Multiple Neighborhood Cellular Automata (MNCA) - Slackermanz

[cgoler2]

I think a good example of a rule would be B2/S/Nm+B1/S34/Nvn.

[Slackermanz]

I think a good example of a rule would be B2/S/Nm+B1/S34/Nvn.

Is that valid notation for an existing program?

I like the format, but it does have a problem that would exclude a lot of patterns - there's no way to specify the order of the transitions independently from the neighborhood. Sometimes good results require that the order of transitions doesn't necessarily respect the order of NH assessments.

Yours looks something like this:

Code: Select all

NH0
Transition_N0_0
Transition_N0_1

NH1
Transition_N1_0
Transition_N1_1

But I've found using the transitions out-of-order can be important in getting certain results:

Code: Select all

NH0
NH1

Transition_N0_0
Transition_N1_0
Transition_N1_1
Transition_N0_1

[bprentice]

I think a good example of a rule would be B2/S/Nm+B1/S34/Nvn.

You could test this example using one of the RuleTreeGen files.

there's no way to specify the order of the transitions independently from the neighborhood.

Any transition order can be coded in a RuleTreeGen file to test whether the proposed rule produces interesting results.

Brian Prentice

[Slackermanz]

Brian Prentice

dvgrn

There's further implementation details in a post I just published here, as well as a ShaderToy implementation.

https://slackermanz.com/understanding-m ... -automata/

If you were wanting to give MNCA implementations a try, this might add some clarity (I'd hope), and if not, let me know where it's confusing.

[Slackermanz]

New video showing mutation and evolution across 225 generations:

https://www.youtube.com/watch?v=xyx5C40HpQM

[ihatecorderships]

The behavior of these rules are interesting. Just to clarify, MNCA is still on a square grid, just with different neighborhoods, right?
Also, do you think something like non-totalistic MNCA would be possible?

[Slackermanz]

The behavior of these rules are interesting. Just to clarify, MNCA is still on a square grid, just with different neighborhoods, right?
Also, do you think something like non-totalistic MNCA would be possible?

Yes, these still take place on a square grid in my implementations, but that doesn't need to be the case. I imagine the same would go for totalistic/non-totalistic CA: The MNCA technique itself augments the rulespace, so it should offer benefits to any similar domains it can be applied to.

[Slackermanz]

Some recent developments in my exploration of MNCA:

Complex groups with emergent symmetry, movement, and rotation:
https://www.youtube.com/watch?v=O2IYLhXGN8s

Pretty crazy/lifelike-looking soliton aggregation that self-assembles into coherent structures:
https://twitter.com/slackermanz/status/ ... 5915354119