Author Topic: Multiple ties  (Read 3404 times)

Offline Botsareus

  • Society makes it all backwards - there is a good reason for that
  • Bot God
  • *****
  • Posts: 4483
    • View Profile
Multiple ties
« on: March 30, 2009, 10:56:22 AM »
If RobotA creates a tie with "tie id 4" to RobotB what happens if RobotB then tries to make a tie with RobotA using the same id "tie id 4”? Is this command simply ignored?

P.S.

I know two robots can only have one tie between them... But does this mean the tie command is ignored after one tie between two robots but the original "tie id 4" remains?

Thankyou for help.
« Last Edit: March 30, 2009, 10:57:26 AM by Botsareus »

Offline Ta-183

  • Bot Destroyer
  • ***
  • Posts: 105
    • View Profile
Multiple ties
« Reply #1 on: March 30, 2009, 12:03:05 PM »
I haven't really done anything with ties yet so I'm not entirely sure, but bots can only have three ties so I don't think 'tieID4' is even a valid tie ID. But I might be wrong. Actually, make that likely wrong.

Offline Moonfisher

  • Bot Overlord
  • ****
  • Posts: 592
    • View Profile
Multiple ties
« Reply #2 on: March 30, 2009, 05:25:26 PM »
Hmmm I'm pretty sure bots can have more than 3 ties, but they may only be able to form 3 of them at the time (With the exception of birth ties)
I think the birth tie has ID 0 in the parent but ID 1 in the offspring.
My guess would be that ties not formed by yourself start at one and increase. (Since I always use *.numties .deltie store to clear ties and it seems to work)
Also if you write .tie inc you can form several ties, so even if you don't give them an ID I'm asuming they still recieve a unique ID (But I may be wrong).
But what I'm saying is that the tie can have different ID's at both ends... how exactly the ID is determined I'm not sure.

You'll run into another problem though... tie physics and tie angles. Multibots are very stiff an ackward and have problems facing the direction they want, and if they reproduce or make a new tie while tied the tie angles get all messed up by the birth tie somehow. I think 2.44 may have made some improvements, but there are some anoying bugs in that version IMO (Using 2.43.L)
Just saying, multibots are tough to work with... you can make a snake or an inchworm fairly easy, but once it gets more complicated the ties just tend to get in the way.
Personaly I think the most powerfull MB you can make (For F1 rules) would be a fungus tie feeder of some sort , like the evolved version of fruitflies that flourished on IM, wich survived by tieing the alge together and swarming anything hostile (Not sure how they could tell the difference, but they remained "dormant" feeding slowly from the alge till something tryed to attack the alge).

Offline Numsgil

  • Administrator
  • Bot God
  • *****
  • Posts: 7742
    • View Profile
Multiple ties
« Reply #3 on: March 31, 2009, 07:51:41 AM »
There used to be limits on ties, but the current program let's you have unlimited ties.

If a ties is being created between two bots and another tie already exists between those bots, the old die is deleted and replaced with the new tie.  This is the reason you can snap birth ties by forming a new tie with the parent and deleting that one.

Offline Ta-183

  • Bot Destroyer
  • ***
  • Posts: 105
    • View Profile
Multiple ties
« Reply #4 on: March 31, 2009, 09:34:58 AM »
Haha, I remember flyfruit. There was also an evolved Seasnake that had apparently been infected with flyfruit DNA as well as it's own DNA multiple times, and grew into a lethal fungus that whooped the ass out of even the most advanced man-made seasnake. I miss IM. It doesn't work right in 2.44, but I haven't tried 2.43L recently.


I may develop a fungal Slam-Funk strain though, I have one version that uses teamwork to hunt (but can't tell foe from food yet).

Offline Prsn828

  • Bot Destroyer
  • ***
  • Posts: 139
    • View Profile
Multiple ties
« Reply #5 on: March 31, 2009, 02:52:10 PM »
Quote from: Ta-183
Haha, I remember flyfruit. There was also an evolved Seasnake that had apparently been infected with flyfruit DNA as well as it's own DNA multiple times, and grew into a lethal fungus that whooped the ass out of even the most advanced man-made seasnake. I miss IM. It doesn't work right in 2.44, but I haven't tried 2.43L recently.


I may develop a fungal Slam-Funk strain though, I have one version that uses teamwork to hunt (but can't tell foe from food yet).

IM crashes DB almost immediately when I try to run it in DB 2.43.L
2.44.1 just crashes in general, lol.

I regret never being able to experience it with more than just myself, but oh well...

So I didn't realize there wasn't a limit to the number of ties; does that apply to 2.44.1, or to 2.43.L, or is it for both?
So, what will it be? Will you submit to my will, or must I bend reality to suit my needs?
Better answer before I do BOTH!

Offline Moonfisher

  • Bot Overlord
  • ****
  • Posts: 592
    • View Profile
Multiple ties
« Reply #6 on: March 31, 2009, 05:44:09 PM »
After EricL shut down his sims IM is pretty quiet. Not sure where the teleporter populations are uploaded to, I hope it wasn't something Eric was hosting.
And the Seasnake you mentioned is the one I ment... flyfruit was a simple blind fleeing alge with a strong anti viral defence (Easy to make for an alge since store costs don't compare to 40nrg a cycle), it had a virus version of fruitflies in it as well as 50 other viruses (Or maybe less, think I copy pasted a large bunch a few times to lower the odds that enemy viruses would land as the last gene).
It also just selected the virus to launch randomly from it's gene count, so if it got hit by a new virus it would have a 49/50 (And increasing for each infection) chance to just acumulate the virus and use it aswell.
(The alge only had copy errors enabled, to mutate new offspring and viruses slightly)
Stole a lot of viruses wherever I could find them to make it interesting, and a seasnake got infected by several viruses including fruitflies. Main reason I started working on the thing was because Eric had issued the 9K bots on IM challenge (I don't recall seeing the infected seasnakes pass 7000 though, but I can't be sure if other species where infected aswell, in the end I'm not sure it would realy count anyway, but fruitflies was having a hard time spreading through teleporters by conventional means.)

Anyway the best version I saw pop into my sim whas this one I think :
Also I figured out what triggers the attacks... poison... so bots that don't use poison can actualy beat it, bots that do use poison get slaughtered from what I've seen. I'm guessing it's predators on IM all used poison.
And it's not actualy a real MB since it doesn't set the multi tag, but it seems to stiffen ties and wigles around to end up tieing all the alge in a large blob.
(If running them against bots I recomend lowering the alge count and maybe just inserting a single one of them, then the population should allow a propper framerate.)

[div class=\'codetop\']CODE[div class=\'codemain\' style=\'height:200px;white-space:pre;overflow:auto\']
''''''''''''''''''''''''  Gene:  1 Begins at position  1  '''''''''''''''''''''''
 cond
 *111 431 !=
 start
 99 .sharenrg store
 .dnalen store
 300 .memloc store
 336 973 store
 *.body 5 >
 *.nrg 80 >
 dropbool
 *.maxvel .dn store
 50 .repro store
 *.totalmyspecies 1000 <
 *.nrg 200 >
 and
 60 mult not
 *.body 1 >
 and
 *.body 0 sub 10 drop 10 91 floor 312 *.robage !=
 0 *.numties 1 -3 >
 .tienum inc
 *.numties .up store
 >
 or
 and
 *.numties .deltie store
 dupbool
 *973 and
 =
 1 453 .tieloc store
 1000 .tieval store
 not
 and
 =
 2 *973 dupbool
 452 .tieloc store
 -1 .tieval store
 dropbool
 *973 2 >
 and
 8 .tieval store
 1 *.body =
 0 *.numties =
 and
 dupbool
 *.eye5 0 =
 *.memval *.dnalen dec
 =
 or
 and
 350 .aimright store
 *.eye5 0 *.dnalen !%=
 and
 *.refbody 20 >
 *.totalmyspecies 600 >
 or
 and
 *.refxpos *.refypos floor .setaim store
 .tie inc
 clearbool
 dec
 *.thisgene *.genes sgn add >=
 *.genes 777 *777 1 sub -1 mult sqr dup *777 .genes dec
 *.thisgene 1 sub *.thisgene 1 sub sgn mult *.thisgene 1 add *.thisgene 1 sub sgn 1 ceil dec
 stop
''''''''''''''''''''''''  Gene:  1 Ends at position  179  '''''''''''''''''''''''

''''''''''''''''''''''''  Gene:  2 Begins at position  180  '''''''''''''''''''''''
 cond
 *104 431 2 !=
 start
 .eye5width store
 336 .eye3width store
 336 .tmemloc store
 *.body 80 >
 and
 *.maxvel store
 402 -133 1000 <
 *.nrg 200 >
 and
 ^ 60 *.body 10 mult sub .strbody inc
 not
 *.body 10 mult 10 sub 0 floor .fdbody store
 *.numties 0 div *.robage 1 >
 and
 *.tiepres .tienum store
 *.tmemval *.dnalen !=
 *.numties 8 >
 or
 and
 *.numties .deltie store
 *.tmemval *.dnalen !=
 452 .tieloc store
 -1 .tieval store
 *.numties 0 =
 and
 dupbool
 *.eye5 0 =
 *.memval dec
 *.dnalen =
 or
 and
 350 .aimright store
 dropbool
 *.memval *308 !=
 and
 *683
''''''''''''''''''''''''  Gene:  2 Ends at position  278  '''''''''''''''''''''''
''''''''''''''''''''''''  Gene:  3 Begins at position  279  '''''''''''''''''''''''
 cond
 .setaim store
 .tie inc
 clearbool
 *.genes
''''''''''''''''''''''''  Gene:  3 Ends at position  285  '''''''''''''''''''''''
''''''''''''''''''''''''  Gene:  4 Begins at position  286  '''''''''''''''''''''''
 cond
 .genes dec
 *.thisgene *.genes sgn add .mkvirus store
 *.genes 777 *777 1 sub overbool
 mult sqr dup store
 mult *777 store
 *.thisgene 1 sub sgn 1 *.thisgene *.shootval mult sqr *.thisgene add 7 1 sub clear add sgn -1 mult mult add .delgene store
 else
''''''''''''''''''''''''  Gene:  4 Ends at position  331  '''''''''''''''''''''''


I think this one was infected twice by the fruiflies virus, and possibly also by a second smaller virus... all mutated together to create this new behavior. It seemed to do well as an evo bot because of the many offspring, not actualy something I had expected. But personaly found it very interesting that random mutations had no problems improving fruitflies for IM... all the bots to pop into my sim had this behavior on some level, so it was clearly working out better than the original design.

Offline Numsgil

  • Administrator
  • Bot God
  • *****
  • Posts: 7742
    • View Profile
Multiple ties
« Reply #7 on: April 01, 2009, 07:30:41 AM »
IM uploads to an FTP subdomain of the darwinbots.com site.  So as long as the forums work IM mode will have somewhere to upload to.  But yeah, without Eric's constantly running comps IM is pretty pointless.