Aerospace Instrument-Making Annotation << Back Intellectual means in processes of diagnosing of refusals of onboard control systems of aircraft in the conditions of environment influence. Part II Drogaytsev V.S. Kulikov R.E. The formalistic approach to creation of a neural networks of decision-making in pro-cesses of detection and diagnosing of refusals of onboard control systems by power plants of aircraft in the course of their test in the conditions of influence of factors of environment is considered. The fragment of structure of model of the neural network intended for decision-making in processes of detection and diagnosing of refusals of onboard control systems. The technology of construction and training of a neural network is based on use of intellectual systems of support of formal procedures of methods of the solution of problem tasks. The regulation made by each layer of a neural network of decisions is provided with a logical formalism of data and knowledge bases of required intellectual systems. The offered method allows to define models of neural networks, to reveal demanded lists of neurons in network layers, to synthesize the reference images distinguished by neurons and to form means of neurons of activation functions. The output parameters providing identification of the set list of refusals of onboard control systems are characterized by word length of binary codes in the form of activation functions for an output layer of a neural network. This approach to creation of a neural networks allows to reduce time intervals of cycles of construction and training of networks and to increase reliability of made decisions. Key words: neural network, reference image, activation function, synaptic communications, hidden layer, neuron. Contacts: E-mail: va.ushakov@mail.ru Pp. 03-09.