This link will help the expert get started on what to do: www.vicstarenterprise.com/week5gettingstarted.mp4
This assignment has two parts:
1. In Part 1, you will write an agent that learns how to play Connect 3 using the Q-learning algorithm – repeatedly playing the game against a random opponent and learning from the results. Use initial Q-values of zero, and a discount factor γ=1γ=1. Choices of other parameters are up to you.
To demonstrate your results, you will produce a learning curve, a graphical plot showing the learning rate of your algorithm, comparing its performance to a random agent. You will also need to write a modified version of your code which produces a simplified version of the graph which we can run.
Detailed instructions are available in the notebook. This part is worth 80% of the assignment.
2. In Part 2, you will write a short amount of text explaining your implementation, any decisions or extensions you made, and what parameter values you used. This part is worth 20% of the assignment.
Get Free Quote!
321 Experts Online