forked from Urinx/ReinforcementLearning
-
Notifications
You must be signed in to change notification settings - Fork 0
/
MCTS_Gomoku.py
198 lines (157 loc) · 5.94 KB
/
MCTS_Gomoku.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
from math import *
import random
import numpy as np
class GameState:
def __init__(self):
self.player_just_moved = 2
def clone(self):
st = GameState()
st.player_just_moved = self.player_just_moved
return st
def move(self, action):
self.player_just_moved = 3 - self.player_just_moved
def actions(self):
""" Get all possible moves from this state.
"""
def win(self, player):
""" Get the game result from the viewpoint of player.
"""
def end(self):
""" Whether the game is end or not
"""
def __repr__(self):
pass
class Gomoku(GameState):
def __init__(self, w=8): # 15x15
self.player_just_moved = 2
self.board = [] # 0 = empty, 1 = player 1 (X), 2 = player 2 (O)
self.w = w
for y in range(w):
self.board.append([0] * w)
def clone(self):
st = Gomoku()
st.player_just_moved = self.player_just_moved
st.board = [self.board[i][:] for i in range(self.w)]
st.w = self.w
return st
def move(self, action):
a, b = action
assert 0 <= a <= self.w and 0 <= b <= self.w and self.board[a][b] == 0
self.player_just_moved = 3 - self.player_just_moved
self.board[a][b] = self.player_just_moved
def actions(self):
return [(i, j) for i in range(self.w) for j in range(self.w) if self.board[i][j] == 0]
def check_five(self, i, j, player):
if 2 <= i < self.w-2 and 2 <= j < self.w-2 and self.board[i-2][j-2] == self.board[i-1][j-1] == self.board[i][j] == self.board[i+1][j+1] == self.board[i+2][j+2] == player:
return 1
elif 2 <= j < self.w-2 and self.board[i][j-2] == self.board[i][j-1] == self.board[i][j] == self.board[i][j+1] == self.board[i][j+2] == player:
return 1
elif 2 <= i < self.w-2 and 2 <= j < self.w-2 and self.board[i+2][j-2] == self.board[i+1][j-1] == self.board[i][j] == self.board[i-1][j+1] == self.board[i-2][j+2] == player:
return 1
elif 2 <= i < self.w-2 and self.board[i-2][j] == self.board[i-1][j] == self.board[i][j] == self.board[i+1][j] == self.board[i+2][j] == player:
return 1
return 0
def win(self, player):
for i in range(self.w):
for j in range(self.w):
if self.check_five(i, j, player):
return 1
elif self.check_five(i, j, 3-player):
return 0
if self.actions() == []: return 0.5
return -1
def end(self):
return self.win(1) >= 0
def __repr__(self):
row = '{:>2} ' + ' | '.join(['{}'] * self.w) + ' '
line = '\n ' + ('----' * self.w)[:-1] + '\n'
s = ' ' + '%2d ' * self.w % tuple(range(self.w)) + '\n'
s += line.join([row.format(i, *map(lambda j: [' ', 'X', 'O'][j], self.board[i])) for i in range(self.w)])
return s
class Node:
def __init__(self, action=None, parent=None, state=None):
self.action = action
self.parent = parent
self.childs = []
self.W = 0
self.N = 0
self.untried_actions = state.actions()
self.player_just_moved = state.player_just_moved
def select(self):
s = sorted(self.childs, key = lambda c: c.U())[-1]
return s
def add_child(self, a, s):
n = Node(a, self, s)
self.untried_actions.remove(a)
self.childs.append(n)
return n
def update(self, result):
self.N += 1
self.W += result
def U(self):
if self.parent:
return self.W / self.N + sqrt(2 * log(self.parent.N) / self.N)
return 0
def __repr__(self):
return "[A: %s, U: %.2f, W/N: %.1f/%d, Untried: %s]" \
% (self.action, self.U(), self.W, self.N, self.untried_actions)
def show_node_tree(self, indent=0):
print("| " * indent + str(self))
for c in self.childs:
c.show_node_tree(indent+1)
def show_children_nodes(self):
print('\n[*] Child Nodes')
for c in self.childs: print(c)
def UCT(rootstate, itermax, verbose=False):
rootnode = Node(state=rootstate)
for i in range(itermax):
node = rootnode
state = rootstate.clone()
# Select
while node.untried_actions == [] and node.childs != []:
node = node.select()
state.move(node.action)
# Expand
if node.untried_actions != []:
action = random.choice(node.untried_actions)
state.move(action)
node = node.add_child(action, state)
# Rollout
while state.actions() != []:
state.move(random.choice(state.actions()))
# Backpropagate
while node != None:
node.update(state.win(node.player_just_moved))
node = node.parent
if verbose: rootnode.show_node_tree()
else: rootnode.show_children_nodes()
return sorted(rootnode.childs, key = lambda c: c.N)[-1].action
def random_play(game):
return random.choice(game.actions())
def human_play():
t = input('[*] Your turn (i j): ')
a, b = t.split(' ')
i, j = int(a), int(b)
return (i, j)
def play_game():
game = Gomoku()
while not game.end():
print(game)
if game.player_just_moved == 1:
# action = UCT(game, 1000) # Player O
action = random_play(game)
else:
action = UCT(game, 10000) # Player X
# action = human_play()
game.move(action)
print("[*] Player %s move: %s\n" % (['X', 'O'][game.player_just_moved-1], action))
print(game)
r = game.win(game.player_just_moved)
if r == 1:
print("[*] Player %s win" % ['X', 'O'][game.player_just_moved-1])
elif r == 0:
print("[*] Player %s win" % ['X', 'O'][2-game.player_just_moved])
else:
print("[*] Player draw")
if __name__ == "__main__":
play_game()