-
Notifications
You must be signed in to change notification settings - Fork 0
/
zbot_minimax.py
225 lines (197 loc) · 8.32 KB
/
zbot_minimax.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
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
import chess
import random
import utils
import math
from time import time
from zbot_eval import eval
import numpy as np
import json
from joblib import Parallel, delayed
MINUS_INF = -math.inf
POS_INF = math.inf
seed = 91
def timer_func(func):
# This function shows the execution time of
# the function object passed
def wrap_func(*args, **kwargs):
t1 = time()
result = func(*args, **kwargs)
t2 = time()
print(f'Function {func.__name__!r} executed in {(t2-t1):.4f}s')
return result
return wrap_func
class Zbot_minimax():
def __init__(self, eval: eval, depth = 4, verbose = False):
self.depth = depth
self.eval = eval
self.nodes = 0
self.verbose = verbose
self.wins = 0
def get_move_random(self, board):
moves = [m for m in board.legal_moves]
if len(moves) > 0:
return moves[random.randrange(0, len(moves))]
def get_move(self, board: chess.Board):
if board.turn == chess.WHITE:
move, eval = self.maxi(board, d = self.depth, alpha = MINUS_INF, beta = POS_INF)
if self.verbose:
print(self.nodes)
self.nodes = 0
if move != None:
return move
else:
return self.get_move_random(board)
elif board.turn == chess.BLACK:
move, eval = self.mini(board, d = self.depth, alpha = MINUS_INF, beta = POS_INF)
if self.verbose:
print(self.nodes)
self.nodes = 0
if move != None:
return move
else:
return self.get_move_random(board)
def mini(self, board, d, alpha, beta):
self.nodes += 1
outcome = board.outcome()
non_quiet_moves, rest = utils.get_sorted_legal_moves(board.copy())
if outcome != None:
if outcome.termination == chess.Termination.CHECKMATE:
if outcome.winner == chess.BLACK:
return None, MINUS_INF
elif outcome.winner == chess.WHITE:
return None, POS_INF
else:
return None, 0
if d == 0 and non_quiet_moves == []:
return None, self.eval.evaluate(board)
best_eval = POS_INF
best_move = None
search_moves = non_quiet_moves + rest if d >= 0 else non_quiet_moves
for m in search_moves:
# for m in board.legal_moves:
board.push(m)
move, move_eval = self.maxi(board, d-1, alpha, beta)
if move_eval < best_eval:
best_eval = move_eval
best_move = m
board.pop()
if best_eval <= alpha:
return best_move, best_eval
if best_eval < beta:
beta = best_eval
return best_move, best_eval
def maxi(self, board, d, alpha, beta):
self.nodes += 1
outcome = board.outcome()
non_quiet_moves, rest = utils.get_sorted_legal_moves(board.copy())
if outcome != None:
if outcome.termination == chess.Termination.CHECKMATE:
if outcome.winner == chess.BLACK:
return None, MINUS_INF
elif outcome.winner == chess.WHITE:
return None, POS_INF
else:
return None, 0
if d == 0 and non_quiet_moves == []:
return None, self.eval.evaluate(board)
best_eval = MINUS_INF
best_move = None
search_moves = non_quiet_moves + rest if d >= 0 else non_quiet_moves
for m in search_moves:
# for m in board.legal_moves:
board.push(m)
move, move_eval = self.mini(board, d-1, alpha, beta)
if move_eval > best_eval:
best_eval = move_eval
best_move = m
board.pop()
if best_eval >= beta:
return best_move, best_eval
if best_eval > alpha:
alpha = best_eval
return best_move, best_eval
class Zbot_genetic():
def __init__(self, population_size = 10, generations = 5, d = 1):
self.population_size = population_size
self.generations = generations
self.depth = d
self.population = []
for i in range(self.population_size):
# W = np.random.default_rng(seed = i + seed).uniform(-1, 1, (12, 64))
# V = np.random.default_rng(seed = i + 2*seed).uniform(-1, 1, 12)
M = np.random.default_rng(seed = i + 2*seed).uniform(-100, 100, 12)
self.population.append(Zbot_minimax(eval = eval(M), depth = self.depth))
# Wa = np.load('./weights/Wa.npy')
# Va = np.load('./weights/Va.npy')
# self.population.append(Zbot_minimax(eval = eval(Wa, Va), depth = self.depth))
# Wb = np.load('./weights/Wb.npy')
# Vb = np.load('./weights/Vb.npy')
# self.population.append(Zbot_minimax(eval = eval(Wb, Vb), depth = self.depth))
# for i in range(2, self.population_size):
# e = eval.breed(self.population[0].eval, self.population[1].eval)
# e.mutate(mutation_rate = 0.1)
# self.population.append(Zbot_minimax(eval = e, depth = self.depth))
# Ma = np.load("./weights/Ma.npy")
# Mb = np.load("./weights/Mb.npy")
# self.population.append(Zbot_minimax(eval = eval(Ma), depth = self.depth))
# e = eval(Mb)
# while np.array_equal(Ma, Mb):
# e.mutate(mutation_rate = 0.5, mutation_factor = 0.5)
# print(Ma, "\n", Mb)
# self.population.append(Zbot_minimax(eval = e, depth = self.depth))
# for i in range(2, self.population_size):
# e = eval.breed(self.population[0].eval, self.population[1].eval)
# e.mutate(mutation_rate = 0.3)
# self.population.append(Zbot_minimax(eval = e, depth = self.depth))
def play(self, i, j):
if i != j:
A = self.population[i]
B = self.population[j]
result = utils.simulate_game(A, B, verbose = False)[0]
# print(i, j, result)
if result == 1:
# A.wins += 1
return i
elif result == -1:
# B.wins += 1
return j
def simulate_evolution(self):
for gen in range(self.generations):
# for i in range(self.population_size):
# for j in range(self.population_size):
# self.play(i, j)
result = Parallel(n_jobs = 6)(delayed(self.play)(i, j) for i in range(self.population_size) for j in range(self.population_size))
for i in result:
if i != None:
self.population[i].wins += 1
self.population.sort(key = lambda eval: eval.wins, reverse = True)
# np.save("./weights/Wa.npy", self.population[0].eval.W)
# np.save("./weights/Va.npy", self.population[0].eval.V)
# np.save("./weights/Wb.npy", self.population[1].eval.W)
# np.save("./weights/Vb.npy", self.population[1].eval.V)
while np.array_equal(self.population[0].eval.M, self.population[1].eval.M):
self.population[1].eval.mutate(mutation_rate = 0.3)
np.save("./weights/Ma.npy", self.population[0].eval.M)
np.save("./weights/Mb.npy", self.population[1].eval.M)
print([self.population[i].wins for i in range(self.population_size)])
self.population[0].wins = 0
self.population[1].wins = 0
for i in range(2, self.population_size):
e = eval.breed(self.population[0].eval, self.population[1].eval)
e.mutate(mutation_rate = 0.3 - 0.2*gen/self.generations, \
mutation_factor = 0.3 - 0.2*gen/self.generations)
self.population[i] = Zbot_minimax(eval = e, depth = self.depth)
# t1 = time()
# G = Zbot_genetic(population_size = 7, generations = 20, d = 1)
# G.simulate_evolution()
# t2 = time()
# print(f'{(t2-t1):.4f}')
# A = Zbot_minimax(eval = eval(W, V), depth = 2)
# W = np.random.default_rng().uniform(-1, 1, (12, 64))
# V = np.random.default_rng().uniform(-1, 1, 12)
# B = Zbot_minimax(eval = eval(W, V), depth = 2)
# t1 = time()
# board_history = utils.simulate_game(A, B, verbose = False)[1]
# t2 = time()
# print(f'{(t2-t1):.4f}')
# utils.view_game(board_history)