Eliza's Journal Page

9/17/03

Web pages are set up, although there isn't much on them yet.

9/23/03

I went to BGSU library last weekend. The library I needed was closed, but I found out that the book is what I need. I'm doing reasonably well on getting info, but need to start looking at it again.

9/24/03

I got the tentative bibliography done, and went down to Wildman Science Library and got the articles I need photocopied. That means that I have everything I intended to get except the book, which is on order.

10/2/03

The book finally came yesterday, probably too late to be much use. I did my survey presentation yesterday, which I think went reasonably well. I still need to write up the paper. I'm beginning to get ideas on the project: something with game theory, using computers to do the brute force checking.

10/8/03

I got my survey paper written last night. It went as easily as I'd hoped; the first draft was done in about two easy hours, and a couple successive proof-reading and citation passes got it into shape. I'm planning on doing some preliminary coding for the project over break. I'm going home tonight!

10/13/03

I got the code started for my pegboard on Friday. I have the node class finished and am working on debugging the board class. It's going pretty well, except for the usual pointer problems.

10/23/03

I've been meaning to add another entry for a couple days. I'm doing test runs of a first implementation of the game tree program for Hique with just one peg missing in the center. I'm also working on a more efficient version, starting with a 4x4 board with one peg missing. I'm doing okay, although I'm having trouble with return values, and I can't figure out why. The proposal is due next week...

10/26/03

Well, I've done some pretty major changes to the code. I went back to a 4x4 game board, and once I had that working, using a 4x4 array and a single number for each move, which references a row in the moves array, I shifted it back to the Hique board, this time using one 7x7 array with not quite all the spots used. It's working quite well, I got the return stuff working, and it's a major speed improvement over the last version. Using the -O compilation option, I think the entire game tree for the Hique board with one peg missing in the middle should take less than 24 hours. I've even got the rough draft of the proposal all done!

11/10/03

Where does the time go? I'm currently running the "standard" tree; I have the initial position generator working; and I'm working on the symmetry filter version of the initial position generator and a sort program to split the output of the initial position generating program into useful categories.

11/17/03

I've gotten a version of the program with the boards formally declared as classes (they were basically classes, just not declared as such). It does sorting and symmetry checking, which took figuring out how to get the program to output its boards to a file, depending on the value of the board, then read them back in to check against the already evaluated boards. I've also got a very slightly different version which instead of evaluating ALL the moves of a new board, does the first level of moves, then checks the resulting board against the output files from the program run with one less pegs than the current run, i.e. the number of pegs of the current run, minus the one removed in the move. Checking takes a while, but it makes the evaluation asymptotically quadratic instead of exponential. I've got output for boards starting with 1-5 pegs, and 6 pegs is currently running. This is a little behind where I'd hoped, but then the output is half or more analysed before I look at it, which makes it a lot easier to grasp.

12/1/03

I haven't gotten all that much done, unfortunately; the computer running is going slowly. The analysis is going fairly well.

12/3/03

Well, I've got one winning sequence of moves for the standard board! The output files that my program is generating are using the second half of my quota and bombing, so I don't have a complete list, but I know that list is very, very, very big, especially since the last stuff I've seen in the output file must be within the first quarter of the results. I have gotten somewhat of a feel for some of the output, having looked at the files, and think I'm going to quit about here on this part. The running is going slowly, as always, but at least eni1 is back up and really working for the first time.

12/12/03

I've done the final presentation. It went okay; there were things that could have been better, but it was decent. This means I'm done! Except for the final draft of the final paper...