Модели и методы анализа системы передачи информации в компьютерных сетях в условиях немарковского трафика

Abstract

Thesis for granting Degree of Candidate of Technical sciences in specialty 05.13.05 – Computer Systems and Components. – National Technical University "Kharkiv Polytechnic Institute", Kharkov, 2016. This thesis developed a set of mathematical models and methods that together solve important scientific and technical analysis of the problem of transmission of information via computer networks in a non-Markovian traffic. Using a simulation model evaluated the effectiveness of the service system with self-similar input stream. The dependences of the basic characteristics of the system on the value of self-similarity, and system load. The method of analyzing the effectiveness of information transfer routes from the primary to the end nodes via intermediate nodes using a probability-time graphs. The implementation methodology provides a Pareto-optimal set of routes according to the criteria and the probability of the average time of delivery. The technology of finding the optimal route, taking into account the dynamics of traffic intensity. To calculate the predicted values of the input rate used by the correlation function of the flow. Evaluating the effectiveness of routes in the network carried out using probability-time graphs. In this two-criterion assessment of the quality of routes is carried out by constructing a Pareto-optimal sets them. The problem of finding the best route is solved by genetic algorithm. Using a genetic algorithm in constructing an optimal algorithm in branching system of roads is ensured by the procedure of decomposition.