Practical uses for cellular automata?

[LuveelVoom]

I was wondering about what real-world applications cellular automata might have. I’ve spent quite a while looking at 2(occasionally more) state 2d CA, but mostly in the exploratory domain (rulegolfing, etc…) with a little bit of engineering. Has anyone looked into practical uses for CA? Are there any papers or threads about this? What about research groups? There’s also a journal of cellular automata (except it costs 40$ per article.) Any ideas? (Wasn’t sure where to put this so I’m putting it in sandbox.)

[confocaloid]

(Attention conservation notice: this question is a multifaceted one and hopefully others will add more ideas. If you are looking for some ideas of CA-inspired or CA-based screensavers, check the forum thread: viewtopic.php?f=12&t=6463)

I was wondering about what real-world applications cellular automata might have. I’ve spent quite a while looking at 2(occasionally more) state 2d CA, but mostly in the exploratory domain (rulegolfing, etc…) with a little bit of engineering. Has anyone looked into practical uses for CA? [...]

Cellular automata can be helpful as toy models, or as ways to illustrate ideas. I see this as a real-world practical application. CA can be explored in many ways, with different ideas, techniques, mathematics. While many of those ideas are not directly/immediately CA-related and have wide consequences, some "target" is needed and cellular automata are an interesting choice for a "target".
Conway's Life, with all reactions and patterns that were found and made by various people, can help illustrate the basic idea that simple rules can lead to complex consequences which are determined by the rules. Although the same basic idea also can be illustrated with many other examples, it is helpful to have multiple examples which can be less or more accessible to different people.
Patterns with different growth rates can help illustrate different functions.
Discoveries of oscillators of every period can help illustrate the possibility of proving something about all positive integers, by proving that it is true for all large enough integers (stable Herschel tracks) and covering the remaining cases one-by-one.
I think one can simplify to "mathematics is useful; cellular automata are useful because they can help illustrate and explore mathematics".

Crossposting (what I think is) relevant discussion from a few other forum threads:

[...] Occasionally you stumble upon something interesting, and the it's the thrill of trying to find something that is one better than the last discovery.

Personally I don't have much interest in the practical applications of patterns. I like the challenge of finding patterns within certain constraints, hence why I'm interested in long-lived methuselahs and die hards. For example, I think there's something said for die hards such that the longest bounding box edge is equal to or less than the starting population. Can we find a die hard with initial population 8 and the bounding box 8x8 or smaller? What about population 9 and BB of 9x9 or smaller? I think there are some interesting ones to be found and it's just a matter of finding them. And when we do find one, can we find one that lasts longer?
[...]

[...] I prefer to state a rule, e.g. in the case of B23/S3 still lifes, "white dots must be in clusters of 2 or 3, black dots must not be in clusters of 3." But just by changing the rules, you can handle any 1-step constraint on any 2-state Moore neighborhood CA--not limited to totalistic assuming you are willing to memorize neighborhood shapes.

Here's a constellation of two phases of glider, a blinker, and a block. [...] I think it is very easy to pick out the deaths and fairly easy to pick out the births. It would need to be tested along with other designs on someone who hasn't been playing around with Life for many years.

Finally, here's a big complicated still life that came out of lifesrc. [...] I think it is quite easy for anyone to convince themselves it's a still life (aside from those who already think it's trivial). This moves the fan-out work away from your brain or visual cortex and outsources it to the tiles but leaves constraint checking in your head.

So where am I going with this? Well, I think the whole idea of doing computation with tiling constraints is intrinsically interesting (even if they're not pure geometric constraints) but I have more specific (though it's a stretch to say practical) motives:

• Can we make this a basis of puzzles and games?
  • E.g. a simple puzzle would be to build a still life out of a particular multi-set of tiles, such as the 9 tiles in a block. 4 corner pieces, 4 edge pieces, 1 interior piece.
  • Would a 2-player Scrabble-like game be playable where the goals is to build still lifes instead of words.
• A tool for analyzing/developing intuition about Life or other 2-state CAs offline. Sometimes it helps to get away from a screen.
• A decorative scheme along the lines of Truchet tiles. This is a stretch. E.g. the big still life may not seem pretty, but more symmetric patterns might be, and forcing them to be still lifes gives it a recognizable motif.

[...]

[...]
There are actually two main issues here disguised as a single issue: one is the difficulty (it's much harder to design a real-life universal constructor than one in CGoL), and another is the utility (a real-life universal constructor would have unfathomably many practical applications). That said, building a universal constructor in CGoL was already much more difficult (and lifelike) than doing the same in von Neumann's original 29-state rule, for example, so this is incremental progress in the correct direction.

Although of course none of the actual structures built in CGoL will be applicable to creating a real-life universal constructor, some of the general principles may be useful. One aspect of the RCT project was the idea of bootstrapping: where a simple construction arm builds a larger but more efficient construction arm. [...]
[...]
I expect that when a real-life universal constructor is eventually engineered, it will proceed in a similar way: bootstrapping from a biological universal constructor (i.e. ribosomes + tRNA + all of the other machinery required for DNA transcription and translation) to firstly build machines out of proteins, which then can be used to manipulate molecules and assemble them into more efficient constructors.

[d/dx]

What "journal" are you referring to exactly?