Deterministic $k$-Median Clustering in Near-Optimal Time
Abstract
The metric -median problem is a textbook clustering problem. As input, we are given a metric space of size and an integer , and our task is to find a subset of at most `centers' that minimizes the total distance from each point in to its nearest center in . Mettu and Plaxton [UAI'02] gave a randomized algorithm for -median that computes a -approximation in time. They also showed that any algorithm for this problem with a bounded approximation ratio must have a running time of . Thus, the running time of their algorithm is optimal up to polylogarithmic factors. For deterministic -median, Guha et al.~[FOCS'00] gave an algorithm that computes a -approximation in time, where the degree of the polynomial in the approximation is unspecified. To the best of our knowledge, this remains the state-of-the-art approximation of any deterministic -median algorithm with this running time. This leads us to the following natural question: What is the best approximation of a deterministic -median algorithm with near-optimal running time? We make progress in answering this question by giving a deterministic algorithm that computes a -approximation in time. We also provide a lower bound showing that any deterministic algorithm with this running time must have an approximation ratio of , establishing a gap between the randomized and deterministic settings for -median.
Cite
@article{arxiv.2504.15115,
title = {Deterministic $k$-Median Clustering in Near-Optimal Time},
author = {Martín Costa and Ermiya Farokhnejad},
journal= {arXiv preprint arXiv:2504.15115},
year = {2026}
}
Comments
Accepted at ICALP 2025, Results for $k$-Means are updated in the new version