Aerospace Instrument-Making Annotation << Back USING NEURAL NETWORK METHODS FOR THE MATHEMATICAL DESCRIPTION OF THE ACTIONS OF TARGETED SYSTEMS V.V. Lavlinskiy, S.N. Yanshin The developed mathematical models for describing the actions of various targeted systems have some uncertainty of qualitative and quantitative parameters. In view of this, the aim of the work is to find approaches to the use of specific methods that allow taking into account this kind of interaction. Methods: mathematical apparatus of neural networks. Results: a model for describing the actions of targeted systems based on the MATLAB mathematical package is developed. Practical relevance: the proposed model allows us to evaluate the actions of systems based on descriptions of changes in continuous signals of inputs and outputs, with direct communication, two interconnected systems with discrete inputs and outputs. Keywords: actions, system, goal-oriented system, neural network, mathematical model. DOI: 10.25791/aviakosmos.6.2021.1227 Pp. 58-66.