question, i like your video i was wondering if u can help me with a question i came up with an idea lets say there are two pc's ,in pc 1 there are three box's that are out putting letter from a-z box=f box=r box=w the next set of boxs box=h box=l box=x box 2 a-z box-3 a-z is it possible to see what pc 1 is out puting then take that info and manually in put it into pc 2 then could pc 2 know where pc 1 is going and what pc 1s next out put would be before it comes out
i wish there was some more variables to make it more interesting maybe higher predators and more variables for the pray maybe to help avoid flocking, and maybe an environment thats not so uniform
I'm new to genetic programing. but am quickly learning. One idea i had while watching this would probably be very difficult to implement, but would have a very stunning result. If you could some how send the data on the surrounding cells, up to like 2 away, then they will learn more than to just run, they'd actually learn to run AWAY! xD but that would be an increase of 12 new pieces of information, might get some monsters of programs...
The implementation of that wouldn't be too complicated, but a naiive implementation of an algorithm that finds nearby elements for each element is naturally an order N^2 problem.
is this array based or vector based? If it is array based, which i what i assumed in my comment, You could pass it's current location and then have it search a square like a 5 by 5 sqaure around it only. that would greatly reduce the amount of processor power needed. how ever, if it is vector based... I'll get back to you on that. xD
well, the evolution pattern is repeating over and over...only those who run constantly have a small chance to survive - until a hunter from other direction hits them. those who stall are eaten. not a single group of hunters nor single group of prey survived longer.
What you can do is have an array, about 100 long for each organism, which randomly mutates when they split, and have different commands, by going to different subroutines if there are different numbers in the array.
I did something somehat similar at some point. Instead of predators you had to kill the beings by clicking on them.
The movement pattern was made up of two vector field defined by matrices of forces. One of the matrices would be centered in the mouse cursor, the other one in the center of the world.
Results were pretty unimpressive though" The world matrix was random, the cursor matrix was an outward pointing circle.
I hope it's clear... Anyway nice work! Reminded me of that project.
I think if anything you need more organisms in the small space. You can see the populations of successful hunters die off from lack of energy after they consume the school of prey they're chasing. It's like there's a big, energy-less void between each group of organisms.
I think for more genetic continuity it may be wise to make the energy requirements for movement less, or put more energy into the system.
There is an oxygenlimit, if it is raised more creatures may exist. But this needs much computingpower. And there is the danger of no evolution at all because both reproducing for prey and hunting for hunters gets more easy.
Expand the space, then add another species or two or twenty, for variety and competition, then maybe add a little more genetic information to each, then add some environmental factors maybe? I want to see these simulations get complicated as hell!!!
You know, I would love to learn the code you used to write this sufficiently enough to cooperate with you on it. I've had an idea floating around of creating an open-source simulation that's constantly having more variables added and spends thousands of generations evolving freely. This is how I theorize AI will be developed - the same way actual intelligence was.
Nice simulation! How is the movement behavior encoded into the genome?
AsymptoteInverse 11 months ago
this is not evolution as neith the preditor nor the prey developes in any way from thier original state
vincentmagician 1 year ago
@vincentmagician Patterns of movement develop,
deathsminion25 4 months ago
@deathsminion25 oooohhhh they developed running in the downward direction... didnt notice that...
vincentmagician 4 months ago
is this a cellular autonoma?
mooglebee 1 year ago
@mooglebee Nope.
sevenclev 1 year ago
This has been flagged as spam show
manattan342 1 year ago
can you send a download link to me?
laputahayom 2 years ago
@laputahayom Sorry, I lost all sources in a system crash last year.
sevenclev 2 years ago
i wish there was some more variables to make it more interesting maybe higher predators and more variables for the pray maybe to help avoid flocking, and maybe an environment thats not so uniform
Prontest 2 years ago
This brings to mind sand games where you try to set powder on fire whilst it's falling.
matty406 2 years ago
Is the formation of lines genetic or just how you introduce the next generation?
I might be interesting if the preditor had to eat much more prey to stay alive, like birds and bugs.
sjh7132 2 years ago
They form lines by themselves. Sometimes they even form corners, if they run diagonally,
sevenclev 2 years ago
I see... so as the prey runs from the hunters, the hunters follow, and increase in numbers and then decrease as they run out of prey?
Plozen 2 years ago
I'm new to genetic programing. but am quickly learning. One idea i had while watching this would probably be very difficult to implement, but would have a very stunning result. If you could some how send the data on the surrounding cells, up to like 2 away, then they will learn more than to just run, they'd actually learn to run AWAY! xD but that would be an increase of 12 new pieces of information, might get some monsters of programs...
snkatk 2 years ago
The implementation of that wouldn't be too complicated, but a naiive implementation of an algorithm that finds nearby elements for each element is naturally an order N^2 problem.
matt1746 2 years ago
is this array based or vector based? If it is array based, which i what i assumed in my comment, You could pass it's current location and then have it search a square like a 5 by 5 sqaure around it only. that would greatly reduce the amount of processor power needed. how ever, if it is vector based... I'll get back to you on that. xD
snkatk 2 years ago
well, the evolution pattern is repeating over and over...only those who run constantly have a small chance to survive - until a hunter from other direction hits them. those who stall are eaten. not a single group of hunters nor single group of prey survived longer.
although visual effect is astonishing ;)
renoor 2 years ago
Yes, that is why I called it "tiny". Not much evolution is going on...
sevenclev 2 years ago
how did you implement the "genetic component"?
Is it a matrix? How does it tell the dots how to move?
LogicRising 3 years ago
Its a simple array of direction-values (North, Northwest etc,). The creeps continuously iterate through the array and walk according to these values.
I believe for this simulation every array consisted of 16 values. I also tried variable lengths, but that resulted in absolute chaos.
sevenclev 2 years ago
What you can do is have an array, about 100 long for each organism, which randomly mutates when they split, and have different commands, by going to different subroutines if there are different numbers in the array.
pasadema1 2 years ago
I did something somehat similar at some point. Instead of predators you had to kill the beings by clicking on them.
The movement pattern was made up of two vector field defined by matrices of forces. One of the matrices would be centered in the mouse cursor, the other one in the center of the world.
Results were pretty unimpressive though" The world matrix was random, the cursor matrix was an outward pointing circle.
I hope it's clear... Anyway nice work! Reminded me of that project.
cristicbz 2 years ago
I think if anything you need more organisms in the small space. You can see the populations of successful hunters die off from lack of energy after they consume the school of prey they're chasing. It's like there's a big, energy-less void between each group of organisms.
I think for more genetic continuity it may be wise to make the energy requirements for movement less, or put more energy into the system.
I don't know, just speculation ,but hella cool.
OverusedChewToy 3 years ago
Depends.
There is an oxygenlimit, if it is raised more creatures may exist. But this needs much computingpower. And there is the danger of no evolution at all because both reproducing for prey and hunting for hunters gets more easy.
sevenclev 3 years ago
Expand the space, then add another species or two or twenty, for variety and competition, then maybe add a little more genetic information to each, then add some environmental factors maybe? I want to see these simulations get complicated as hell!!!
GBart 3 years ago
Yeah I wanted to extend it, but time is sparse...
"Species" is a relative term, since most of the creatures have diffrent genetic codes and hence would be described as diffrent species.
If I ever extend this simulation, I will try to add senses, so creatures cann learn to use information in their surroundings.
sevenclev 3 years ago
You know, I would love to learn the code you used to write this sufficiently enough to cooperate with you on it. I've had an idea floating around of creating an open-source simulation that's constantly having more variables added and spends thousands of generations evolving freely. This is how I theorize AI will be developed - the same way actual intelligence was.
GBart 3 years ago
The idea is intresting, but my Code would not be a good basis for that.
Anyway, its written in Java and I could make it available, if you want it.
sevenclev 3 years ago
sweet
damianpoirier 3 years ago