TicTacToe 和 Minimax
TicTacToe and Minimax
我是一名年轻的程序员，正在学习 python 并努力实现 AI（使用 minimax）来玩 TicTacToe。我开始在网上看教程，但是教程在 JavaScript 上，因此无法解决我的问题。我也看过这个问题 ( Python minimax for tictactoe )，但它没有任何答案，而且实现方式与我的有很大不同。
编辑：您将在下面找到的代码是其中一个答案 (@water_ghosts) 建议的编辑。
编辑 #2：我删除了 possiblePositions，因为 AI 应该从 possiblePositions 中选择一个自由字段而不是一个位置（这在实现 minimax 时不会让它变得那么聪明 :)）
现在代码完全没有给出任何错误并且可以正常运行，但有一件小事：AI 总是选择下一个可用字段。例如，在我远离获胜的情况下，它不会阻止我的获胜选项，而是选择下一个空闲位置。
如果您想知道 elements dict 在那里做什么：我只是想确保程序选择了最佳索引...
这是我的代码：
class TicTacToe:
def __init__(self):
self.board = [" ", " ", " ", " ", " ", " ", " ", " ", " "]
self.playerSymbol = ""
self.playerPosition = []
self.aiSymbol = ""
self.aiPosition = []
self.score = 0
self.winner = None
self.scoreBoard = {
self.playerSymbol: -1,
self.aiSymbol: 1,
"tie": 0
}
self.turn = 0
self.optimalMove = int()
def drawBoard(self):
print(self.board[0] + " | " + self.board[1] + " | " + self.board[2])
print("___" + "___" + "___")
print(self.board[3] + " | " + self.board[4] + " | " + self.board[5])
print("___" + "___" + "___")
print(self.board[6] + " | " + self.board[7] + " | " + self.board[8])
def choice(self):
answer = input("What do you want to play as? (type x or o) ")
if answer.upper() == "X":
self.playerSymbol = "X"
self.aiSymbol = "O"
else:
self.playerSymbol = "O"
self.aiSymbol = "X"
def won(self):
winningPositions = [{0, 1, 2}, {3, 4, 5}, {6, 7, 8}, {0, 4, 8}, {2, 4, 6}, {0, 3, 6}, {1, 4, 7}, {2, 5, 8}]
for position in winningPositions:
if position.issubset(self.playerPosition):
self.winner = self.playerSymbol
print("Player Wins :)")
return True
elif position.issubset(self.aiPosition):
self.winner = self.aiSymbol
print("AI wins :(")
return True
if self.board.count(" ") == 0:
self.winner = "tie"
print("Guess it's a draw")
return True
return False
def findOptimalPosition(self):
bestScore = float("-Infinity")
elements = {} # desperate times call for desperate measures
for i in range(9):
if self.board[i] == " ":
self.board[i] = self.aiSymbol # AI quasi made the move here
if self.minimax(True) > bestScore:
bestScore = self.score
elements[i] = bestScore
self.board[i] = " "
return max(elements, key=lambda k: elements[k])
def minimax(self, isMaximizing):
if self.winner is not None:
return self.scoreBoard[self.winner]
if isMaximizing:
bestScore = float("-Infinity")
for i in range(9):
if self.board[i] == " ":
self.board[i] = self.aiSymbol
bestScore = max(self.minimax(False), bestScore)
self.board[i] = " "
return bestScore
else:
bestScore = float("Infinity")
for i in range(9):
if self.board[i] == " ":
self.board[i] = self.playerSymbol
bestScore = min(self.minimax(True), bestScore)
self.board[i] = " "
return bestScore
def play(self):
self.choice()
while not self.won():
if self.turn % 2 == 0:
pos = int(input("Where would you like to play? (0-8) "))
self.playerPosition.append(pos)
self.board[pos] = self.playerSymbol
self.turn += 1
self.drawBoard()
else:
aiTurn = self.findOptimalPosition()
self.aiPosition.append(aiTurn)
self.board[aiTurn] = self.aiSymbol
self.turn += 1
print("\n")
print("\n")
self.drawBoard()
else:
print("Thanks for playing :)")
tictactoe = TicTacToe()
tictactoe.play()
我来自 java 背景，对此不习惯 :(
任何帮助将不胜感激
我乐于接受改进我的代码和解决此问题的建议和方法。
在此先感谢并保持健康，
克里斯蒂
改变这部分，你的实现将 return optimalMove 即使它没有进入 if statement，并且 optimalMove 不会在那个时候分配, 所以把 return 放在里面。
if score > sampleScore:
sampleScore = score
optimalMove = i
return optimalMove
play() 中的 optimalMove = 0 和 findOptimalField() 中的 optimalMove = i 声明了两个不同的变量，每个变量都是声明它的函数的局部变量。
如果您希望多个函数访问同一个变量，您可以使用 global 关键字，但这通常被认为是一种不好的做法。它会使代码难以推理（例如 var = x 是创建一个新的局部变量还是覆盖全局变量的值？）并且它不会阻止您在变量声明之前意外使用它。
由于您来自 Java 背景，您可以将其转变为 class 以获得更符合您预期的行为，从而消除对全局变量的需求：
class TicTacToe:
def __init__(self):
self.board = [" ", " ", " ", " ", " ", " ", " ", " ", " "]
self.playerSymbol = ""
self.playerPosition = []
self.aiSymbol = ""
self.aiPosition = []
self.score = 0
self.playerSymbol = None
self.aiSymbol = None
...
def drawBoard(self):
print(self.board[0] + " | " + self.board[1] + " | " + self.board[2])
...
def choice(self):
answer = input("What do you want to play as? (type x or o) ")
if answer.upper() == "X":
self.playerSymbol = "X"
self.aiSymbol = "O"
...
每个方法现在都采用一个显式 self 参数来引用当前实例，您可以使用它来访问属于 class 实例而不是特定方法的任何变量。如果您不在变量前包含 self. ，该变量对于声明它的方法仍然是局部的。在这种情况下，drawBoard() 方法将无法访问 choice() 中定义的 answer 变量。
您可以在 class 的任何方法中创建新的 self. 变量，但最佳做法是在 __init__ 构造函数方法中初始化所有变量，使用 None 作为还没有值的变量的占位符。
我将此作为答案发布，以防万一将来有人遇到同样的问题:)
我遇到的主要问题（除了我糟糕的编程风格之外）是我忘记更新列表 playerPosition 和 aiPosition 的内容。
您可以查看工作代码中的其余更改：
class TicTacToe:
def __init__(self):
self.board = [" ", " ", " ", " ", " ", " ", " ", " ", " "]
self.playerSymbol = ""
self.playerPosition = []
self.aiSymbol = ""
self.aiPosition = []
self.winner = None
self.scoreBoard = None
self.turn = 0
self.optimalMove = int()
def drawBoard(self):
print(self.board[0] + " | " + self.board[1] + " | " + self.board[2])
print("___" + "___" + "___")
print(self.board[3] + " | " + self.board[4] + " | " + self.board[5])
print("___" + "___" + "___")
print(self.board[6] + " | " + self.board[7] + " | " + self.board[8])
def choice(self):
answer = input("What do you want to play as? (type x or o) ")
if answer.upper() == "X":
self.playerSymbol = "X"
self.aiSymbol = "O"
else:
self.playerSymbol = "O"
self.aiSymbol = "X"
self.scoreBoard = {
self.playerSymbol: -1,
self.aiSymbol: 1,
"tie": 0
}
def availableMoves(self):
moves = []
for i in range(0, len(self.board)):
if self.board[i] == " ":
moves.append(i)
return moves
def won_print(self):
self.won()
if self.winner == self.aiSymbol:
print("AI wins :(")
exit(0)
elif self.winner == self.playerSymbol:
print("Player Wins :)")
exit(0)
elif self.winner == "tie":
print("Guess it's a draw")
exit(0)
def won(self):
winningPositions = [{0, 1, 2}, {3, 4, 5}, {6, 7, 8},
{0, 4, 8}, {2, 4, 6}, {0, 3, 6},
{1, 4, 7}, {2, 5, 8}]
for position in winningPositions:
if position.issubset(self.playerPosition):
self.winner = self.playerSymbol
return True
elif position.issubset(self.aiPosition):
self.winner = self.aiSymbol
return True
if self.board.count(" ") == 0:
self.winner = "tie"
return True
self.winner = None
return False
def set_i_ai(self, i):
self.aiPosition.append(i)
self.board[i] = self.aiSymbol
def set_clear_for_ai(self, i):
self.aiPosition.remove(i)
self.board[i] = " "
def set_i_player(self, i):
self.playerPosition.append(i)
self.board[i] = self.playerSymbol
def set_clear_for_player(self, i):
self.playerPosition.remove(i)
self.board[i] = " "
def findOptimalPosition(self):
bestScore = float("-Infinity")
elements = {} # desperate times call for desperate measures
for i in self.availableMoves():
self.set_i_ai(i)
score = self.minimax(False)
if score > bestScore:
bestScore = score
elements[i] = bestScore
self.set_clear_for_ai(i)
if bestScore == 1:
print("you messed up larry")
elif bestScore == 0:
print("hm")
else:
print("whoops i made a prog. error")
return max(elements, key=lambda k: elements[k])
def minimax(self, isMaximizing):
if self.won():
return self.scoreBoard[self.winner]
if isMaximizing:
bestScore = float("-Infinity")
for i in self.availableMoves():
self.set_i_ai(i)
bestScore = max(self.minimax(False), bestScore)
self.set_clear_for_ai(i)
return bestScore
else:
bestScore = float("Infinity")
for i in self.availableMoves():
self.set_i_player(i)
bestScore = min(self.minimax(True), bestScore)
self.set_clear_for_player(i)
return bestScore
def play(self):
self.choice()
while not self.won_print():
if self.turn % 2 == 0:
pos = int(input("Where would you like to play? (0-8) "))
self.playerPosition.append(pos)
self.board[pos] = self.playerSymbol
self.turn += 1
self.drawBoard()
else:
aiTurn = self.findOptimalPosition()
self.aiPosition.append(aiTurn)
self.board[aiTurn] = self.aiSymbol
self.turn += 1
print("\n")
print("\n")
self.drawBoard()
else:
print("Thanks for playing :)")
if __name__ == '__main__':
tictactoe = TicTacToe()
tictactoe.play()
但如前所述，代码可能有效，但在逻辑和结构方面存在很多问题，因此请不要直接复制粘贴 :))
我是一名年轻的程序员，正在学习 python 并努力实现 AI（使用 minimax）来玩 TicTacToe。我开始在网上看教程，但是教程在 JavaScript 上，因此无法解决我的问题。我也看过这个问题 ( Python minimax for tictactoe )，但它没有任何答案，而且实现方式与我的有很大不同。
编辑：您将在下面找到的代码是其中一个答案 (@water_ghosts) 建议的编辑。
编辑 #2：我删除了 possiblePositions，因为 AI 应该从 possiblePositions 中选择一个自由字段而不是一个位置（这在实现 minimax 时不会让它变得那么聪明 :)）
现在代码完全没有给出任何错误并且可以正常运行，但有一件小事：AI 总是选择下一个可用字段。例如，在我远离获胜的情况下，它不会阻止我的获胜选项，而是选择下一个空闲位置。
如果您想知道 elements dict 在那里做什么：我只是想确保程序选择了最佳索引...
这是我的代码：
class TicTacToe:
def __init__(self):
self.board = [" ", " ", " ", " ", " ", " ", " ", " ", " "]
self.playerSymbol = ""
self.playerPosition = []
self.aiSymbol = ""
self.aiPosition = []
self.score = 0
self.winner = None
self.scoreBoard = {
self.playerSymbol: -1,
self.aiSymbol: 1,
"tie": 0
}
self.turn = 0
self.optimalMove = int()
def drawBoard(self):
print(self.board[0] + " | " + self.board[1] + " | " + self.board[2])
print("___" + "___" + "___")
print(self.board[3] + " | " + self.board[4] + " | " + self.board[5])
print("___" + "___" + "___")
print(self.board[6] + " | " + self.board[7] + " | " + self.board[8])
def choice(self):
answer = input("What do you want to play as? (type x or o) ")
if answer.upper() == "X":
self.playerSymbol = "X"
self.aiSymbol = "O"
else:
self.playerSymbol = "O"
self.aiSymbol = "X"
def won(self):
winningPositions = [{0, 1, 2}, {3, 4, 5}, {6, 7, 8}, {0, 4, 8}, {2, 4, 6}, {0, 3, 6}, {1, 4, 7}, {2, 5, 8}]
for position in winningPositions:
if position.issubset(self.playerPosition):
self.winner = self.playerSymbol
print("Player Wins :)")
return True
elif position.issubset(self.aiPosition):
self.winner = self.aiSymbol
print("AI wins :(")
return True
if self.board.count(" ") == 0:
self.winner = "tie"
print("Guess it's a draw")
return True
return False
def findOptimalPosition(self):
bestScore = float("-Infinity")
elements = {} # desperate times call for desperate measures
for i in range(9):
if self.board[i] == " ":
self.board[i] = self.aiSymbol # AI quasi made the move here
if self.minimax(True) > bestScore:
bestScore = self.score
elements[i] = bestScore
self.board[i] = " "
return max(elements, key=lambda k: elements[k])
def minimax(self, isMaximizing):
if self.winner is not None:
return self.scoreBoard[self.winner]
if isMaximizing:
bestScore = float("-Infinity")
for i in range(9):
if self.board[i] == " ":
self.board[i] = self.aiSymbol
bestScore = max(self.minimax(False), bestScore)
self.board[i] = " "
return bestScore
else:
bestScore = float("Infinity")
for i in range(9):
if self.board[i] == " ":
self.board[i] = self.playerSymbol
bestScore = min(self.minimax(True), bestScore)
self.board[i] = " "
return bestScore
def play(self):
self.choice()
while not self.won():
if self.turn % 2 == 0:
pos = int(input("Where would you like to play? (0-8) "))
self.playerPosition.append(pos)
self.board[pos] = self.playerSymbol
self.turn += 1
self.drawBoard()
else:
aiTurn = self.findOptimalPosition()
self.aiPosition.append(aiTurn)
self.board[aiTurn] = self.aiSymbol
self.turn += 1
print("\n")
print("\n")
self.drawBoard()
else:
print("Thanks for playing :)")
tictactoe = TicTacToe()
tictactoe.play()
我来自 java 背景，对此不习惯 :( 任何帮助将不胜感激
我乐于接受改进我的代码和解决此问题的建议和方法。 在此先感谢并保持健康， 克里斯蒂
改变这部分，你的实现将 return optimalMove 即使它没有进入 if statement，并且 optimalMove 不会在那个时候分配, 所以把 return 放在里面。
if score > sampleScore:
sampleScore = score
optimalMove = i
return optimalMove
play() 中的 optimalMove = 0 和 findOptimalField() 中的 optimalMove = i 声明了两个不同的变量，每个变量都是声明它的函数的局部变量。
如果您希望多个函数访问同一个变量，您可以使用 global 关键字，但这通常被认为是一种不好的做法。它会使代码难以推理（例如 var = x 是创建一个新的局部变量还是覆盖全局变量的值？）并且它不会阻止您在变量声明之前意外使用它。
由于您来自 Java 背景，您可以将其转变为 class 以获得更符合您预期的行为，从而消除对全局变量的需求：
class TicTacToe:
def __init__(self):
self.board = [" ", " ", " ", " ", " ", " ", " ", " ", " "]
self.playerSymbol = ""
self.playerPosition = []
self.aiSymbol = ""
self.aiPosition = []
self.score = 0
self.playerSymbol = None
self.aiSymbol = None
...
def drawBoard(self):
print(self.board[0] + " | " + self.board[1] + " | " + self.board[2])
...
def choice(self):
answer = input("What do you want to play as? (type x or o) ")
if answer.upper() == "X":
self.playerSymbol = "X"
self.aiSymbol = "O"
...
每个方法现在都采用一个显式 self 参数来引用当前实例，您可以使用它来访问属于 class 实例而不是特定方法的任何变量。如果您不在变量前包含 self. ，该变量对于声明它的方法仍然是局部的。在这种情况下，drawBoard() 方法将无法访问 choice() 中定义的 answer 变量。
您可以在 class 的任何方法中创建新的 self. 变量，但最佳做法是在 __init__ 构造函数方法中初始化所有变量，使用 None 作为还没有值的变量的占位符。
我将此作为答案发布，以防万一将来有人遇到同样的问题:)
我遇到的主要问题（除了我糟糕的编程风格之外）是我忘记更新列表 playerPosition 和 aiPosition 的内容。 您可以查看工作代码中的其余更改：
class TicTacToe:
def __init__(self):
self.board = [" ", " ", " ", " ", " ", " ", " ", " ", " "]
self.playerSymbol = ""
self.playerPosition = []
self.aiSymbol = ""
self.aiPosition = []
self.winner = None
self.scoreBoard = None
self.turn = 0
self.optimalMove = int()
def drawBoard(self):
print(self.board[0] + " | " + self.board[1] + " | " + self.board[2])
print("___" + "___" + "___")
print(self.board[3] + " | " + self.board[4] + " | " + self.board[5])
print("___" + "___" + "___")
print(self.board[6] + " | " + self.board[7] + " | " + self.board[8])
def choice(self):
answer = input("What do you want to play as? (type x or o) ")
if answer.upper() == "X":
self.playerSymbol = "X"
self.aiSymbol = "O"
else:
self.playerSymbol = "O"
self.aiSymbol = "X"
self.scoreBoard = {
self.playerSymbol: -1,
self.aiSymbol: 1,
"tie": 0
}
def availableMoves(self):
moves = []
for i in range(0, len(self.board)):
if self.board[i] == " ":
moves.append(i)
return moves
def won_print(self):
self.won()
if self.winner == self.aiSymbol:
print("AI wins :(")
exit(0)
elif self.winner == self.playerSymbol:
print("Player Wins :)")
exit(0)
elif self.winner == "tie":
print("Guess it's a draw")
exit(0)
def won(self):
winningPositions = [{0, 1, 2}, {3, 4, 5}, {6, 7, 8},
{0, 4, 8}, {2, 4, 6}, {0, 3, 6},
{1, 4, 7}, {2, 5, 8}]
for position in winningPositions:
if position.issubset(self.playerPosition):
self.winner = self.playerSymbol
return True
elif position.issubset(self.aiPosition):
self.winner = self.aiSymbol
return True
if self.board.count(" ") == 0:
self.winner = "tie"
return True
self.winner = None
return False
def set_i_ai(self, i):
self.aiPosition.append(i)
self.board[i] = self.aiSymbol
def set_clear_for_ai(self, i):
self.aiPosition.remove(i)
self.board[i] = " "
def set_i_player(self, i):
self.playerPosition.append(i)
self.board[i] = self.playerSymbol
def set_clear_for_player(self, i):
self.playerPosition.remove(i)
self.board[i] = " "
def findOptimalPosition(self):
bestScore = float("-Infinity")
elements = {} # desperate times call for desperate measures
for i in self.availableMoves():
self.set_i_ai(i)
score = self.minimax(False)
if score > bestScore:
bestScore = score
elements[i] = bestScore
self.set_clear_for_ai(i)
if bestScore == 1:
print("you messed up larry")
elif bestScore == 0:
print("hm")
else:
print("whoops i made a prog. error")
return max(elements, key=lambda k: elements[k])
def minimax(self, isMaximizing):
if self.won():
return self.scoreBoard[self.winner]
if isMaximizing:
bestScore = float("-Infinity")
for i in self.availableMoves():
self.set_i_ai(i)
bestScore = max(self.minimax(False), bestScore)
self.set_clear_for_ai(i)
return bestScore
else:
bestScore = float("Infinity")
for i in self.availableMoves():
self.set_i_player(i)
bestScore = min(self.minimax(True), bestScore)
self.set_clear_for_player(i)
return bestScore
def play(self):
self.choice()
while not self.won_print():
if self.turn % 2 == 0:
pos = int(input("Where would you like to play? (0-8) "))
self.playerPosition.append(pos)
self.board[pos] = self.playerSymbol
self.turn += 1
self.drawBoard()
else:
aiTurn = self.findOptimalPosition()
self.aiPosition.append(aiTurn)
self.board[aiTurn] = self.aiSymbol
self.turn += 1
print("\n")
print("\n")
self.drawBoard()
else:
print("Thanks for playing :)")
if __name__ == '__main__':
tictactoe = TicTacToe()
tictactoe.play()
但如前所述，代码可能有效，但在逻辑和结构方面存在很多问题，因此请不要直接复制粘贴 :))