ng-dat/GoGame.Minimax

A project for testing Go game AI players' strategies with Minimax, Qlearning and related methods

A project for testing Go game AI players' strategies with Minimax, Qlearning and related methods

Running

Play the game, run cmd:

python play.py

#Todo:

• Step 1: Env
• Step 2: Random player
• Step 3: Greedy player
• Step 4: Minimax player 1 – Win loose score metrics
• Step 5: Minimax player 2 – alphabeta-pruning
• Step 6: Minimax player 3 – Custom score metrics
• Step 7: Minimax player 4 – Stored param
• Step 8: Negamax player
• Step 9: PVS player
• Step 10: Starting moves
• Step 11: Qlearner – Qtable
• Step 12: Qlearner – DQLearning
• Step 13: Testing & Visualize
  • Folder description
  • A sample gif of a Game
  • A Comparision table
    • Win/lose & Time spent between players
    • Win/lose between extra-strategies(score metric, game start)
    • Win/lose & Time spent between hyper-params config(depth search, branching factor) of each player