Box2d Classic control MuJoCo Toy text Examples Testing What's new Basics There are two basic concepts in reinforcement learning: the environment (namely, the outside world) and the agent (namely, the algorithm you are writing). The agent sends actions to the environment, and the environment ...