Undergrad Research Project - Game-theoretic learning in distributed information settings

Spring 2017

Anirudh Sridhar
Soummya Kar
Project description

Game theory serves as an effective model for many problems in economics, politics, social science and engineering to name a few fields. The goal of this research is to study when a state of equilibrium is reached when (1) players' actions are inertial and/or myopic and (2) players have limited information about the world, and only learn through communication with and observations of their neighbors. This model of a game is particularly useful when there is a very large number of acting agents and communication is constrained and thus limited to a small number of neighbors.

One main challenge is to study when agents learn equilibrium strategies if they only observe their neighbors instead of communicating. As a special case, we are studying leader-follower networks and their convergence properties.

Return to project list