def run(states, sensible_moves, currentPlayer, lastMove):
n = 5
width, height = 8, 8
board = Board(width=width, height=height, n_in_row=n)
board.init_board()
board.states = states
board.availables = sensible_moves
board.current_player = currentPlayer
board.last_move = lastMove
best_policy = PolicyValueNetNumpy(width, height, policy_param)
mcts_player = MCTSPlayer(best_policy.policy_value_fn,
c_puct=5,
n_playout=400)
nextmove = mcts_player.get_action(board)
return nextmove
def run(states, sensible_moves, currentPlayer, lastMove):
#胜利所需要连续的子
n = 5
#棋盘宽度，高度
width, height = 8, 8
board = Board(width=width, height=height, n_in_row=n)
board.init_board()
board.states = states
board.availables = sensible_moves
board.current_player = currentPlayer
board.last_move = lastMove
#策略价值网络
best_policy = PolicyValueNetNumpy(width, height, policy_param)
#纯蒙特卡洛搜索
mcts_player = MCTSPlayer(best_policy.policy_value_fn,
c_puct=5,
n_playout=400)
#从蒙特卡洛搜索中返回下一步要走的地方
nextmove = mcts_player.get_action(board)
return nextmove
