Darwinbots Forum

Bots and Simulations => Bestiary => F2 bots => Topic started by: bacillus on October 07, 2008, 10:49:58 PM

Title: Hydra(F2MB)(Bacillus)(08-10-2008)
Post by: bacillus on October 07, 2008, 10:49:58 PM: An attempt to make a multibot that uses more than two ties in an organized fashion, whilst trying to keep it simple and effective. You be the judge if it worked or not
Code: [Select]
'Hydra 'A multibot by Bacillus 'Uses no steering system, instead relies on fixed cells acting as 'pivot for other bots. Very basic in its workings, basically a 'branching animal multi. Pretty stupid, but can get VERY nasty 'very quickly if it finds a field of veggies. def type 990 def head 1 def main 2 cond *.robage 0 = start .tie inc .head .type store stop start 500 *.body sub dup .strbody store - .fdbody store 10 30 rnd *.vel sub .up store 333 .out1 store 333 .tout1 store *.type .out2 store *.type .tout2 store 50 .sharenrg store 50 .sharewaste store *.in1 *.out1 sub sgn abs *.eye5 30 sub sgn 0 floor mult .fixpos store stop cond *.nrg 6000 > *.numties 3 < *.eye5 30 < start 50 .repro store .main .type store stop cond *.in1 *.out1 != *.eye5 30 > start 16 .shootval store *.refxpos *.refypos angle .setaim store -6 .shoot store *.velup 2 sub .dn store stop cond *.in1 *.out1 = *.shflav 0 = start 314 .aimdx store stop cond *.in1 *.out1 = *.eye5 0 = or *.shflav 0 != start *.shang .aimdx store 0 .shflav store stop cond *.type .main = *.waste 100 > start *.waste .shootval store -4 .shoot store stop
Title: Hydra(F2MB)(Bacillus)
Post by: De-Maskus on October 12, 2008, 01:51:21 PM: Quote from: bacillus
An attempt to make a multibot that uses more than two ties in an organized fashion, whilst trying to keep it simple and effective. You be the judge if it worked or not
Code: [Select]
'Hydra 'A multibot by Bacillus 'Uses no steering system, instead relies on fixed cells acting as 'pivot for other bots. Very basic in its workings, basically a 'branching animal multi. Pretty stupid, but can get VERY nasty 'very quickly if it finds a field of veggies. def type 990 def head 1 def main 2 cond *.robage 0 = start .tie inc .head .type store stop start 500 *.body sub dup .strbody store - .fdbody store 10 30 rnd *.vel sub .up store 333 .out1 store 333 .tout1 store *.type .out2 store *.type .tout2 store 50 .sharenrg store 50 .sharewaste store *.in1 *.out1 sub sgn abs *.eye5 30 sub sgn 0 floor mult .fixpos store stop cond *.nrg 6000 > *.numties 3 < *.eye5 30 < start 50 .repro store .main .type store stop cond *.in1 *.out1 != *.eye5 30 > start 16 .shootval store *.refxpos *.refypos angle .setaim store -6 .shoot store *.velup 2 sub .dn store stop cond *.in1 *.out1 = *.shflav 0 = start 314 .aimdx store stop cond *.in1 *.out1 = *.eye5 0 = or *.shflav 0 != start *.shang .aimdx store 0 .shflav store stop cond *.type .main = *.waste 100 > start *.waste .shootval store -4 .shoot store stop

I tried it, and I wasn't very impressed with it's behavior. Mainly in it's lack of coordinated locomotion. However, I was impressed by a certain unforeseen effect of accidentally leaving mutations enabled. It developed more eccentric movement routines in only 1 total mutation, which was really impressive. I'm gonna let that sim run a while longer and see if anything else develops.

EDIT: While I was making this post, four more mutations happened, creating a particularly virulent strain. I haven't smiled this wide in a while.
Title: Hydra(F2MB)(Bacillus)
Post by: bacillus on October 13, 2008, 12:11:20 AM: It is a very stupid bot, but I wanted to avoid the mess of figuring out and dealing with the turning effects of tie physics with more than two ties, so I just skipped it altogether.
Post the mutation, I really want to see a version of this gone good
Title: Hydra(F2MB)(Bacillus)
Post by: De-Maskus on October 15, 2008, 10:01:22 PM: Quote from: bacillus
It is a very stupid bot, but I wanted to avoid the mess of figuring out and dealing with the turning effects of tie physics with more than two ties, so I just skipped it altogether.
Post the mutation, I really want to see a version of this gone good

I wouldn't say 'gone good', it just hauls relatively more ass than the regular version.

Hydra Velocans

Code: [Select]
'Hydra Velocans 'A multibot by Bacillus 'Mutation observed by DeMaskus 'Creates erratic movement '#generation: 8 '#mutations: 3 cond *.robage 0 = start 348 inc 1 990 store stop '''''''''''''''''''''''' Gene: 1 Ends at position 11 ''''''''''''''''''''''' start 500 *.body sub dup .strbody store - 312 *.vel sub .up store 333 .out1 store 333 .tout1 store *990 .out2 store *990 .tout2 store 50 .sharenrg store 50 .sharewaste store *.in1 *.out1 sub sgn abs *.eye5 30 sub sgn 0 floor mult .fixpos store stop '''''''''''''''''''''''' Gene: 2 Ends at position 57 ''''''''''''''''''''''' cond *.nrg 6000 > *.numties 3 < *.eye5 30 < start 50 .repro store 2 990 store '''''''''''''''''''''''' Gene: 3 Ends at position 74 ''''''''''''''''''''''' else '''''''''''''''''''''''' Gene: 4 Ends at position 75 ''''''''''''''''''''''' cond *.in1 *.out1 != *.eye5 30 > start 16 .shootval store *.refxpos *.refypos angle .setaim store -6 .shoot store *.vel 2 sub .dn store stop '''''''''''''''''''''''' Gene: 5 Ends at position 100 ''''''''''''''''''''''' cond *.in1 *.out1 = *.shflav 0 = start 314 .aimright store stop '''''''''''''''''''''''' Gene: 6 Ends at position 112 ''''''''''''''''''''''' cond *.in1 *.out1 = *.eye5 0 = or *.shflav 0 != start *.shang .aimright store 0 .shflav store stop '''''''''''''''''''''''' Gene: 7 Ends at position 131 ''''''''''''''''''''''' cond *990 2 = *.waste 100 > start *.waste .shootval store -4 .shoot store stop '''''''''''''''''''''''' Gene: 8 Ends at position 146 ''''''''''''''''''''''' '#hash: <EQs7R%40B}#+$ha7%x$
EDIT: Oh wow, I didn't completely look at the DNA before now. It seems as if a few genes were exposed to a major deletion. Perhaps if you combine the two DNA files into one, just back-to-back and copy-pasted it a few times into the file to ensure DNA integrity you might be able to evolve a better functioning version. Overall, it seems to be pretty mutation proof when you look at it's behavior. By mutation proof of course I mean maintaining its general characteristics. A few plug-and-play genes, a few hundred thousand cycles of evolution, and we may have a valid competitor here.
Title: Hydra(F2MB)(Bacillus)
Post by: bacillus on October 15, 2008, 11:47:06 PM: Looks pretty unstable to me, not even a multibot (in theory it is, but it doesn't have any permanent ties). Got mashed by the original version.
Title: Hydra(F2MB)(Bacillus)
Post by: De-Maskus on October 16, 2008, 07:00:38 PM: Quote from: bacillus
Looks pretty unstable to me, not even a multibot (in theory it is, but it doesn't have any permanent ties). Got mashed by the original version.
True. It seems I may have mistaken it as a multibot when it simply didnt get rid of birth ties.

What if you re-inserted the tie-ing gene into the mutated DNA and ran it?
Title: Hydra(F2MB)(Bacillus)
Post by: bacillus on October 16, 2008, 11:56:40 PM: It doesn't make a difference, it just snaps the ties.
Title: Hydra(F2MB)(Bacillus)
Post by: De-Maskus on October 17, 2008, 06:16:41 PM: Quote from: bacillus
It doesn't make a difference, it just snaps the ties.
That's strange.