English

Approximation Algorithms for Continuous Clustering and Facility Location Problems

Data Structures and Algorithms 2022-09-05 v3

Abstract

We consider the approximability of center-based clustering problems where the points to be clustered lie in a metric space, and no candidate centers are specified. We call such problems "continuous", to distinguish from "discrete" clustering where candidate centers are specified. For many objectives, one can reduce the continuous case to the discrete case, and use an α\alpha-approximation algorithm for the discrete case to get a βα\beta\alpha-approximation for the continuous case, where β\beta depends on the objective: e.g. for kk-median, β=2\beta = 2, and for kk-means, β=4\beta = 4. Our motivating question is whether this gap of β\beta is inherent, or are there better algorithms for continuous clustering than simply reducing to the discrete case? In a recent SODA 2021 paper, Cohen-Addad, Karthik, and Lee prove a factor-22 and a factor-44 hardness, respectively, for continuous kk-median and kk-means, even when the number of centers kk is a constant. The discrete case for a constant kk is exactly solvable in polytime, so the β\beta loss seems unavoidable in some regimes. In this paper, we approach continuous clustering via the round-or-cut framework. For four continuous clustering problems, we outperform the reduction to the discrete case. Notably, for the problem λ\lambda-UFL, where β=2\beta = 2 and the discrete case has a hardness of 1.271.27, we obtain an approximation ratio of 2.32<2×1.272.32 < 2 \times 1.27 for the continuous case. Also, for continuous kk-means, where the best known approximation ratio for the discrete case is 99, we obtain an approximation ratio of 32<4×932 < 4 \times 9. The key challenge is that most algorithms for discrete clustering, including the state of the art, depend on linear programs that become infinite-sized in the continuous case. To overcome this, we design new linear programs for the continuous case which are amenable to the round-or-cut framework.

Keywords

Cite

@article{arxiv.2206.15105,
  title  = {Approximation Algorithms for Continuous Clustering and Facility Location Problems},
  author = {Deeparnab Chakrabarty and Maryam Negahbani and Ankita Sarkar},
  journal= {arXiv preprint arXiv:2206.15105},
  year   = {2022}
}

Comments

24 pages, 0 figures. Full version of ESA 2022 paper https://drops.dagstuhl.de/opus/volltexte/2022/16971 . This version adds a link to the conference version and fixes minor formatting issues