Code center > Darwinbots3
Organelles, better multicellular organisms, and more
NotLegalTender:
--- Quote from: Peter on April 19, 2013, 04:05:37 PM ---Could you give a few examples on what a organel in db should do?
What could be improved with multibots?
--- End quote ---
Multibots - Make them behave more like multicellular organisms, or atleast, make it possible to do so.
Organelles - Metabolize food, photosynthesize, respire, chemosynthesize, store water, thermosynthesize, produce shell, produce poison, and so on.
Peter:
What should be changed/added so it would behave more like multicellular organisms?
NotLegalTender:
--- Quote from: Peter on April 19, 2013, 04:11:13 PM ---What should be changed/added so it would behave more like multicellular organisms?
--- End quote ---
Some multibots are incapable of reproducing on their own without pulling off parts of them. Not only that, they are the bots that break the easiest.
Peter:
There have been some changes with tie mechanics in the past with DB2. Breaking/making unstable quite some multibots.
I think there were ideas for metabolism(organel kind of involved) and specialization(makes multibot more appealable ). But I've been a bit out of the loop. So, I think it's smarter letting Numsgil explain it all. :P
Numsgil:
--- Quote from: NotLegalTender on April 19, 2013, 03:31:53 PM ---So I've heard that veggies are gonna have chloroplasts. We should allow for all organisms to have organelles. Each organelle could serve a different purpose. Multibots/multicellular organisms should be made better, and add the possibility of them evolving naturally on their own.
--- End quote ---
I'd like to add organelles, but I don't want to add them just for the sake of having them.
Chloroplasts make for interesting decisions, because as you build more and more to start generating your own energy, the increased volume, mass and intertia make it very difficult to steal energy from others. And the bigger you are, the less likely you are to be able to fend off attacks from others. This creates an interesting dynamic that's equally accessible to evolution and bot writers, and there's no single right answer. Different strategies can use different levels of chloroplasts.
By comparison, an old idea I had was to have substances in the world, each with their own given free energy level, and with random substances tagged as shell, slime, etc. This would force bots to navigate a dense web to figure out how to metabolize one substance to turn it in to another. The problem with this is that it's not necessarily an interesting problem. There's an optimal metabolic chain to build from the given substances, and while it would be interesting to see an evo sim try to evolve that optimal pathway, it's not at all an interesting thing for bot writers to play with. They'd always copy+paste the optimal solution, and the optimal solution could be found with something like Dijkstra's algorithm. So it's complicated and boring. Bleh.
I'm sort of thinking of having something like cilia for movement, where having more would let you move faster, but building more costs resources. And every time you divide you have to also divide up the cilia among the daughter cells. Likewise for various other things a bot can do. But I need to think through all the implications, and make sure it's not just extra pieces to manage for the sake of having more stuff to do. It needs to fundamentally add interesting strategic decisions to the sim. I think it will, but I want to be very deliberate about this sort of thing.
--- Quote ---Astronomy
--- End quote ---
Not high on my list, but there should be some interesting things you can do with the fluid sim I'm looking to have in DB3. Periodically swirl things around a bit, like a tide, or alter the day/night pattern in different areas of the sim.
--- Quote ---Possession
--- End quote ---
Also not high on my list, but very doable. It would just consist of tying a game pad or keyboard in to calling certain DNA codules. I'd probably add network play just for the heck of it, although that's a whole can of worms (determinism - getting the sim to run EXACTLY the same on two different machines given the same input - is really hard. Determinism is how most multiplayer games, like Starcraft, work. But others cheat and send over an updated state every second or two)
--- Quote ---Spores and pollen should be forms of reproduction added, and sexrepro should be made easier to code. Also, vegetables should be capable of producing fruit and seeds, to distribute them further across the map.
--- End quote ---
There's no explicit spore or pollen feature, but it wouldn't be hard for a very small bot to either hibernate until an opportune moment or fertilize a large bot. In terms of smaller bots hitching a ride on a larger one, that will be possible. Bots will be able to just hard attach themselves to each other, like barnacles.
--- Quote ---Also, chemosynthesis and thermosynthesis should be added. However, this might require stuff like hydrothermal vents, and volcanism.
--- End quote ---
This would have to tie in with the fluid sim to be really useful, but I'm not 100% sure I know how to do diffusion in a way that's efficient. But something I'm keeping in the back of my mind.
Navigation
[0] Message Index
[#] Next page
[*] Previous page
Go to full version