Connectivity problem in wireless networks

We study the complexity of the following connectivity problem in wireless networks: for a given placement of n nodes in the plane, the goal is to compute a channel and power assignment that forms strongly connected communication structure spanning all nodes. The complexity measure is the total number of assigned channels, and the goal is to minimize this number. We work with two signal inference models: Geometric Radio Networks (GRN) and Signal to Interference Plus Noise Ratio (SINR). We show a generic polynomial-time transformation from the wide class of separable assignments in GRN to assignments in the SIRN model. This transformation preserves asymptotic complexity, i.e., the number of channels used in the assignments. In this way we show an assignment, constructed in polynomial-time, guarantying connectivity in the SINR model by using only O(log n) channels, which is an improvement over the best previous result O(log² n) presented in [21].