Core Capacity of Wireless Ad Hoc Networks*

In this paper, we model energy-limited wireless ad hoc networks as non-transferable-utility cooperative network flow games. We define the %ore capacity region” of a wireless ad hoc network to be the set of Pareto-optimal utility vectors that cannot be blocked by any proper subset of the node set. We show that the core capacity region is non-empty under the linear utility model. Under the many-to-one traffic model in which all the nodes have traffic demands for a single base station, we show that the only utility vector in the core capacity region is the one achieved by the cellular uplink topology. Under the one-to-one traffic model in which each node generates a traffic demand for another randomly picked node, we demonstrate by simulation the growth of core sum capacity as a function of the number of nodes.