Question: Can we create a model that describes how 1-dimensional cellular automata calculate?

Details: I know cellular automata can calculate the solution to certain problems. These calculations can be described using a particle-model. But can we use these particles to program a cellular atomata by hand?

Sub-questions:

  • Can we deduce the rules that will produce particles given the local neighborhood?
  • How can we know how particles will interact? I.e. can we construct the particle catalog?
  • How can we connect what we know from the answer to the two questions above to construct a high-fitness solution?

Strategy:

  • Try replicating the density classification task, but using adaptive neighborhood search instead of a genetic algorithm.

Guide for rules:

111 110 101 100 011 010 001 000 // Radius = 1 0 0 1 0 0 1 1 0

100 101 010 110 001 011 100

10111 01111 11110 11101 11011 01011 10110 01101 11010 10101 // Radius = 2 0 1 1 0 1 0 1 1 1 1

11110 01101 01111 10110