Cellular Automata with Irregular Structure: A Compact Representation

Cellular Automata (CA) are a standard theoretical model of uniform parallel computation on a grid of cells. They can be looked upon as a discrete type of dynamic systems. As such, they are important tools for modeling spatially distributed processes of different kind – from ecology, through biology, and artificial life, to economics. They are also a useful theoretical model for studying classes of computational complexity. We are interested in the ability of CA to converge to a fixed point with interesting properties, which are quantized by an objective function. In this way, CA qualify as a promising embryogenic representation for Evolutionary Design. Moreover, by modifying the original CA concept by allowing irregular mesh, we can achieve a more flexible and compact representation resulting in faster evolutionary progress.