Repository for Catan Bot

First iteration of creating a Catan Bot. The goal is to create a Deep-Q Neural Network to play Catan.

Limitations of 1st Iteration:

• Random Starting Positions (arguably the most important decisions a player makes)
• One trade offer per round (limits the action space of the DQN - includes trading with the bank)
• 4 player games (this is the general method of play)

Steps

• Implement Catan ("Optimized" for self play and RL)
• Create DQN

Future Plans:

• Implement a Hierarchichal DQN structure (I have to read more papers about the concept before implementing)