The paper discusses the problem of ensuring the viability of one of the classes of software systems - systems with knowledge bases. The software tools designed for the development of systems of this class are considered, the analysis of mechanisms of achieving this important property of systems with knowledge bases is carried out. The mechanisms of increasing their viability based on the each component development according to its declarative model created on a model description language are proposed. A three-level extensible architecture of tools that implements all the proposed mechanisms is proposed.

In this study, a prototype of an autonomous system was developed and investigated to provide cyber security and quality of service for multi-cloud platforms. Based on the developed system is a mathematical model of traffic analysis. The mathematical model is based on the neural network. A hybrid neural network based on a multi-layer perceptron and a self-organizing Kohonen network was designed. This approach allowed to more accurately classify and detect malicious traffic. The conducted experimental researches have shown that using the proposed approach allows to increase the effectiveness of detection of such attacks as denial of service. At the same time, during the attack, the required quality of service is maintained in the multi-cloud platform network.

In the article, research results of applicability of visual analytics tools for solving problems of practical data analysis are represented. Usage of a complex characteristic of visual representation function was offered and substantiated; it allows to increase effectiveness of visualization while using it in expert systems.

The article discusses the development of a conceptual model and knowledge bases to support integrated research in infrastructure logistics. Concepts of the conceptual model are used for creating fact templates of the rule knowledge bases. The research process is considered as a consistent interpretation of knowledge bases, while the sequence can be set arbitrarily by the user. The description of the research methodology; an overview of the transport and logistics ontologies used to create the conceptual model and the classification of knowledge bases by types of solved problems are presented in the article.

The most popular approaches to the problem of searching objects on images are: a linguistic approach, within which syntactic recognition of objects of strictly predetermined structure is carried out; artificial neural networks; descriptive image algebras; approaches that describe objects using predicate logic; CBIR (Content-Based Image Retrieval) technologies, based on descriptive logics (DL). The article presents an ontological approach based on descriptive logic with an extension to the spatial domain of data represented in the form of attributive graphs. The process of image analysis is controlled by a strategy containing: a preliminary morphological analysis stage; stage of the formation of a hypothesis about the category of the found object; stage of confirmation of the hypothesis by a logical conclusion about the class of the object. In the course of the analysis, a decision tree is constructed on the categories of objects. After selecting a hypothesis, a derivation tree is formed about the correspondence of the object to any DL-definition from this category. With an unsatisfactory degree of similarity and differences, a transition to another branch of the decision tree about categories occurs. Examples of the work of the image analysis system are shown.

The paper considers the application of machine learning methods in the analysis of database workload monitoring data. An approach is proposed for clustering queries performed in the database based on the input coefficient, which reflects the relation of the work done to the returned result. An example of analysis of load history using heat maps is given and the ratio of complexity parameters of actual queries to the introduced coefficient is considered.

Monitoring of ixodidae helps to detect changes in the ecological situation in the region under investigation and predict possible waves of transport of pathogenic viruses and microorganisms. In such statistics, the Ministry of Health, Sanitary-Epidemiological Station, and other public and private services are in need at the moment. The situation is complicated by the need for highly qualified specialists and the difficulty in determining by directories. To solve existing problems, software tools are needed to simplify the identification of ixodidae ticks.
The purpose of this research is the development of an expert system for the identification of ixodid ticks from photographs using neural networks. As a result, neural networks such as AlexNet and VGG were used. Acceptable results were obtained for use by specialists in real objectives. To provide client access to the server part, a public RESTful API is created

Three control algorithms are considered for nonlinear objects, the construction of which is due to different types of uncertainties in the description. All algorithms are based on unified principles of synergetic control theory and control methods on manifolds. It is assumed that the initial object is given in the form of a system of ordinary nonlinear differential / difference equations with the presence of unknown components (including unknown deterministic / stochastic perturbations) on the right side of the equations responsible for the dynamics of a poorly formalized object. Examples of applied problems, solved on the basis of the presented algorithms, are given.

The simulation of electrical circuits by Monte-Carlo analysis requires the powerful computing systems. Based on the Russian CAD, there was created the simulating system, which allows effectively parallelizing the simulation process by the Monte-Carlo analysis. In present work we show the results of analysis and we take appraisal of capabilities the simulating electrical circuits on the computing system that is consist six workstations with number processors core - 168

In the work, models with different characteristics, developed and implemented by the team of the authors of the chair of informatics and mathematical modeling of the Irkutsk State Agrarian University for many years, were considered in order to plan the receipt of food products in areas with developed agriculture and rich food forest resources. Three groups of models are distinguished. The first of them allow optimizing the production of agricultural products. This includes the tasks of obtaining optimal plans for sowing, the production of crop, livestock products and their combination, multi-stage tasks for locating crops, models for optimizing the receipt of agricultural products under risk conditions, and ecological and mathematical models. The second group of models allows optimizing the harvesting of food wild-growing products, commercial wild game meat and their combination. The third group of models is a synthesis of the first two models. It describes the production of agricultural products in conjunction with the procurement of food forest resources and commercial wild game meat. The article describes the directions of improvement of the described models: detailed production processes in the form of additional restrictions, risk taking into account the natural, climatic, and economic features of the territory. Based on models for optimizing the harvesting of food-borne wild products and commercial wild game meat, it is proposed to build and implement models of mathematical programming for the preparation of a combination of these two types of resources.

The article deals with the analysis of algorithmic and technological issues of geometric modeling in solving multi-dimensional initial-boundary problems. The paper describes the basic geometric objects of the computational domain and operations on them. A principle of organization of information interfaces using object-oriented programming tools and text formats, language data management XML are proposed

We offer the technology of developing an intelligent multi-agent solver for non-procedural statements of computational problems on a distributed computational model of the subject domain. A microservices-oriented approach is used to organize computations based on the semantic interaction of solver-applied agents. A finite-automaton model of the dynamics of the functioning of the solver's agents is given. The offed technology is demonstrated using the example of constructing a distributed solver for a study the behavior of trajectories of autonomous binary dynamical systems.

Described are the tools developed for intelligent scientific Internet resources that are designed to visualize and process data stored in external sources. The architecture of the access system to external sources, the scheme of its functioning and the approaches used for its creation, as well as the methodology and examples of the proposed means using are presented.

The simple predicate program of reversing a list as datatype object is presented. Deductive verification of this program is simple. The imperative program of reversing a linked list is obtained by the set of program transformations for the predicate program.

While studying results of experimental measurements related to most of areas of scientific researches, one of the problems with high relevance was creation of analysis tools that are applicable for studying large amount of initial data. Existing approaches to the analysis of experimental data have a number of disadvantages, which include significant resource capacity and high training requirements for specialists involved in the analysis. In this article some results of solving the problem of creating visual analysis tools for experimental data, aimed at increasing the effectiveness of empirical data research, have been showed.

The paper describes a developed software package that allows to minimize labor costs for the production of agricultural products in an enterprise belonging to one of the groups: micro, small, medium and large. The basis of the software package is a database created on the basis of information from annual reports of agricultural enterprises of various levels over a long period of time. The data model consists of 11 entities. Mathematical software of the software complex allows to group agricultural enterprises by the number of employees whose number can vary over a period of many years, and the methods of probability theory and mathematical statistics used to evaluate statistical properties in the variability of each group of farms and specific enterprises. The identification of regularities in the variability of production and economic characteristics determines the type of the extreme task for modeling labor inputs for the production of crop and livestock products. To optimize labor costs, linear and parametric programming models are proposed, optimization problems with interval and random estimates, as well as variants with deterministic and uncertain characteristics. They allow us to determine the direction of increasing the efficiency of using labor for the production of the main types of products. The proposed models using the database are implemented for the enterprises of the Irkutsk region.

As part of this study, the structure of the diagnostic system for industrial frequency converters (FC) has been developed, taking into account the addition of the existing composition of the system to the new components, the availability of which will make it possible to detect faults in the converter in real time. The author of this article proposes an approach to realizing the program reading of the values FC of the IF parameters in real time. A detailed concept has been considered and the results of algorithmization and software implementation of the reading process has been presented.