English

Approximation algorithms for non-sequential star packing problems

Data Structures and Algorithms 2024-11-19 v1 Discrete Mathematics

Abstract

For a positive integer k1k \ge 1, a kk-star (k+k^+-star, kk^--star, respectively) is a connected graph containing a degree-\ell vertex and \ell degree-11 vertices, where =k\ell = k (k\ell \ge k, 1k1 \le \ell \le k, respectively). The k+k^+-star packing problem is to cover as many vertices of an input graph GG as possible using vertex-disjoint k+k^+-stars in GG; and given k>t1k > t \ge 1, the k/tk^-/t-star packing problem is to cover as many vertices of GG as possible using vertex-disjoint kk^--stars but no tt-stars in GG. Both problems are NP-hard for any fixed k2k \ge 2. We present a (1+k22k+1)(1 + \frac {k^2}{2k+1})- and a 32\frac 32-approximation algorithms for the k+k^+-star packing problem when k3k \ge 3 and k=2k = 2, respectively, and a (1+1t+1+1/k)(1 + \frac 1{t + 1 + 1/k})-approximation algorithm for the k/tk^-/t-star packing problem when k>t2k > t \ge 2. They are all local search algorithms and they improve the best known approximation algorithms for the problems, respectively.

Keywords

Cite

@article{arxiv.2411.11136,
  title  = {Approximation algorithms for non-sequential star packing problems},
  author = {Mengyuan Hu and An Zhang and Yong Chen and Mingyang Gong and Guohui Lin},
  journal= {arXiv preprint arXiv:2411.11136},
  year   = {2024}
}

Comments

Accepted for presentation in WALCOM 2025