Self Regulating Mechanisms for Network Immunization

ABSTRACT. Many real network, e.g. internet or communication systems, can be modeled as scale-free network. An important property of these type of networks is that the diffusion of information to all nodes is very fast. Therefore, immunization strategies are necessary for this type of networks to prevent that a virus can infect, starting from some nodes, the entire network. In recent years, various immunization strategies has been developed and they can be divided into two main class: centralized and distributed strategies. Both of these approaches have a limitation of having the need to know the number of nodes of the network that must be immunized. In this paper, we propose an AOC based immunization strategy modification in which the entities self regulate their population inside the network. In this strategy the entities are deployed in a network with a threshold of coverage of the graph to be achieved. The entities, while collectively search the nodes with the highest degree, estimate the coverage of the graph through the information in their local environment and reproduce accordingly. The self regulating mechanism is evaluated on a set of public benchmark networks, even with community-based structures, and experimental results about convergence, coverage efficiency, computational cost and scalability are presented.