Keras Deep Learning Cookbook

DQN agent class

The DQNAgent class has the following methods that embody, model, remember, and so on, which we discussed earlier:

class DqnAgent:    def __init__(self, state_size, action_size):            def _build_model(self):    def remember(self, state, action, reward, next_state, done):           def act(self, state):    def replay(self, batch_size):    def load(self, name):    def save(self, name):

Each method has the following functions:

• __init__ (self, state_size, action_size): Initializes the class with state_size and action_size parameters:
  • state_size = 4
  • action_size = 2
• _build_model(self): Builds the neural network model using the Keras sequential model and returns it. This model has two hidden layers with 24 neurons each. The last output layer has an output ...