Evolution in Gameplay

Welcome back, Falthron.

i think that idea is maybe too lamarckian, and as a system for a way to evolve it seems too restraining.

PaperGrape wrote:

This video might help explain it better.It really talks about the issues we speak of at the end. [url= http://il.youtube.com/watch?v=_m97_kL4ox0] http://il.youtube.com/watch?v=_m97_kL4ox0[/url] Polyworld is an attempt to get smart AIs through evolution.

Wait- is Polyworld an "official" project? Like, is Google running it? Could we somehow download Polyworld? Is it OpenSource?

PaperGrape wrote:

This video might help explain it better.It really talks about the issues we speak of at the end. http://il.youtube.com/watch?v=_m97_kL4ox0 http://il.youtube.com/watch?v=_m97_kL4ox0 Polyworld is an attempt to get smart AIs through evolution.

Fixed.

InvaderZim wrote:

PaperGrape wrote:

This video might help explain it better.It really talks about the issues we speak of at the end. [url= http://il.youtube.com/watch?v=_m97_kL4ox0] http://il.youtube.com/watch?v=_m97_kL4ox0[/url] Polyworld is an attempt to get smart AIs through evolution.

Wait- is Polyworld an "official" project? Like, is Google running it? Could we somehow download Polyworld? Is it OpenSource?

If you are thinking what I am thinking you are thinking (I think I am, B2!), then no. Just no.

~sciocont wrote:

Welcome back, Falthron.

i think that idea is maybe too lamarckian, and as a system for a way to evolve it seems too restraining.

Actually is sort of isn't, I'm trying what you said to, mix the two. You see it adds up variables and after X generations the one who survived are those who developed the new features that were usefukl to your life sytle. The more generations in between the more change you can have within it.

Lamarck is the theory that the organisms actions that influeced its abilities will immediately passed on to the next generation.

The difference within mine is that it is the ones who survived using "your" lifestyle developed features to make the lifestyle that was setas the norm easier.

Falthron wrote:

Actually is sort of isn't, I'm trying what you said to, mix the two. You see it adds up variables and after X generations the one who survived are those who developed the new features that were usefukl to your life sytle. The more generations in between the more change you can have within it.

Lamarck is the theory that the organisms actions that influeced its abilities will immediately passed on to the next generation.

The difference within mine is that it is the ones who survived using "your" lifestyle developed features to make the lifestyle that was setas the norm easier.

Falthron is speaking the truth, it is probably as close to Darwinian as you can get with equations and such. I think that this was how ParadoxJuice planned to run SuperAutoEvo way back when he existed.

Bashinerox wrote:

If you are thinking what I am thinking you are thinking (I think I am, B2!), then no. Just no.

Oh. Darn.

US_of_Alaska wrote:

Falthron wrote:

Actually is sort of isn't, I'm trying what you said to, mix the two. You see it adds up variables and after X generations the one who survived are those who developed the new features that were usefukl to your life sytle. The more generations in between the more change you can have within it.

Lamarck is the theory that the organisms actions that influeced its abilities will immediately passed on to the next generation.

The difference within mine is that it is the ones who survived using "your" lifestyle developed features to make the lifestyle that was setas the norm easier.

Falthron is speaking the truth, it is probably as close to Darwinian as you can get with equations and such. I think that this was how ParadoxJuice planned to run SuperAutoEvo way back when he existed.

I think you are right. It's a near-perfect combination of the two.

I vote guided mutations for Organism mode, and Direct edit for God Mode. (although you could use guided mutations in god mode if you wanted to)

I say go Bashi's way, still. Guided mutations would be good as well, i think i talked bout a way to do that in the previous forums.

~sciocont wrote:

I say go Bashi's way, still. Guided mutations would be good as well, i think i talked bout a way to do that in the previous forums.

Do you mean with the direction editor?

No, it was where you chose from a few different offspring that were generated by auto evo. The offspring you choose would be the ones that pop up in game, therefore, you would enhance natural selection before offspring were born.

Say, the question of the evolution of the other organisms was never answered. How will they evolve?

DragonEye4 wrote:

Say, the question of the evolution of the other organisms was never answered. How will they evolve?

Most likely through a Lamarckian-Darwinian hybrid.

We should still test all methods in the actual engine and then choose the most effective one. Nothing can substitute conjecture like actual testing.

GamerXA wrote:

We should still test all methods in the actual engine and then choose the most effective one. Nothing can substitute conjecture like actual testing.

Well the engine won't exist for a while, i'm afraid.

We might be able to test the concept with minor prototypes.

Indeed, we might. And I too am leaning heavily toward darwinian evoution; So far that I am nearly falling off the chair. Metaphorically, of course.*Shifty eyes* I would like to note that quite often, a parent's lifestyle will affect their child's a bit. Just a tad.
-Waap.

GamerXA wrote:

We should still test all methods in the actual engine and then choose the most effective one. Nothing can substitute conjecture like actual testing.

That's why Bashi is making a highly modular engine. We can, during testing, switch between different modes of evolution.

Well, that is certainly good. Very good.
-Waap.

US_of_Alaska wrote:

DragonEye4 wrote:

Say, the question of the evolution of the other organisms was never answered. How will they evolve?

Most likely through a Lamarckian-Darwinian hybrid.

But I thought that the OP states that evolution without player interaction we would not be able to do?

DragonEye4 wrote:

US_of_Alaska wrote:

DragonEye4 wrote:

Say, the question of the evolution of the other organisms was never answered. How will they evolve?

Most likely through a Lamarckian-Darwinian hybrid.

But I thought that the OP states that evolution without player interaction we would not be able to do?

No, it states that it would be very difficult. If these animals follow both darwinian and lamarckian rules, they should evolve with the same simplicity that you do, although, since they have little player interaction, they will probably evolve slower.

if something that makes you reject Darwinian evolution is the complexity of the world then we should add complexity, like ice ages that change the water levels and temp, droughts, floods, meteor collisions( that actually wipe out species not the spore like meteors that don't even kill you even if fallen in your head) massive coronal ejections, slow atmosphere composition and temperature changes, volcanos and super volcanos, UFOs (that actually change the biosphere with new, removing old or changing current species), earthquakes, hurricanes and .... CONTINENTAL DRIFT. movement of the tectonic plates is probably one of the biggest mutation factors cause it isolates species, makes and destroys habitats and much more.


most of this can be selected with a random math expression only the last one is harder but it's only needed till the space stage and only in your homeworld.


that will improve natural selection work...

eumesmo wrote:

if something that makes you reject Darwinian evolution is the complexity of the world then we should add complexity, like ice ages that change the water levels and temp, droughts, floods, meteor collisions( that actually wipe out species not the spore like meteors that don't even kill you even if fallen in your head) massive coronal ejections, slow atmosphere composition and temperature changes, volcanos and super volcanos, UFOs (that actually change the biosphere with new, removing old or changing current species), earthquakes, hurricanes and .... CONTINENTAL DRIFT. movement of the tectonic plates is probably one of the biggest mutation factors cause it isolates species, makes and destroys habitats and much more.


most of this can be selected with a random math expression only the last one is harder but it's only needed till the space stage and only in your homeworld.


that will improve natural selection work...

We're going to have all that sort of things. They don't solve the problem of complexity.

Imagine, for a moment, there is a butterfly in Thailand. It flaps its wings. Now, in the real world this will set up a chain of events that could lead to something on a vrey large scale. in a game, we simply can't keep track of everything happening. We're not going to have constant fluid dynamics simulations running, we're not going to factor in the gravitational pull of Jupiter on the butterfly's left leg. Now do you understand the problem better? If we want completely realistic evolution, we need a completely realistic world , which is, at this point itn technology, impossible. We can't even get close on a good modern computer.

I believe that's called Chaos Theory, and I don't believe that's been proven to be completely as it says. Even if we can not simulate the universe exactly we may not need to...

Speaking of which.., are we going to keep track of space debris that are large enough to get a sizable effect on the planets life? I believe that we should.

GamerXA wrote:

I believe that's called Chaos Theory, and I don't believe that's been proven to be completely as it says. Even if we can not simulate the universe exactly we may not need to...

Speaking of which.., are we going to keep track of space debris that are large enough to get a sizable effect on the planets life? I believe that we should.

anything big enough to causse noticeable changes will be there, but "keeping track of" isn't really the right term. We'll recognize it when the game decides it's time for an impact or something.