This device is not compatible.
PROJECT
Create a Custom Reinforcement Learning Environment
In this project, we’ll learn how to create a custom reinforcement learning environment from scratch for a gridworld and a shower system.
You will learn to:
Develop custom Gymnasium environments.
Integrate reinforcement learning algorithms in custom environments.
Train the REINFORCE agents.
Visualize the agents’ learning curve.
Skills
Python Programming
Reinforcement Learning
Object Oriented Programming
Prerequisites
Familiarity with object-oriented programming in Python
Basic understanding of reinforcement learning
Knowledge of machine learning
Technologies
Python
Gymnasium
Matplotlib
Project Description
In many reinforcement learning projects online, the environment is already created, and we only train the agents in those environments. However, in the real world, we must create an environment customized for our application.
In this project, we will learn how to create two custom reinforcement learning environments: one for training an agent to navigate a gridworld and the other for obtaining optimal temperature in a shower system. We will use the Gymnasium and Matplotlib libraries to tackle such a real problem and create a reinforcement learning environment step by step to solve it.
Project Tasks
1
Gridworld Environment
Task 0: Introduction
Task 1: Develop the Environment Class
Task 2: Develop the init() and reset() Methods
Task 3: Develop the step() Method: Change Agent’s State
Task 4: Develop the step() Method: Reward and Termination
Task 5: Train the Agent
2
Shower Environment
Task 6: Develop the Shower Environment Class
Task 7: Develop the init() and reset() Methods
Task 8: Develop the step() Method: Change the Agent's State
Task 9: Develop the step() Method: Reward and Termination
Task 10: Train the Agent
Congratulations!
Relevant Courses
Use the following content to review prerequisites or explore specific concepts in detail.