Approximating Transitivity in Directed Networks

TL;DR

This paper investigates the problem of approximating transitive reductions in directed graphs, introducing extensions with constraints and edge labels, and provides approximation algorithms with proven bounds and hardness results.

Contribution

It introduces new variants of the transitive reduction problem with constraints and edge labels, and offers approximation algorithms with specific ratios and hardness proofs.

Findings

Polynomial-time approximation for minimization within ratio 1.5

Approximate maximization within ratio 2

MAX-SNP hardness for cycle length limited to 5

Abstract

We study the problem of computing a minimum equivalent digraph (also known as the problem of computing a strong transitive reduction) and its maximum objective function variant, with two types of extensions. First, we allow to declare a set $D\subset E$ and require that a valid solution $A$ satisfies $D\subset A$ (it is sometimes called transitive reduction problem). In the second extension (called $p$-ary transitive reduction), we have integer edge labeling and we view two paths as equivalent if they have the same beginning, ending and the sum of labels modulo $p$. A solution $A\subseteq E$ is valid if it gives an equivalent path for every original path. For all problems we establish the following: polynomial time minimization of $|A|$ within ratio 1.5, maximization of $|E-A|$ within ratio 2, MAX-SNP hardness even of the length of simple cycles is limited to 5. Furthermore, we believe that the combinatorial technique behind the approximation algorithm for the minimization version might be of interest to other graph connectivity problems as well.