Reinforcement learning is a machine learning algorithm which allows software agents (or bots) to automatically determine the ideal behavior in an environment by providing negative and positive rewards to those agents based on their actions. Agents learn from trial and error to achieve their goal of maximizing the reward in the given environment.
An example of reinforcement learning is an AI mouse in a maze. The mouse is the agent, the environment is the maze, and the objective is for the mouse to navigate the maze in the shortest amount of steps to find the cheese at the end. The image above shows an AI mouse trained using reinforcement learning trying to nagivate the maze.