def run_maze():
step = 0
for episode in range(5):
# initial observation
observation = env.reset()
while True:
# fresh env
#env.render()
# RL choose action based on observation
action = DQN.choose_action(observation)
# RL take action and get next observation and reward
observation_, reward, done = env.step(action)
DQN.store_transition(observation, action, reward, observation_)
if (step > 200) and (step % 5 == 0):
DQN.learn()
# swap observation
observation = observation_
# break while loop when end of this episode
if done:
print('game over')
print(env.balance)
break
step += 1
ep_r = 0
while True:
env.render()
action = RL.choose_action(observation)
observation_, reward, done, info = env.step(action)
# the smaller theta and closer to center the better
x, x_dot, theta, theta_dot = observation_
r1 = (env.x_threshold - abs(x))/env.x_threshold - 0.8
r2 = (env.theta_threshold_radians - abs(theta))/env.theta_threshold_radians - 0.5
reward = r1 + r2
RL.store_transition(observation, action, reward, observation_)
ep_r += reward
if total_steps > 1000:
RL.learn()
if done:
print('episode: ', i_episode,
'ep_r: ', round(ep_r, 2),
' epsilon: ', round(RL.epsilon, 2))
break
observation = observation_
total_steps += 1
#RL.plot_cost()
def run_model(_train_episode=100,
_learn_threshold=200,
_learn_interval=5,
_base_currency='USD',
_trade_log_mode='NONE',
_trade_log_raw=False,
_trade_log_to_file=False,
_show_checkout=True,
_show_step=True):
step = 0
for episode in range(_train_episode):
observation, TI_initial, initial_time = env.reset()
TI_initial_balance = copy.deepcopy(TI_initial)
train_name = TI_initial.account_name
while True:
action = DQN.choose_action(observation)
observation_, reward, done, TI_end, end_time = env.step(
action, print_step=_show_step)
DQN.store_transition(observation, action, reward, observation_)
if (step > _learn_threshold) and (step % _learn_interval == 0):
DQN.learn()
observation = observation_
if done:
print('$' * 20 + ' GAME OVER ' + '$' * 20)
TI_initial_balance.account_name = 'Initial_Checkout_Review' + ' (episode: ' + str(
episode + 1) + ')'
TI_initial_balance.checkout_all_in(initial_time,
_base_currency)
TI_end_balance = copy.deepcopy(TI_end)
TI_end_balance.account_name = 'End_Checkout_Review' + ' (episode: ' + str(
episode + 1) + ')'
TI_end_balance.checkout_all_in(end_time, _base_currency)
print('MI: Initial Time: {}; End Time: {}'.format(
initial_time, end_time))
if _show_checkout == True:
TI_initial_balance.account_review()
TI_end_balance.account_review()
if _trade_log_mode == False:
pass
elif _trade_log_mode == 'ALL':
TI_end_balance.trade_log_review(raw_flag=_trade_log_raw)
elif _trade_log_mode == 'TWOENDS':
TI_end.trade_log_review(tar_action_id=0,
raw_flag=_trade_log_raw)
TI_end.trade_log_review(tar_action_id='LAST',
raw_flag=_trade_log_raw)
else:
print(
'Invalid _trade_log_mode input ({}). Must be \'ALL\', \'TWOENDS\', or False'
.format(_trade_log_mode))
return -1
if _trade_log_to_file:
trade_log_base_dir = './logs/trade_logs/'
if not os.path.exists(trade_log_base_dir):
os.makedirs(trade_log_base_dir)
trade_log_file_name = trade_log_base_dir + str(train_name)
log_file_readable = open(trade_log_file_name + '.txt',
'w+')
with contextlib.redirect_stdout(log_file_readable):
TI_end.trade_log_review()
print(
"### READABLE trade_log of {} successfully exported to: ###\n\t\t{}"
.format(str(train_name), trade_log_file_name + '.txt'))
log_file_readable.close()
log_file_raw = open(trade_log_file_name + '.json', 'w+')
json.dump(TI_end.trade_log, log_file_raw, indent=4)
print(
"### RAW trade_log of {} successfully exported to: ###\n\t\t{}"
.format(str(train_name),
trade_log_file_name + '.json'))
log_file_raw.close()
break
step += 1
