Weekly Journal: Week 2 (May28 to June 3)


Victoria Manfredi
vmanfred@cs.smith.edu



Amy's at a conference this week, so I am on my own with a list of things to do :-). I spent all of Monday making graphs of the outputs using gnuplot. I made graphs of the number of pricebots of each algorithm type, the prices of each, and the profits of each, as these changed as the number of generations changed. I have been getting some interesting results (so interesting that they may be wrong :-) so I want to make sure these results are actually correct. I have also been doing some reading on introductory game theory which is helping me understand a little better the program and the project. It is also pretty interesting!

When I checked the results, everything seemed to be being done correctly, so I will have to assume they're correct and wait for Amy to return from her conference so she can have a look at them.

I spent the rest of the week reading about Q-learning and trying to figure that out, and looking at the simulator code to see how I could fix the Q-learning algorithm that was implemented so that it runs correctly. This has definitely been a week of Q-learning.



Smith CS Page / Brown CS Page / Vicky's Homepage / Amy's Homepage / Weekly Journal