Little evolver challenge (self-copieing organism ! ;))

[Skybuck Flying]

Hello,

Here is a little evolver challenge:

The challenge is to:

1. Evolve a warrior that copies itself through the core.

Restrictions to the evolver's code:

2. The evolver cannot copy warriors itself, it can only make modifications.

3. The evolver may spawn/create as many copies in the initial round as it
wants but as indicated in rule 2 it may not copy anything beyond initial
round.

The idea behind this challenge is to see if it's possible to create a
warrior that can actually copy itself, since most of darwin's evolution
theory is based on "selection of copies".

For there to be copies "nature" first has to "invent" the copy machine it
seems... a cell which can copy itself or something like that.

I suspect in nature there were first many many initial lifeforms... being
modified by some process only, perhaps some radiation or whatever... and
then eventually some kind of copieing machine was invented ! ;) :)

Can this be mimmicked inside a computer simulation ?

I consider this a pretty important test ! ;) :)

If for some reason it cannot be achieved then this was cast some doubts on
evolution theory... or there is some part missing in this theory ! ;)

Bye,
Skybuck :)

[Skybuck Flying]

"
That is different issue that is not covered by theory of evolution.
What you attempt to model is called "abiogenesis" and while there are
lot of research and hypotheses about it but nothing of it is too clear yet.
https://en.wikipedia.org/wiki/Abiogenesis
"

Thanks for pointing this out.

I believe a black hole might have something to do with this. I wrote a
posting about that a while ago ! ;)

>
> I suspect in nature there were first many many initial lifeforms... being
> modified by some process only, perhaps some radiation or whatever... and
> then eventually some kind of copieing machine was invented ! ;) :)
>
> Can this be mimmicked inside a computer simulation ?
>
> I consider this a pretty important test ! ;) :)
>
> If for some reason it cannot be achieved then this was cast some doubts on
> evolution theory... or there is some part missing in this theory ! ;)

"

The part how life *started* about 4 billions years ago on our planet
is indeed missing from that theory. There are no clear theory and
modeling some millions of years of chemical processes happening on a
planet is quite expensive. Perhaps take some simpler unresolved problem
first. For example three-body problem:
https://en.wikipedia.org/wiki/Three-body_problem
;)
"

It doesn't need to be a chemical process simulation.

It could be simply a computer program, and generating random instructions or
slightly mutating those instructions, to see if some "copieing program" /
"copy instruction sequence" occurs.

This could then give some indication how easy or hard it is... and thus what
the likelyhood is of this occuring in some other system, chemical or what
not, it would give some kind of indication of what kind of mixing/mutating
process is required for something like this.

Number of instances, number of mutations, number of rounds that kind of
thing/information, instruction sequence length.

At least in corewars it only seems to take a few instructions, though those
instructions do have certain offsets and modifiers and addressing modes.

So I am still unsure how easy or hard it would be for a random
generation/modification process to ultimately generate something like that
by chance...

Bye,
Skybuck.