A resiliency analysis of information distribution policies over mobile ad hoc networks

Mobile ad hoc networks (MANETs) have been widely used for information distribution tasks. The radio communication channels that are established dynamically between the different terminals of these networks, like wireless sensors, ground vehicles, and autonomous assets (e.g. UAVs), allow for a flexible and efficient information distribution process. However, given the dynamic topology and limited bandwidth of MANETs, the performance of such systems is often limited by their resiliency against connection interruptions caused by different actors, such as adverse weather conditions, random component breakdowns, physical barriers, and adversarial attacks. Whenever this type of connection interruption occur, the MANET losses its ability to distribute the information efficiently, causing delays that often reduce the value of the information being sent over the network. In this paper, we are interested in analyzing the performance of this type of networks in the presence of such interruptions. To this end, we first transform the information distribution process into a flow problem over a time-expanded directed acyclic graph, setting as the objective function the maximization of the time-discounted value of the distributed data. To analyze the resiliency of the resulting information distribution plans, we take the stance of an adversary and apply network interdiction techniques to estimate the worst possible deterioration that could be inflicted by targeting some connections of the MANET. The interdiction scheme we introduce can be adapted to model various situations and contexts. Finally, we test our framework on instances of different sizes and analyze some interesting properties observed in the experiments.