Search academic texts

Registration Number

0223U001134, 0121U107499 , R & D reports

Title

Development of methods and tools for analysis and prediction of dynamic behavior of nonlinear objects

popup.stage_title

Head

Liovkin Valerii M.,

Registration Date

25-01-2023

Organization

Zaporizhzhya Polytechnic National University

popup.description2

 The object of research is the process of diagnosing the condition and behavior of complex objects. The purpose of the work is to develop and study methods and tools for decision-making for data processing in intelligent systems for determining the future state and dynamic behavior of complex objects and processes. Research methods include methods of computational intelligence, machine learning, forecasting, methods of decision theory, parallel computing. Methods of synthesis of predictive models are analyzed. Methods of coding and further use of information about neural networks are analyzed and improved. A method of structural-parametric synthesis of neural network models based on reinforcement training has been developed. In particular, a method based on the asynchronous method of the actor-critic has been developed to support genetic diversity during the neuroevolutionary synthesis of predictive models based on artificial neural networks. The results of research are recommended for use in solving problems of technical and medical diagnosis and forecasting, support the decision-making process and more.

Product Description

popup.authors

Hofman Yevhen O.
Kornienko Olexandr V
Leoshchenko Serhii D.
Lovkin Vаlerii M.
Skrupska Iryna