Re: Lineage selection

> 1) The most widely accepted definition of "species" for sexually
> reproducing organisms is based on "reproductive isolation" [*]: an
> elephant and an ant belong to different species because they cannot
> produce offspring. A Pekinese and a Great Dane belong to the same
> species because they can produce offspring [**].
Yes there is maybe clear line between ants and elephants. But it is
biologically unclear
when some lineage can reproduce "in theory", can reproduce "in practically",
just don't newer reproduce because distinction in _space_ or _time_ e.t.c I
see concept so unclear anyway that don't mind whole concept much. It however
has been too little use in EA:s even crossovering is widely used .
And how you can have crossovering if you don't have species consept, that I
see like too non biological. And typical EA user anyway gets better results
without crossover for that reason.

> 2) Any scheme for pairing individuals not based on reproductive
> isolation has no claim to call itself "biologically realistic".
And what it was again this reproductive isolation? Do apes and humans have
reproductive isolation?
Or horses and mules?

> 3) Mating according to "lineage" is not biologically realistic because
> in nature it is possible for subpopulations with relatively recent
> common ancestry to become reproductively isolated.
I partially agree. I don't mean it is very biologically realistic. But I see
it as biologically realistic than any other EA method.

> 4) Mating closely related individuals is not likely to produce much
> novelty, as the offspring will closely resemble the parents. Therefore
> crossover becomes ineffective as a search mechanism.
This is whole point. This is missunderstanding. This is place where I and
propably many EA/GA people disagree. So I simply disagree that. I have one
another great tool which I call "Structural alphabet" which is not totally
my invention but consept kind is. It is kind of enhancement for "hot points"
e.t.c if you have heard such term .

> 5) A highly fit genotype very quickly becomes a common ancestor to the
> entire population. All other "species" in the population just go extinct
> [***]
Not, as I told I have other tool for that. This is also biologically
realistic that it happends if enviroment
has no different kind of resources to different species evolve. For that
part I if needed add resource consept. So that for example multi subproblem
cases, best species for only one problem have good
possibility to survive.

> [*] Ok, there are some grey areas here... so-called "ring species", for
> example.
I agree, and in my case we are only talkin EA/GA:s.

> [**] Ouch!
Jou.

> [***] At least if all population members compete with each other for
> selection on a "level playing field".
> --
> Lionel B

t. Harri


.

Relevant Pages

  • Re: The Paradox of Speciation
    ... reproductive isolation and reproductive incompatibility. ... So your species concept is not the usual one in biology. ... laboratory populations, and there are many experiments that show this. ... never lose reproductive compatibility, ...
    (talk.origins)
  • Re: Evolution is not a fact
    ... I have a paper under review in which I argue that species ... and others' work on this the BSC ... that there's other ways of achieving reproductive isolation ... sympatric speciation (i.e., diverging selection means hybrid forms have ...
    (talk.origins)
  • Re: Evolution is not a fact
    ... absolutely necessary for a speciation event." ... What worries me is that by defining species in such a way that only ... "You don't have reproductive isolation ... sympatric speciation (i.e., diverging selection means hybrid forms have ...
    (talk.origins)
  • Re: The Paradox of Speciation
    ... reproductive isolation and reproductive incompatibility. ... So your species concept is not the usual one in biology. ... laboratory populations, and there are many experiments that show this. ... never lose reproductive compatibility, ...
    (talk.origins)
  • Re: The Paradox of Speciation
    ... reproductive isolation and reproductive incompatibility. ... So your species concept is not the usual one in biology. ... I don't find your "reproductive compatibility" to be at all biologically ... laboratory populations, and there are many experiments that show this. ...
    (talk.origins)