Keras Deep Learning Cookbook

How to do it...

Let's go ahead and implement the Dueling DQN agent-based Cartpole playing program. Perform the following steps to get the agent in place:

1. Initialize the Open AI gym environment env
2. Define the number of actions from env
3. Create a sequential neural network
4. Initialize the SequentualMemory with a limit of 100 and window_length of 1
5. Initialize the BoltzmannQPolicy instance policy
6. Create DQNAgent, as follows:

dqn = DQNAgent(model=model, nb_actions=nb_actions, memory=memory, nb_steps_warmup=10,               enable_dueling_network=True, target_model_update=1e-2, policy=policy)

7. Compile the DQNAgent with the optimization method as Adam and the loss function as mean absolute error (MAE)
8. Call dqn.fit to find the rewards:

ENV_NAME = 'CartPole-v0' ...