Monthly Archives: April 2007

modeling spite

Traditional approaches to decision analysis assume that decision makers are rational, but in reality, we’re fairly irrational. That’s what makes us human. What happens in certain applications when decision makers are likely to be spiteful? That’s what one of my students addressed in his presentation for my class in multiobjective decision analysis (all students have to present a research article).

The Ultimatum Game is a test of whether decision makers play in a “traditional” manner in bargaining games. This game involves two players, the Proposer and the Responder. The game is played only once, and the players remain anonymous to one another. The Proposer is given a certain amount of money (say, $20) and the Proposer divides this amount into two to give to each participant (say, $15 for the Proposer and $5 for the Responder). The Responder can either accept the amount or reject it, resulting in neither player receiving any money. The Game Theory solution to this problem is for the Responder to accept the amount that the Proposer gives–regardless of the amount of the reward–because some money is better than none.

However, in many experiments, it has been shown that if Proposers offer less than 25% of the total amount to the Responder, the Responder nearly always rejects it. This is nuts–they are rejecting free money! This was verified in countries all over the world. When the study was done in rural Indonesia, the Responders (poor workers) would reject money that would amount to several weeks worth of wages — out of spite! I bet there are a million applications where this would be useful.

Camerer, Ho, Chong. “A cognitive hierarchy of games,” Quarterly Journal of Economics, 2004.
Cameron. “Raising the stakes in the Ultimatum Game: Experimental evidence from Indonesia,” Economic Inquiry, 1999.
Henrich. “”Does culture matter in economic behavior?” American Economic Review, 2000.
Slonim, Roth. “Learning in high stakes Ultimatum Games: An experiment in the Slovak Republic,” Econometrica, 1998.


we are all Hokies

I am shocked, disgusted, and saddened by yesterday’s shootings at Virginia Tech. It is a painful reminder about how precious and fragile life is. My heart goes out to the victims and their families — I cannot imagine their losses.

Edelman Award

INFORMS announced the finalists for the Edelman Award for Achievement in Operations Research and the Management Sciences. Every finalist for this award is a model of efficiency and management strategy.

According to INFORMS, the finalists are:
1. Coca-Cola Enterprises, for a project entitled “Optimizing Distribution at Coca-Cola Enterprises”
2. The U.S. Coast Guard, for “Maximizing Aircraft Availability by Managing Aircraft Maintenance Throughput at the U.S. Coast Guard Aircraft Repair and Supply Center”
3. Hewlett-Packard, for “Procurement Risk Management at Hewlett-Packard Company”
4. Chrysler Group and J.D. Power and Associates for “PIN Incentive Planning System: A Decision Support System for Planning Automobile Manufacturers’ Pricing and Promotions”
5. Memorial Sloan-Kettering Cancer Center for “Operations Research Answers to Cancer Therapeutics”

All of the finalists look interesting. Is it wrong to wish that Coca-Cola loses? I don’t like pop (I am a Chicago native, so it’s pop, not soda) nor do I like high fructose corn syrup. I am not happy that the average person in the United States drinks about two cans of pop per day (even teenagers and children!!!). It’s no wonder since Coca-Cola spends more on advertising than almost every other company. I’d like to see an enormous advertising campaign for broccoli instead. But then again, I drink about two cups of coffee per day, so maybe I’m no better (Does coffee count as a vegetable? It’s made from beans, so I think I could make the case).

Humanitarian logistics

Yesterday, I went to a great talk given by Dave Goldsman from Georgia Tech called “The Role of Simulation in Humanitarian Logistics.” The seminar was organized by my colleague Jill Hardin at VCU as part of the Operations Research Seminar Series. The talk summarized some of Dr. Goldman’s experiences in applying simulation to problems in health clinic management/flow and the spread of Guinea worms in Africa. Guinea worms are particularly interesting. Guinea worms are parasites that are contracted by drinking contaminated water. Once contracted, they eat their way out of the body and reproduce. Breaking the chain results in extermination. This can be easily accomplished by filtering water. However, because of the civil war in Sudan, many soldiers walk home, contracting Guinea worms along the way. Finding a strategy for targeting resources is important for completely eradicating the Guinea worm. The research was funded by the Carter Center, and apparently, Jimmy Carter’s work has nearly eradicated the Guinea worm already.

Equity saves lives

Clearly, advances in medicine and health care save lives. But not everyone has access to advanced medical procedures — or even basic health care — in America today. This leads to the following question: if you had a dollar to spend on health care, should you use it to advance medicine or should you use it to ensure that people have access to existing health care? How many lives would be saved if everyone had eual access to health care?

Steven Woolf of VCU answers this question in a recent article published in the April issue of the American Journal of Public Health (read the press release). He finds that between 1996 and 2002, “eight times as many lives could have been saved by eliminating the higher death rates experienced by Americans with inadequate education.” I found this number to be overwhelming and compelling, but sadly, few people seem interested in talking about it. It is compounded by the fact that nearly all research dollars for the medical field are spent to improve medicine, not to make sure medical resources are used “fairly.”

OR has the opportunity to make a difference here. Many OR applications optimize over the “expected value.” In other words, the solution is optimal on average, but there are some clear winners and losers in the solution. This does not take into account the real-world discrepencies in health care in America — the same groups of people always seem to come out ahead or behind. Looking at expected value isn’t “bad” as long as we understand that the expected value solution may not always be equitable and address the extremes in expected value models.

Opt Art

Last month, Robert Bosch gave a fascinating talk on operations research and art here at Virginia Commonwealth University. His talk focused on showcasing how optimization, in particular, using entire sets of dominos to reconstruct paintings and images, such as the Mona Lisa and Marilyn Monroe. You can see his masterpieces at DominoArtwork. It was fascinating to see an illustration of how art connects to optimization.

Some links:
Bob’s OR/MS Today article
Images using knots
Science for kids article featuring Bob’s art

Murder and the Operations Researcher

This week, I read “An O.R. Missionary’s Visits to the Criminal Justice System” by Alfred Blumstein that appeared in the February issue of Operations Research (Read a synopsis of his article here). I remember the talk in 2005, which was an inspiration to me and a big reason why I started this blog on public policy OR. Dr. Blumstein’s contributions to the criminal justice system are nothing short of amazing–he was even awarded the 2007 Stockholm Prize in Criminology for his research on criminal careers.

What I find interesting about Dr. Blumstein’s approach to the criminal justice system is that he takes a systems-level approach and analyzes the impacts of each decision (i.e., sentencing guidelines) on the entire system to get counter-intuitive results.

You can download the podcast of an interview with Dr. Blumstein that appeared on FOCUS 580, a radio program out of Champaign-Urbana, Illinois.