## Exploring cellular automata 2 — my lovely maggot children

In some other post I linked to the Wolfram “atlas” of 1-dimensional elementary cellular automata, the list of 256 rules that can apply to a row of “on” or “off” pixels to determine a successive row, in which a pixel’s successive value is only dependent on its own and those of its two neighbors. There are only 256 of these because of this limitation.

We can extend this to a 2-dimensional space. This seems to be easier to follow what’s going on — maybe just by the nature of looking at a computer screen. It’s the space that the Game of Life lives in. A grid of pixels can be on or off, and a successor grid follows, with each pixel’s state determined by its initial state and those of its neighbors. In order to visualize this, you could lay each successor on top of its parent grid to make a 3D object (analogous to forming a 2D image for a 1D array of pixels). But this would be really hard to look through. Instead, just play each image after its successor as an animation.

A ruleset can be described with a series of images, like the 1-D automaton, but again, you’d have to either stack the start example on top of the resulting center pixel, or place them side by side. In addition, there are going to be many more images necessary for a rule, as there are more configurations of on-and-off neighboring cells in 2 dimensions.

This also introduces the uncertainty in what a neighborhood is in 2D. We could consider the 4 cells in the cardinal directions as the only neighbors (the von Neumann neighborhood), or we could tack on the corners, too (the Moore neighborhood).

For the von Neumann neighborhood, each rule alone would require $2^5 = 32$ images. With a choice of on or off for the resulting cell for each image, there are 1024 elementary cellular automata. Not a bewildering number, but still a lot.

The Moore neighborhood would have 9 squares, replacing the center square, and so one would need $2^9 = 512$ icons for each rule. There are $512^2 = 262144$ rules of this type. Yikes! There are the same number of squares in the “extended” von Neumann neighborhood, so these numbers apply there as well.

Both of these neighborhoods are probably characterized completely — tracking a few of the properties of 263000 systems doesn’t seem to be out of the realm of possibility, especially when considering that this number can be reduced by a factor of due to horizontal-mirroring, vertical-mirroring, centerpoint-mirroring, and color-inverting of each of these. So, the 512 in the 5-cell neighborhood reduces to only 64 with unique behaviors, and the 9-cell neighborhood reduces to 32,768.

What seemed exciting was to look at larger systems. A Super von Neumann neighborhood of 21 squares will have 4,398,046,511,104 rulesets of $2^{21} =$ 2,097,152 icons each, and the Super Moore neighborhood of 25 squares has 1,125,899,906,842,624 rulesets of $2^25 =$ 33,554,432 icons each. Now we’re having fun.

It felt like there was some chance of picking something weird out of here.

## A ruleset based on an oscillator

One contruction I didn’t mention in the last post was an oscillator — a repeating, unmoving pattern. Some oscillators can be seen in the wakes of the dirty puffers.

Oscillators are emergent in GoL; there isn’t anything about the rule that explicitly reflects an intent of repetition. I was interested in making a rule that was based on an oscillator, and to see if something like spaceships would emerge.

### Maggotworld

The simplest oscillator I could think of that seemed interesting was a clockwise spinning line. 3 activated cells, with a diagonal between the first two, would appear to spin around in a circle. If there were three cells in this configuration, the next epoch would deactivate the cell in the back, and activate a new one in the next step around a circle. This resulted in an 8-period oscillator. I originally called the idea Spinworld, but it soon seemed appropriate to call it Maggotworld.

Instead of matching the Super Moore neighborhood around every pixel with one of the 33.5 million in the ruleset, the maggotworld propagator just checks if there are three activated pixels in the right configuration nearby to turn on. There are eight ways this could happen — corresponding to the eight squares that are activated throughout the cycle. This doesn’t mean there are only eight icons of the 33.5 million that lead to an activated central pixel. The pixels toggle; a dead pixel toggles on if there are one, or three maggots approaching it. It toggles off if there are two or four approaching it. The same is likewise true for the tail ends of a maggot — two maggots with the same tail will leave the tail point on, as it is toggled twice. So there are many ways pixels could be left on or off depending on the situation. This seems ripe for emergent behavior!

I ended up trying a bunch of different setups to see how they behaved way back in April, and wrote most of the above then, but haven’t gotten around to this showcase until now.

#### Random starts

I began by just slapping random starting configurations down and seeing what happened.

After some initial discouragement, some neat patterns did emerge. In the last animation you can spot a few figure 8s, figure triple-Os(?), and maggots passing straight through each other. Not a spaceship, but at least something a bit more complicated.

#### Not so random starts

If I couldn’t get a spaceship out of this, I could at least look for something cool. Time to get systematic. I placed maggots at varying relative phases and positions to start and let them go.

This was a bit discouraging — none of the ones that touched made any lasting patterns beyond a single oscillator. Anything interesting would have to be made of more than 2 of these, or have some stationary gunk waiting around to push them around a bit.

In one of the random setups, I saw this little friend, which I’ll call a slingshot oscillator.

It’s got a period of 14. Having a different period is… useful? How is any of this useful? I dunno whatever.

Another neat result would be to get a pattern that turns on cells outside of the normal oscillator radius. Here’s one. It looks like a really low resolution hand-style cursor. Let’s call it the flinger.

The last thing of this type is a little u. I like this because, while it seems like these rules easily reduce the number of lit cells, this generates an extra sixth cell by the end of its process.

#### Behavior in grids

If I couldn’t get a spaceship flying away in empty space, maybe I could get something neat with stationary patterns placed ahead of time. I started by plopping oscillators in a lattice of lit points. It’s real easy to get some ongoing domino action this way.

Vertical motion. There’s a lot of trash left behind, but the key to see here is that only one or two lit squares next to a lit bar drives the motion. The lit side determines which way it travels.

#### Trams and fuses

The last one looked neat. I tried to make more “trams” following a track. Here are a bunch. There are also configurations that rip the track apart. Let’s call em fuses.

Some of the styles are repeated with a hole in the track to show how they deal with that. Holes are also placed at different locations along the track for the same style of tram to show off how they travel: trams hop two spaces at a time, not one. An “oddly” placed hole will force it to reflect or scatter differently than an “evenly” placed one.

I was taken with the self-repairing track on the bottom.

This seemed like a way to create a caterpillar-like spaceship of indefinite length that scootches forward a couple squares each cycle. I wasn’t able to figure how to get it to repeat, thought.

Another fuse had an end that, after some acrobatics, sent another tram that flipped the orientation of the fuse.

#### Timers

One last thing and I’ll let you go, I swear. A row of lit squares can be used as a timer of indefinite length. This could go up above with the other oscillators, but the fact that this can be stretched to as long as you need it puts it in its own class.

These can also take a turn. These turns could be used to stop the motion, or to initiate the timer in the first place:

Maybe these are more appropriately called fuses? But I already used that word.

Showing these little tricks and behaviors is interesting enough alone for me, but it would be really swell if we could figure out how to produce oscillators / timers of truly any period. I’m also still wondering if there is an elusory spaceship, one without a grid or track to support it, hiding in these rules. The holy grail after that is to get these things to do binary logic, but let’s save that for another day.

# I Lied

I said I had only one last thing but I totally lied to you. You must be feeling pretty silly right now! Just one more bit. I promise. 🤞

Before finishing up with the maggotworld animations, I had tried a lot of different rulesets to see how they’d act. One of them was simple: A cell would turn on if it had an odd number of neighbors (in the Moore neighborhood), and off otherwise.

What a pretty pattern! I was pleased to learn later that I had recreated something called Fredkin’s Replicator. You can see some explanation in Numberphile’s Terrific Toothpick Patterns (starting at 14:20). A lecture also by Neil Sloane gives more insight into analyzing this and other “odd-ruled” cellular automata. Do yourself a favor and check out at least the Numberphile vid.

## Exploring cellular automata 1

Conway’s game of life is maybe the most well known ruleset for cellular automata. Each cell in a grid has some possible set of states, and set of rules for changing state based on its own and it’s (usually) immediate neighbor cells. Time elapses with a number of discrete steps, sometimes called epochs. In Game of Life, a cell can either be “alive” or “dead” (on or off), and will live for the next time step only if it has one or two alive neighbors.

Some incredible patterns emerge, including spaceships, guns (creating gliders and spaceships), breeders (creating guns), other puffersrakes, long-lived methuselahs (this example creating some gliders).

Other rulesets lead to neat patterns. Here’s a fun one with many cell states. It lets you rate the results as it scans some starting parameter space.

So, we’re certainly not limited to one ruleset. Coming up with one is easy — finding if there are any neat patterns maybe not so much. It might be better to first explore simpler rules — ones in only 1 dimension. In this case, a linear space is segmented into cells, and each cell has only two neighbors. Given some configuration of three cells in a row, the middle cell’s state in the next epoch will be dependent on its own state as well as the two alongside. When considering how to change any cell, since there are only three cells to consider, and only two possible states for each, there are only $2^3 = 8$ arrangements to consider when changing any cell. Because of this small number, it’s no problem to display the rule of the automaton as a list of diagrams, each showing one of the eight starting arrangements and the resulting state of the middle cell in the next epoch.

With some initial state, for example:

The next line is produced by scanning over this line and matching the top row with the one from the rule icons.

Note that this wraps around, connecting the two edges. This process then continues indefinitely for successive lines. We’re free to pick any length of line and any arrangement of on and off cells as an initial condition, but once this is done, the rest is automated. This particular initial condition for this rule ends up looking like this:

This is kind of neat, but these processes don’t get to show off unless they’ve got some room to work with. Here’s the same rule applied to an initial state of a single black cell in a long row of white.

The repeating triangles are a pattern characteristic of some of the more interesting rules. Rules 30, 90, 110, and 184 are supposedly the most cited in scientific or mathematics articles.

Because there are only 8 arrangements of cells we need to specify to determine a 1-D ruleset, and each arrangement offers the choice of the middle cell turning on or off, there are $2^8 = 256$ possible rules of this type. You can see links to them all here. One could argue there are only 128 rules, 64, or even 32, since negating all the colors, taking a mirror image, or negating the mirror image doesn’t change the behavior much. This means each rule is one of a set of four that operate in the same way.

I’ve already linked to the Wolfram Atlas a few times. The images above were taken from there or made using Mathematica. Stephen Wolfram, of Mathematica fame, is also well-known for his work on cellular automata. He’s argued that these things may be more fundamental to nature than mathematics. It’s considered controversial. I haven’t read A New Kind of Science, a book where he lays this all out, but it’s free and online. Maybe worth checking out. It seems like some of the argument hinges on the fact that some of these automata are Turing complete — they can be set up to operate and produce results like any computer program.

I’d just like to note that my interest in cellular automata is based pretty much entirely on having fun. It’s 11 o’ clock on a Saturday night.

A step after checking out the 1-D rules is to explore the 2-D rules. Next I’ll show off a rule and try to do something interesting with it.