Author Topic: Only tiny bots in my evosim.  (Read 6543 times)

Offline Testlund

  • Bot God
  • *****
  • Posts: 1574
    • View Profile
Only tiny bots in my evosim.
« on: October 24, 2006, 07:03:04 AM »
I'm not sure if it is a bug, but I think it's strange that my bots only produce very tiny bots. Written bots seem to produce larger bots. So the numbers are dwindling in my sim and I'm afraid I will have to turn off all costs for them to survive.
The internet is corrupt and controlled by criminally minded people.

Offline Henk

  • Bot Destroyer
  • ***
  • Posts: 110
    • View Profile
Only tiny bots in my evosim.
« Reply #1 on: October 24, 2006, 10:00:44 AM »
Have they got .strbody commands? Because if they don't build up new body the size of bots will decrease as generations pass...
cond
*.DBbugs 0 =
start
.rejoice inc
stop

Offline EricL

  • Administrator
  • Bot God
  • *****
  • Posts: 2266
    • View Profile
Only tiny bots in my evosim.
« Reply #2 on: October 24, 2006, 10:34:27 AM »
In sims with a lot of energy flying around, namely sims using corpse mode and the no decay option on nrg shots, is more advantageous to use nrg to reproduce then to grow.  This is becuase reproducing increases your cross section and thus your probability of intercepting a shot to a much larger degree than growing does.

Additionlly, if using dynamic costs, a super high reproduction rate can be used as a weapon since it spikes CostX and make life harder for competitors.  Even though your mortality rate will be huge, with so many more individuals, the smaller, faster reproducing bots will out compete the slower, larger guys over time in a high CostX environment since they can cover much more area and thus have a much lower probability that the entire population will go extinct due to lack of food.  If you only have a few, larger, slower reproducing individuals, the chances that there will be a time where they can't see food for a while is higher, which is deadly even to larger, stronger bots in a high CostX environment.  Putting yoru eggs in a thousand baskets out competes putting them in 10.

So basically, the environment today favors the evolution of fast reproducing bacteria in many sims.  We could damp this down considerably if we added gestation and maturity times so that a bot could not reproduce for X cycles after birth and for Y cycles after giving birth.  The longer we made these, the more it would encourage larger, more long lived bots.  I favor adding these since I think most of us are interested in breeding smarter macro behaviour, not bacteria, something that is more likely to emerge for longer lived, macro organisms....
Many beers....

Offline Testlund

  • Bot God
  • *****
  • Posts: 1574
    • View Profile
Only tiny bots in my evosim.
« Reply #3 on: October 24, 2006, 11:43:59 AM »
But I was thinking that the size of the offspring depends on how much energy the parent wants to give to it's child. Alga minimalis for instance is designed to give 50 % of it's energy to it's child, making them both the equal size. In my sim I think it whould be completely random how much energy the offspring will get, depending on how the mutations turn out. But it's allways a very tiny potion of energy that the parent gives to it's child.

I don't know if they have .strbody commands. In any case I have set it so 10 % of energy recived goes into body. They are all autotrophs by the way. Also I have it set to energy per kilobody which means the larger they are the faster they will get energy. That together with an ageing cost of 20 starting at 20000 cycles should make it much easier for larger bots to survive and reproduse imo. Though Eric's explanation makes sense too, it whould be more likely for a heterotroph population, I think.

I'm not sure about adding a maturity function to the sim. I think about the bots as bacteria that is dividing when they reproduce, and real bacteria doesn't have any maturity time as far as I know. Alga Minimalis for instance only reproduce when it has built up a lot of energy, which takes many cycles. Sometimes bots just turn cancerous though and I don't understand what's triggering that. Logically it should only happen if it is written in the dna.
The internet is corrupt and controlled by criminally minded people.

Offline EricL

  • Administrator
  • Bot God
  • *****
  • Posts: 2266
    • View Profile
Only tiny bots in my evosim.
« Reply #4 on: October 24, 2006, 02:51:14 PM »
No matter what percentage a parent gives an offspring, the resulting two bots will present more cross sectional area than the parent alone.  Given this, it is not surprising that in an nrg shot rich sim, selection favors parents giving offspring tiny amounts so they can produce as many offspring as possible, present the most cross sectional area and increase the probability that some offspring will happen upon a food source.

Unless the nrg inflow is huge, tiny autotrophs with kilobody feeding are essentially hetrotrophs.  The nrg they get via autotrophy is negligable compared to other sources, particularly with lots of nrg shots flying around.  With that combo, the price of having 99 out of 100 offspring die quickly isn't as high as it sound since the nrg isn't lost to the survivors.  They are close enough to absorb a significant percentage of the corpse decay.

Even in a energy rich environment, it still takes some amount of time for bacteria to divide and some amount of time before they can do it again.  There are morphological processes at work that need time to happen even if there is plenty of energy around.  We are not simulating this.  Bots can reproduce every cycle if they want.
« Last Edit: October 24, 2006, 03:26:42 PM by EricL »
Many beers....

Offline Jez

  • Bot Overlord
  • ****
  • Posts: 788
    • View Profile
Only tiny bots in my evosim.
« Reply #5 on: October 24, 2006, 05:31:55 PM »
I have to agree with the size perspective, I remember reading a book once where a planet had been overtaken by AI 'bots. They had evolved into billions of tiny dust like particles rather than enormous war machines.

The largest part of animal biomatter on our planet, if I remember correctly and may Shwarz slap me if I get it wrong, is very very small.
If you try and take a cat apart to see how it works, the first thing you have in your hands is a non-working cat.
Douglas Adams