Kuban State University
Author list of organization
List of articles written by the authors of the organization
-
08.00.13 Mathematical and instrumental methods of Economics
DescriptionTraditionally, control decisions are made by solving repeatedly the forecasting problem for different values of control factors and choosing a combination of them that ensures the transfer of the control object to the target state. However, real control objects are affected by hundreds or thousands of control factors, each of which can have dozens of values. A complete search of all possible combinations of values of control factors leads to the need to solve the problem of forecasting tens or hundreds of thousands or even millions of times to make a single decision, and this is completely unacceptable in practice. Therefore, we need a decision-making method that does not require significant computing resources. Thus, there is a contradiction between the actual and the desired, a contradiction between them, which is the problem to be solved in the work. In this work, we propose a developed algorithm for decision-making by solving the inverse forecasting problem once (automated SWOT analysis), using the results of cluster-constructive analysis of the target states of the control object and the values of factors and a single solution of the forecasting problem. This determines the relevance of the topic. The purpose of the work is to solve the problem. By decomposing the goal, we have formulated the following tasks, which are the stages of achieving the goal: cognitive-target structuring of the subject area; formalization of the subject area (development of classification and descriptive scales and gradations and formation of a training sample); synthesis, verification and increasing the reliability of the model of the control object; forecasting, decision-making and research of the control object by studying its model. The study uses the automated system-cognitive analysis and its software tools (the intelligent system called "Eidos") as a method for solving the set tasks. As a result of the work, we propose a developed decision-making algorithm, which is applicable in intelligent control systems. The main conclusion of the work is that the proposed approach has successfully solved the problem
-
05.13.18 Mathematical modeling, numerical methods and software complexes
DescriptionThe article presents ways to estimate the socio-economic condition of the subjects of the Russian Federation using fuzzy production systems, discriminant analysis and neural networks. This study is based on the thirteen indicators of the socio-economic condition of the regions of the Russian Federation identified by RIA Rating rating agency. The goal of the work is to develop a software package combining the estimation methods above and allowing to get three alternative estimation options for a subject of the Russian Federation, using these thirteen indicators. The developed program assigns each region an estimation from the following set: AAA– the highest socio-economic condition, AA - very high, A– high, BBB– above average, BB– average, B– below average, CCC– low, CC– very low, C– the lowest indicators of the socio-economic conditions in the region of the Russian Federation. The software package is developed in Matlab and has an intuitive interface. The results can be used as an alternative to current methods of estimation of the socio-economic indicators