The circuit of congestion control using feedback by the sign of function of sensitivity to telecommunications network performance was considered. To determine a given function, the use of a simple neural network model of a dynamic system was proposed. Control over the existence or a threat of congestion is executed based on the analysis of the length of a queue at the side of information receiver. To analyze the system, the cost function was determined as the objective function of congestion existence. The proposed algorithm of optimal control ensures the formation of a control signal in such a way that the system output should maximally match the pre-established features – the key indicators for network efficiency. The congestion control circuit with the feedback based on the sign of sensitivity of the function of system performance was developed. The sign of performance sensitivity provides an optimal direction to configure the data source rate. The neural model for a multi-step prediction of the state of the queue at the side of the telecommunication network receiver was proposed. If the neural network is configured to monitor the dynamics of the system and shows that the quadratic error is negligible, it is believed that the executed step corresponds to the system output, predicted in advance. The algorithm of additive increase/multiple decrease, which determines the change of the data source rate, depending on the sign of function of sensitivity of performance indicator was proposed. This algorithm is an alternative system of congestion prediction and flow control based on the threshold queue filling. A comparative analysis of the effectiveness of controlling circuits for congestion detection based on queues and on the function of sensitivity of telecommunication network performance was performed. It was shown that the magnitude of the queue and fluctuation in the source rate is smaller than that for the queue-based circuit. Results from modeling the performance of the proposed circuit show that the circuit based on a sensitivity function has better key performance indicators in comparison with the conventional circuit of queue threshold selection


    Zugriff

    Download


    Exportieren, teilen und zitieren



    Titel :

    DEVELOPMENT OF THE METHOD TO CONTROL TELECOMMUNICATION NETWORK CONGESTION BASED ON A NEURAL MODEL


    Beteiligte:

    Erscheinungsdatum :

    2019-04-30


    Anmerkungen:

    oai:zenodo.org:3384629
    Eastern-European Journal of Enterprise Technologies 2(9 (98)) 67-73



    Medientyp :

    Aufsatz (Zeitschrift)


    Format :

    Elektronische Ressource


    Sprache :

    Englisch



    Klassifikation :

    DDC:    629




    NEURAL NETWORK MODELS FOR TRAFFIC CONTROL AND CONGESTION PREDICTION

    Gilmore, John F. / Abe, Naohiko | Taylor & Francis Verlag | 1995


    NEURAL NETWORK MODELS FOR TRAFFIC CONTROL AND CONGESTION PREDICTION

    Hobeika, Antoine G. | Taylor & Francis Verlag | 1995


    Road intersection congestion prediction method based on time point process neural network model

    WANG JIANLONG / YI CHONGZHENG / ZHANG LI et al. | Europäisches Patentamt | 2023

    Freier Zugriff

    Convolutional neural network for recognizing highway traffic congestion

    Cui, Hua / Yuan, Gege / Liu, Ni et al. | Taylor & Francis Verlag | 2020


    Prediction of Traffic Congestion through Convolutional Neural Network

    K, Gayathri / B, Gopalakrishnan / S, Renuga Devi | IEEE | 2021