[R] Genetic Algorithms with rbga.bin using AIC as the evalFunc

[Bob Farmer]

Hi all.
I would like to use the rbga.bin function (from the "genalg" package) as 
part of a model selection process to whittle down a list of ~40 
potential explanatory variables to only the most important ones.

Unfortunately, despite my working knowledge of R for linear modeling and 
basic statistics, I cannot understand the two examples provided with the 
documentation for rbga.bin 
(http://finzi.psych.upenn.edu/R/library/genalg/html/rbga.bin.html). 
What's particularly confusing is that I can't see how the evaluation 
function (evalFunc) can vary with each iteration of the rbga.bin.

Can anybody help me design a command that incorporates the AIC() value 
of a linear model (where the inclusion of selected variables in the 
model is dependent on the current iteration of the rbga.bin function) as 
the evalFunc?  Otherwise, is there any additional documentation that 
would help me understand the provided examples?  It's been slow going 
trying to dig through the help() function and various introductory 
references for what is very convoluted syntax.

Thanks.
--Bob Farmer
MSc. candidate
Department of Integrative Biology
University of Guelph
Guelph, Ontario, Canada
rfarmer at uoguelph.ca