This algorithm uses Q-Learn algorithm for learn how to play t-rex game. The algorithm get only x and y coodinates as reference as result of this, when game start to be very fast the agent makes mistakes.
The reference T-rex game used: http://www.trex-game.skipser.com/