design. The results of the model came back over and over again saying two things: there were too many distribution centers, a result everyone believed, and it was far better for each distribution center to specialize in particular products, rather than have every center handle every product. The latter decision went deeply against the grain of the organization, and objection after objection was raised against the model. It would have been easy to put in a constraint “Every distribution center has to handle every product”. But there was no justification for this constraint except the ingrained belief of the client. In fact, we could show that adding the constraint was costing the organization a significant amount of money. Eventually, at least some of the organization bought into the decisions and began devising specialized distribution centers, but it was gut-wrenching, and perhaps career threatening. After all the discussion and fighting against the decisions, I am convinced those were the right choices: the organization had to change, not just improve.

“Operations Research: The Sort of Decisions That Will Get You Fired” doesn’t have the ring of “The Science of Better”. But the insights OR can get you may lead to radically different solutions than the incremental changes the SoB campaign implied. And those are the changes that can fundamentally change firms and organizations. And careers.