![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}
Chemical evolution for Cellular Automata
Mar. 11th, 2019 07:58 pmSomething I've wanted to do years ago was to take a cellular automata, probably Conway's Game of Life, and do a particular set-up with it. First, take a very large but finite board, non-wrapping, with spaces outside the board treated as dead. Then, once every X ticks, with X being some number preferably larger than 3, but easily larger, the whole board is shifted in one direction, let's say down so it's like they're falling, by one cell, with the cells at the top being filled in randomly and the ones at the bottom simply being discarded.
Then, you have a system where patterns can only remain on the board if they move in the opposite direction of the shift, at the same speed! Given that, I would like to try having the shift be large enough to allow for larger/more complex ships. Then, as the iterations continue, there would be a continued input of new patterns into an environment with a clear selective pressure, which to me seemed comparable in concept to simple genetic algorithms, so I figured this would somehow lead to more sophisticated patterns developing over time.
However, at the time I somehow wasn't able to really invest myself in wrapping my head around scripting for Golly, so I let the idea sit there for a while. Recently, though, I saw a video on chemical evolution ( https://www.youtube.com/watch?v=mRzxTzKIsp8 ). The model of chemical evolution, with fatty acids accumulating in water due to neither floating or sinking, reminded me of the idea of having a CA setup that selects for patterns that neither rise nor sink on a board that constantly scrolls.
Also, I'm now slightly better/more experience dealing with scripting, and think I might be able to pull this off now! So I'll want to try it out soon, though I'd also like to keep writing all these ideas I've mostly kept in my head for a while.
Then, you have a system where patterns can only remain on the board if they move in the opposite direction of the shift, at the same speed! Given that, I would like to try having the shift be large enough to allow for larger/more complex ships. Then, as the iterations continue, there would be a continued input of new patterns into an environment with a clear selective pressure, which to me seemed comparable in concept to simple genetic algorithms, so I figured this would somehow lead to more sophisticated patterns developing over time.
However, at the time I somehow wasn't able to really invest myself in wrapping my head around scripting for Golly, so I let the idea sit there for a while. Recently, though, I saw a video on chemical evolution ( https://www.youtube.com/watch?v=mRzxTzKIsp8 ). The model of chemical evolution, with fatty acids accumulating in water due to neither floating or sinking, reminded me of the idea of having a CA setup that selects for patterns that neither rise nor sink on a board that constantly scrolls.
Also, I'm now slightly better/more experience dealing with scripting, and think I might be able to pull this off now! So I'll want to try it out soon, though I'd also like to keep writing all these ideas I've mostly kept in my head for a while.
