2022

Постійне посилання на фонд

https://ela.kpi.ua/handle/123456789/50069

Переглянути

Перегляд 2022 за Ключові слова "algorithm"

Зараз показуємо 1 - 2 з 2
  • Ескіз
    ДокументВідкритий доступ
    Heuristic algorithms for finding the minimum Steiner tree in the problem of optimizing the deployment and motion control of several flying information and telecommunication robots
    (National Technical University of Ukraine "Igor Sikorsky Kyiv Polytechnic Institute", 2022) Lysenko, Olexandr; Valuiskyi, Stanislav; Novikov, Valeriy; Sushyn, Ihor; Chumachenko, Serhii; Guida, Oleksandr
    Background. The article explores the problem of combining the motion control of existing FITRs and the deployment of new FITRs so that the number of new FITRs deployed to support the communication of terrestrial subscribers can be minimized. This problem is formulated as the Steiner Minimum Tree Problem (SMT) with existing mobile Steiner points with a constraint on the edges length of the network graph. Objective. Improve the mathematical model for ensuring the connectivity of episodic radio networks using FITRs and improve the algorithms for providing the connectivity of episodic radio networks using FITRs. Methods. The two algorithms (deploying new FITRs before moving existing FITRs, and moving existing FITRs before deployment of new FITRs) separate the problem and solve the deployment problem, the movement one after the other. In contrast, the algorithm for deploying new FITRs while moving existing FITRs optimizes the deployment problem and the control of movement across and solves these two problems simultaneously. Results. A proposed method includes three heuristic algorithms for placing new FITRs, taking into account the movement of existing FITRs (that is, considering scenarios for moving existing FITRs: deploying new FITRs before, after, or during the movement of existing FITRs) for the SMT problem with existing mobile Steiner points with a constraint on the edges length of the network graph. Conclusions. Evaluation of the effectiveness of the proposed algorithms in various scenarios shows that algorithms taking into account the movement of existing FITRs are always more efficient (in terms of the number of newly added FITRs) than an algorithm without taking into account the movement of existing FITRs.
  • Ескіз
    ДокументВідкритий доступ
    Static and dynamic assessments of information signs in recognition of sources and objects of observation in the process of radio monitoring
    (National Technical University of Ukraine "Igor Sikorsky Kyiv Polytechnic Institute", 2022) Ilnytskyi, Аnatoliy; Tsukanov, Oleg
    Background. The current state and problems of the surveillance and radio monitoring systems of Ukraine require fundamentally new approaches to increasing their efficiency and the level of informatization. At the same time, the informatization of the radio monitoring system should be understood as the process of implementation and application in various areas of their activity of methods and means of collecting, transmitting, processing, saving and using information in order to increase the effectiveness of conducting radio monitoring and meet the needs of national security based on the formation and use of information resources. Objective. The purpose of the paper is to increase the effectiveness of radio monitoring by using the calculation of estimates of dynamic and static informational features when recognizing sources and objects of radio radiation and determining their phase (operational) state and level of possible danger. Methods. Recognition is based on the method of least squares by calculating the degree of "similarity" (similarity coefficient) of the recognized object with objects whose classes are known. Both the researched and reference objects are presented as a set of values of informational features of various nature, some of which are unchanged over the entire period of observation, that is, static, while others change dynamically. Results. The structure of the automated system of classification and recognition of surveillance objects and the recognition algorithm based on the calculation of static and dynamic information features and the similarity coefficient are proposed. Conclusions. A distinctive feature of deciding whether an object or a source of information belongs to one or another class feature is the calculation of the degree of "similarity" (similarity coefficient) of the recognized object to objects whose classes are known. To eliminate recognition errors associated with a violation of the synchronicity of measurements of the values of dynamic informational features of reference objects and objects to be recognized, a calculation is required taking into account possible time shifts.