Actually I'm just evolving for max average height, and slowly raising the time they're being tested in. They actually sorta look like a guy! But with only one leg. Anyway every time I leave it be for it to start getting "perfected" they always balance for a while and then do a jump or flail at the end to squeeze a bit of max height out before they'd fall over.
Random question, is possible or will it be possible to evolve organisms that can react and respond to changing enviornment? (As in, change direction, follow something, cooperate, etc.)
Random suggestion, didn't know where else to put it- I'm evolving a creature that balances, and it always decides to do something stupid right before the test ends because it knows it won't pay for it. :P Maybe there should be an option to slightly alter the runtime of each test (scaled) so they don't evolve stupid end-behaviors?
That's interesting. Are you evolving for maximizing jumping height, by any chance? I've seen the sort of thing you describe, but only for jumpers - they do one really good jump and then do just silly stuff until time runs out.
It all depends on the selective pressure the program applied. If our environment prefered hopping then more animals would hop. This is only a simulation not a recreation. Perhaps if it were allowed to run for many more generations walking would evolve, but again it all depends on how he programmed the creatures and thier environment.
Yeah, plus the physics of 3DVCE is not tweaked to match or even approximate reality very well. I suspect the gravity is a little low, making it too easy to evolve a jumping movement, tempting or steering the population away from other possible strategies.
Actually I'm just evolving for max average height, and slowly raising the time they're being tested in. They actually sorta look like a guy! But with only one leg. Anyway every time I leave it be for it to start getting "perfected" they always balance for a while and then do a jump or flail at the end to squeeze a bit of max height out before they'd fall over.
SolaceAvatar 2 years ago
That sounds cool!
kjlg74 2 years ago
... I seem to have killed that file somehow, sorry. T_T;
SolaceAvatar 2 years ago
Random question, is possible or will it be possible to evolve organisms that can react and respond to changing enviornment? (As in, change direction, follow something, cooperate, etc.)
Asurnasurpal 2 years ago
I hope eventually I can add some complexity in that general direction. Eventually
kjlg74 2 years ago
Random suggestion, didn't know where else to put it- I'm evolving a creature that balances, and it always decides to do something stupid right before the test ends because it knows it won't pay for it. :P Maybe there should be an option to slightly alter the runtime of each test (scaled) so they don't evolve stupid end-behaviors?
SolaceAvatar 2 years ago
That's interesting. Are you evolving for maximizing jumping height, by any chance? I've seen the sort of thing you describe, but only for jumpers - they do one really good jump and then do just silly stuff until time runs out.
kjlg74 2 years ago
It all depends on the selective pressure the program applied. If our environment prefered hopping then more animals would hop. This is only a simulation not a recreation. Perhaps if it were allowed to run for many more generations walking would evolve, but again it all depends on how he programmed the creatures and thier environment.
dlandon2000 2 years ago
good point. it might also be that my genetic representation happens to lend itself more easily to jumping bodies and brains than to walking ones.
kjlg74 2 years ago
Because we evolved from swimmers. And even then, many animals have reverted to jumping (frogs, grasshoppers, etc.).
shanedk 2 years ago
The energy is more or less free. Some kind of penalty or cost for using energy is something I should implement one day.
kjlg74 2 years ago
Good question. I have yet to see anything that I'd really call walking. It's a little disappointing.
kjlg74 2 years ago
Maybe walking just isn't that efficient. Look at how long it took to get robots to walk.
shanedk 2 years ago
Yeah, plus the physics of 3DVCE is not tweaked to match or even approximate reality very well. I suspect the gravity is a little low, making it too easy to evolve a jumping movement, tempting or steering the population away from other possible strategies.
kjlg74 2 years ago
They look like a herd of geometric gazelles. :-)
papafox 2 years ago
I like how some of them occasionally go, "Wheeeeee!!!"
shanedk 2 years ago
I think those guys might be abusing the physics engine. They get some pretty good height!
kjlg74 2 years ago
Haha! I love Neos!
shanedk 2 years ago