English

Fully Dynamic $k$-Clustering in $\tilde O(k)$ Update Time

Data Structures and Algorithms 2023-10-27 v1

Abstract

We present a O(1)O(1)-approximate fully dynamic algorithm for the kk-median and kk-means problems on metric spaces with amortized update time O~(k)\tilde O(k) and worst-case query time O~(k2)\tilde O(k^2). We complement our theoretical analysis with the first in-depth experimental study for the dynamic kk-median problem on general metrics, focusing on comparing our dynamic algorithm to the current state-of-the-art by Henzinger and Kale [ESA'20]. Finally, we also provide a lower bound for dynamic kk-median which shows that any O(1)O(1)-approximate algorithm with O~(poly(k))\tilde O(\text{poly}(k)) query time must have Ω~(k)\tilde \Omega(k) amortized update time, even in the incremental setting.

Keywords

Cite

@article{arxiv.2310.17420,
  title  = {Fully Dynamic $k$-Clustering in $\tilde O(k)$ Update Time},
  author = {Sayan Bhattacharya and Martín Costa and Silvio Lattanzi and Nikos Parotsidis},
  journal= {arXiv preprint arXiv:2310.17420},
  year   = {2023}
}

Comments

Accepted at NeurIPS 2023

R2 v1 2026-06-28T13:02:48.308Z