The purpose of this project is to illustrate how decision trees may be used to model international conflict, as well as to see how closely the dollar auction models real conflicts between nations.

It will be due on Tuesday, March 18, immediately after Spring Break. You may work alone or with one other person. The assignment will have two parts: first, you will submit something that can be posted to the course web page (preferably in HTML or PDF format); second, you will briefly explain your project in class, using your web page for reference.

Much of this will be similar to the North Korea nuclear escalation exercise, to which you should refer for an example.

• Choose a conflict between two nations to model. Pick a conflict that is fairly short, or a short episode in a long conflict, so that each country only has a few choices. The idea is to find something where leaders of nations choose whether to escalate the situation or to seek peaceful alternatives. You should find something where you can understand the preferences or motives of each side, even if you don't agree with them; for example if you think one side is motivated by pure evil you will not be able to do this correctly.

  You might choose something current or historical. It could be a situation which led to war, or a situation which could have led to war but didn't. Try to find something where intervention by a third nation is less important. Here are some examples:

  • U.S.-U.S.S.R.: The Cuban missile crisis, 1962 (but not the entire Cold War)
  • U.S.-Vietnam: from the Gulf of Tonkin resolution to the Tet offensive (not the entire war)
  • U.S.-Iraq: the present conflict, or the buildup to the 1991 war
  • U.S.-China: the spy-plane incident of 2001
  • Israel-Palestine: the recent conflict after the collapse of the Oslo accords
  Obviously you should not choose the recent U.S.-North Korea conflict.
• List the options available to each country. The list should have at least three options for each but not many more. Simplify the actual conflict, without removing important alternatives.
• Explain the alternatives that each nation prefers. Which alternatives are preferred may depend on what the other nation has already done; take this into account. This part will involve assumptions and may be subject to debate: use your judgment.
• Once you've done the previous step, the rest should not be subject to any debate: your assumptions should completely determine a decision tree.
• Prune the decision tree using each country's preferences. Determine the optimal strategy that both nations will use, assuming rationality, and the ultimate sequence of events.
• If this is a historical conflict, compare your prediction to what actually happened. If it is a current event, make an educated guess as to whether your prediction is likely to be followed.
• If your prediction is not accurate, explain whether you think it is because your assumptions could be flawed, or whether the participants are not acting rationally.
• Write up your explanation and conclusions, and draw a decision tree. Save your work as an HTML document or convert it to a PDF, and send it to my SAS account: scpresto@sas.upenn.edu.
• In class on Tuesday, March 18, you will describe to the class briefly what you did and what your prediction was. The idea here is to get people interested in reading what you wrote, while referring them to the web page for more details.
• If you need help choosing a topic or researching it, please feel free to contact me for suggestions. I may also be able to help with file format conversions.