Randomized target search and its convergence in dynamic multi-layer networks
Qing Hui, Chen Peng · 2016
Modeling and analyzing a randomized target search problem for surveillance and protection in time-dependent, switching multi-layer networks will be considered in this paper. We use a tensor-based modeling approach to characterize time evolution of graph topologies for multi-layer networks. With this multilinear algebraic model, we conduct a global convergence analysis for the proposed randomized target search algorithm in dynamic multi-layer networks. Sufficient conditions to guarantee the global convergence of such algorithms are derived via the new notion of subset matrix paracontraction and some latest results developed based on this notion. This technique can help to determine computationally efficient, stable strategies for protecting hierarchical critical infrastructural networks and enhance its resilience in dynamically changing environments under disruptive events.