Fuzzy Modeling
Our research group focuses on fuzzy modeling, which is a set of methods based on the following theories:
- Theory of fuzzy sets and fuzzy relations,
- Theory of fuzzy natural logic,
- Theory of fuzzy transformation,
- Theory of artificial neural networks,
- Theory of probability and mathematical statistics.
Our goal is to develop these theories and and special methods based on them for solving complex problems in various areas of science and technology, especially where the available information is vague or imprecisely specified. Fuzzy modeling has been successfully applied in the following areas:
- Robust control of complex processes – our algorithms have demonstrated high resilience to significant random disturbances.
- Multi-criteria decision-making – we enable working with criteria that cannot be quantified and are expressed only in natural language.
- Utilization of fuzzy relation compositions – we have successfully applied robust models processing expert information in decision-making systems and classification models based on expert knowledge.
- Computer vision – our models are used in the development of computer vision methods.
- Efficient function approximation – our methods achieve comparable or higher accuracy than classical approaches, while being less dependent on initial conditions.
- Solving differential and integral equations – we propose robust fuzzy methods for the numerical solution of these equations.
- Time series modeling and forecasting – we develop models for the analysis and description of time series behavior and forecasting their development.
Updated: 19. 03. 2025
