Automata of High Complexity and Methods of Increasing Their Reliability by Redundancy

A natural question regarding the limits of automation is: What are the limits of automata, or how closely can an automaton imitate the human being--what kind of specific processes can an automaton do better, more quickly, and more reliably than the human being? Aspects of these questions may be treated from the standpoint of automata theory. We will give brief summaries of two far-reaching theories in this field: the Turing theory of the possibilities of computing machines, and the yon Neumann theory of self-reproducing automata. The question of reliabili ty becomes rather serious when the complexity of the automaton becomes large. An attempt towards a theory of how redundancy should be applied on different levels for reducing the influence of temporary component errors in a universal computing machine is given.