Weekly Journal: Week 6 (June 25 to July 1)
Victoria Manfredi
vmanfred@cs.smith.edu
Well I spent Monday and Tuesday making the program work so that after a single pricebot calculates a price, all the pricebots go through a round of buying and get new profits. Also, because there are 4 algorithms, there are lots of sets of 4 variables, one for each algorithm. This was driving me crazy, so I finally made lots of arrays of size 4 instead so I could use a for loop to go through the variables. I think that most of the bugs are out of the algorithms now (gt, my, and df, but maybe not ql :-) and so Wednesday I started writing up my results and graphs, trying to explain what I think is happening in the results, and also trying to make sure that my results and graphs are correct.
Well, I spent Thursday and Friday working on the paper and making some new graphs to further explain old graphs. There are a few things that I haven't figured out about some of the graphs. But since I think (hope?) that the results when I just compete one algorithm against itself are right, I think that the results of several different algorithms competing against each other are also correct. I understand most of what is happening when I have several different algorithms working against each other, but there are a few things I am not quite sure of (like why the derivative following prices and profits are more spread out when they compete agains non-df pricebots: this may be due to that the df are always losing (decreasing in number) against the other algorithms; I am not sure why the df lose either: I know that the df is probably the worst algorithm of the group but I don't know yet what mechanism exactly is causing the df to lose). So I am going to see what I can figure out for myself, and ask Amy MOnday whatever I can't figure out. Figure 12.
Smith CS Page
/
Brown CS Page
/
Vicky's Homepage
/
Amy's Homepage
/
Weekly Journal