-
Notifications
You must be signed in to change notification settings - Fork 0
/
MCTS.py
169 lines (138 loc) · 6.27 KB
/
MCTS.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
import logging
import math
import heapq as hq
import random
EPS = 1e-8
log = logging.getLogger(__name__)
from china_chess.algorithm.china_chess_board import *
class MCTS:
"""
This class handles the MCTS tree.
"""
def __init__(self, game, nnet, args):
self.game = game
self.nnet = nnet
self.args = args
self.Qsa = {} # stores Q values for s,a (as defined in the paper)
self.Nsa = {} # stores #times edge s,a was visited
self.Ns = {} # stores #times board s was visited
self.Ps = {} # stores initial policy (returned by neural net)
self.Es = {} # stores game.getGameEnded ended for board s
self.Vs = {} # stores game.getValidMoves for board s
def getActionProb(self, canonicalBoard, iter_number, episodeStep, temp=1):
"""
This function performs numMCTSSims simulations of MCTS starting from
canonicalBoard.
Returns:
probs: a policy vector where the probability of the ith action is
proportional to Nsa[(s,a)]**(1./temp)
"""
for i in range(self.args.numMCTSSims):
self.search(canonicalBoard, i, [], 0, 0, iter_number)
s = self.game.stringRepresentation(canonicalBoard)
counts = [self.Nsa[(s, a)] if (s, a) in self.Nsa else 0 for a in range(self.game.getActionSize())]
if temp == 0:
bestAs = np.array(np.argwhere(counts == np.max(counts))).flatten()
bestA = np.random.choice(bestAs)
probs = [0] * len(counts)
probs[bestA] = 1
return probs
counts = [x ** (1. / temp) for x in counts]
counts_sum = float(sum(counts))
probs = [x / counts_sum for x in counts]
return probs
def _top_k(self, data_list, number):
y = sorted(data_list, key=lambda x: x[0], reverse=True)
if len(y) < number:
return y[0][1]
y = random.choices(y[:number])[0][1]
return y
@staticmethod
def is_draw(continue_steps):
if len(continue_steps) == 12:
if continue_steps[0] == continue_steps[4] and continue_steps[4] == continue_steps[8] and \
continue_steps[1] == continue_steps[5] and continue_steps[5] == continue_steps[9] and \
continue_steps[2] == continue_steps[6] and continue_steps[6] == continue_steps[10] and \
continue_steps[3] == continue_steps[7] and continue_steps[7] == continue_steps[11]:
return True
return False
def search(self, canonicalBoard, i, continue_steps, is_eat_param, times, iter_number):
"""
This function performs one iteration of MCTS. It is recursively called
till a leaf node is found. The action chosen at each node is one that
has the maximum upper confidence bound as in the paper.
Once a leaf node is found, the neural network is called to return an
initial policy P and a value v for the state. This value is propagated
up the search path. In case the leaf node is a terminal state, the
outcome is propagated up the search path. The values of Ns, Nsa, Qsa are
updated.
NOTE: the return values are the negative of the value of the current
state. This is done since v is in [-1,1] and if v is the value of a
state for the current player, then its value is -v for the other player.
Returns:
v: the negative of the value of the current canonicalBoard
"""
if is_eat_param >= 60:
return 0
if MCTS.is_draw(continue_steps):
return 0
s = self.game.stringRepresentation(canonicalBoard)
if s not in self.Es:
self.Es[s] = self.game.getGameEnded(canonicalBoard, 1)
if self.Es[s][0]:
# terminal node
return -self.Es[s][1]
if s not in self.Ps:
# leaf node
self.Ps[s], v = self.nnet.predict(canonicalBoard)
valids = self.game.getValidMoves(canonicalBoard, 1)
self.Ps[s] = self.Ps[s] * valids # masking invalid moves
sum_Ps_s = np.sum(self.Ps[s])
if sum_Ps_s > 0:
self.Ps[s] /= sum_Ps_s # renormalize
else:
# if all valid moves were masked make all valid moves equally probable
# NB! All valid moves may be masked if either your NNet architecture is insufficient or you've get overfitting or something else.
# If you have got dozens or hundreds of these messages you should pay attention to your NNet and/or training process.
log.error("All valid moves were masked, doing a workaround.")
self.Ps[s] = self.Ps[s] + valids
self.Ps[s] /= np.sum(self.Ps[s])
self.Vs[s] = valids
self.Ns[s] = 0
return -v
valids = self.Vs[s]
cur_best = -float('inf')
best_act = -1
temp_list = []
# pick the action with the highest upper confidence bound
for a in range(self.game.getActionSize()):
if valids[a]:
if (s, a) in self.Qsa:
u = self.Qsa[(s, a)] + self.args.cpuct * self.Ps[s][a] * math.sqrt(self.Ns[s]) / (
1 + self.Nsa[(s, a)])
else:
u = self.args.cpuct * self.Ps[s][a] * math.sqrt(self.Ns[s] + EPS) # Q = 0 ?
temp_list.append((u, a))
if u > cur_best:
cur_best = u
best_act = a
a = best_act
# a = self._top_k(temp_list, 2)
next_s, next_player, is_eat = self.game.getNextState(canonicalBoard, 1, a)
next_s = self.game.getCanonicalForm(next_s, next_player)
if len(continue_steps) == 12:
del continue_steps[0]
continue_steps.append(a)
if is_eat:
is_eat_param = 0
else:
is_eat_param += 1
v = self.search(next_s, i, continue_steps, is_eat_param, times + 1, iter_number)
if (s, a) in self.Qsa:
self.Qsa[(s, a)] = (self.Nsa[(s, a)] * self.Qsa[(s, a)] + v) / (self.Nsa[(s, a)] + 1)
self.Nsa[(s, a)] += 1
else:
self.Qsa[(s, a)] = v
self.Nsa[(s, a)] = 1
self.Ns[s] += 1
return -v