def play(self):
for i in range(self.G):
self.initiate_game_manager()
self.current_state = self.game_manager.initial_state()
self.root = Node(self.current_state, None)
self.simulator = GameSimulator(self.M, self.game_manager)
self.starting_player = self.game_manager.current_player
print(f"Starting player:\t P{self.game_manager.current_player}")
if (self.verbose):
print(self.game_manager.get_start_position_as_string())
self.single_game()
print("--------------------------------------------------------------")
print(f"Total games:\t\t\t\t {self.G}")
print(f"Games won by Player 1:\t\t\t {self.p1_winner_count}")
print(
f"Games where Player 1 had first move:\t {self.p1_starting_count}")
print(f"Games won by Player 2:\t\t\t {self.p2_winner_count}")
print(
f"Games where Player 2 had first move:\t {self.p2_starting_count}")
print(
f"Percentage of games won by the player who had first move: {100*self.correct_winner_count/self.G} %"
)
print("--------------------------------------------------------------")
def clean(self):
result = cherrypy.request.json
nickname = cherrypy.request.headers['Nickname']
game_simulator = self.games.get_game(nickname)
game_simulator.clean()
game_simulator = GameSimulator()
return result
def nickname(self):
nickname = cherrypy.request.json
print('nickname')
print(nickname)
if self.games.exists(nickname):
raise cherrypy.HTTPError(message='Invalid nickname')
self.games.set_game(nickname, GameSimulator())
return nickname
def sim(depth, conf):
print "Trying for depth " + str(depth) + "..."
simulator = GameSimulator(Game)
winner, conf_arr = simulator.simulate(conf, depth)
if (winner == 0):
print "Found winning strategy for FIRST player."
elif (winner == 1):
print "Found winning strategy for SECOND player."
else:
print "No winning strategy found"
print ""
print_strategy(conf_arr)
return False
print ""
print_strategy(conf_arr)
return True
for idx,var in enumerate(tfVars[0:total_vars//2]):
op_holder.append(tfVars[idx+total_vars//2].assign((var.value()*tau) + ((1-tau)*tfVars[idx+total_vars//2].value())))
return op_holder
def updateTarget(op_holder,sess):
for op in op_holder:
sess.run(op)
def saveScore(score):
my_file = open(reward_savefile, 'a') # Name and path of the reward text file
my_file.write("%s\n" % score)
my_file.close()
###########################################
game = GameSimulator()
game.initialize()
ACTION_COUNT = game.get_action_size()
gpu_options = tf.GPUOptions(per_process_gpu_memory_fraction=0.33)
SESSION = tf.Session(config=tf.ConfigProto(gpu_options=gpu_options))
if LOAD_MODEL:
EPSILON_MAX = 0.25 # restart after 20+ epoch
agent = Agent(memory_cap = MEMORY_CAP, batch_size = BATCH_SIZE, resolution = RESOLUTION, action_count = ACTION_COUNT,
session = SESSION, lr = LEARNING_RATE, gamma = GAMMA, epsilon_min = EPSILON_MIN, trace_length=TRACE_LENGTH,
epsilon_decay_steps = EPSILON_DECAY_STEPS, epsilon_max=EPSILON_MAX, hidden_size=HIDDEN_SIZE)
def __init__(self):
self.games = GameSession()
self.game_simulator = GameSimulator()
