Homework 3 -- due Wednesday March 11

Written questions

1. [10 points] Answer Question 5.2 of the textbook
2. [10 points] Answer Question 6.2 of the textbook
3. [15 points] Answer Question 6.3 of the textbook
4. [15 points] Answer Question 6.15 of the textbook

Lisp questions

For this question you will use the cognac game (available in the aima software in search/domains/cognac.lisp) and the alpha-beta-ttt-agent (available in the aima software in search/agents/ttt-agent.lisp). Look in search/test-search.lisp to see how to run a game, how to specify the agents playing in the game, and the size of the board. For instance, to run the random-game-agent against the alpha-beta-ttt-agent on a game with n=7 and k=4 you need to do: (run-game (make-cognac-game :n 7 :k 4) :agents '(random-game-agent alpha-beta-ttt-agent))

Answer the following questions:

1. [5 points] Play the alpha-beta-ttt-agent against yourself in a cognac game with n=7 and k=4. To play use the human-game-agent, which is defined in search/agents/ps-agents.lisp. When you are asked to play enter a list with the x and y coordinates of the square you want to play in. Show a game in which you play first and and one in which you play second.
2. [15 points] Study the effect of the order in which players play on their ability to win depending on the size of the board. Build a table of results of playing two default alpha-beta-ttt-agent against each other for values of n=3,4,5,6,7 with values of k=3 and k=4.
3. [20 points] Study the effect of changing the evaluation function used by the alpha-beta-ttt-agent on games with different values for n and k. Build a table of results of playing the default alpha-beta-ttt-agent against an agent with your evaluation function for values of n=3,4,5,6,7 with values of k=3 and k=4. Using the results from the previous question, count how many times the agent with your evaluation function does better than the default agent.