This document summarizes research on using genetic algorithms to learn strategies in repeated matrix games against other learning algorithms and humans. It introduces the problem of modeling multi-agent systems using matrix games and the motivation to incorporate human input. It describes prior work on adaptive learning in multi-agent systems and interactive artificial learning. The document outlines the evaluation environment and discusses genetic algorithm variations like propagating history, stabilization conditions, and dynamic mutation. It presents results and analysis on the performance of genetic algorithms in repeated matrix games with and without human input.