MATHEMATICAL ASPECTS OF NEURO-FUZZY TECHNOLOGY APPLICATION IN PROJECT MANAGEMENT

  • Elena Kiseleva
  • Olga Prytomanova
  • Sergiy Zhuravel

Abstract

The combination of neural networks and fuzzy logic provides the core of
neuro-fuzzy technologies, are fundamentally different mathematical constructions. Artificial
neural networks are considered as universal models of the human brain, capable of
learning the recognition of unknown regularities. Artificial neural networks are built on
the principle of organization and functioning of their biological analogues (networks of
human brain neurons). In recent years, neural networks have become a practice wherever
problems of identification, prediction, classification, management need to be addressed.
Unlike neural networks, in which unstructured numerical data is used to find a solution
to a problem through training and tuning, fuzzy technologies (fuzzy systems) use expert information
about the regularities found in available experimental data in the form of natural
language rules “IF-TO “. These rules are formalized with the help of fuzzy logic, they allow
to build identification models with relatively small datasets.

Published
2018-01-27
Section
Mathematical Modeling of Systems and Processes