What you'll learn

Reinforcement learning can be challenging without the right tools and a clear understanding of how agents interact with their environments. In this course, Getting Started with Gymnasium, you will learn how to use Gymnasium to design and run simulation environments, train reinforcement learning agents, and apply decision-making strategies to practical problems. First, you will explore how to create and configure Gymnasium environments while understanding the roles of agents and environments and how they communicate through actions, observations, and rewards. Next, you will build and train agents using both policy gradient and value-based methods, including a Deep Q-Network to solve the CartPole environment, and learn how to evaluate and visualize their performance with Gymnasium tools and wrappers. Finally, you will learn to apply a multi-armed bandit approach to a real-world inspired A/B testing scenario using Thompson Sampling, and use debugging, logging, and monitoring techniques to ensure robust and reliable agent behavior. When you finish this course, you will have the skills and knowledge needed to confidently use Gymnasium to develop, test, and refine reinforcement learning agents for a range of applications.