RL, known as a semi-supervised learning model in machine learning, is a technique to allow an agent to take actions and interact with an environment so as to maximize the total rewards. RL is usually modeled as a Markov Decision Process (MDP).
Imagine a baby is given a TV remote control at your home (environment). In simple terms, the baby (agent) will first observe and construct his/her own representation of the environment (state). Then the curious baby will take certain actions like hitting the remote control (action) and observe how would the TV response (next state). As a non-responding TV is dull, the baby dislike it (receiving a negative reward) and will take less actions that will lead to such a result(updating the policy) and vice versa. The baby will repeat the process until he/she finds a policy (what to do under different circumstances) that he/she is happy with (maximizing the total (discounted) rewards).
The study of RL is to construct a mathematical framework to solve the problems. For example, to find a good policy we could use valued-based methods like Q-learning to measure how good an action is in a particular state or policy-based methods to directly find out what actions to take under different states without knowing how good the actions are.
However, the problems we face in the real world can be extremely complicated in many different ways and therefore a typical RL algorithm has no clue to solve. For example, the state space is very large in the game of GO, environment cannot be fully observed in Poker game and there are lots of agents interact with each other in the real world. Researchers have invented methods to solve some of the problems by using deep neural network to model the desired policies, value functions or even the transition models, which therefore is called Deep Reinforcement Learning. This article makes no distinction between RL and Deep RL.