Breaking O(nr) for Matroid Intersection

Abstract: We present algorithms that break the $\tilde O(nr)$-independence-query bound for the Matroid Intersection problem for the full range of $r$; where $n$ is the size of the ground set and $r\leq n$ is the size of the largest common independent set. The $\tilde O(nr)$ 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 $r$ ($r=\omega(\sqrt{n})$), first by the $\tilde O(n^{1.5}/\epsilon^{1.5})$-query $(1-\epsilon)$-approximation algorithm of CLSSW [FOCS'19], and subsequently by the $\tilde O(n^{6/5}r^{3/5})$-query exact algorithm of BvdBMN [STOC'21]. No algorithm, even an approximation one, was known to break the $\tilde O(nr)$ bound for the full range of $r$. We present an $\tilde O(n\sqrt{r}/\epsilon)$-query $(1-\epsilon)$-approximation algorithm and an $\tilde O(nr^{3/4})$-query exact algorithm. Our algorithms improve the $\tilde O(nr)$ bound and also the bounds by CLSSW and BvdBMN for the full range of $r$.