Project information

• Category: Algorithm Design
• Location: Champaign, IL
• Project date: Dec 2023
• Project URL: temporarily unavailable

The player uses up, down, left and right to control the snake which grows in length (when it eats the food pellet), with the snake body and walls around the environment being the primary obstacle. Exploration of reinforement.

• Implemented a reinforcement Q learning agent in Python for a snake game, using a Markov Decision Process to maximize rewards through actions.
• Applied Temporal-Difference Q-learning algorithm and an exploration policy to balance between learning new states and exploiting known high-reward actions.
• Rigorously trained and tested the agent, ensuring adherence to specified parameters and achieving benchmark performance goals.