def monte_carlo(self, board, depth=None):
if depth is None:
depth = []
ttt = TicTacToe()
ttt.copy(board)
# ランダムに自分の一手を行う
position = self.pick_one(ttt)
ttt.set(position, self.t)
if ttt.eval() == TicTacToe.ON_PROGRESS:
# 敵の手もランダムに一手を行う
t_enemy = TicTacToe.O if self.t == TicTacToe.X else TicTacToe.X
position_enemy = self.pick_one(ttt)
ttt.set(position_enemy, t_enemy)
depth = depth + [position]
# 結果を判定。決着がつかないなら再帰的に繰り返す
result = ttt.eval()
if result != TicTacToe.ON_PROGRESS:
return {
"position": depth[0],
"game_result": result,
"moves": depth
}
return self.monte_carlo(ttt.board, depth)
def monte_carlo(self, board, t, depth=None):
self.order_count += 1
if depth is None:
depth = []
ttt = TicTacToe()
ttt.copy(board)
# ランダムに自分の一手を行う
position = _pick_one(ttt)
ttt.set(position, t)
t_enemy = TicTacToe.O if t == TicTacToe.X else TicTacToe.X
depth = depth + [(t, position)]
# 結果を判定。決着がつかないなら再帰的に繰り返す
status = ttt.eval()
if status != TicTacToe.ON_PROGRESS:
result = {
"game_result": status,
"moves": depth,
TicTacToe.WIN_O: 0,
TicTacToe.WIN_X: 0,
TicTacToe.DRAW: 0,
}
result[status] += 1
return result
result = {
"position": position,
TicTacToe.WIN_O: 0,
TicTacToe.WIN_X: 0,
TicTacToe.DRAW: 0,
}
mc_results = [
self.monte_carlo(ttt.board, t_enemy, depth) for _ in range(1)
]
for mc_result in mc_results:
result[TicTacToe.WIN_O] += mc_result[TicTacToe.WIN_O]
result[TicTacToe.WIN_X] += mc_result[TicTacToe.WIN_X]
result[TicTacToe.DRAW] += mc_result[TicTacToe.DRAW]
return result
