The optimal action to take in a simulated game with a non-negative reward

In this section, we will understand the way in which we can take the right action for a simulated game. Note that this exercise will primarily help you to grasp how reinforcement learning works.

Getting ready

Let's define the environment we are operating in this simulated setting.

You have three boxes, on which two players are playing a game. Player 1 marks a box with 1 and player 2 marks one with 2. The player who is able to mark two consecutive boxes wins.

The empty board for this game looks as follows:

For the problem we just defined, only player 1 has an opportunity to win the game. The possible scenarios in which player 1 wins are either...