Lections/gym-tic-tac-toe
2
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Typo in win check

Browse Source
This commit is contained in:
Nicolai Czempin 2017-04-04 18:04:18 +02:00
parent a209d02d8b
commit acd515bc9a
1 changed files with 13 additions and 12 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

17
gym_tic_tac_toe/envs/tic_tac_toe_env.py
View File

@ -7,7 +7,7 @@ class TicTacToeEnv(gym.Env):
def __init__(self): def __init__(self):
self.action_space = spaces.Tuple((spaces.Discrete(2), spaces.Discrete(9))) self.action_space = spaces.Tuple((spaces.Discrete(2), spaces.Discrete(9)))
self.observation_space = spaces.Discrete(3)#Tuple(spaces.Discrete(3), spaces.Discrete(9)) self.observation_space = spaces.Discrete(3) # Tuple(spaces.Discrete(3), spaces.Discrete(9))
def _step(self, action): def _step(self, action):
done = False done = False
reward = 0 reward = 0
@ -16,9 +16,9 @@ class TicTacToeEnv(gym.Env):
# p = p*2 - 1 # p = p*2 - 1
# check move legality # check move legality
proposed = self.state['board'][square] board = self.state['board']
proposed = board[square]
om = self.state['on_move'] om = self.state['on_move']
print ("on move: ", om)
if (proposed != 0): # wrong player, not empty if (proposed != 0): # wrong player, not empty
print("illegal move ", action, ". (square occupied): ", square) print("illegal move ", action, ". (square occupied): ", square)
done = True done = True
@ -28,21 +28,22 @@ class TicTacToeEnv(gym.Env):
done = True done = True
reward = -2 * om # player who did NOT make the illegal move reward = -2 * om # player who did NOT make the illegal move
else: else:
self.state['board'][square] = p board[square] = p
self.state['on_move'] = -p self.state['on_move'] = -p
# check game over # check game over
for i in range(3): for i in range(3):
if (self.state['board'][i * 3] == p and self.state['board'][i*3 + 1] == p and self.state['board'][i*3+2] == 2): # horizontals and verticals
if ((board[i * 3] == p and board[i * 3 + 1] == p and board[i * 3 + 2 ] == p)
or (board[i + 0] == p and board[i + 3] == p and board[i + 6] == p)):
reward = p reward = p
done = True done = True
break break
#TODO other cases
return np.array(self.state), reward, done, {} return np.array(self.state), reward, done, {}
def _reset(self): def _reset(self):
self.state = {} self.state = {}
self.state['board'] = [0,0,0,0,0,0,0,0,0] self.state['board'] = [0, 0, 0, 0, 0, 0, 0, 0, 0]
self.state['on_move'] = 1 self.state['on_move'] = 1
return self.state return self.state
def _render(self, mode='human', close=False): def _render(self, mode='human', close=False):
@ -56,7 +57,7 @@ class TicTacToeEnv(gym.Env):
moves = [] moves = []
for i in range (9): for i in range (9):
if (self.state['board'][i]== 0): if (self.state['board'][i] == 0):
p = self.state['on_move'] p = self.state['on_move']
m = [p, i] m = [p, i]
moves.append(m) moves.append(m)