English

Faster Matroid Intersection

Data Structures and Algorithms 2019-11-26 v1 Discrete Mathematics

Abstract

In this paper we consider the classic matroid intersection problem: given two matroids \M1=(V,\I1)\M_{1}=(V,\I_{1}) and \M2=(V,\I2)\M_{2}=(V,\I_{2}) defined over a common ground set VV, compute a set S\I1\I2S\in\I_{1}\cap\I_{2} of largest possible cardinality, denoted by rr. We consider this problem both in the setting where each \Mi\M_{i} is accessed through an independence oracle, i.e. a routine which returns whether or not a set S\IiS\in\I_{i} in \indep\indep time, and the setting where each \Mi\M_{i} is accessed through a rank oracle, i.e. a routine which returns the size of the largest independent subset of SS in \Mi\M_{i} in \rank\rank time. In each setting we provide faster exact and approximate algorithms. Given an independence oracle, we provide an exact O(nrlogr\indep)O(nr\log r \indep) time algorithm. This improves upon the running time of O(nr1.5\indep)O(nr^{1.5} \indep) due to Cunningham in 1986 and O~(n2\indep+n3)\tilde{O}(n^{2} \indep+n^{3}) due to Lee, Sidford, and Wong in 2015. We also provide two algorithms which compute a (1ϵ)(1-\epsilon)-approximate solution to matroid intersection running in times O~(n1.5/\eps1.5\indep)\tilde{O}(n^{1.5}/\eps^{1.5} \indep) and O~((n2r1ϵ2+r1.5ϵ4.5)\indep)\tilde{O}((n^{2}r^{-1}\epsilon^{-2}+r^{1.5}\epsilon^{-4.5}) \indep), respectively. These results improve upon the O(nr/\eps\indep)O(nr/\eps \indep)-time algorithm of Cunningham as noted recently by Chekuri and Quanrud. Given a rank oracle, we provide algorithms with even better dependence on nn and rr. We provide an O(nrlogn\rank)O(n\sqrt{r}\log n \rank)-time exact algorithm and an O(nϵ1logn\rank)O(n\epsilon^{-1}\log n \rank)-time algorithm which obtains a (1\eps)(1-\eps)-approximation to the matroid intersection problem. The former result improves over the O~(nr\rankt+n3)\tilde{O}(nr \rankt+n^{3})-time algorithm by Lee, Sidford, and Wong. The rank oracle is of particular interest as the matroid intersection problem with this oracle is a special case of the submodular function minimization problem with an evaluation oracle.

Keywords

Cite

@article{arxiv.1911.10765,
  title  = {Faster Matroid Intersection},
  author = {Deeparnab Chakrabarty and Yin Tat Lee and Aaron Sidford and Sahil Singla and Sam Chiu-wai Wong},
  journal= {arXiv preprint arXiv:1911.10765},
  year   = {2019}
}

Comments

38 pages. Preliminary version appeared in FOCS 2019