Cut-Equivalent Trees are Optimal for Min-Cut Queries

TL;DR

This paper proves that cut-equivalent trees are essentially optimal for min-cut queries, establishing near-linear time construction equivalence and introducing robust approximation techniques that improve max-flow computations.

Contribution

The paper shows the equivalence between near-linear time construction of cut-equivalent trees and efficient min-cut query data structures, and introduces approximation methods for faster max-flow algorithms.

Findings

Cut-equivalent trees can be constructed in near-linear time if and only if efficient min-cut query data structures exist.

A $(1+ ext{epsilon})$-approximate flow-equivalent tree can be built in $n^{2+o(1)}$ time using approximation algorithms.

This leads to the first near-quadratic time $(1+ ext{epsilon})$-approximate All-Pairs Max-Flow algorithm.

Abstract

Min-Cut queries are fundamental: Preprocess an undirected edge-weighted graph, to quickly report a minimum-weight cut that separates a query pair of nodes $s,t$. The best data structure known for this problem simply builds a cut-equivalent tree, discovered 60 years ago by Gomory and Hu, who also showed how to construct it using $n-1$ minimum $st$-cut computations. Using state-of-the-art algorithms for minimum $st$-cut (Lee and Sidford, FOCS 2014) arXiv:1312.6713, one can construct the tree in time $\tilde{O}(mn^{3/2})$, which is also the preprocessing time of the data structure. (Throughout, we focus on polynomially-bounded edge weights, noting that faster algorithms are known for small/unit edge weights.) Our main result shows the following equivalence: Cut-equivalent trees can be constructed in near-linear time if and only if there is a data structure for Min-Cut queries with near-linear preprocessing time and polylogarithmic (amortized) query time, and even if the queries are restricted to a fixed source. That is, equivalent trees are an essentially optimal solution for Min-Cut queries. This equivalence holds even for every minor-closed family of graphs, such as bounded-treewidth graphs, for which a two-decade old data structure (Arikati et al., J.~Algorithms 1998) implies the first near-linear time construction of cut-equivalent trees. Moreover, unlike all previous techniques for constructing cut-equivalent trees, ours is robust to relying on approximation algorithms. In particular, using the almost-linear time algorithm for $(1+\epsilon)$-approximate minimum $st$-cut (Kelner et al., SODA 2014), we can construct a $(1+\epsilon)$-approximate flow-equivalent tree (which is a slightly weaker notion) in time $n^{2+o(1)}$. This leads to the first $(1+\epsilon)$-approximation for All-Pairs Max-Flow that runs in time $n^{2+o(1)}$, and matches the output size almost-optimally.