def run_game_loop(board, agent1, agent2, do_print=False):
game = Game(board)
while not game.is_finished():
dice = game.roll_dice()
if do_print:
board.print()
print("Dice for AI1 were {}.".format(dice))
ai_moves = agent1.move(dice)
if do_print:
print("Resulting moves: {}.".format(", ".join(
[str(move) for move in ai_moves])))
game.apply(game.PLAYER1, ai_moves)
if game.is_finished():
break
dice = game.roll_dice()
if do_print:
board.print()
print("Dice for AI2 were {}.".format(dice))
ai_moves = agent2.move(dice)
if do_print:
print("Resulting moves: {}.".format(", ".join(
[str(move) for move in ai_moves])))
game.apply(game.PLAYER2, ai_moves)
return game.get_winner()
def run_game_loop(board, agent1, agent2, do_print=False):
game = Game(board)
while not game.is_finished():
agent1.learn()
#agent1.backprop()
dice = game.roll_dice()
if do_print:
board.print()
print("Dice for AI1 were {}.".format(dice))
ai_moves = agent1.move(dice)
if do_print:
print("Resulting moves: {}.".format(", ".join(
[str(move) for move in ai_moves])))
game.apply(game.PLAYER1, ai_moves)
if game.is_finished():
break
agent2.learn()
#agent2.backprop()
dice = game.roll_dice()
if do_print:
board.print()
print("Dice for AI2 were {}.".format(dice))
ai_moves = agent2.move(dice)
if do_print:
print("Resulting moves: {}.".format(", ".join(
[str(move) for move in ai_moves])))
game.apply(game.PLAYER2, ai_moves)
winner = game.get_winner()
if winner > 0:
agent1.learn(np.array([1.0, 0.0]))
agent2.learn(np.array([0.0, 1.0]))
#agent1.backprop(np.array([1.0, 0.0]))
#agent2.backprop(np.array([0.0, 1.0]))
else:
agent1.learn(np.array([0.0, 1.0]))
agent2.learn(np.array([1.0, 0.0]))
#agent1.backprop(np.array([0.0, 1.0]))
#agent2.backprop(np.array([1.0, 0.0]))
return winner, game.get_num_moves()