In the DeepRacer league, the objective of the competition is to complete a race track as fast as possible and reinforcement learning is used to teach the robot to create the best model possible to do so. By creating the best reward function and training the robot as long as possible, it will be able to complete the track faster each iteration.