English

Breaking O(nr) for Matroid Intersection

Data Structures and Algorithms 2021-05-13 v1

Abstract

We present algorithms that break the O~(nr)\tilde O(nr)-independence-query bound for the Matroid Intersection problem for the full range of rr; where nn is the size of the ground set and rnr\leq n is the size of the largest common independent set. The O~(nr)\tilde O(nr) bound was due to the efficient implementations [CLSSW FOCS'19; Nguyen 2019] of the classic algorithm of Cunningham [SICOMP'86]. It was recently broken for large rr (r=ω(n)r=\omega(\sqrt{n})), first by the O~(n1.5/ϵ1.5)\tilde O(n^{1.5}/\epsilon^{1.5})-query (1ϵ)(1-\epsilon)-approximation algorithm of CLSSW [FOCS'19], and subsequently by the O~(n6/5r3/5)\tilde O(n^{6/5}r^{3/5})-query exact algorithm of BvdBMN [STOC'21]. No algorithm, even an approximation one, was known to break the O~(nr)\tilde O(nr) bound for the full range of rr. We present an O~(nr/ϵ)\tilde O(n\sqrt{r}/\epsilon)-query (1ϵ)(1-\epsilon)-approximation algorithm and an O~(nr3/4)\tilde O(nr^{3/4})-query exact algorithm. Our algorithms improve the O~(nr)\tilde O(nr) bound and also the bounds by CLSSW and BvdBMN for the full range of rr.

Keywords

Cite

@article{arxiv.2105.05673,
  title  = {Breaking O(nr) for Matroid Intersection},
  author = {Joakim Blikstad},
  journal= {arXiv preprint arXiv:2105.05673},
  year   = {2021}
}

Comments

17 pages; also at ICALP 2021

R2 v1 2026-06-24T02:02:22.502Z