[ExI] ‘Survival of fittest’ is disputed
Mike Dougherty
msd001 at gmail.com
Thu Aug 26 00:41:06 UTC 2010
On Wed, Aug 25, 2010 at 9:23 AM, The Avantguardian
<avantguardian2020 at yahoo.com> wrote:
> Yes although there are some key differences that make genetic algorithms more
> like the selective breeding of domesticated organisms than the evolution of
> organisms in the wild. For example in nature, the solution space is several
> orders of magnitude higher in dimensionality. Also the local maxima are moving
> targets and there is a feedback of the solutions on the problem constraints. As
> well as the tendency of the natural algorithms to eat one another. :-)
sound _like_ GA
Selective breeding is an example of a reduced dimensionality and
constrained fitness function. Effectively GA are selectively breeding
for an answer to a specific purpose. Wild natural environments are
more complex but there is no way to prove that a goal exists or if
it's simple/complex. (A simple answer with an unlimited runtime may
generate unimaginable complexity)
More information about the extropy-chat
mailing list