def train_dqn(self):
# get size of state and action from environment
state_size = 4
action_size = 2
agent = DQNAgent(state_size, action_size)
scores, episodes = [], []
for episode in range(1, EPISODES + 1):
# run reinforcement learning for every episode
done = False
score = 0
self.reset()
state, _, _, _ = self.step(-1)
state = np.reshape(state, [1, state_size])
rospy.loginfo("Episode %d: starting", episode)
while not done:
action = agent.get_action(state)
next_state, reward, done, _, = self.step(action)
rospy.loginfo("Episode %d: action: %d pitch: %f", episode,
action, next_state[0])
next_state = np.reshape(next_state, [1, state_size])
# if an action make the episode end, then gives penalty of -100
reward = reward if not done or score == 499 else -100
# save the sample <s, a, r, s'> to the replay memory
agent.append_sample(state, action, reward, next_state, done)
# every time step do the training
agent.train_model()
score += reward
state = next_state
if done or score >= 500:
# every episode update the target model to be same with model
agent.update_target_model()
# every episode, plot the play time
score = score if score == 500 else score + 100
scores.append(score)
episodes.append(episode)
pylab.plot(episodes, scores, "b")
pylab.savefig("./cartpole_dqn.png")
print(
"episode:",
episode,
" score:",
score,
" memory length:",
len(agent.memory),
" epsilon:",
agent.epsilon,
)
break
# if the mean of scores of last 10 episode is bigger than 490
# stop training
if np.mean(scores[-min(10, len(scores)):]) > 490:
break
# save the model
if episode % 20 == 0:
agent.model.save_weights("./cartpole_dqn.h5")
rospy.loginfo("Episode %d: completed", episode)
