English

Separating $k$-Median from the Supplier Version

Data Structures and Algorithms 2024-01-26 v1

Abstract

Given a metric space (V,d)(V, d) along with an integer kk, the kk-Median problem asks to open kk centers CVC \subseteq V to minimize vVd(v,C)\sum_{v \in V} d(v, C), where d(v,C):=mincCd(v,c)d(v, C) := \min_{c \in C} d(v, c). While the best-known approximation ratio of 2.6132.613 holds for the more general supplier version where an additional set FVF \subseteq V is given with the restriction CFC \subseteq F, the best known hardness for these two versions are 1+1/e1.361+1/e \approx 1.36 and 1+2/e1.731+2/e \approx 1.73 respectively, using the same reduction from Max kk-Coverage. We prove the following two results separating them. First, we show a 1.5461.546-parameterized approximation algorithm that runs in time f(k)nO(1)f(k) n^{O(1)}. Since 1+2/e1+2/e is proved to be the optimal approximation ratio for the supplier version in the parameterized setting, this result separates the original kk-Median from the supplier version. Next, we prove a 1.4161.416-hardness for polynomial-time algorithms assuming the Unique Games Conjecture. This is achieved via a new fine-grained hardness of Max-kk-Coverage for small set sizes. Our upper bound and lower bound are derived from almost the same expression, with the only difference coming from the well-known separation between the powers of LP and SDP on (hypergraph) vertex cover.

Keywords

Cite

@article{arxiv.2401.13819,
  title  = {Separating $k$-Median from the Supplier Version},
  author = {Aditya Anand and Euiwoong Lee},
  journal= {arXiv preprint arXiv:2401.13819},
  year   = {2024}
}

Comments

20 pages; To appear at IPCO 2024

R2 v1 2026-06-28T14:26:26.608Z