Reinforcement Learning (RL) is, together with Supervised Learning and Unsupervised Learning, one of the three fundamental learning paradigms in Machine Learning. The goal in RL is to enhance the manipulation of a controlled system by using data from past experiments. It differs from supervised learning and unsupervised learning because the algorithms are not trained with a fixed sampled dataset. Rather, they learn through trial and error.
Take a look the detailed explanation at DyCon Blog