Emmanuel Medina showed many mappings of relationships between the elementary cellular automata when started from a single black cell. He investigated three different types of grouping: CAs that could emulate each other, CAs who could be "composed" by updating each cell as the results of applying two different rules sequentially before printing the next row, and CAs which produced a different rule when some function was applied to the first or last cell of the update neighborhood to produce another rule. His results showed that he was grouping similar-looking CA together very well. He could eliminate any rule numbers he wanted from his graph, and the overall clustering was preserved.
Comments