An algorithm for K-connectivity under pure SINR model

Abstract

In this paper we describe a deterministic algorithm to compute a (vertex disjoint) k-connected subgraph (k ≥ 1) of an undirected complete weighted Euclidean graph. The weight of the computed subgraph is within an O(k log n) factor of the weight of an optimum kconnected subgraph for a given graph, where n is the number of nodes. In particular, the algorithm was designed for pure SINR (Signal-toInterference-plus-Noise Ratio) model. Variations of SINR models are currently considered the most appropriate ones to design and analyse algorithms for wireless networks when interference is taken into consideration. To the best of our knowledge, we describe the first algorithm to compute a k-connected vertex-disjoint subgraph under SINR. It has O(log g) runtime complexity, where g is the granularity of the network.