Algorithmic model of formal neuron synthesis and minimization of logical function

The mathematical methods of neuron synthesis are based on the correlation principle. Being actually a completed mathematical method, VLSI neuron synthesis apparatus has nevertheless some disadvantages caused by the development of scientific-technical progress, One of these disadvantages is that the synthesis has been worked out for logical functions (LF) of a comparatively limited number of arguments. Any increase in number of LF arguments leads to the situation when manual conversion is impossible. However the majority of the existing methods of synthesis have insufficient degree of formalization which makes the development of effective computer aided design systems rather difficult. FN is a new step in the development of threshold cells for which methods of computer identification of threshold and non-threshold LF have not yet been worked out, preventing developing new methods of FN synthesis. Methods based on the combinatory properties of LF with the help of which the realization of the conversions for LF of many variables in computers are described. Besides for a limited number of variables these methods secure much more effective solution than traditional methods of analogous conversions can do