A very effective density classifier two-dimensional cellular automaton with memory
Ramón Alonso‐Sanz, Larry Bull · Journal of Physics A Mathematical and Theoretical · 2009
Conventional cellular automata (CA) are memoryless, i.e. the new state of a cell depends on the neighborhood configuration solely at the preceding time step. This paper considers an extension to the standard framework of CA by implementing memory capability in cells. It is shown that the HPP rule, one of the most important block automaton rules, endowed with the memory of the most frequent recent state, behaves as an excellent classifier of the density in the initial configuration, which surpasses the performance of the best two-dimensional density classifier reported in the literature.