from capstone.game.games import Connect4
from capstone.game.players import RandPlayer
from capstone.game.utils import play_series
class MyPlayer(object):
def choose_move(self, game):
return game.legal_moves()[0]
my = MyPlayer()
game = Connect4()
players = [my, RandPlayer()]
play_series(game, players, n_matches=1000)
# Problem Set J (Challenger)
# Problem 273, Page 79
# Red wins with E2
# A B C D E F G
challenger = [
[' ', ' ', ' ', ' ', ' ', ' ', ' '], # 6
[' ', ' ', ' ', ' ', ' ', ' ', ' '], # 5
[' ', ' ', ' ', ' ', ' ', 'O', ' '], # 4
[' ', ' ', ' ', ' ', ' ', 'X', ' '], # 3
[' ', ' ', ' ', ' ', ' ', 'O', ' '], # 2
[' ', ' ', 'O', 'X', 'O', 'X', 'X']
] # 1
# c42pdf(filename, c4.board)
game = Connect4(challenger)
mdp = GameMDP(game)
env = Environment(mdp)
qnetwork = QNetwork(move_mapper,
n_input_units=42,
n_hidden_layers=1,
n_output_units=7,
n_hidden_units=100,
learning_rate=0.01)
egreedy = EGreedy(provider=env.actions,
qfunction=qnetwork,
epsilon=1.0,
selfplay=True)
qlearning = ApproximateQLearning(env=env,
qfunction=qnetwork,
policy=egreedy,
from capstone.game.players import AlphaBeta
from capstone.game.utils import c42pdf
from capstone.rl import FixedGameMDP, Environment
from capstone.rl.learners import QLearning
from capstone.rl.policies import RandomPolicy
from capstone.rl.utils import QValuesPlotter
from capstone.rl.value_functions import TabularQ
seed = 23
board = [['X', 'O', 'O', ' ', 'O', ' ', ' '],
['X', 'O', 'X', ' ', 'X', ' ', ' '],
['O', 'X', 'O', 'X', 'O', 'X', 'O'],
['O', 'X', 'O', 'X', 'O', 'X', 'O'],
['X', 'O', 'X', 'O', 'X', 'O', 'X'],
['X', 'O', 'X', 'O', 'X', 'O', 'X']]
game = Connect4(board)
mdp = FixedGameMDP(game, AlphaBeta(), 1)
env = Environment(mdp)
qlearning = QLearning(env=env,
qfunction=TabularQ(random_state=seed),
policy=RandomPolicy(env.actions, random_state=seed),
learning_rate=0.1,
discount_factor=1.0,
n_episodes=1000)
qlearning.train(callbacks=[
QValuesPlotter(state=game,
actions=game.legal_moves(),
filepath='figures/c4_ql_tab_qvalues.pdf')
])
####################
from capstone.game.games import Connect4
from capstone.game.players import RandPlayer
from capstone.game.utils import play_match
game = Connect4('XO-----' 'XO-----' 'OXOXOXO' 'OXOXOXO' 'XOXOXOX' 'XOXOXOX')
players = [RandPlayer(), RandPlayer()]
play_match(game, players)