English

On the $d$-Claw Vertex Deletion Problem

Discrete Mathematics 2022-03-15 v1 Data Structures and Algorithms

Abstract

Let dd-claw (or dd-star) stand for K1,dK_{1,d}, the complete bipartite graph with 1 and d1d\ge 1 vertices on each part. The dd-claw vertex deletion problem, dd-CLAW-VD, asks for a given graph GG and an integer kk if one can delete at most kk vertices from GG such that the resulting graph has no dd-claw as an induced subgraph. Thus, 1-CLAW-VD and 2-CLAW-VD are just the famous VERTEX COVER problem and the CLUSTER VERTEX DELETION problem, respectively. In this paper, we strengthen a hardness result in [M. Yannakakis, Node-Deletion Problems on Bipartite Graphs, SIAM J. Comput. (1981)], by showing that CLUSTER VERTEX DELETION remains NP-complete when restricted to bipartite graphs of maximum degree 3. Moreover, for every d3d\ge 3, we show that dd-CLAW-VD is NP-complete even when restricted to bipartite graphs of maximum degree dd. These hardness results are optimal with respect to degree constraint. By extending the hardness result in [F. Bonomo-Braberman et al., Linear-Time Algorithms for Eliminating Claws in Graphs, COCOON 2020], we show that, for every d3d\ge 3, dd-CLAW-VD is NP-complete even when restricted to split graphs without (d+1)(d+1)-claws, and split graphs of diameter 2. On the positive side, we prove that dd-CLAW-VD is polynomially solvable on what we call dd-block graphs, a class properly contains all block graphs. This result extends the polynomial-time algorithm in [Y. Cao et al., Vertex deletion problems on chordal graphs, Theor. Comput. Sci. (2018)] for 2-CLAW-VD on block graphs to dd-CLAW-VD for all d2d\ge 2 and improves the polynomial-time algorithm proposed by F. Bonomo-Brabeman et al. for (unweighted) 3-CLAW-VD on block graphs to 3-block graphs.

Keywords

Cite

@article{arxiv.2203.06766,
  title  = {On the $d$-Claw Vertex Deletion Problem},
  author = {Sun-Yuan Hsieh and Hoang-Oanh Le and Van Bang Le and Sheng-Lung Peng},
  journal= {arXiv preprint arXiv:2203.06766},
  year   = {2022}
}
R2 v1 2026-06-24T10:11:42.111Z