Enumeration of Far-Apart Pairs by Decreasing Distance for Faster Hyperbolicity Computation

TL;DR

This paper introduces a memory-efficient method for enumerating far-apart node pairs in graphs, significantly improving the feasibility of computing hyperbolicity for large graphs by reducing memory usage and computational complexity.

Contribution

The authors develop a new data structure that enables enumeration of far-apart pairs without storing the entire distance matrix, allowing hyperbolicity computation on larger graphs than previously possible.

Findings

Memory consumption reduced by at least two orders of magnitude.

Enabled hyperbolicity computation on several large graphs previously infeasible.

Only a small fraction of far-apart pairs need consideration for hyperbolicity.

Abstract

Hyperbolicity is a graph parameter which indicates how much the shortest-path distance metric of a graph deviates from a tree metric. It is used in various fields such as networking, security, and bioinformatics for the classification of complex networks, the design of routing schemes, and the analysis of graph algorithms. Despite recent progress, computing the hyperbolicity of a graph remains challenging. Indeed, the best known algorithm has time complexity $O(n^{3.69})$, which is prohibitive for large graphs, and the most efficient algorithms in practice have space complexity $O(n^2)$. Thus, time as well as space are bottlenecks for computing hyperbolicity. In this paper, we design a tool for enumerating all far-apart pairs of a graph by decreasing distances. A node pair $(u, v)$ of a graph is far-apart if both $v$ is a leaf of all shortest-path trees rooted at $u$ and $u$ is a leaf of all shortest-path trees rooted at $v$. This notion was previously used to drastically reduce the computation time for hyperbolicity in practice. However, it required the computation of the distance matrix to sort all pairs of nodes by decreasing distance, which requires an infeasible amount of memory already for medium-sized graphs. We present a new data structure that avoids this memory bottleneck in practice and for the first time enables computing the hyperbolicity of several large graphs that were far out-of-reach using previous algorithms. For some instances, we reduce the memory consumption by at least two orders of magnitude. Furthermore, we show that for many graphs, only a very small fraction of far-apart pairs have to be considered for the hyperbolicity computation, explaining this drastic reduction of memory. As iterating over far-apart pairs in decreasing order without storing them explicitly is a very general tool, we believe that our approach might also be relevant to other problems.