Software & Systems

Articles of journal № 3 at 2019 year.

Order result by: Public date | Title | Authors

11. On the formalization of functional requirements in information system projects [№3 за 2019 год]
Authors: R.D. Gutgarts , E.I. Provilkov
Visitors: 7152
There are a lot of scientific and applied research devoted to discussion of problems in IT project man-agement. Theн mostly focus on the financial aspects and duration of a project. However, functional as-pects are often overlooked. This might be explained by the fact that such indicators as money and time can be calculated using appropriate methods and algorithms taking into account various risks. Mean-while, a reasonable numerical equivalent to determine project functionality still does not exist. When considering such specialized IT project as the design and development of an information sys-tem (or its separate module), the implemented functionality is a fundamental factor which affects all other project indicators. However, fundamental literature and periodical scientific publications pay in-sufficient attention to functionality as a semantic item for various reasons. Scientists and specialists are mostly interested in managing information systems requirements including functional ones. The for-malization of functional requirements is a subject for discussion in a scientific community and is con-sidered in various aspects. However, there are no standardized or unified solutions. The paper considers the issues related to functional requirements, features of their initial formula-tion, presentation for discussion with a customer and with formalization for software implementation in a project. The authors briefly analyze approaches to the formalization of requirements and propose an ap-proach to the formalization of functional requirements, which can be applied to certain types of tasks that are included in the software that represents an information system as a software product. This may be the first step to create prerequisites for the development of an algorhythmic component. Thus, for a more correct calculation of project complexity, and for more accurate planning of its financial and time costs.

12. Rating assessment of academic staff in a university based on an automated information system [№3 за 2019 год]
Authors: E.N. Natochaya, Zubkova, T.M.
Visitors: 4655
The paper describes an automated information system (AIS) for assessing the efficiency of professional activities of academic staff in educational organizations of higher education. The proposed system is necessary to form a report on performance indicators of academic staff ac-tivities, based on which a head can quickly make management decisions. When applying AIS, the head of department, faculty or educational organization has an opportunity of selecting a suitable candidate for vacant positions; he is also able to form a so-called top of academic staff whose activities coincide with higher education development trends as much as possible. The authors present a formalized representation of the assessment process of an academic staff rat-ing as IDEF0 model and a decomposition scheme. During a decomposition of a software purpose model the authors allocate the following functions: maintaining data on the academic staff, rating calculation, forming report forms. Academic staff performance indicators are united into groups: educational, or-ganizational and methodical, research, scientific and organizational, educational and social work, repu-tation and image activity. The results of total rating assessment might be considered at competitive selection for a vacant po-sition and a subsequent conclusion of an effective contract, at funding acquisition of the equipment for scientific research, conference trips, training, etc., when determining the size of commercial incentive to salary. The used mathematical model of rating calculation is a method of qualimetric assessment of academic staff activity quality. The calculation includes basic, current and private ratings, as well as production and creative activity rating. The paper presents the developed algorithm to solve the problem, a data structure; shows AIS work and the obtained results.

13. Predicting an object state based on applying the Kalman filter and deep neural networks [№3 за 2019 год]
Authors: A.Yu. Puchkov , Dli M.I., E.I. Lobaneva , M.A. Vasilkova
Visitors: 6697
The paper presents an algorithm for predicting an object state based on data from different sources (for example, video cameras) coming in the form of images aimed at critical technological zones. The pro-posed algorithm is based on the consistent use of a deep artificial neural network and the Kalman filter. A neural network is designed to reduce the input data dimension (images) performing the function of an encoder, which gives of an observation vector of the object state on the output. Based on these ob-servations, the object state is evaluated by a recurrent filter. Using the filter directly for images would lead to a large dimension of the problem; it would be impossible to perform it practically due to com-putational difficulties. The program that implements the proposed algorithm was developed in Python 3.6 using the Spyder integrated environment from the Anaconda assembly for the Linux operating environment. The choice of a programming language is due to the availability of powerful libraries for machine learning Tensor-Flow from Google, as well as the convenient Keras framework for creating and working with deep neu-ral networks. The paper describes the results of a model experiment on using the proposed algorithm for predict-ing an object state, which consisted in attributing the obtained observations to a particular class. The experiment also involved generating sets of images belonging to different classes, differing in their tex-ture. A line-by-line horizontal pixel shift simulated the noise in the images. The comparative analysis of the predicted results with and without using the Kalman filter has shown that filtering reduces the number of false classifications. The developed algorithm might be used in decision support systems and automated process control systems.

14. Design of the QVT Operational Mappings interpreter for UML Refactoring in terms of the model driven architecture approach [№3 за 2019 год]
Authors: O.A. Deryugina , E.V. Kryuchkova
Visitors: 3953
The paper discusses the MDA (Model Driven Architecture) approach, which has been introduced by the OMG consortium and is aimed at the automation of the software development process. MDA pro-poses the following steps of the software development: design of the Platform Independent Model (PIM), design of the Platform Specific Model (PSM), development of the Code Model. The paper provides an overview of the MDA standards: XMI (XML Metadata Interchange), which unifiers model and metamodel interchange between software products; QVT (Que-ry/View/Transformation), which describes model query languages. The paper is aimed at the design of the QVT Operational Mappings language Interpreter for the UML Refactoring tool. The UML Refactoring tool provides the UML class diagram analysis and trans-formation. Typically, UML class diagrams are used to describe the software object-oriented architec-ture. UML Refactoring tool provides object-oriented metrics calculation (Avg, DIT, Avg. NOC, Avg. CBO, etc.) and searching for the transformations (Interface Insertion, Façade, Strategy) minimizing the refactoring fitness function value, which has been chosen by a user. Based on the information about the QVTo language, the Interpreter class has been designed for the UML Refactoring tool. This class translates QVT commands to the sequence of the transformations of the UML class diagram including add class transformation, add attribute to class, add method to class, add interface, add attribute to interface, add method to interface, add package, add class to package, add interface to package, add package to package. For each transformation, there is a newly designed class to extend Refactoring.java class. This class is an input for the Transformator.java class, which calls method execute() of the Refactoring.java class.

15. The prototype of an intelligent e-book based on technology of knowledge direct imposition [№3 за 2019 год]
Authors: G.B. Bronfeld , D.I. Kirov , V.V. Kondratyev
Visitors: 5622
The paper considers the process of introduction of intelligent e-books (IEB). It briefly discusses the ba-sics of creating the IEB in the form of an elinga. Elinga is based on the technology of direct imposition of knowledge (TDIK). TDIK includes applying a new model of knowledge representation – molinga, which in fact repre-sents text sentences as short semantic networks. The developed knowledge bases include a large set of molingas. The use of TDIK makes the knowledge base to contain only sentences with different seman-tic meanings. Molinga corresponds to the structure of production models, however it has a core con-taining a simple sentence with a code description, indicating a confidence factor and postconditions. These postconditions might contain graphical images, data files, or calculation models. The technology is developed in the framework of expert systems design, however each component is implemented differently. As a result, a software package – elinga – has unique capabilities in com-parison with conventional expert systems. Molingas allow applying TDIK introduced by J. Gray to nu-meric data as well. A logical inference is based on using the modified modus ponens rule. The process of finding a solu-tion is based on the dialog-associative search in the human-computer discourse using intermediate re-sults obtained during the logical inference. The paper describes basic functions of the elinga prototype and its operation modes. The elinga ac-tually implements V. Bush’s dream. Based on a fundamentally new technology, this approach allows users to solve various problems that were unsolvable or difficult to solve earlier more effectively using different basis of knowledge integration.

16. Development of operation algorithms of a mathematical model of an airship anti-stealth radar [№3 за 2019 год]
Author: S.V. Susha
Visitors: 8317
The paper describes a complex mathematical model of an airship anti-stealth radar system. The pur-pose of the study was to justify the technical appearance, application features, to assess the effective-ness of the operation and combat (information) capabilities of the complex. The development result in-cludes a number of simulation models (a target environment model, an Earth model, an on-board sys-tems model including a radar station model, an on-board control system model and a navigation system functioning model, a ground control center model including a model for displaying information about detected and tracked targets, a model board control), as well as functionally complete blocks (systems for processing and analyzing results). When modeling, all simulation models in the complex mathematical model are constructed accord-ing to a single principle. The functioning dynamics of the simulated complex is simulated by succes-sive changing of their states at some time intervals. The paper provides a block diagram of a general algorithm of a complex mathematical model in a simulation mode. The modeling process assumes stepwise changing of the model time by a step size. There are algorithms for the main units and their relationship as part of a general algorithm for the op-eration of a complex mathematical model of an airship radar system in a simulation mode. The operation algorithms of the target environment model include both the aerospace target and ra-dio-electronic environments. The spatial position and orientation of targets with respect to a stationary point of the airship radar system and the radiation of all on-board electronic target means is determined by targets parameters, as well as the direction of their arrival and radiation intensity. The operation algorithms of the navigation system model include source data of a carrier position – its location error vectors. The data values of these vectors are determined by the navigation system characteristics. A radar station model is based on calculating the detection parameter using the radar equation and calculating a signal propagation process. This model includes algorithms for primary and secondary processing of radar information. The implementation of the presented algorithms in a complex mathematical model allows reasona-ble describing of operation processes of an airship radar system when detecting, tracking, and recog-nizing subtle air targets. It will provide an assessment of the effectiveness of the options for building the complex and its information capabilities.

17. Development of a database and a converter for retrieval and analysis of specialized data from a medical device [№3 за 2019 год]
Authors: Eremeev, A.P. , S.A. Ivliev
Visitors: 7135
When developing expert systems, there may be difficulties associated with storage or data exchange formats. There may be situations when the data is stored in a proprietary format, or exchange files for such systems have proprietary format. This makes automated data analyze difficult, since they have to be manually entered into an expert system. However, there are methods that allow converting data into an easy-to-use format. The paper considers the analysis of database binary files of a medical apparatus for studying com-plex vision impairment in order to extract biophysical studies data for further analysis. Since the standard software does not allow information exchange with external systems in open formats, it is necessary to develop additional methods and software to determine data physical structure for subse-quent conversion to an open format. The initial data for the analysis is information about what data are stored in a medical device data-base, as well as general principles of physical data representation in computer systems. Converter de-veloping follows determining the structure of data files. Converter output files can be used in further neural network training. This approach allows quick creating of a database of samples (precedents) eliminating the need for manual data transfer. The proposed approach can further serve as a basis for data analysis in other similar situations.

18. Graphical shell development for parallel calculations based on the OpenFOAM platform [№3 за 2019 год]
Author: Chitalov, D.I.
Visitors: 5196
Graphical interfaces are an essential element of human-computer interaction. They form input data for programs and visualize results. Commercial software mainly comes with built-in graphical interaction tools. At the same time, some open software solutions (the OpenFOAM platform in particular) do not have built-in interaction tools. This problem remains relevant since the existing graphical interfaces have disadvantages. This paper presents a step-by-step process of developing a graphical shell for implementing user in-teraction with the OpenFOAM platform, i.e for setting up numerical experiments applied to problems of continuum mechanics in the parallel execution mode. To achieve this goal, there is a defined list of tasks and necessary tools: Python 3.5 programming language, PyQt5 interface element description framework, PyCharm integrated development environment. The paper gives diagrams showing the interconnection of program modules and a program operation mechanism. The results of the application development and testing are formed by the example of a pro-ject of one of the standard tasks of continuum mechanics, which are a part of the OpenFOAM platform distribution kit. The authors provide scientific novelty of the research, in particular: serialization of numerical experiment parameters using the Python Pickle module and database tables in the SQLite format; the ability to create different versions of files with continuum mechanics task parameters; the ability to run console commands using bash scripts. They also identify the practical value of the work and future prospects.

19. Developing a self-learning method for a spiking neural network to protect against DDoS attacks [№3 за 2019 год]
Authors: E.V. Palchevsky, O.I. Khristodulo
Visitors: 8377
The paper is devoted to the development of a specialized training method for a spiking neural network, which allows speeding up the detection and elimination of attacks by external unauthorized traffic. The paper considers the problem of protecting information availability and teaching neural net-works. It also justifies the need for mathematical analysis to develop new methods of self-learning of neural networks. The paper introduces the developed self-learning spiking neural network that is nec-essary to protect against DDoS attacks. A new self-learning method for a spiking neural network is based on the uniform distribution of neurons across all cores of each processor in a cluster. This allows a neural network to learn from scratch in a short time (530 minutes). As a result, it quickly and effec-tively eliminates DDoS attacks. The authors tested the developed spiking neural network in two modes: combat and normal. The tests gave load values for physical resources of each physical server in a cluster. Long-term testing of a spiking neural network shows fairly low load on a central processor, RAM and solid-state drive during DDoS attacks. Naturally, optimal load increases the availability of each physical server, and makes it possible to simultaneously run resource-intensive computational processes without any disruption of the working environment. Testing was conducted on computing cluster servers in one of Moscow data centers. The spiking neural network has shown stable operation and effective protection against DDoS attacks.

20. Implementation of a fuzzy model of interaction between objects in complex technical systems based on graphs [№3 за 2019 год]
Author: Muntyan, Е.R.
Visitors: 5801
The paper describes the process of developing a fuzzy model based on graphs, which allows investigat-ing the joint action of objects in complex technical systems using the example of a part of an extended perimeter security system. The objects of such system are stationary and mobile objects of a protected perimeter, uninhabited intelligent interacting mobile robotic platforms, a decision maker who controls their movement (here it is a computer), and potential intruders entering a protected perimeter territory. To represent objects of complex technical systems, the model uses the concept of “actor” according to the Actor-network theory of B. Latour. The results of simulating an extended perimeter protection system are analyzed on the example of three models that take into account various types of communi-cation in a graph: model 1 with the same type of connections, model 2 with the same and different types of connections, model 3 with the same type, different types of edges and multiple connections in a vector form. To simulate and study joint actions of objects in complex technical systems, the author proposes a fuzzy model of interaction of actors, which makes it possible to take into account the complex of vari-ous types of communication between graph nodes according to the specifics of the subject domain. These models are implemented in the software module developed by the author. Experimental research have shown the advantage of the model 3 proposed in the article (a multiple link graph).

◄ ← Preview | 1 | 2 | 3 | Next → ►