Kuban State Technological University
Author list of organization
List of articles written by the authors of the organization
-
USING NEURAL NETWORKS FOR THE PROJECT TEAM SELECTION
05.13.10 Management in social and economic systems
DescriptionThe study was carried out with the financial support of the RHNF as part of the research project of the RFBR 17-02-00475-OGN "Application of metaheuristic algorithms for solving direct and inverse problems of optimizing the management of spatially distributed complexes"). The article considers the issue of automated selection of specialists for the project. It is proposed to use artificial neural networks as a decisive core of the system. We have considered several solutions to the problem with a basic version based on a cascade of two neural networks
-
FRAME EXPERT SYSTEMS USING NEURAL NETWORKS
05.13.18 Mathematical modeling, numerical methods and software complexes
DescriptionThe study was carried out with the financial support of the RHNF as part of the research project of the RHF RFBR 17-02-00475а "Application of metaheuristic algorithms for solving direct and inverse problems of optimizing the management of spatially distributed complexes"). The article discusses the use of artificial neural networks in frame expert systems, which in some cases make it possible to bypass a number of limitations of the standard version of frame expert systems. Several options for using neural networks in frame expert systems were considered and a number of tasks to be solved for each
-
Description
The article has examined the application and found the processes that need to be automated. The advantages of this method of automation of the technical support service are considered. Using the unified modeling language (UML), an information system model is constructed based on the information received. Using the developed model, the method of accounting for channel services and equipment will be automated, as well as the process of interaction with customers
-
METHODOLOGY OF OPTIMAL PLACEMENT OF TRADE NETWORK OBJECTS
05.13.18 Mathematical modeling, numerical methods and software complexes
DescriptionThe article outlines the idea of a methodology for locating distribution centers of a spatially distributed distribution network without restrictions on the territory. In the past fifteen years in Russia, the profitability of offices and retail space has been significantly higher than logistic complexes. At present, it is possible to talk about a change in the investment attractiveness of the segment of distribution centers and storage facilities. The method consists in solving three problems: determining the number of distribution centers that need to be placed using the method of comparing options; determination of the best locations for placement of distribution centers using the ant colony algorithm; identification of the best location from the previously determined ant colony algorithm using the penalty function method. This method of optimal placement of objects of a spatially distributed complex can be applied not only to the distribution network, but also to any transport company with distribution centers, for example, a logistics company, delivery services, etc.
-
Description
This article is dedicated to the study of the parameters of the artificial immune system for solving the polymorphic viruses’ detection problem. The goal is to define a vector of the immune system parameters that would ensure the minimum number of errors of the first kind, the minimum number of errors of the second kind and the maximum percentage of polymorphic viruses’ detection. That is, the most accurate classification of them as a malicious code, in relation to any theoretically possible vector of parameters of the artificial immune system. A distinctive feature of the studied artificial immune system is the use of a class of genetic algorithms that provide more efficient training of detectors. The configurable parameters of the system are: the algorithm for determining the proximity of the detector and the pathogen, which can be realized by determining the Levenshtein distance or by the method of adjacent bits; as well as the method of implementing the crossing-over operator, the method of implementing the mutation operator, the method of implementing the selection operator, the algorithm for determining the proximity of the detector lines. In addition, the article considers the expediency of using a distributed network of several nodes, each of which will have an immune system that will exchange data with other nodes of the network. As a result of the research, a set of optimal parameters was obtained in which the system achieves the maximum accuracy of recognition of polymorphic viruses
-
Description
The article is devoted to the analysis of the use of large data processing technology in information systems of territorially distributed complexes
-
Description
The article formulates and solves the task of discrete control in the thermophilic stage of the composting process. It is shown that considering the relay control entity to maintain specified process conditions requires the organization of the sliding mode. We have solved the problem of minimizing the temperature deviation of the substrate from the set values and the deviation of the oxygen concentration in the gas phase of the bioreactor from the specified values. The article shows the algorithm to compute the discrete control of the composting process in the thermophilic stage. This work was prepared in the framework of the scientific project 16-48-230441 a(R) "Mathematical modeling of the processes occurring in the automated installation for year-round production of organic fertilizers in the conditions of the Krasnodar region", financed by RFBR and the administration of the Krasnodar region
-
METHODIC OF KNOWLEDGE ASSESSMENT WITHIN THE FRAMEWORK OF THE AUTOMATED TESTING SYSTEM
DescriptionIn this article, we consider approaches to the transfer of knowledge to students and an objective semiautomatic assessment of knowledge. The characteristic features of the application and the possibility of using cognitive training methodologies and complex systems for testing skills and the theoretical base of trainees are analyzed. The problems of development of this direction and possible ways of their solution are described. The basic concepts are introduced and the existing methods of calculating the average score for checking the student's knowledge are considered, and a new approach to solving this problem is proposed. Based on the conducted researches it is offered to use the complex system of testing of end users, which includes testing, monitoring, collecting, analyzing and displaying the results of students/groups/ course. The main requirements for the creation of such a complex and the rules to be followed are formulated for a more objective assessment of knowledge. A model of an integrated modular system for objective semi-automatic testing of knowledge through testing is described
-
THE TASK OF FORECASTING OF PUBLICATION ACTIVITY
DescriptionIn the article questions of forecasting of publication activity and a problem of planning of actions on management of publicity activity of scientific collective are considered
-
THE STUDY OF LEARNING ALGORITHMS OF NEURO-FUZZY SYSTEMS CONTROL OF BIOTECHNOLOGICAL PROCESS
DescriptionThe subject of study of this work was learning algorithm of neuro-fuzzy systems with different membership functions. In the prior works there are no published studies of such studies, making it difficult synthesis of neuro-fuzzy control system with new objects in the application of biotechnology, including technological agribusiness entities. A comparative analysis of learning algorithms of neuro-fuzzy system with different membership functions using the method of error back propagation and а hybrid method. For this we used a training sample that contains data of temperature and concentration of dissolved gas in the culture liquid: oxygen (pO2), carbon dioxide (pCO2) of a biotechnological process. It is shown that the hybrid method carries out training of a neural network for the number of periods is 23 times smaller than the algorithm back-propagation errors. The studies found that the two-sided Gaussian membership function provides the smallest learning error of the network δ equal of 3,28•10–3, compared to the other, giving the largest error of training the neural network δ=0,138. Therefore, the task of running the fermentation process effective is the use a hybrid method of education and two-sided Gaussian membership functions. According to the research, we can conclude that for the adaptation of neuro-fuzzy network ANFIS and fuzzy inference system Sugeno zero order to solve biotechnological process control tasks microbiological production efficiency is to use a hybrid method of education and bilateral Gaussian membership functions