JonF
Posts: 634 Joined: Feb. 2005
|
Quote (Daniel Smith @ Feb. 05 2009,19:08) | Quote (JonF @ Feb. 04 2009,17:31) | Quote (Daniel Smith @ Feb. 04 2009,19:06) | Quote (Wesley R. Elsberry @ Feb. 04 2009,14:56) | Quote (Daniel Smith @ Feb. 03 2009,19:00) | [snip] Wesley,
It's easy to point to something that works, and works well, and postulate that such a system would be advantageous and selected for. Of course it would be! Why wouldn't it? Life's systems are optimized.
The hard part is finding the precursors - those systems that necessarily didn't work as well in the environment - and finding the path from there to the existing, refined system.
With the TSP, we have the luxury of viewing all the precursors and seeing exactly why they didn't work as well as the final solution.
If biological evolution worked like the TSP, why has such detail eluded us? |
Perhaps I was unclear. The TSP example shows that in either evolutionary computation or biological evolution, there is no evaluation of "potential", only the evaluation of which of existing candidate "solutions" is better than other existing candidate solutions. The wrasse example, which you've ignored, also makes that point plus it demonstrates that which candidates are relatively better is often tied directly to physical, chemical, and physiological constraints.
I can't make heads or tails of your last question, as it rests on a false assumption. |
Wesley,
Just to be clear, when I speak of "potential", I'm talking about multi-step, or multi-layered systems for which the majority of pieces have to be in the correct place in order for the system to work. "Potential" in that regard is a "psuedo-system" or a "psuedo-pathway" - one that doesn't actually work but is a necessary precursor to one that does.
In the electrical circuit example it was all the circuits that didn't actually work but were selected because they showed potential towards the ultimate goal. |
You weren't unclear ... you're just wrong again.
No circuits were selected because they showed potential towards the ultimate goal. Circuits were selected because, as Wesley just said, "there is no evaluation of "potential", only the evaluation of which of existing candidate "solutions" is better than other existing candidate solutions.".
Until you figure this out you're mired in error. |
Define "better" as it applies to the circuits. |
The original paper defined "better" precisely. Go back to that.
|