Hands-On Q-Learning with Python

Setting a state manually

We can manually set the state of the environment by changing the state variable. Note that you should not set the state manually when running this task; this is bad practice and is not encouraged by OpenAI Gym. We are demonstrating it here for illustration purposes only:

env.env.s = 20env.render()

The following is the result for the preceding code:

Generally, in a game loop, we will be sent to the state that results from taking an action selected by the Q-learning algorithm and will not need to set the state manually. In this demonstration, it can be useful to verify for yourself that each value corresponds to a fixed ...