This report presents a comparative analysis of Reinforcement Learning (RL) and Genetic Algorithms (GA) in solving the pole balancing problem. The study evaluates the performance of both methods under identical training and testing conditions, highlighting their strengths and weaknesses. 📊
The report is part of the final project for the course of "Natural Computation Methods for Machine Learning" at Uppsala University. 🎓
It explores the comparative advantages and disadvantages of using a genetic algorithm versus a reinforcement learning approach for the pole balancing problem. The results demonstrate that reinforcement learning outperformed the genetic algorithm in this scenario, emphasizing its simplicity of implementation and superior environmental comprehension. 🌟
- Filippo Balzarini filomba01
- Iason Kaxiras Iason-kaxiras
- Melvin Gode Klackson
You can download the report here.