score vs raw objective, stats and more HELP

Theresa Radke theresaradke at
Wed Aug 2 21:30:15 EDT 2000


I am a senior computer science student taking a genetic algorithm 
course this summer semester.

Our final project (is also an inter-class contest of sorts) involves 
evolving magic squares with galib.

I have my program running, and creating squares, my issues however 
are as follows.

When a value is returned from my objective function as a "score", how 
is that value "scaled" and converted to a score for the statistics 

For example, I am unable to tell at which generation the program is 
producing the first magic square.  Because one of my classmates might 
perhaps be subscribed to this group, I don't want to give too much 
away concerning my fitness algorithm, but I can say that it returns a 
value ...if the value is larger the better the individual fitness is.

My issue is how to reference the best individual within the 
terminator function to ensure the exit upon first magic square is 

I have examined the flushed data, and thought I'd found a value that 
the best of generation reaches when a magic square is encountered, 
however when i use that value in the terminator function as a 
comparitor it does not exit.

I could solve this issue by checking maxEver's fitness (summing 
columns and rows and diagonals) at each generation end to see if it 
sums to the magic number, however I cannot determine how to access 
the gene of the best individual to do the manual computation within 

I am using the GA2DArrayGenome object for my chromosome.  Any help or 
insight that anyone might be able to offer would be greatly 

Theresa Radke
"great spirits often encounter violent opposition from mediocre minds"
--Albert Einstein
Created by -
Awarded PCMagazine's Editors' Choice

More information about the galib mailing list