STAT 155: Game Theory (Spring 2007)

Instructor: Asaf Nachmias.

Email: a s a f n a c h @ m a t h

Office hours: Tue 3:30-4PM, Wed 2-3PM, Thu 3:30-4PM, 341 Evans.

Class time: Tuesday and Thursday, 2:00 - 3:30 (room 330, Evans Hall).


Announcements




Homework




Possible topics for presentations




Course description

Game theory is a fascinating subject, which sheds light on diverse topics including economics, statistical decision theory, voting and even evolutionary biology. We will first study combinatorial games and zero sum games for which there is a satisfying general theory, and continue with more general games where the key notion is Nash equilibrium, and the problem of deciding between Nash equilibria is a topic of active research. To gain a feeling for the strategies involved, we will try some of the games in class. Additional topics: 1. The notion of Shapley value helps identify the power of an agent when multiple coalitions are possible. 2. We will compare voting systems such as simple majority vote versus the electoral college. 3. We'll consider hide and seek games related to the assigment problem. 4. We will discuss Arrow's impossibility theorem for a "rational" scheme of jointly deciding between more than two options. We will also explore some of the beautiful mathematical tools that give game theory its power: von Neumann's minimax theorem will be established using supporting hyperplanes for convex sets; we'll prove the existence of Nash equilibria via the Brouwer fixed point theorem, following Nash's original argument.


Text

We will use the online texts: An excellent additional resource, where you can find other lecture notes and a glossary of game theory concepts, and also play some games, is gametheory.net

More information about random turn selection games can be found here .


Prerequisite

Undergraduate courses in linear algebra and probability. If you can solve linear equations and know what is the expectation of a random variable, you're in.

Grading Students taking the course for credit will be graded based on


Each student will be graded based on his/her performance, rather than on how it compares to the work of other students. Hence collaboration in studying, and thinking about homeworks, is encouraged and expected.