Hnatushenko V. Models and Methods of Improving the Quality of Traffic Transfer in Wireless Networks

The thesis is devoted to a topical applied scientific problem of providing a given level of service quality to customers of wireless networks by taking into account properties of the network traffic, with a subsequent choice of the transmission distance and limits of connectivity. The method of generating traffic with a given degree of self-similarity and long-term dependence on the basis of the ON-OFF model, which, unlike the known methods, provides desired settings based on changes in data rate, was developed. For the first time, a mathematical model of the network based on selection of a fixed rate of request denials as QoS was developed. The use of this model makes it possible to effectively manage the traffic in a network regardless of the number of its elements. A model of elements of a wireless sensor network node was developed, which, due to the intensity of traffic analysis, makes it possible to track energy consumption at each node and further optimize the energy efficiency of the whole network. A mathematical model of placement and movement of cluster centers on the base of determining the optimal renovation time of the location of sensors in wireless sensor networks to maximize coverage was developed. For the first time a mathematical model of the traffic in wireless sensor networks based on dynamic selection of parameters (replication level, transmission range and schedules of activity / sleeping of nodes) with the provision of QoS and limitations of connections that allows one to maximize the service life of a WSN was proposed. A method of calculating energy consumption for transfer, processing and reading of data at the nodes in the propagation of traffic based on probabilistic analysis of the distribution of lifetime was developed, which allows for selection of an efficient operation mode of WSN elements. Combined implementation of models and methods introduced in this work allows one to build wireless networks on the basis of self-similar traffic characteristics, which can increase the quality of serving the users.