Spaces:
Runtime error
Runtime error
File size: 7,021 Bytes
b3accf9 |
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 |
import tqdm
from classes.game_model import tic_tac_toe_model
from classes.Qlearningagent import QlearningAgent
class environment():
def __init__(self, tic_tac_toe: tic_tac_toe_model, q_agent: QlearningAgent, train: bool, show_stats: bool):
self.board = tic_tac_toe
self.q_agent = q_agent
self.train = train
self.show_stats = show_stats
def play_one_game(self, piece):
game_over = False
win_piece = 0
self.board.reset_matrix()
if piece == 1:
piece_enemy = 2
states_x = []
rewards_x = []
while not game_over:
w = self.board.check_win()
if w != 4:
state = self.board.matriz.copy()
reward_x = self.board.reward_piece(piece)
state_x = (tuple(state.flatten()),-1,piece)# -1 for terminal state
states_x.append(state_x)
rewards_x.append(reward_x)
win_piece = w
break
#x move and state/reward/move dynamic
state = self.board.matriz.copy()
avaible_moves = self.board.get_avaible_moves()
action_x = self.q_agent.choose_move(state,avaible_moves,piece)
i, j = self.board.number_ij(action_x)
self.board.move(i,j,piece)
reward_x = self.board.reward_piece(piece)
state_x = (tuple(state.flatten()),action_x,piece)
states_x.append(state_x)
rewards_x.append(reward_x)
w = self.board.check_win()
if w != 4:
state = self.board.matriz.copy()
reward_x = self.board.reward_piece(piece)
state_x = (tuple(state.flatten()),-1,piece)# -1 for terminal state
states_x.append(state_x)
rewards_x.append(reward_x)
win_piece = w
break
self.board.get_random_move(piece_enemy)
self.q_agent.update_q_value(states_x,rewards_x)
return win_piece
else:
piece_enemy = 1
states_o = []
rewards_o = []
while not game_over:
self.board.get_random_move(piece_enemy)
w = self.board.check_win()
if w != 4:
state = self.board.matriz.copy()
reward_o = self.board.reward_piece(piece)
state_o = (tuple(state.flatten()),-1,piece)# -1 for terminal state
states_o.append(state_o)
rewards_o.append(reward_o)
win_piece = w
break
state = self.board.matriz.copy()
avaible_moves = self.board.get_avaible_moves()
action_o = self.q_agent.choose_move(state,avaible_moves,piece)
i, j = self.board.number_ij(action_o)
self.board.move(i,j,piece)
reward_o = self.board.reward_piece(piece)
state_o = (tuple(state.flatten()), action_o, piece)
states_o.append(state_o)
rewards_o.append(reward_o)
w = self.board.check_win()
if w != 4:
state = self.board.matriz.copy()
reward_o = self.board.reward_piece(piece)
state_o = (tuple(state.flatten()),-1,piece)# -1 for terminal state
states_o.append(state_o)
rewards_o.append(reward_o)
win_piece = w
break
self.q_agent.update_q_value(states_o,rewards_o)
return win_piece
def play_ia_vs_ia(self):
game_over = False
self.board.reset_matrix()
ia_x = 1
ia_o = 2
states_x = []
rewards_x = []
states_o = []
rewards_o = []
while not game_over:
w = self.board.check_win()
if w != 4:
win_piece = w
break
state_x = self.board.matriz.copy()
avaible_moves_x = self.board.get_avaible_moves()
action_x = self.q_agent.choose_move(state_x,avaible_moves_x,ia_x)
i, j = self.board.number_ij(action_x)
self.board.move(i,j,ia_x) # x play
#x state/reward
reward_x = self.board.reward_piece(ia_x)
state_x = (tuple(state_x.flatten()),action_x,ia_x)
states_x.append(state_x)
rewards_x.append(reward_x)
w = self.board.check_win()
if w != 4:
win_piece = w
break
state_o = self.board.matriz.copy()
avaible_moves_o = self.board.get_avaible_moves()
action_o = self.q_agent.choose_move(state_o,avaible_moves_o,ia_o)
i, j = self.board.number_ij(action_o)
self.board.move(i,j,ia_o) # o play
reward_o = self.board.reward_piece(ia_o)
state_o = (tuple(state_o.flatten()),action_o,ia_o)
states_o.append(state_o)
rewards_o.append(reward_o)
if win_piece == 1:
state = self.board.matriz.copy()
reward_o = self.board.reward_piece(ia_o)
state_o = (tuple(state.flatten()),-1,ia_o)# -1 for terminal state
states_o.append(state_o)
rewards_o.append(reward_o)
elif win_piece == 2:
state = self.board.matriz.copy()
reward_x = self.board.reward_piece(ia_x)
state_x = (tuple(state.flatten()),-1,ia_x)# -1 for terminal state
states_x.append(state_x)
rewards_x.append(reward_x)
self.q_agent.update_q_value(states_x,rewards_x)
self.q_agent.update_q_value(states_o,rewards_o)
return win_piece
def run(self, n):
wins_x = []
wins_o = []
wins_ia = []
if self.show_stats:
print(f'Playing {n} games with X')
for i in tqdm.tqdm(range(0,n)):
wins_x.append(self.play_one_game(piece=1))
print(f'Playing {n} games with O')
for i in tqdm.tqdm(range(0,n)):
wins_o.append(self.play_one_game(piece=2))
print(f'Playing {n} games ia vs ia')
for i in tqdm.tqdm(range(0,n)):
wins_ia.append(self.play_ia_vs_ia())
else:
for i in range(0,n):
wins_x.append(self.play_one_game(piece=1))
for i in range(0,n):
wins_o.append(self.play_one_game(piece=2))
for i in range(0,n):
wins_ia.append(self.play_ia_vs_ia())
return wins_x,wins_o, wins_ia |