Approximations of the Densest k-Subhypergraph and Set Union Knapsack problems

Abstract

For any given $\epsilon>0$ we provide an algorithm for the Densest $k$-Subhypergraph Problem with an approximation ratio of at most $O(n^{\theta_m+2\epsilon})$ for $\theta_m=\frac{1}{2}m-\frac{1}{2}-\frac{1}{2m}$ and run time at most $O(n^{m-2+1/\epsilon})$, where the hyperedges have at most $m$ vertices. We use this result to give an algorithm for the Set Union Knapsack Problem with an approximation ratio of at most $O(n^{\alpha_m+\epsilon})$ for $\alpha_m=\frac{2}{3}[m-1-\frac{2m-2}{m^2+m-1}]$ and run time at most $O(n^{5(m-2)+9/\epsilon})$, where the subsets have at most $m$ elements. The author is not aware of any previous results on the approximation of either of these two problems.