Software & Systems

Articles of journal № 2 at 2015 year.

Order result by: Public date | Title | Authors

11. Dynamic structures in relational databases [№2 за 2015 год]
Author: A.A. Poltavtsev
Visitors: 7519
There is a set of problems in the development of applications for work with databases. The principal cause of these problems is in non-optimized SQL queries and stored procedures. To achieve good productivity, it is necessary to make SQL queries correctly, create (or delete) additional indexes, denormalize a database (in certain cases), shift a part of logic on triggers and stored procedures. It is necessary to maintain agreed algorithms structuring methods, structuring methods of data used in these algorithms and structuring methods (storage schemes construction) for these data in databases based on a rel a-tional model. This paper is devoted to generation and a manipulation data structures in relational databases whose compo-nents in programming languages are connected by explicit indexes. The author considers a specific, rather simple data stru c-ture. It is a linear unidirectional coherent list. Lists as trees are among the most basic data models used in computer programs. In a sense, lists are simple forms of trees because one can think of a list as a binary tree where every left child is a leaf. How-ever, lists also present some aspects that are not special cases of what we have learned about trees. For instance, we shall talk about operations on lists such as pushing and popping that have no common analog for trees.

12. Formation of agent’s representations of the domain in a situation of choice [№2 за 2015 год]
Authors: Vinogradov G.P., Shmatov G.P., Borzov D.A.
Visitors: 8955
The work is relevant due to the fact that the IT effectiveness in managing a multi -agent system under uncer-tainty depends on how the agents with their help shape their own subjective understanding of the properties of choice situ a-tion. To create a “correct impression” the agent (subject) collects information for “removing” various uncertainties and form-ing a hypothetical model of the situation of choice. The objective of the work is development of modeling techniques aimed at identifying mental strategies of a specific agent by analysing its behavior, speech patterns, non-verbal reactions in situations of purposeful state, selection and operating ways. This involves interaction of a researcher and an agent. In the process of this interaction specific cognitive, linguistic and behavioral skills are revealed. They are used by the subject in the situations of choice to achieve desired results. They, in turn, reflect a value system, norms, potentials, beliefs and knowledge of a specific subj ect. These patterns are represented in shaping decisions and implementing them in the behavior. In this connection, it is of interest to develop mathematical models that take into account the agent's behavior based on the theory of fuzzy systems and the th eory of relations as well as to prove the possibility of application of fuzzy measures for including subjective evaluations in the selection model. The authors chose following research methods: general methodology and methods of system analysis, analytical model-ing and methods of theories of set, theory of active systems, decision making theory, theory of artificial intelligence, theo ry of fuzzy systems and fuzzy logic, game theory. The paper investigated the structure and types of agent’s representations about the purposeful state situation. Based on subjective perceptions of reality the authors formalized a model of choice. It is shown that the description of concepts through agent’s language form a model. The agent assesses this model using the terms of the linguistic variable “conviction of the utility”. This assessment expresses the agent’s confidence in usefulness of the model-presentation for transition from one purposeful state to another. The paper shows how this assessment changes as new information is being received. To cal-culate the membership function it is proposed to use the Zadeh’s compositional rule. There are evaluates to describe the level of agent’s confidence and a degree of doubt in the usefulness of its ideas. It is shown that when the agent evaluates its repre-sentation using a set of parameters, the representation quality under uncertainty can be evaluated using fuzzy integral. This technique allowed the authors to propose the algorithm of a certain level of agent’s conviction in the us efulness of its ideas. There is a description of the algorithm application in medical diagnostics.

13. Decision-making in high-tech products production systems [№2 за 2015 год]
Authors: Burdo G.B., Vinogradov G.P., Isaev A.A.
Visitors: 8855
The features of multiproduct production (MP) of high-tech products are determined by simultaneous performing a large number of orders while solving its development (evolution) problems. Filling of orders shall be performed within a conc eptual scheme that defines the long-range and tactical MP goals. MP functions under uncertainty when the manager is unable to accurately determine the external environment state. As a consequence, he has a hard time choosing a course of action. To remove uncertainty in decision making it is appropriate to use subjective assessment of the situation and a solution vari ant obtained on the basis of agreement of all concerned parties. This problem is primarily related to the difference in the perception of choice when fill ing the orders by production staff. To solve this problem the authors suggest an information theory of hierarchical systems and the theory of active systems as a theoretical basis. However, the development of practically useful decision-making models in hierarchical sys-tems, where there are no antagonistic contradictions due to common goals and interests, is far from practical completion. According to the authors, the solution of this problem is in the specificity of modern high -tech production that operates under risk, uncertainty and dynamics of both external and internal environment. The paper presents the mathematical statement and description of the agreed decision-making problem. They suppose that the agents which control their subsystems know and “feel” their controlled objects as they see them. They allow “playing” different pro-duction situations. The paper shows a model and an algorithm for making coordinated decisions which become the basis for an ex-ample of using the method in the typical company.

14. The architecture of a smart home management system integrated into the сloud environment [№2 за 2015 год]
Author: Nikolaev P.L.
Visitors: 10730
The article describes a flexible architecture of a smart home management system integrated into the cloud en-vironment. This cloud environment provides users with the infrastructure for remote control actuators and for monitoring and processing of data from different sensors that form the system of smart home. The paper presents the benefits for end users from integration of their smart houses into the cloud. The remote control is possible with mobile and stationary devices either via Internet or home LAN. The presented architecture corresponds to the concept of the “Internet of things”. The article pre-sents the typical hardware architecture of a smart home management system consisting of three levels of automation. Then the author offers the hardware scheme of a smart home management system integrated into the cloud and consisting of five levels of automation. The paper describes in detail all components of a smart home management system (the cloud server, the home server, the controller, control devices and terminal equipment) and ways of their interaction with each other. In addi-tion to the hardware component of the smart home management system architecture, the article also considers its software. For the interaction of control devices, cloud and home servers it is is necessary to apply web-services technology. The carried out analysis has revealed that the preferable option is to use a RESTful web service: a request and data transmission are car-ried in JSON via HTTP-requests (GET and POST). This paper also provides the execution algorithm of remote management function for smart home system.

15. Forecasting stock price growth using feedforward neural networks [№2 за 2015 год]
Author: Lesik I.A.
Visitors: 16161
The article considers the problem of predicting stock prices using feedforward neural networks. The author used a single-layer Perceptron (without the hidden layer) and a multilayer perceptron (by Rumelhart) as a basis for the algo-rithms. He also implemented a backpropagation algorithm for training neural networks and optimum value selection of a learning rate parameter for algorithms. The paper specifies the conditions of algorithm convergence (in the sense of rms con-vergence to some constant). There is an estimate for the asymptotic complexity of algorithms. The paper also includes a test and analysis of forecast results using time series of share prices changes dynamics in of Russian power company OAO "Gaz-prom". Traditional methods of estimating prediction accuracy are based on the measurement of the forecast deviation from real data (for example, the calculation of the average absolute percentage assessment (MAPE) or average absolute error (MSE). They do not always make possible to assess forecast quality for stock value dynamics adequately. Stocks are characterized by high volatility since such data needs to determine when the growth and fall will be, and most importantly it is difficult to pre-dict leaps separating stationary periods of value changes. Therefore, this article uses assessment methodology of forecast quality which is an alternative to such methods as MAPE and MSE counting. The results of prediction using the calculation of prediction accuracy and the analysis of the forecast diagram allow mak-ing a conclusion about the possibility of obtaining a reliable forecast of stock growth using feedforward neural networks.

16. Software package for cloud virtual environments development [№2 за 2015 год]
Authors: Gribova, V.V., Fedorischev L.A.
Visitors: 8611
The article describes implementation of the IRVIS software package (Internet development of virtual interactive envi-ronments). IRVIS is designed for development and maintenance of virtual environments based on IACPaaS (Intellectual Applic a-tions, Control and Platform as a Service) cloud platform. The main principles are: automation of a development process; invol ving different experts and designers into the process; cloud services programs. Based on mentioned above principles the authors stated some requirements to IRVIS software package and to its implementing methods. The paper describes the main services of the IRVIS. They are: a structure editor, a graphic editor and an interpreter. The struc-ture editor is applied to develop a declarative model of a virtual environment. The article describes the method of implement ing a structure editor as an output service of IWE (Improved Web Editor) of cloud platform. The graphic editor is aimed to form visual el-ements of the virtual environment model and later add it to the declarative view of the model. The paper considers a method of im-plementing the graphical editor as a client-server cloud service of the IACPaaS platform and a method of applying the multiagent approach with messages system included in this service. The interpreter is designed for starting and operating of the develop ed virtu-al environments, each of them becomes an individual service of the cloud platform. Similar to graphic editor, the paper shows a method of implementing the interpreter as a client-server cloud service of the platform using a multiagent approach. At the end of the paper there are some examples of using the IRVIS software package for creating cloud virtual environments: a computer training simulator with classic ophthalmology methods, a virtual chemical laboratory and a virtual editor of a city district.

17. An experimental study of the performance impact of jemalloc memory profiler on execution time of multithreaded applications [№2 за 2015 год]
Authors: Ivanov E.Yu., Toropov A.V., Kosyakov M.S.
Visitors: 8016
This paper examines the performance impact of jemalloc memory profiler on execution time of multithreaded applications. Using popular synthetic tests the authors implemented a multithreaded application which allows varying different parameters related to memory allocation. A special criteria is suggested to com-pare and estimate profiling results. Tbricks algorithmic trading system is used to demonstrate that popular syn-thetic tests are not adequate enough to analyze memory profiling impact on operational characteristics of the ap-plication. Studies have shown that inactive profiling doesn’t affect the characteristics while with active profiling its values depend on sampling interval and can decrease manyfo ld. According to our analyses of jemalloc library source code related to memory profiling, the part which takes stacktraces is the main reason of extra overhead during profiling. Thus, we compare performance impact on operational characteristics of the app lication using libunwind and prof-libgcc stacktrace libraries: results of experiments made with libunwind stacktrace library are in average 20 % worse than with prof-libgcc. Based on experimental study we recommend using sampling inte r-val which allows of capturing information of at least 90 % of all allocated memory, while execution time of highly multithreaded applications decreases by less than 5 % in comparison with the disabled profiling. This al-lows analyzing memory allocations of applications while executing in production environment.

18. The limitation method for verifiable models [№2 за 2015 год]
Authors: Korablin Yu.P., Shipov A.A.
Visitors: 9988
Nowadays software systems become more and more complex and sophisticated, so we need to have some use-ful instruments to check if they fit target specifications, especially if we are talking about large and distributed software sys-tems. The article demonstrates a new method which is the separate stage of verification process of information systems. It is intended to make verification more effective. Since the method is a separate verification step for Model Checking method and can be automated easily, its implementation to an existing technology is not a difficult task. The method is based on the principle saying that the model of all systems contains redundant computational sequences. They are not necessary to be analyzed during the properties verification process and it does not affect the final result. The main goal of this method is to exclude these computational sequences from the original model before certain properties veri-fication. The article provides a detailed description of the proposed method and its efficiency estimation. The theoretical material of the article includes illustrative examples of the method, in particular for distributed software system. The authors believe that using of this method can significantly reduce the number of analyzed sequences in verification process for a given system computational model, and thereby improve the performance of the process.

19. Formal model development for a process of finding solutions by the modified Rete algorithm for fuzzy expert systems [№2 за 2015 год]
Authors: Zaw Min Htike, Mikhailov I.S.
Visitors: 7521
The paper considers the basic concepts of fuzzy production expert systems. This type of expert systems is based on a set of rules presented in terms of linguistic variables. The authors propose a developed Rete algorithm modification for fuzz y rule base. It allows creating rules and solutions in the limited natural language and it provides system acceleration due to a single compu-ting the same conditions in the various rules. The authors proposed a decision tree formal model of a modified Rete algorithm for a fuzzy production knowledge base. The model consists of a set of vertex-conditions, vertex-solutions, relations between vertices and relations describing the fuzzy expert system rules. Rete algorithm modification graph is formed in such a way that in each ca se not a correct rule condition value is calculated, but linguistic variable values in the rule.The proposed algorithm processes rules from the fuzzy rule base and converts them into the formal model of a modified Rete algorithm decision tree. The difference between Rete al-gorithm modification and the classical algorithm is application to fuzzy variables. Therefore, the building of fuzzy truth values of decision tree vertices is performed by fuzzy operators at each stage of the algorithm. It allows formulating the conditions a nd conse-quences in the rule base as well as the solutions in the limited natural language. The same conditions are combined during decision tree construction. It provides acceleration of decision tree processing comparing to the sequential viewing of the expert sys tem rules.

20. Expert system for monitoring of technological processes [№2 за 2015 год]
Authors: Burov A.N., Kalabin, A.L., Kozlov A.V., Pakshver E.A.
Visitors: 12898
The authors describe an expert system for technologists who work in polyacrylonitrile fibers production. It can help to process and analyze data. The expert system is a part of a larger software system for monitoring technological pro-cesses. The aim of the monitoring system is to gather, store and process data about the state of technological process parame-ters. The expert system analyses data stored by the monitoring system. It consists of two parts: the module for rule creation and the module for rule checking. The module for rule creation allows technologists to store their knowledge as a set of rules in a database. This feature helps to store technologists experience and pass it to other specialists. Every rule includes IF and THEN parts. The IF part contains various logical conditions. The THEN part has conclusion that should be made if logical conditions are true. The system allows creating complex rules based on other simple rules. The module for checking rules al-lows specialists to apply some rule to data set and check logical conditions against that data. The article describes system ar-chitecture and its place in the monitoring system. The work of the system is demonstrated on data from the real process of correlation assessment.

◄ ← Preview | 1 | 2 | 3 | Next → ►