English

Vertex Sparsifiers for Hyperedge Connectivity

Data Structures and Algorithms 2022-07-13 v2

Abstract

Recently, Chalermsook et al. [SODA'21(arXiv:2007.07862)] introduces a notion of vertex sparsifiers for cc-edge connectivity, which has found applications in parameterized algorithms for network design and also led to exciting dynamic algorithms for cc-edge st-connectivity [Jin and Sun FOCS'21(arXiv:2004.07650)]. We study a natural extension called vertex sparsifiers for cc-hyperedge connectivity and construct a sparsifier whose size matches the state-of-the-art for normal graphs. More specifically, we show that, given a hypergraph G=(V,E)G=(V,E) with nn vertices and mm hyperedges with kk terminal vertices and a parameter cc, there exists a hypergraph HH containing only O(kc3)O(kc^{3}) hyperedges that preserves all minimum cuts (up to value cc) between all subset of terminals. This matches the best bound of O(kc3)O(kc^{3}) edges for normal graphs by [Liu'20(arXiv:2011.15101)]. Moreover, HH can be constructed in almost-linear O(p1+o(1)+n(rclogn)O(rc)logm)O(p^{1+o(1)} + n(rc\log n)^{O(rc)}\log m) time where r=maxeEer=\max_{e\in E}|e| is the rank of GG and p=eEep=\sum_{e\in E}|e| is the total size of GG, or in poly(m,n)\text{poly}(m, n) time if we slightly relax the size to O(kc3log1.5(kc))O(kc^{3}\log^{1.5}(kc)) hyperedges.

Keywords

Cite

@article{arxiv.2207.04115,
  title  = {Vertex Sparsifiers for Hyperedge Connectivity},
  author = {Han Jiang and Shang-En Huang and Thatchaphol Saranurak and Tian Zhang},
  journal= {arXiv preprint arXiv:2207.04115},
  year   = {2022}
}

Comments

submitted to ESA 2022

R2 v1 2026-06-25T00:46:16.667Z