Improved Girth Approximation and Roundtrip Spanners

In this paper we provide improved algorithms for approximating the girth and producing roundtrip spanners of $n$-node $m$-edge directed graphs with non-negative edge lengths. First, for any integer $k \ge 1$, we provide a deterministic $\tilde{O}(m^{1+ 1/k})$ time algorithm which computes a $O(k\log\log n)$ multiplicative approximation of the girth and a $O(k\log\log n)$ multiplicative roundtrip spanner with $\tilde{O}(n^{1+1/k})$ edges. Second, we provide a randomized $\tilde{O}(m\sqrt{n})$ time algorithm that with high probability computes a 3-multiplicative approximation to the girth. Third, we show how to combine these algorithms to obtain for any integer $k \ge 1$ a randomized algorithm which in $\tilde{O}(m^{1+ 1/k})$ time computes a $O(k\log k)$ multiplicative approximation of the girth and $O(k \log k)$ multiplicative roundtrip spanner with high probability. The previous fastest algorithms for these problems either ran in All-Pairs Shortest Paths (APSP) time, i.e. $\tilde{O}(mn)$, or were due Pachocki et al (SODA 2018) which provided a randomized algorithm that for any integer $k \ge 1$ in time $\tilde{O}(m^{1+1/k})$ computed with high probability a $O(k\log n)$ multiplicative approximation of the girth and a $O(k\log n)$ multiplicative roundtrip spanners with $\tilde{O}(n^{1+1/k})$ edges. Our first algorithm removes the need for randomness and improves the approximation factor in Pachocki et al (SODA 2018), our second is constitutes the first sub-APSP-time algorithm for approximating the girth to constant accuracy with high probability, and our third is the first time versus quality trade-offs for obtaining constant approximations.