Considerations for programming a genetic algorithm on parallel computers:

I decided to make it very object oriented, breaking it into class Problem, 
class Solution, etc. with the hope that if I was solving a much different problem with a
genetic algorithm, I could reuse a lot of code.

How to divide the program amongst several computers?
I have a different population on each computer.  Killing and breeding a given percent of 
populations and mutating a given percent of populations is done separately on each computer
as if the others werent there.  Then, once a generation mix the populations of all the 
different computers with each other.

At first I wanted the GAProb object to manage the fitness and keep it completely out of the
Population object, but then I changed it and put the Problem within the Population object 
as well.  I don't want to have a problem defined for each solution because that would be a lot 
of wasted memory.