Polishchuk O. Models and methods of complex evaluation of complex network and hierarchically network systems

Current state of researches dedicated to complex network systems and different inter-actions between systems was analysed in this thesis. Main directions of those researches were defined and conclusions were driven that mostly those researches consist in studying structural features of systems and interactions. The flow models of complex network system were developed and main local and global flow characteristics of its elements were defined. Those characteristics are applied for determining actual structure of the system, analysing its development on all lifecycle stages and finding alternative flow directions that bypass isolated areas of network. Main structural and functional features for classifi-cation of different interactions between systems are defined. The impact of interactions on operation of interconnected complex network systems is demonstrated. Influence and bet-weenness parameters of components of network systems and monoflow partially overlap-ped multilayered systems were defined. Using those parameters, the conditions were stu-died under which system is resistant to negative internal and external influence. Importan-ce of different interactions within system and between systems is determined. The scena-rios are proposed of deliberate attacks on network systems, in order to define the most at-tractive targets of those attacks and to develop efficient ways for protecting those targets. Methods are proposed for reducing complexity of network systems and different interac-tions between systems which allows to create relevant models of those systems and inter-actions. The notion of flow core of complex network systems and monoflow partially overlapped multilayered network systems is defined, that allows to distinguish the most important components of system structure in terms of operation, to downscale substantial-ly system model while tracking the extent of preserving its relevance and to solve the pro-blem of finding communities within the network.