Brych M. Models and algorithms of heterogeneous mobile networks operation.

The thesis is devoted to solving the actual scientific problem of developing the models and algorithms for adaptive utilization of licensed and unlicensed spectrum resources by LTE mobile networks in the presence of several competing network operators and high heterogeneity of radio access technologies, device types and parameters of data transmission quality. The current state and state-of-the-art trends in 5G mobile networks development have been analyzed in details. The key constraints of the network capacity have been identified, such as spectrum scarcity, problems of spectrum allocation and licensing of the wireless communication standards in Ukraine, absence of a unified management platform for a heterogeneous network infrastructure, etc. To alleviate these above mentioned issues, several solutions to improve the future 5G mobile network infrastructure have been proposed. First of all, the method of carrier sense multiple random access has been further developed to adjust media access control to the LTE frame structure. This solution enables the adaptive utilization of unlicensed radio frequency resources to increase the total throughput of the heterogeneous mobile network. In addition, a new model for the unlicensed spectrum sharing by several mobile network operators has been proposed. The novelty of this model is in the new game theoretical framework with proportional payoff distribution among the operators by calculation of the Shapley value, which prevents their selfish behavior. In the proposed framework, Shapley value determines how much spectrum should be allocated for each mobile network operator depends on their payoffs. High payoff means that spectrum utilization is close to the initial spectrum allocation. Low payoff means that network operator allocates much more or less spectrum than is actually required. The key advantage of this approach is in the flexibility to the traffic fluctuation within internal operator’s network. Proposed model takes into account the instantaneous bandwidth demand of the operator, which enables the increasing of the spectrum utilization efficiency in conditions of simultaneous operation of multiple LTE operators in the unlicensed spectrum. Further in this thesis, a new approach for cooperative spectrum sensing for the scenario of device-to-device assisted cellular network. The main idea of this approach is to reduce the number of total spectrum sensing attempts for each user by enabling their cooperation in clusters via device-to-device communication. Each cluster is formed by two or more devices in the vicinity of each other, which sense the carrier occupancy alternately and share the obtained information with other cluster members. Thus, the number of sensing attempts is decreased proportionally to the number of users in a single cluster. To further improve the bandwidth utilization, cluster members configure the size of contention window to minimize the idle time between their access to the unlicensed channel. The simulation of a heterogeneous mobile network operation in the licensed spectrum has been conducted and advantages of the proposed solutions have been confirmed.