def testAgainstSecondMoveHeuristikAgentIn100Testgames(self):
heuristicSearchAgentWins = 0
tdqAgent1000Wins = 0
for testGameCount in range(100):
ttt = TicTacToe(4)
tdqAgent1000 = TicTacToeTDQLearningAgent(
TICTACTOE_4x4_TDQ_AGENT_1000_NAME, 4)
while not ttt.is_terminal():
action = tdqAgent1000.suggestAction(ttt)
print action
ttt.make_move(action)
if not ttt.is_terminal():
HeuristicSearchAgentTicTacToe.processAction(ttt)
print ttt.printable_game_matrix()
if ttt.is_victory() and ttt.get_player_which_moved_last() == 'X':
tdqAgent1000Wins += 1
elif ttt.is_victory() and ttt.get_player_which_moved_last() == 'O':
heuristicSearchAgentWins += 1
print 'Second move heuristic search agent wins: ' + str(
heuristicSearchAgentWins
) + ' games against TD-Q-Agent-1000 in 16 field Tic Tac Toe!'
print 'First move TD-Q-Agent-1000 wins: ' + str(
tdqAgent1000Wins
) + ' games against heuristic search agent in 16 field Tic Tac Toe!'
self.assertTrue(tdqAgent1000Wins >= 50)
def test_three_times_three_printable_game_matrix(self):
action_sequenz = [(0, 0), (1, 0), (2, 0), (0, 2), (2, 2), (2, 1),
(1, 1), (0, 1), (1, 2)]
tictactoe = TicTacToe(3)
tictactoe.initialize_game_matrix_with_action_sequence(action_sequenz)
test_game_matrix_string = 'X | O | O\n---------\nO | X | X\n---------\nX | O | X\n'
self.assertEqual(tictactoe.printable_game_matrix(),
test_game_matrix_string)
def testAgainstFirstMoveRandomAgentIn100Testgames(self):
randomAgentWins = 0
tdqAgent1000Wins = 0
for testGameCount in range(100):
ttt = TicTacToe(3)
tdqAgent1000 = TicTacToeTDQLearningAgent(TICTACTOE_3x3_TDQ_AGENT_1000_NAME, 3)
while not ttt.is_terminal():
RandomAgent.processTicTacToeAction(ttt)
if not ttt.is_terminal():
ttt.make_move(tdqAgent1000.suggestAction(ttt))
print ttt.printable_game_matrix()
if ttt.is_victory() and ttt.get_player_which_moved_last() == 'X':
randomAgentWins += 1
elif ttt.is_victory() and ttt.get_player_which_moved_last() == 'O':
tdqAgent1000Wins += 1
print 'First Move random agent wins: ' + str(
randomAgentWins) + ' games against TD-Q-Agent-1000 in 9 field Tic Tac Toe!'
print 'Second Move TD-Q-Agent-1000 wins: ' + str(
tdqAgent1000Wins) + ' games against random agent in 9 field Tic Tac Toe!'
self.assertTrue(tdqAgent1000Wins >= 50)
def testAgainstFirstMoveRandomAgentIn9FieldTicTacToe(self):
randomAgentWins = 0
heuristicSearchAgentWins = 0
for testGameCount in range(100):
ttt = TicTacToe(3)
while not ttt.is_terminal():
RandomAgent.processTicTacToeAction(ttt)
if not ttt.is_terminal():
HeuristicSearchAgentTicTacToe.processAction(ttt)
print ttt.printable_game_matrix()
if ttt.is_victory() and ttt.get_player_which_moved_last() == 'X':
randomAgentWins += 1
elif ttt.is_victory() and ttt.get_player_which_moved_last() == 'O':
heuristicSearchAgentWins += 1
print 'First Move random agent wins: ' + str(
randomAgentWins
) + ' games against heuristic search agent in 9 field Tic Tac Toe!'
print 'Second Move heuristic search agent wins: ' + str(
heuristicSearchAgentWins
) + ' games against random agent in 9 field Tic Tac Toe!'
self.assertTrue(heuristicSearchAgentWins >= 60)
score += 0.1
if value == 3:
score += 0.3
if self.__calculate_player_turn(state) == player_token:
score += score * 2
return score
def __calculate_player_turn(self, tictactoe_state=TicTacToe(4)):
if tictactoe_state.count_of_game_tokens_in_game() % 2 == 0:
return 'X'
elif tictactoe_state.count_of_game_tokens_in_game() % 2 == 1:
return 'O'
# state = result(result(result(result(result(result(ttt_state, (0, 0)), (0, 1)), (0,2)), (1,0)), (1,2)), (1,1))
ttt_state = TicTacToe(4)
zobrist_hasing = TicTacToeZobrist()
pvs = PrincipalVariationSearch(zobrist_hasing)
# ttt_state = pvs.result(
# pvs.result(pvs.result(pvs.result(pvs.result(pvs.result(ttt_state, (2, 1)), (2, 0)), (1, 2)), (0, 0)), (2, 2)),
# (1, 1))
#ttt_state = pvs.result(pvs.result(pvs.result(pvs.result(ttt_state, (2, 1)), (2, 0)), (1, 2)), (0, 0))
ttt_state = pvs.result(pvs.result(pvs.result(pvs.result(pvs.result(pvs.result(pvs.result(pvs.result(ttt_state, (1, 1)), (2, 0)), (2, 1)), (0, 0)), (0, 1)), (3, 1)), (2, 2)), (1, 0))
#ttt_state = pvs.result(pvs.result(ttt_state, (1, 1)), (2, 0))
# print ttt_state.printable_game_matrix()
# time_before_funciton_call = time.time()
# print pvs.zobrist_alpha_beta_search(ttt_state)
# print 'Time in milliseconds: ' + str(int((time.time() - time_before_funciton_call) * 1000))
# print count
print ttt_state.printable_game_matrix()
print pvs.evaluate(ttt_state)
elif state.count_of_game_tokens_in_game() % 2 == 1:
return 'O'
def terminal_test(state):
if state.is_victory():
return True
elif not state.is_victory() and state.count_of_game_tokens_in_game(
) == state.get_maximal_amount_of_game_tokens():
return True
else:
return False
def utility(state):
if player(state) == 'X' and state.is_victory():
return -1
elif player(state) != 'X' and state.is_victory():
return 1
elif state.count_of_game_tokens_in_game(
) == state.get_maximal_amount_of_game_tokens() and not state.is_victory():
return 0
ttt = TicTacToe(4)
ttt.make_move((1, 1))
ttt.make_move((2, 1))
ttt.make_move((2, 2))
ttt.make_move((1, 2))
print ttt.printable_game_matrix()
print iterative_deepening(ttt, 1)