We consider (combinatorial) networks constructed by using unreliable gates with a given error probability. We show that for almost all Boolean functions f there are networks realizing f, having almost the same error probability as the gates and having almost the same complexity as the minimal (unreliable) networks realizing f in case no gate has failed (having a very great error probability). This may be contrasted with results of 1.) von Neumann (1952), 2.) Dobrushin/Ortyukov (1977), 3.) Pippenger (1985) to the effect that the number of gates needed 1.) for minimal (reliable) networks is larger by at most a logarithmic factor than the number needed for unreliable networks [5], 2.) for some Boolean functions is larger by at least a logarithmic factor, 3.) for almost all Boolean functions is a (very great) multiple of the number of gates for unreliable realizations.
CITATION STYLE
Uhlig, D. (1987). Reliable networks from unreliable gates with almost minimal complexity. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 278 LNCS, pp. 462–469). Springer Verlag. https://doi.org/10.1007/3-540-18740-5_101
Mendeley helps you to discover research relevant for your work.