Langton's Ant unit cells?

For general discussion about Conway's Game of Life.
Post Reply
BokaBB
Posts: 2973
Joined: December 30th, 2019, 11:55 am
Location: Serbia

Langton's Ant unit cells?

Post by BokaBB » April 14th, 2021, 10:30 am

Has anyone ever made an unit cell in Conway's game of life or another life - like cellular automaton that can simulate Langton's Ant?
Have a good day!

BokaBB
777
I CAN APGSEARCH NOW!


Sure, I was a bad person, but I have changed myself.
I'd love to befriend anybody who's interested.
Have a good day!

BokaBB

GUYTU6J
Posts: 2200
Joined: August 5th, 2016, 10:27 am
Location: 拆哪!I repeat, CHINA! (a.k.a. 种花家)
Contact:

Re: Langton's Ant unit cells?

Post by GUYTU6J » April 14th, 2021, 11:33 am

As far as I know, there is an example of Langton's Ant unit cell in Wireworld, that comes with the golly package. Beside that, I haven't heard of one.

One such thing can definitely be built in Game of Life, indeed, given the vast available tools from both signal and periodic (especially HashLife-friendly) circuitry.

Compared with metapixels, the hypothetical unit cell may be simpler in design because there's no need to collect neighbouring cell's states and update every cell.

But there could be some other issues. How would the grid of unit cells handle multiple active elements (i.e. more than one ant)? And should the cell display the ant's location and orientation that makes visualization more complex with states beyond binary black/white?

Also... constructing a unit cell that can be programmed to simulate arbitrary 2-dimensional Turing machine (turmite) on a square grid is theoretically possible, but the interactions between the head and the rest of pattern become sophisticated.

User avatar
dvgrn
Moderator
Posts: 10565
Joined: May 17th, 2009, 11:00 pm
Location: Madison, WI
Contact:

Re: Langton's Ant unit cells?

Post by dvgrn » April 14th, 2021, 1:09 pm

GUYTU6J wrote:
April 14th, 2021, 11:33 am
But there could be some other issues. How would the grid of unit cells handle multiple active elements (i.e. more than one ant)? And should the cell display the ant's location and orientation that makes visualization more complex with states beyond binary black/white?
Yup, just wiring up some kind of standard stable or HashLife-friendly metapixel plus a memory cell, and allowing for glider inputs from four different directions and doing the right thing in each case, is a reasonably simple wiring task (as these things go).

If you have to also deal with the possibility of multiple ants, things get way more complicated -- you might need a lot of extra safety devices like universal regulators and timing circuits, and/or arrange all the paths so that an ant's travel time from one metacell to the next is always exactly the same. (Otherwise multiple ants could get out of synch with each other, depending on what they're doing.)

Just limit the design to "Only One Ant Allowed At A Time", and it could be a nice entertaining design project. I'd be very happy to help out, but will politely refuse to do all the work.

Post Reply