The Beer Game is a simulation game commonly used for supply chain management education, aiming to demonstrate the dynamics and complexities within a supply chain. This project has implemented a Beer Game based on a gym environment, utilizing two reinforcement learning algorithms: Deep Q-Networks (DQN) and Proximal Policy Optimization (PPO) to train intelligent agents. This report will introduce the main classes and functions in the project, including the BeerGame class, Agent class, the implementation of the DQN algorithm, the implementation of the PPO algorithm, and the overall workflow of the project.