PulseExploreJournal ClubDebatesTrendingResearchersJournals
Instagram
HomeExploreJournal ClubTrending
Synapse
⌘+K
Synapse
March 6, 20241 citationsOpen Access

Parameterized Algorithms for Balanced Cluster Edge Modification Problems

View Full Paper
JMJayakrishnan MadathilKMKitty Meeks

Key Points

Key points are not available for this paper at this time.

Abstract

We introduce Cluster Edge Modification problems with constraints on the size of the clusters and study their complexity. A graph G is a cluster graph if every connected component of G is a clique. In a typical Cluster Edge Modification problem such as the widely studied Cluster Editing, we are given a graph G and a non-negative integer k as input, and we have to decide if we can turn G into a cluster graph by way of at most k edge modifications -- that is, by adding or deleting edges. In this paper, we study the parameterized complexity of such problems, but with an additional constraint: The size difference between any two connected components of the resulting cluster graph should not exceed a given threshold. Depending on which modifications are permissible -- only adding edges, only deleting edges, both adding and deleting edges -- we have three different computational problems. We show that all three problems, when parameterized by k, admit single-exponential time FPT algorithms and polynomial kernels. Our problems may be thought of as the size-constrained or balanced counterparts of the typical Cluster Edge Modification problems, similar to the well-studied size-constrained or balanced counterparts of other clustering problems such as k-Means Clustering.

Ask AI
Helpful
Bookmark
Share
View Full Paper

Cite This Study

Madathil et al. (2024) studied this question.

synapsesocial.com/papers/68e757abb6db6435876cf627https://doi.org/10.48550/arxiv.2403.03830
Ask AI
Helpful
Bookmark
Share
View Full Paper