Neighborhood detection and rule selection from cellular automata patterns

Using Genetic Algorithms (GAs) to search for cellular automation (CA) rules from spatio-temporal patterns produced in CA evolution is usually complica ted and time-consuming when both the neighborhood structure and the local r ule are searched simultaneously. The complexity of this problem motivates t he development of a new search which separates the neighborhood detection f rom the GA search. In this paper, the neighborhood is determined by indepen dently selecting terms from a large term set on the basis of the contributi on each term makes to the next state of the cell to be updated. The GA sear ch is then started with a considerably smaller set of candidate rules pre-d efined by the detected neighborhood. This approach is tested over a large s et of one-dimensional (1-D) and two-dimensional (2-D) CA rules. Simulation results illustrate the efficiency of the new algorithm.