GALIB

[CANAVAN]

To whom it concerns,
		I am a final year student of computer science in Maynooth 
University, in Ireland. I am currently working on a thesis based on evolutionary 
algorithms. For this I am using the examples you have generated. I am primarily 
using "ex3.C" as my main example. It is Copyright (c) 1995-1996  Massachusetts 
Institute of Technology and appears to be written by "mbwall". The date on the 
coding is 28jul94. It is a simple Genetic Algorithm that tries to copy a smiley 
face, "smiley.txt". I am investigating whether the ga works better when it is 
divided up into segments, for example, I divided the "smiley.txt" file into 
eight, 7x4 files, rather than one 14x16 file. I have found that it proves to be 
far more accurate when performed this way. However, I am now trying to assess 
whether the time difference between the two is too large. This is where I hope 
you can help me.
		I am not sure where I should start and finish my timing fuction 
in order to provide accurate and comprehensive results. I initially started and 
finished it in the "ex3.C" program, but this proved inaccurate as time was spent 
opening and closing the file, which is not what I am looking for. I am simply 
looking to see how long the program spends on the in the actual ga method, ie, 
where the comparisons and generations take place. I have tried manipulating a 
number of files, including GASimpleGA.C(where I tried timing the "step()" 
function), and GABaseGA.h(where the "evolve()" method is evoked), but neither 
provided me with any results. I was hoping that you could assist me with this. I 
hope the problem is clear adn hope to hear from you soon.
							regards,
								Darragh Canavan