Random hypernets in reliability analysis of multilayer networks

The general approach to constructing structural models of non-stable multi-level networks is proposed. This approach is based on hypernets—relatively new mathematical object, which is successively used for modeling different multilevel networks in the Institute of Computational Mathematics and Mathematical Geophysics SB RAS, Russia, for last 30 years. Hypernets allow standard description of neighboring levels interconnection in a mathematically correct way. Using this mathematical object allows easy modifications of data with model changing and/or development and efficiently organize data search for different computational or optimization algorithms. Optimization of mapping of secondary (logical) network onto structure of unreliable primary (physical) network is considered as example.