Hands-On Machine Learning for Algorithmic Trading

Training a Q-learning agent using Python

In this section, we will demonstrate how to build a Q-learning agent using the 3 x 4 grid of states from the previous section. We will train the agent for 2,500 episodes, use a learning rate of α= 0.1, and an ε=0.05 for the ε-greedy policy (see the gridworld_q_learning notebook for details):

max_episodes = 2500alpha = .1epsilon = .05

Then, we will randomly initialize the state-action value function as a NumPy array with the dimensions number of states x number of actions:

Q = np.random.rand(num_states, num_actions)skip_states = list(absorbing_states.keys())+[blocked_state]Q[skip_states] = 0

The algorithm generates 2,500 episodes that start at a random location and proceed according to the ε-greedy ...