FUZZY CONTROL SYSTEM OF TELECOMMUNICATION NETWORK RESOURCES

Abstract

It is shown in this paper that the automatic control of the basic parameters of a telecommunication network occurs in conditions of uncertainty, that is connected with the lack of sufficient statistics characterizing the control object (in this case, the system of information distribution or the transmission, path). Therefore, when managing the resources of a telecommunication network, it is necessary to take into account some factors that are not formalized or difficult to formalize. It is shown that in the process of managing the resources of telecommunication networks there are a number of problems that cannot be described by well-known management methods, for example: it is not possible to reliably determine the state of the information distribution system in a given node of the network, and in the process of operational management there is not enough time to collect and process the necessary statistical data; it is not possible to really assess the impact of interference on the quality of the transmission path; it is not possible to estimate the operator's losses caused by deviations of the parameters of the paths or their malfunction, etc. It is proposed to solve these problems in the theory and practice of network management by constructing and using fuzzy algorithms in mathematical models of resource management. At the same time, the level of complexity of such control systems is so high that the use of only known deterministic and stochastic models does not provide desirable characteristics. In these cases, adequate mathematical models of control objects (information distribution systems, transmission paths) should be built on the basis of the fuzzy set theory apparatus, and the resource management system acquires the features of an intelligent control system. Representing performance indicators using fuzzy numbers can provide more information than accurate deterministic interval estimation.The core of such a control system is the fuzzy controller. It is established that the main problem with creating fuzzy controllers is that changing the parameters of a control object requires modifying the control rules with their subsequent iterative adjustment, which takes time.