Reinforcement learning is one of the types of Machine learning that uses hit and trial method. For success, a positive point will be awarded and a negative point for failure. The machine learns from both positive and negative experiences for choosing better action in that situation. Reinforcement learning makes machines take the best decision possible in each situation.
Example: Consider a football player trying to score a goal. He shoots the ball and goalkeeper stops. Then he will get a negative point and if he scores he gets one point as a reward. From all these rewards and negative points, he learns how to play in different situations.
You can watch this video to get an overview of reinforcement learning: