On Feb 24, 2010, at 8:47 AM, Uday Kamath wrote:
> I wanted to confirm that If I have N Sub Population of Species ,
> each having its own dynamics of individuals, mutation, crossover but
> working on same Problem, then this is parallel evolution, i.e. they
> wont interfere with each other, or same as running N runs with
> single species?
Assuming you're NOT using the coevolution package, and your Problem
class doesn't have any shared data (make sure your Problem.clone()
method copies everything you added, particularly arrays) then the
answer is: yes.
> I am implementing parameter optimization where Problem is same, but
> can be approached in different ways and hence different species, so
> wanted to make sure each would evolve without interference from
> other species or get affected by others.
Should be fine.