Information dissemination dynamics in delay tolerant network: a bipartite network approach

In this paper, we present a model of a delay tolerant network(DTN) and identify that this model can be suitably reformulated as a bipartite network and that the major predictions from the former are equivalent to that of the latter. In particular, we show that the coverage of the information dissemination process in the DTN matches accurately with the size of the largest component in the suitably thresholded one-mode projection of the corresponding bipartite network. In the process of this analysis, some of the important insights gained are that (a) arbitrarily increasing the number of agents participating in the dissemination process cannot increase the coverage once the system has reached the stationary state for a given buffer time (i.e., the time for which a message resides in the buffer of the places visited by the agents), (b) the coverage varies inversely with the square of the number of places in the system and directly with the square of the average social participation of the agents and (c) it is possible to design an optimal buffer time for a desired cost of coverage. To the best of our knowledge, this is the first such work that employs the rich theoretical backbone of bipartite networks as a "proxy" for the analysis of the otherwise intractable DTN dynamics thus allowing for various novel analytical estimates.