Candidate of
Economic Sciences Omelianenko O.P.
applicant Zinchenko
MM
Kiev National
University of Civil Engineering and Architecture
DEVELOPMENT OF THE NEWEST METHODS OF ANALYSIS AND
FORECASTING IN THE CONTEXT OF MANAGEMENT OF ECONOMIC SYSTEMS (WITH THE USE OF
FUZZY LOGIC TOOLS)
The
importance of developing and attracting new methods of analysis
and forecasting for management of economic systems
is also conditioned
by a significant inconsistency with the current conditions
of widely used classical approaches, resulting in a significant increase in the
number and magnitude of economic
crises throughout the world during
the last decade. Accordingly, the scientific research is devoted
to the development
of theoretical and methodological provisions and economic and mathematical
models for the analysis and
forecasting of the development of economic systems
with the use of fuzzy
logic tools.
Investment
activity forms an important place
in the formation
of economic security. In order
to increase the adequacy of
modeling and analysis of these
systems, it is essential to
apply fuzzy logic methods, which is a methodology
and a mathematical apparatus, which enables to set
and mathematically reasonably solve even such problems,
for which there is no
any complete statistics, or in the case
, when among the informative factors there are
only qualitative indicators, while ensuring the possibility
of adapting economic and mathematical
models to changing economic conditions. For the application of models built
on fuzzy logic, it is
not mandatory to adhere to
the hypothesis of compliance with
normal distribution or statistical homogeneity of random processes, which is especially
important for young markets that
are actively developing, and in particular, Ukrainian.
It
is shown that the common
quantitative methods for identifying nonlinear dependencies are not adapted
for using expert information about the object
of research in the form
of logical rules, as well
as in the
case of unclear
or linguistically specified parameters of the object.
In order to present these
rules in mathematical form and to ensure
the possibility of adjusting the
model parameters, it is suggested
to use methods
of fuzzy logic. It is
shown that the bases of
fuzzy knowledge are a convenient means of formalizing
the causal relationships of the "input-output" of the object
of modeling in the natural
language with the help of
fuzzy logical statements. These expressions combine input and output
indices given in the form
of linguistic variables with fuzzy terms. Thus,
fuzzy logic is a convenient tool that allows
economists and financiers who are not mathematical
modelers to establish logical connections between explanatory and dependent variables in financial and
economic systems, using natural language
expressions, and perform mathematically-based analysis.
and forecasting the development of these systems.
In order to fine-tune the
fuzzy knowledge bases constructed expertly, it is
proposed to adjust the parameters
of fuzzy models with the
use of methods
for optimizing neural networks. The principles of optimization of models based
on fuzzy logic using algorithms
"Error Back Propagation" and "Extended Delta-Bar-Delta" are
presented, as well as recommendations
for the application
of these algorithms depending on the type
of membership functions, based on which the
model is constructed.