When we talk about deep learning, reinforcement learning is a much amazing branch of it. Without the use of any predefined database. It learns from the application environment and trains its agent to get the maximum rewards for the objective. It’s a trending thesis/research topic these days. Though its roots are buries since 1977 when Temporal difference learning was proposed. Later on, Markov Decision Process proved to be the nail in the modern Deep Reinforcement Learning. The usage of reinforcement learning can be understand by an example.