五子棋 AI 的贪心策略
算法1：落子策略
计算每个空格点对自己的价值，自己价值最大的格点位置 cPos 与价值 cVal
计算每个空格点对对手的价值，对手价值最大的格点位置 pPos 与价值 pVal
落子位置为 cVal > pVal ？ cPos ： pPos
算法2：计算空格点价值
遍历棋盘所有空格点：
如果没有空格点
平局，返回
当前点的价值 = 0
遍历当前格点的四个方向：右、右上、上、左上
如果在该方向上容许下满5子（没有出界或被对手占领）
统计该方向上自己棋子数 N
当前点的价值 += 10^N
如果相邻格点是自己棋子
当前点的价值 += 1
# coding:utf-8
from __future__ import print_function
import random
class Button:
def __init__(self, taken):
self.taken = taken
class Game:
def __init__(self, board):
self.NUMS = 19
self.buttons = {}
self.trans = ['.', 'c', 'p']
for i in range(self.NUMS):
for j in range(self.NUMS):
self.buttons[(i, j)] = Button(board[i][j])
self.nearpos = ((1, 0), (1, 1), (0, 1), (-1, 1))
def display(self):
print(' ', end='')
for j in range(self.NUMS):
print(j % 10, end='')
print()
for i in range(self.NUMS):
print(i % 10, end='')
for j in range(self.NUMS):
print(self.trans[self.buttons[(i, j)].taken + 1], end='')
print()
def compute(self, player):
bsites=[p for p,b in self.buttons.iteritems() if b.taken == -1]
bsites_values=[]
choice_num = len(bsites)
if choice_num == 0:
print('no one wins...game over')
return None
for bpos in bsites:
value = 0
for step_pos in self.nearpos:
_space = 0
for i in range(1,5):
npos = self._add(bpos,self._mul(step_pos,i))
if npos is None or self.buttons[npos].taken == 1-player:
break
else:
_space += 1
for i in range(1,5):
npos = self._add(bpos,self._mul(step_pos,-i))
if npos is None or self.buttons[npos].taken == 1-player:
break
else:
_space += 1
if _space < 4:
continue
max_step = 1
for i in range(-1,-5,-1):
_t = 1
npos = self._add(bpos,self._mul(step_pos,i))
if npos is None or self.buttons[npos].taken == 1-player:
break
for j in range(5):
npos = self._add(npos,step_pos)
if npos is None or self.buttons[npos].taken == 1-player:
break
if self.buttons[npos].taken == player:
_t += 1
if _t > max_step:
max_step = _t
value += 10**(max_step-1)
npos = self._add(bpos,step_pos)
if npos is not None :
if self.buttons[npos].taken == player:
value += 1
npos = self._add(bpos,self._mul(step_pos,-1))
if npos is not None :
if self.buttons[npos].taken == player:
value += 1
if value > 0:
bsites_values.append((bpos,value))
_maxvalue,_maxp=0,None
for p,v in bsites_values:
if v > _maxvalue:
_maxvalue,_maxp = v,p
if _maxp == None:
_maxp = random.choice(bsites)
return _maxvalue,_maxp
def _mul(self, pos1, n):
return (pos1[0] * n, pos1[1] * n)
def _add(self, pos1, pos2):
npos = (pos1[0] + pos2[0], pos1[1] + pos2[1])
if npos[0] >= self.NUMS or npos[1] >= self.NUMS or npos[0] < 0 or npos[1] < 0:
return None
return npos
def check_over(self):
# 横线
for rows in range(self.NUMS):
step_compute, step_person = 0, 0
for cols in range(self.NUMS):
npos = (rows, cols)
if self.buttons[npos].taken == 0:
step_compute += 1
step_person = 0
if step_compute >= 5:
print('c win')
return True
elif self.buttons[npos].taken == 1:
step_compute = 0
step_person += 1
if step_person >= 5:
print('p win')
return True
else:
step_compute, step_person = 0, 0
# 竖线
for cols in range(self.NUMS):
step_compute, step_person = 0, 0
for rows in range(self.NUMS):
npos = (rows, cols)
if self.buttons[npos].taken == 0:
step_compute += 1
step_person = 0
if step_compute >= 5:
print('c win')
return True
elif self.buttons[npos].taken == 1:
step_compute = 0
step_person += 1
if step_person >= 5:
print('p win')
return True
else:
step_compute, step_person = 0, 0
# 上斜线
for pos in [(row, 0) for row in range(self.NUMS)] + [(self.NUMS - 1, col) for col in range(self.NUMS)]:
_step = (-1, 1)
n = 0
step_compute, step_person = 0, 0
while 1:
npos = self._add(pos, self._mul(_step, n))
if npos is None:
break
if self.buttons[npos].taken == 0:
step_compute += 1
step_person = 0
if step_compute >= 5:
print('c win')
return True
elif self.buttons[npos].taken == 1:
step_compute = 0
step_person += 1
if step_person >= 5:
print('p win')
return True
else:
step_compute, step_person = 0, 0
n += 1
# 下斜线
for pos in [(row, 0) for row in range(self.NUMS)] + [(0, col) for col in range(self.NUMS)]:
_step = (1, 1)
n = 0
step_compute, step_person = 0, 0
while 1:
npos = self._add(pos, self._mul(_step, n))
if npos is None:
break
if self.buttons[npos].taken == 0:
step_compute += 1
step_person = 0
if step_compute >= 5:
print('c win')
return True
elif self.buttons[npos].taken == 1:
step_compute = 0
step_person += 1
if step_person >= 5:
print('p win')
return True
else:
step_compute, step_person = 0, 0
n += 1
return False
def next(self, player):
cv, cpos = self.compute(player)
pv, ppos = self.compute(1-player)
if cv > pv:
return cpos
else:
return ppos
if __name__ == '__main__':
# self.buttons[(NUMS/2,NUMS/2)].taken = 0
a=int(input('先手请输入1，后手请输入2：'))
board = [[-1 for _ in range(19)] for _ in range(19)]
if a==1:
board[9][9] = 0
game = Game(board)
game.display()
#print('(9, 9)')
while True:
x = int(input('input x：'))
y = int(input('input y：'))
board[x][y] = 1
game = Game(board)
pos = game.next(0)
board[pos[0]][pos[1]] = 0
game = Game(board)
game.display()
if game.check_over():
break