def deterministic_game(qtable1, qtable2):
wins = 0
loses = 0
ties = 0
for game_num in range(6):
starting_player = game_num % 2
first_mv = game_num % 3
mv_num = 2
pos_num = first_mv
end = False
while (not end) and (mv_num <= 9):
if mv_num % 2 == starting_player:
pos_num = cf.move_against_player(mv_num, pos_num, qtable2)
if boards.winner(BOARDS[mv_num][pos_num]) != 0:
end = True
loses += 1
else:
pos_num = cf.move_against_player(mv_num, pos_num, qtable1)
if boards.winner(BOARDS[mv_num][pos_num]) != 0:
end = True
wins += 1
mv_num += 1
if not end:
ties += 1
return ties, loses, wins
def nondeterministc_game(qtable1, qtable2):
gap1 = qtable1[0][0][:].max() - qtable1[0][0][:].min()
gap2 = qtable2[0][0][:].max() - qtable2[0][0][:].min()
starting_player = randint(0, 1) # losowanie pierwszego gracza
mv_num = 1
pos_num = 0
while mv_num <= 9: # petla symulujaca gre
if mv_num % 2 == starting_player: # ruch pierwszego gracza
pos_num = move_with_gap(mv_num, pos_num, qtable2, gap2)
if boards.winner(BOARDS[mv_num][pos_num]) != 0:
return -1
else: # ruch drugiego gracza
pos_num = move_with_gap(mv_num, pos_num, qtable1, gap1)
if boards.winner(BOARDS[mv_num][pos_num]) != 0:
return 1
mv_num += 1
return 0
def make_action(prev_action, action, mv_num, curr_pos, prev_pos, qtable):
new_pos = cf.position_after_move(mv_num, curr_pos, ACTIONS.index(action))
# uaktualnianie qtabeli w zaleznosci od tego czy gra sie skonczyla
if boards.winner(BOARDS[mv_num][new_pos]) != 0:
next_act = q_update(prev_action, action, mv_num, new_pos, curr_pos,
prev_pos, REWARDS[1], qtable)
end = True
else:
next_act = q_update(prev_action, action, mv_num, new_pos, curr_pos,
prev_pos, REWARDS[0], qtable)
end = False
return new_pos, curr_pos, action, next_act, end
curr_pos, prev_pos, prev_action, action, end = make_action(
prev_action, action, mv_num, curr_pos, prev_pos, QTABLE)
mv_num += 1
with open(file_name, 'wb') as f:
pickle.dump(QTABLE, f)
else:
with open(file_name, 'rb') as f:
QTABLE = pickle.load(f)
# glowna czesc programu - gra z uzytkownikiem
print("Play as O: 1; Play as X: 2; End: 0")
new_game = int(input("Choose your symbol: "))
while new_game != 0:
mv_num = 1
pos_num = 0
end = False # incjacja zmiennych
while (not end) and (mv_num <= 9): # petla dla pojedynczej rozgrywki
active_player = (mv_num + new_game) % 2
# wybieranie i wykonanie akcji
if active_player == 0:
pos_num = cf.player_move(mv_num, pos_num)
else:
pos_num = cf.move_against_player(mv_num, pos_num, QTABLE)
if boards.winner(BOARDS[mv_num][pos_num]) != 0:
end = True
boards.print_board(BOARDS[mv_num][pos_num])
print("-----------------") # wypisanie planszy po ruchu
mv_num += 1
print("Play as O: 1; Play as X: 2; End: 0")
new_game = int(input("Choose your symbol: "))