In this multiagent environment, exist seven agents that belong in two different categories, X and Y. Each category shows a preference to one the two available agent’s actions. These agents are connected in a sparse manner via a connection graph and play a simple twoplayer coordination game, where if they both choose the same action, they get a payoff equal to 1. Otherwise, their payoff is equal to 0. The agents know neither the game they are playing, nor the available actions of the other agents. The purpose of this assignment it to implement a Q-learning, ϵ-greedy agent that learns how to play the game with the other agents he is connected to.
The implementation is found in Qlearning.py.