Designing Domain-Particle Dynamics in Cellular Automata
We develop a method to design one dimensional cellular automata rules that produces repetitive patterns, called domains, and localised dislocations that have behaviour similar to particles in their dynamics.We refer to this as domain-particle dynamics. The two types of domains focused on are cyclic domains and more general domain automata. The designer is comprised of two stages. The first involves producing a template of the CA rule, in a method specific to each type of domain, from which a rule sample can be generated. For cyclic domains with rule radius greater than two, an extra criterion of domain growth is needed. The second stage calculates the domain area and the ratio of thin domain area to whole domain area from the space-time diagrams in order to exclude inappropriate rules. Domain-particle CA can then be selected by eye. The results show the design method works.
Keywords: Designing, filtering, particle dynamics, domains, emergence, one dimensional cellular automata, parallel programming, GPU computing.