Background. The article discusses the formalization of performance indicators of a multiservice network management system based on the use of hybrid neural-fuzzy technology that combines the advantages of fuzzy logic and neural networks. A developed fuzzy neural model was subsequently used to make a decision about the transfer of real time streams through the channel of the multi-servo network. Methods for assessing the effectiveness of multiservice communication networks can be conditionally divided into three large groups: technical, economic and technical and economic. As the technical characteristics of the efficiency of the computer network, various indicators of network performance and reliability are used. Estimates of the costs of designing, installing and maintaining the network are used as economic characteristics. Technical and economic indicators are used for a comprehensive assessment of the project, and include various combinations of technical and economic characteristics. The response time is an integral performance characteristic, the most important for the network subscriber. In general, the response time is defined as the time interval between the occurrence of a subscriber's request for a network service and the receipt of a response to this request. Methods. When implementing in the channel dynamic redistribution of the bandwidth allocated for the transmission of packets of various classes, the model of neuro-fuzzy prediction of the number of dropped packets proposed. The number of discarded packets of a given class from the number that claimed to be transmitted over the channel depends, firstly, on the dynamics of the arrival of packets of this class for transmission over the channel and, secondly, on what the current value of the channel bandwidth is allocated for the transmission of these packets. Studies have shown that to predict the number of packets of a given class, it is advisable to supply data on the number of received packets in the three previous cycles (values of and ) to the input of a fuzzy neural system, as well as the value - the value of the channel bandwidth allocated in the current cycle for transmitting packets of this class over a telecommunication network channel. Findings. This decrease in the allocated bandwidth led to the fact that against the background of the increase in the number of received packets Zi observed in cycles 8 and 9 (see Figure 7), the number of dropped packets in cycle 9 increased compared to cycle 4 (R9> R4). The analysis presented in Figure 8 and Figure 9 of the results shows that the predicted value of R ̌_i, which is calculated by the neuro-fuzzy system in each current cycle i, practically coincides with the real values of the number of dropped packets Ri recorded in the next cycle (i + 1). The prediction accuracy, established as a result of numerous simulation experiments, is 95–97 %. Conclusions. To compare the results obtained with the same initial data, a series of simulation experiments was carried out using a model in which the choice of an intersegment interval was implemented based on the use of a fuzzy inference system. The results of the experiments are presented in graphs showing the dependence of the duration of the data flow and the loss of segments on the average available bandwidth of the telecommunication channel. On these graphs, solid curves show the characteristics of the data stream transmission obtained using neuro-fuzzy selection of the inter-segment interval, and the dashed curves show the results of the management of the inter-segment interval based on the application of the system. The analysis of the presented dependencies shows that when transmitting a data stream over a channel, the available bandwidth of which does not exceed 50 %, the use of a neuro-fuzzy system to select an inter-segment interval provides a decrease in segment losses by 5.2–11.3 % and a decrease in the average time of streaming data by 7.1-12.3 %. Thus, a neuro-fuzzy model has been developed, designed to select the inter-segment interval in a telecommunications network. Unlike the existing ones, this model is based on the use of a fuzzy neural network apparatus. The results of simulation showed that the use of the developed model of the shortage of available channel capacity will ensure a decrease in segment losses and a decrease in the average transmission time of data streams.
1. Bichkov T.D., Goncharovskiy A.А. Vibor marshruta peredacha paketa dannix v telecommunikasionnoy seti pri nechetkix usloviyax // Dinamika sistem, mexanizmov i mashin: Маter V Меjdunar. nauch. – tex. коnf. Оmsk: Izd- vo ОmGTU, Кn. 3, 2004. S. 119 – 122
2. Borisov V.V., Кruglov V.V., Fedulov А.S. Nechetkie modeli I seti.– М.: Goryachaya liniya– Теlecom, 2007.- 284 s.
3. Polschykov К.А., Kubrakova Е. N., Кrasnobayev V.А. Моdel neyro-nechetkogo prognozirovaniya sredney intensivnosti postupleniya zaprosov na peredachu potokov realnogo vremeni po kanalu telekommunikasionnoy seti// Sistemi obrabotki informasi. – 2014. – Vip. 2 (118).– S. 193–197.
4. Popovskiy V.V. Matematicheskie osnovi upravleniya i adaptasiya v telecommunikasionnix systemax/ V.V. Popovskiy, V.F. Oleynic Х.: S 362 s.
5. Polschykov К. The Methodology of Modeling Available for data Traffic Bandwidth Telecommunications Network / K.Polschykov, S. Olexij, N.Rvachova // Proceedings of the X International Conference “Modem Problems of Radio Engineering, Telecommunications and Computer Slavske,2010.-P.158.
6. Polshchikov K.A. Theoretical foundations of packet data transmission in a wireless self-organizing network when providing communication at construction sites // Dissertation for the degree of Doctor of Technical Sciences, Belgorod. — 2016.
7. Marakhimov A.R. System design of information and computer networks in the conditions of fuzzy initial information. - Monograph .: Tashkent, 2006.- 246s.
8. Deart V.Yu. Multiservice communication networks. Transport networks and access networks. - Moscow: Insvyazizdat, 2008. — 168p.
9. Gilyazov R.L., Gitman M.B., Stolbov V.Yu. Construction of a generalized optimality criterion for a given fuzzy hierarchical structure of particular criteria // Proceedings of XXXV Intern. conf. "1T-8 & E'08", Ukraine, Gurzuf, 2008.- P.207-210.
10. Zarubin Yu., Sherboboyeva G.B., Marakhimov A.A Multi-criteria Evaluation Approach in the Automated Information System Design Problems // The 4th International Conference on BIG DATA APPLICATIONS AND SERVICES (BIGDAS2017). ISSN 2466-135X Vol.4. №.1 Hosted by National University of Uzbekistan Korea Big Data Service Society August 15-18. 2017 Tashkent. pp. 9-16.
11. Sherboboyeva G.B. Probability - time characteristics of the multiservice network with multiple access protocol // WCIS-2014. Тashkent. 2014. pp.317-321
12. Sherboboyeva G.B. Fuzzy modeling and multicriteria evaluation of the acceptability design solutions in the automated information system design // WCIS-2014. Тashkent. 2014. pp.458-463.
13. Sherboboeva G.B. Neuro-fuzzy models and algorithms for investigating the effectiveness of control processes in multiservice networks. // Discertation of doctor of philocophy (PhD) on technical sciences, Tashkent-2019.
14. Siddikov I.X., Sherboboyeva G.B. Hybrid model management of information messages in multiservise networks// The advanced science journal, USA, 2014, Vol.2 pp.38-41.
15. William Siler, James J. Buckley. Fuzzy expert systems and fuzzy reasoning. – Hoboken, New-Jersey: John Wiley & Sons, Inc., 2005. – 405 p.
16. Basharin G.P., Gaidamaka Yu.V., Samouylov K.E. Mathematical Theory of Teletraffic and its Application to the Analysis of Multiservice Communication of Next Generation Networks. // Automatic Control and Computer Sciences. – 2013. – 47:2. – Pp. 62-69.
Sherboboyeva, Gulrukh Bakhtiyorovna
"CALCULATION OF THE EFFECTIVENESS OF THE FUNCTIONING OF MULTISERVICE COMMUNICATION NETWORKS TAKING
INTO ACCOUNT DROPPED PACKETS,"
Scientific reports of Bukhara State University: Vol. 4
, Article 1.
Available at: https://uzjournals.edu.uz/buxdu/vol4/iss6/1