Software & Systems

Articles of journal № 2 at 2018 year.

Order result by: Public date | Title | Authors |

21. A method of automated development of the semantic dialogue system database model [№2 за 2018 год]
Author: Posevkin R.V.
Visitors: 7270
The paper considers the problem of intellectual database content analysis for creating a semantic database model. Voice assistants are created to simplify interaction with mobile devices such as smartphones and tablets. A text dialogue system is an analogue of this approach. As a result, user can interact with software using natural language. User interface is a set of software solutions that helps to search, review, obtain and process information from a database that is external storage. Natural language interface is a sort of user interface that accepts and processes natural language queries. This interface can also use natural language in output to show found information to a user. A semantic database model is an important part of the dialogue system. This model includes interconnections and internal structure of a database. Manual creating of a semantic database model significantly increases time and labour costs, as well as development cost of a software system. The main purpose of the article is automation of development of a dialogue system semantic database model. The proposed method uses a set of approaches to automated creating of an existing database semantic model. An object domain thesaurus helps to define semantics and solve the problem of polysemy in text processing. Patterns helps to extract interconnections in a database. The analysis of database field content allows determining data semantic and nature. Locale indication allows decreasing the time for a database content analysis.

22. Adaptive algorithm for optimal route search in time-dependent network [№2 за 2018 год]
Author: Soldatenko A.A.
Visitors: 8057
The Time-Dependent Shortest-Path problem (TDSP) is an extension of the shortest path problem in a graph. TDSP problem arises when designing and operating telecommunications and transport networks. Such networks require considering time and possibility of appearing predictable situations for example traffic jams or traffic reduction. In this case, network is represented with an oriented graph G = (V, E) where for each arc (x, y)  E, two functions are defined. First function is time required for moving along the arc (x, y). Second function is arrival time in vertex y if the movement started from vertex x in time t. Such graph is called time-dependent network. The minimum time for moving from vertex x to vertex y is an optimal route between these vertices. It is known that TDSP for a general time-dependent network without any restrictions on network topology or arrival function is NP-hard. When arrival function satisfies FIFO (First-In First-Out) condition, TDSP problem is polynomially solvable. This paper studies TDSP problem for a polynomial case when arrival functions are monotonous. It is proposed to solve TDSP problem using a two-phased algorithm ALT (A* with Landmarks & Triangle). ALT algorithm is one of the modern least cost routing algorithms originally developed for solving the problem of the shortest path in a graph. In the first phase, ALT algorithm places landmarks in network vertices and calculates potential functions. In the second phase it finds the exact value of an optimal route in a graph with A* algorithm. The paper proposes modification of ALT algorithm which is capable of correct and efficient solving the TDSP problem for a sequence of queries for searching for optimal routes in a time-dependent network. The modification consists in using adaptive heuristic for landmark placement and special formulas for calculating potential functions. This heuristic uses experience of processing all completed queries; adapts current set of landmarks for next queries. There is a description of the modified ALT algorithm and estimation of its working time. The paper also gives a description of the software that implements the proposed algorithm. The results of computational experiments confirm the effectiveness of the modified ALT algorithm.

23. Application of a tree search algorithm and a annealing simulated method in optimization of heat network configuration and structure [№2 за 2018 год]
Authors: V.A. Stennikov, A.A. Chemezov
Visitors: 7497
The paper considers a mathematically complex problem of heat network configuration and structure optimization. It presents a problem statement and methodological approaches and algorithms to solve it. Traditionally in design practice, comparing 2 or 3 pre-planned alternate schemes solve this problem. Significant complication of the scheme and system scale led to the fact that the obtained solutions are far from optimal ones and often lead to insufficient load of network sections and even their inoperability. The development of mathematical programming and optimization methods gives new possibilities for solving practical problems. At the same time, it requires understanding of energy and mathematical features, as well as corresponding connections between them in order to apply mathematical methods correctly. There is a large arsenal of methods. Choosing effective ones requires special research related to capabilities and limited application. The complexity of scheme-structural optimization problem is in the fact that the objective function in the form of reduced costs is convex in terms of flow rate and concave on the heads. Fixing the heads by transformations, it can be reduced to a concave function. The problem is multiextremal as concave programming problems. An optimal solution will look like a tree. Each possible variant of the tree corresponding to the vertex of a constraint polyhedron will correspond to the local minimum of the objective function. This makes this task difficult to formalize, which does not allow finding an analytical solution. The paper considers a historically applied method of tree search, suggests its modifications, as well as other methods that were not previously used to optimize heat networks. The authors compare the proposed methods and algorithms, and assesse their performance in calculation of networks of different sizes and complexity. The algorithm realizations are universal and might be applied to different types of energy systems with a network structure.

24. On information support of decision support [№2 за 2018 год]
Author: Tikhanychev, O.V.
Visitors: 6446
The basis for formalization of information used in the automated management process is the system of classification and coding. It ensures the work of users with software toolsб as well as the interaction of software tools and automated control systems. According to the existing requirements documents, the interaction between automated control systems is organized through specialized interaction protocols: technical, organizational, information, software. However, with the growing scale of managed systems there appears the problem of keeping these protocols up to date, which increases with growing number and types of interacting agents in a management network. Nowadays, there is a wide range of approaches to solving this problem in the world practice. It includes data exchange using XML files, interaction based on HLA technologies, etc. Nevertheless, in general, none of them provides information interaction in distributed multi-agent systems of automated control. After analyzing the capabilities of modern information technologies, the author proposes to solve the problem through abandoning the “hard” principles of information encoding and moving to its classification based on fuzzy set mechanisms, multidimensional OLAP-matrices or RDF-graphs. Features of the organization of automated decision support systems and numerous components of operational information processing in them make interaction based on “fuzzy” information support the most acceptable to solve the problem. There is a conclusion that the proposed approach does not contradict the principles of creating a single information space. It complements them due to the transition from organizational methods of ensuring systems interaction to simpler technological ones. Naturally, this takes into account the limitations for systems that are crucially critical to the error probability in information processing.

25. A method for determining the capabilities of parallel execution of data mining algorithm functions [№2 за 2018 год]
Author: I.I. Kholod
Visitors: 5843
The article describes the method for determining the possibility of parallel execution of data mining algorithm functions. The input parameters of the method are: a data mining algorithm represented as a composition of thread-safe functions, and a mining model represented as an array of unified element trees describing the patterns mined by the algorithm. When determining the possibilities of parallelization, the method takes into account information dependencies between functions, which are determined by a set of used and modifiable elements of a mining model. The method analyzes data connections for each pair of algorithm functions when testing the possibility of task parallelization, as well as for calling functions in loops for different iterations, while verifying the possibility of data parallelization. The analysis includes checking necessary and sufficient conditions for parallel execution for systems with shared and distributed memory. They extend the Bernstein conditions used in compiler theory and are sufficient, but not necessary. The method determines the possibilities of parallel execution of functions for shared memory systems and distributed memory systems. In this case, parallel performance of functions in shared memory systems is more efficient due to the lack of calls for the cloning function and combining of mining models. The result of the proposed method is the parallel form of the source data analysis algorithm. It contains insertions of special functions of higher order that provide parallel execution of algorithm functions that satisfy necessary and sufficient conditions. To illustrate the proposed method, the paper shows parallelizing 1R classification algorithm. It defines algorithm functions that can be paralleled, both by data and by tasks. The result is a parallel form of the 1R algorithm with parallelization function inserts for shared memory systems and distributed memory systems.

26. Algorithmic support of the information system for managing innovation projects in an industry [№2 за 2018 год]
Author: M.V. Chernovalova
Visitors: 7586
Current approaches to project management are focused on the consistent implementation of all phases. They do not take into account the influence of the results obtained earlier and uncertainty factors on innovation project performance indicators and the degree of its feasibility. The paper considers two algorithms that are aimed at solving these problems. The first algorithm takes into account the influence of uncertainty factors on innovation project work duration and implementation costs. A key feature of this algorithm is a system of fuzzy production rules to determine possible deviations of the results of individual project works from the set indicator values. When managing an innovative project, this allows taking into account external and internal uncertainties while minimizing innovation costs. The second algorithm is focused on minimizing the duration of a project or resource costs. The basis for this algorithm is the List Scheduling algorithm. It differs by using fuzzy production rules to provide efficient allocation of available resources depending on the set values of performance indicators, as well as forecasting performance time of a particular work with the allocated resource level. The article also presents the architecture of the proposed information system for project management. It implements the described algorithms and shows the main modules and data flows between them. It is focused on the complex automation of the process of managing innovative projects.

27. Comparison of the adaptive and rigid algorithms of traffic control based on a simulation model in AnyLogic [№2 за 2018 год]
Authors: Ya.I. Shamlichky, A.S. Okhota, S.N. Mironenko
Visitors: 6606
The growth of an automobile park and traffic volume led to increasing traffic intensity. This causes a transport problem in the conditions of cities with historically developed buildings. The problem is especially serious in key points of the street-road network. The problem is related to the increase in transport delays, queues and traffic jams, which in turn causes a reduction in the communication speed, unjustified fuel overexpenditure and increased wear of vehicle units and assemblies. There are two ways to solve the problem: engineering and construction works (widening the carriageway, applying complex road interchanges) and using automated traffic control system (ATCS). The first option is expensive and takes a long time from development to putting the object into use. The second option is less global in terms of changing a road network, but more productive. There are several types of ATCS. A specially trained operator sets up a hard algorithm manually, which is not always convenient and effective. The adaptive algorithm automatically adjusts to the flow, which is more convenient during the so-called “rush hour”. The change in traffic intensity observed during the day requires a corresponding change in a cycle duration and enable signals. The multi-program rigid control is not capable of taking into account short-term random fluctuations in the number of cars approaching an intersection. The use of ATCS based on the adaptive algorithm will solve the problem with increasing traffic in cities. The citywide ATCS should be based on principles that ensure maximum management efficiency and reliability of the system. The improvement and development of ATCS is essential to any major metropolis.

◄ ← Preview | 1 | 2 | 3