Hello Alessandro!
Congratulations on the great job of visualizing the results you've done; it makes everything very clear and easily readable. It was very helpful for me to understand which agent consistently wins against the others. As we studied in the theoretical part of the course, we can see that the Q-learning-based implementation is the best strategy in all cases. I also really like the way in Q-learning you select the best moves, following a very precise strategy and a trial-and-error approach.
Hello Alessandro! Congratulations on the great job of visualizing the results you've done; it makes everything very clear and easily readable. It was very helpful for me to understand which agent consistently wins against the others. As we studied in the theoretical part of the course, we can see that the Q-learning-based implementation is the best strategy in all cases. I also really like the way in Q-learning you select the best moves, following a very precise strategy and a trial-and-error approach.