CS 391 Selected Topics: Game AI Homework #7 - Final Project

Note: This work is to be done individually unless approved by the instructor. 

1. Project Proposal:  Choose your own adventure.  Informally propose a final project that either applies a technique learned in this course, or allows one to learn a different Game AI technique beyond our coverage.  The project idea must be verbally approved by the instructor.  (The instructor's role is primarily to caution against "biting off more than you can chew", suggesting simplifications/abstractions, providing related resources, discussing approaches, etc.)  Suitable game examples:

• Let's Catch the Lion (Dobutsu Shogi)
• GOPS
• Ataxx
• Paper Penguins
• Mancala variants

In general, it's wise to choose an interesting game with relatively simple rules, so that ones main effort is in the quality of the AI rather than in the complexity of the game modeling.

2. Presentation:  Within 15 minutes, you will

• Clearly define your problem.  So often a problem becomes so familiar to the person working on it, that they forget to bring the listener into a good understanding of the problem.  Take a few minutes to ensure all understand your challenge problem.
• Sketch your approach to the problem.  For this portion of the presentation, you may assume your audience has a working familiarity with the material of this course.  (For example, one might say, "I chose to apply UCT to Mancala, creating a single new node per iteration, using a bias coefficient of ____, and using ____ for time management in the game.")
•  Demonstrate your implementation.   Allow time to demonstrate your AI to the class.  You'll want to show the correct functioning of the implementation at a pace where, although you might not have time for a complete game, your audience will see the quality of decision-making clearly.  A human-versus-computer text-based game demo might serve this purpose well.
• Invite a brief question and answer time.

3. Project Submission: Final project code is due at the end of the scheduled final exam period.  There should be clear documentation in a README file describing the significant of each submitted file and how to execute project code. 

In grading, roughly half of the weight will be given each to the presentation, and the final project submission.