International Workshop
Advances & Challenges in Computing (A2C)
A Regular meeting
Date: 25 September 2024 (Wednesday)
Time of event: 18:00-19:00 (East European time, Kyiv)
Place: online
Email: a2c.wokshop@gmail.com
AGENDA:
- Speaker: Volodymyr Manzhula
(Assoc. Prof., Candidate of Technical Sciences,
Associate Professor of Computer Science Department,
West Ukrainian National University)
E-mail: v.manzhula@wunu.edu.ua
Topic: Methods and software tools for the identification of interval models of static systems with nonlinear characteristics
Presentation in Ukrainian
Abstract: A novel comprehensive approach for identification interval models of static systems with nonlinear characteristics has been proposed. This approach combines knowledge represented as an ontological description of the theoretical foundations for identification such models with hybrid methods that integrate global search strategies and local optimization techniques. Together, this ensures guaranteed quadratic convergence and reduces the computational complexity of solving the optimization problem for parametric identification of interval models of static systems to quadratic complexity. As part of this approach, methods for both structural and parametric identification of interval models of static systems with nonlinear characteristics have been developed. These methods are based on solving an optimization problem with a differentiable objective function, using hybrid techniques that incorporate global search strategies alongside quasi-Newton methods for local optimization. Additionally, a software architecture has been designed and implemented for the mathematical modeling of static systems with nonlinear characteristics based on the analysis of interval data. Unlike existing solutions, the proposed architecture includes the implementation of an interval modeling subsystem within a service-oriented environment, optimized through the use of Google Cloud Run, the MapReduce distributed computing model, an intelligent advisory and diagnostic subsystem based on ChatGPT, free software-interpreted tools, and the use of RESTful APIs at all stages of the mathematical modeling process. Altogether, these features streamline the modeling process and make the developed services more accessible to users.
Resume: In 2000, Volodymyr Manzhula graduated from Kamianets-Podilskyi State Pedagogical Ivan Ohiienko University (now Kamianets-Podilskyi National Ivan Ohiienko University). In 2008, he defended his PhD dissertation. In 2011, he was awarded the academic title of Associate Professor. Since 2004, he has been working at the West Ukrainian National University. From 2007 to 2022, he was an Associate Professor in the Department of Computer Science; since 2022, he has been a doctoral candidate in the Department of Computer Science at WUNU. He has more than 70 scientific publications. His research interests include mathematical modeling of static systems, structural and parametric identification, computational methods, interval data analysis, and software engineering.
- Information about research projects and conferences:
Anatoliy Sachenko
(Doctor of Technical Sciences, Professor, scientific advisor of Research Institute for Intelligent Computer Systems).
E-mail: as@wunu.edu.ua