Random Expansion Method for the Generation of Complex Cellular Automata

TL;DR

This paper introduces a novel method for generating complex cellular automata by random specification and refinement, enabling the creation of automata with hundreds of states that exhibit intricate behaviors.

Contribution

It proposes a random expansion approach combined with genetic algorithms to systematically generate and refine complex cellular automata.

Findings

Randomly generated automata can produce complex behaviors.

The method scales to automata with hundreds of states.

Refinement with genetic algorithms enhances complexity.

Abstract

The emergence of complex behaviors in cellular automata is an area that has been widely developed in recent years with the intention to generate and analyze automata that produce space-moving patterns or gliders that interact in a periodic background. Frequently, this type of automata has been found through either an exhaustive search or a meticulous construction of the evolution rule. In this study, the specification of cellular automata with complex behaviors was obtained by utilizing randomly generated specimens. In particular, it proposed that a cellular automaton of $n$ states should be specified at random and then extended to another automaton with a higher number of states so that the original automaton operates as a periodic background where the additional states serve to define the gliders. Moreover, this study presented an explanation of this method. Furthermore, the random way of defining complex cellular automata was studied by using mean-field approximations for various states and local entropy measures. This specification was refined with a genetic algorithm to obtain specimens with a higher degree of complexity. With this methodology, it was possible to generate complex automata with hundreds of states, demonstrating that randomly defined local interactions with multiple states can construct complexity.