Assignment 8 code & information

Code and other information

• Support code is available: qlearning.scm This code is documented in the "Assignment 8 details" handout.

  UPDATED 12/3/99 8:50am: Contrary to what I said in the handout: "the run-learner function takes care of updating the N table", the original code did not, so I have made this correction.

• A sample run is available. Note that because there is a random element in simulating the world, your outputs will not match mine exactly.

  Don't forget about the print-debug-level variable I mentioned in the handout. And the run-epochs function too...

  Also, you should know that although the Q and N tables are passed to your function, they are actually globals, so you can examine them after running some epochs.

• How do you know if your code is working properly? Pay attention to the Q values and the N values. Look at the policy your learner produces and see what it does in the world. Are you generating a reasonable policy?