Simulation Modeling Using Neural Networks to Control Complex Technical Systems

Abstract

As a result of the research, a combined neural network method, models and modeling tools for the operational control of complex technical systems have been developed and investigated. The analysis of existing methods, models and software for operational management of complex technical systems; the requirements for the developed methods, models and software for managing these systems have been substantiated. A combined neural network method for modeling complex technical systems based on a hierarchical neural network model is proposed, combining the capabilities of analytical and neural network approaches and allowing to significantly increase the efficiency of operational management (quality and efficiency of management decisions), as well as the flexibility of modeling and the possibility of operational management of calculation of structural and parametric optimization of the model based on information about the system obtained in real time. A hierarchical neural network model for the operational control of complex technical systems has been developed, consisting at the lower level of a set of interconnected neural network models corresponding to the elements of the modeled system, and a supervisor at the upper level designed to identify structural errors at the lower level of the model. A method of decision support for managing complex technical systems in real and pseudo-real time, based on the proposed hierarchical neural network model, is proposed. Software tools for combined neural network modeling and decision support have been developed for the operational control of complex technical systems.