Minimum Monotone Tree Decomposition of Density Functions Defined on Graphs

TL;DR

This paper investigates the complexity of decomposing density graphs into monotone trees, proving NP-Completeness for various problem variants and offering approximation algorithms, including a 3-approximation for cactus graphs.

Contribution

It proves NP-Completeness for minimum monotone tree decomposition in density graphs and introduces approximation algorithms for specific graph classes.

Findings

Decomposition of density graphs into monotone trees is NP-Complete.

NP-Completeness holds for several problem variants, including the SM-Tree Set.

A 3-approximation algorithm is provided for density cactus graphs.

Abstract

Monotone trees - trees with a function defined on their vertices that decreases the further away from a root node one travels, are a natural model for a process that weakens the further one gets from its source. Given an aggregation of monotone trees, one may wish to reconstruct the individual monotone components. A natural representation of such an aggregation would be a graph. While many methods have been developed for extracting hidden graph structure from datasets, which makes obtaining such an aggregation possible, decomposing such graphs into the original monotone trees is algorithmically challenging. Recently, a polynomial time algorithm has been developed to extract a minimum cardinality collection of monotone trees (M-Tree Set) from a given density tree - but no such algorithm exists for density graphs that may contain cycles. In this work, we prove that extracting such minimum M-Tree Sets of density graphs is NP-Complete. We additionally prove three additional variations of the problem - such as the minimum M-Tree Set such that the intersection between any two monotone trees is either empty or contractible (SM-Tree Set) - are also NP-Complete. We conclude by providing some approximation algorithms, highlighted by a 3-approximation algorithm for computing the minimum SM-Tree Set for density cactus graphs.