Python: Advanced Guide to Artificial Intelligence

Initializing and discretizing for Q-Learning

The observations returned by the pole-cart environment involves the state of the environment. The state of pole-cart is represented by continuous values that we need to discretize.

If we discretize these values into small state-space, then the agent gets trained faster, but with the caveat of risking the convergence to the optimal policy.

We use the following helper function to discretize the state-space of the pole-cart environment:

# discretize the value to a state spacedef discretize(val,bounds,n_states):    discrete_val = 0    if val <= bounds[0]:        discrete_val = 0    elif val >= bounds[1]:        discrete_val = n_states-1    else:        discrete_val = int(round( (n_states-1) *                                   ((val-bounds[0])/ (bounds[1]-bounds[0])) ...