Cut Tree Structures with Applications on Contraction-Based Sparsification

Abstract

We introduce three new cut tree structures of graphs $G$ in which the vertex set of the tree is a partition of $V(G)$ and contractions of tree vertices satisfy sparsification requirements that preserve various types of cuts. Recently, Kawarabayashi and Thorup \cite{Kawarabayashi2015a} presented the first deterministic near-linear edge-connectivity recognition algorithm. A crucial step in this algorithm uses the existence of vertex subsets of a simple graph $G$ whose contractions leave a graph with $\tilde{O}(n/\delta)$ vertices and $\tilde{O}(n)$ edges ($n := |V(G)|$) such that all non-trivial min-cuts of $G$ are preserved. We improve this result by eliminating the poly-logarithmic factors, that is, we show a contraction-based sparsification that leaves $O(n/\delta)$ vertices and $O(n)$ edges and preserves all non-trivial min-cuts. We complement this result by giving a sparsification that leaves $O(n/\delta)$ vertices and $O(n)$ edges such that all (possibly not minimum) cuts of size less than $\delta$ are preserved, by using contractions in a second tree structure. As consequence, we have that every simple graph has $O(n/\delta)$ $\delta$-edge-connected components, and, if it is connected, it has $O((n/\delta)^2)$ non-trivial min-cuts. All these results are proven to be asymptotically optimal. By using a third tree structure, we give a new lower bound on the number of \emph{pendant pairs}. The previous best bound was given 1974 by Mader, who showed that every simple graph contains $\Omega(\delta^2)$ pendant pairs. We improve this result by showing that every simple graph $G$ with $\delta \geq 5$ or $\lambda \geq 4$ or $\kappa \geq 3$ contains $\Omega(\delta n)$ pendant pairs. We prove that this bound is asymptotically tight from several perspectives, and that $\Omega(\delta n)$ pendant pairs can be computed efficiently.