Glushakova G. Methods of syntesis and models of components unconventional neural networks

Dissertation is devoted to development of synthesis methods and models of unconventional, reliable, fault-tolerance and hardy neuronets, wich don't required any special error detecting and correction facilities. A new neuronet paradigm, automate-logical- predicat model of artificial F-neuron and synthesis methods of its state exchange function together with the new mathematical and computer models of F-neuronet components have been developed in proposed work. The author provides the simulation of F-neuronet functioning in order to prove the conformity of developed models. Proposed models should be used in the field where the control systems failure may caused the disastrous effects.