English

An optimal local approximation algorithm for max-min linear programs

Distributed, Parallel, and Cluster Computing 2012-05-15 v1

Abstract

We present a local algorithm (constant-time distributed algorithm) for approximating max-min LPs. The objective is to maximise ω\omega subject to Ax1Ax \le 1, Cxω1Cx \ge \omega 1, and x0x \ge 0 for nonnegative matrices AA and CC. The approximation ratio of our algorithm is the best possible for any local algorithm; there is a matching unconditional lower bound.

Keywords

Cite

@article{arxiv.0809.1489,
  title  = {An optimal local approximation algorithm for max-min linear programs},
  author = {Patrik Floréen and Joel Kaasinen and Petteri Kaski and Jukka Suomela},
  journal= {arXiv preprint arXiv:0809.1489},
  year   = {2012}
}

Comments

16 pages, 3 figures

R2 v1 2026-06-21T11:18:13.048Z