A $\frac{4}{3}$-Approximation Algorithm for the Minimum 2-Edge Connected Multisubgraph Problem in the Half-Integral Case
Sylvia C. Boyd, Joseph Cheriyan, Robert Cummings, Logan Grout, Sharat Ibrahimpur, Zoltán Szigeti, Lu Wang · SIAM Journal on Discrete Mathematics · 2022
Given a connected undirected graph $\overline{G}$ on $n$ vertices and nonnegative edge costs $c$, the $\ensuremath{{2ECM}}$ problem is that of finding a 2-edge connected spanning multisubgraph of $\overline{G}$ of minimum cost. The natural linear program (LP) for $\ensuremath{{2ECM}}$, which coincides with the subtour LP for the traveling salesman problem on the metric closure of $\overline{G}$, gives a lower bound on the optimal cost. For instances where this LP is optimized by a half-integral solution $x$, Carr and Ravi (1998) showed that the integrality gap is at most $\frac43$: they show that the vector $\frac43 x$ dominates a convex combination of incidence vectors of 2-edge connected spanning multisubgraphs of $\overline{G}$. We present a simpler proof of the result due to Carr and Ravi by applying an extension of Lovász's splitting-off theorem. Our proof naturally leads to a $\frac43$-approximation algorithm for half-integral instances. Given a half-integral solution $x$ to the LP for $\ensuremath{{2ECM}}$, we give an $O(n^2)$-time algorithm to obtain a 2-edge connected spanning multisubgraph of $\overline{G}$ with cost at most $\frac43 c^T x$. We also consider a related problem of finding a cheap 2-edge connected spanning subgraph of a 3-regular, 3-edge connected graph $G = (V,E)$ with arbitrary edge costs $c$. We give a polynomial-time Las Vegas algorithm that finds a random 2-edge connected spanning subgraph $H$ of $G$ whose expected cost, $\mathbb{E}\left[{c(H)}\right]$, is at most $\frac45 c(E)$.