Edge covering with budget constrains

Abstract

We study two related problems: finding a set of k vertices and minimum number of edges (kmin) and finding a graph with at least m' edges and minimum number of vertices (mvms). Goldschmidt and Hochbaum \cite{GH97} show that the mvms problem is NP-hard and they give a 3-approximation algorithm for the problem. We improve \cite{GH97} by giving a ratio of 2. A 2(1+\epsilon)-approximation for the problem follows from the work of Carnes and Shmoys \cite{CS08}. We improve the approximation ratio to 2. algorithm for the problem. We show that the natural LP for \kmin has an integrality gap of 2-o(1). We improve the NP-completeness of \cite{GH97} by proving the pronlem are APX-hard unless a well-known instance of the dense k-subgraph admits a constant ratio. The best approximation guarantee known for this instance of dense k-subgraph is O(n^{2/9}) \cite{BCCFV}. We show that for any constant \rho>1, an approximation guarantee of \rho for the \kmin problem implies a \rho(1+o(1)) approximation for \mwms. Finally, we define we give an exact algorithm for the density version of kmin.