A simple mathematical system made of cells arranged on a grid. Cells have a state; all states evolve simultaneously according to a uniform set of rules such that the state at step i+1 depends on the state in step i of the cell in question and of cells in a small neighbourhood. Such a discrete dynamical system may serve to model physical systems; large cellular automata, despite their simplicity at the local level, can show behaviour of substantial complexity. As information processing systems, cellular automata may also be regarded as a subclass of artificial neural networks, in which node connections are of the nearest-neighbour type in two dimensions. see Wolfram86, Raghavan93.