Reverse Greedy is Bad for k-Center

TL;DR

This paper analyzes the reverse greedy algorithm for the k-center problem, showing it has a worst-case approximation ratio between 2k-2 and 2k, indicating limitations of this approach.

Contribution

The paper provides new bounds on the approximation ratio of the reverse greedy algorithm for k-center, highlighting its inefficiency in worst-case scenarios.

Findings

Reverse greedy has approximation ratio between 2k-2 and 2k.

The analysis improves understanding of the algorithm's limitations.

The results guide future algorithm design for k-center.

Abstract

We show the reverse greedy algorithm is between a $(2k-2)$- and a $2k$-approximation for $k$-center.