Attestation committee
Accreditation committee
Expert committee
Dispositions, instructions
Normative acts
Nomenclature
Institutions
Scientific councils
Seminars
Theses
Scientific advisers
Scientists
Doctoral students
Postdoctoral students
CNAA logo

 română | русский | english


Adaptive hardware architectures for neuro-fuzzy systems with self-organization


Author: Cărbune Viorel
Degree:doctor of engineering
Speciality: 05.13.13 - Computers, computational systems and networks
Year:2020
Scientific adviser: Sergiu Zaporojan
doctor, associate professor (docent), Technical University of Moldova
Institution: Technical University of Moldova

Status

Term of presenting of the thesis 18 December, 2020

Abstract

Adobe PDF document0.59 Mb / in romanian

Thesis

CZU 004.3/004.8

Adobe PDF document 3.53 Mb / in romanian
157 pages


Keywords

knowledge extraction methods, decision support system, fuzzy system, neural system, hardware adaptive architectures

Summary

Thesis structure. The Ph.D. thesis comprises the introduction, four chapters, conclusions, and bibliography (120 titles), 4 appendixes, 131 pages of main text, 81 figures and 5 tables. The obtained results are published in 15 scientific articles.

The study domain includes theoretical and practical aspects of human knowledge extraction methologies.

The purpose of research consists in development of new models, methods and algorithms for extracting human expert knowledge, developing adaptive hardware architectures for researching decision-making processes and building decision support systems in industrial applications.

The research objectives include the analysis of general aspects of neuro-fuzzy systems, knowledge extraction methods, research and development of decision support methods and algorithms, design, simulation and analysis of adaptive hardware architectures.

The scientific novelty consists in proposing new models, methods, algorithms for knowledge extraction and parameterized hardware structures. The originality of the proposed solutions consists in approaching and combining intelligent "machine learning" techniques with the behavioral model of the human operator.

The solved scientific problem resides in elaboration and research of the original methods of human operator experience extraction through collecting and processing of automatically generated statistical data, which led to a new approach in the extraction of knowledge.

The theoretical significance of the work consists in the elaboration and development of original methods that can be used to extract the experience of the qualified operator. The presented approach involves development of algorithms for knowledge extracting from the experience of the human operator.

The applied value consists in: proposing models, methods and algorithms for processing data of technological process evolution.

The implementation of results consists in the use of the elaborated models and systems in “Microfir Tehnologii Industriale” S.R.L. company confirmed by the act of implementation.