Computing treedepth in polynomial space and linear fpt time

TL;DR

This paper introduces an efficient algorithm for computing the treedepth of a graph within polynomial space and linear fixed-parameter tractable time, improving upon previous exponential space methods.

Contribution

The authors present a new algorithm that decides treedepth in polynomial space with linear FPT time, and enhance it with randomized techniques for better efficiency.

Findings

Algorithm runs in $2^{O(d^2)}\cdot n^{O(1)}$ time and polynomial space.

Randomization reduces time to $2^{O(d^2)}\cdot n$ and space to $d^{O(1)}\cdot n$.

Improves upon previous exponential space algorithms for treedepth computation.

Abstract

The treedepth of a graph $G$ is the least possible depth of an elimination forest of $G$: a rooted forest on the same vertex set where every pair of vertices adjacent in $G$ is bound by the ancestor/descendant relation. We propose an algorithm that given a graph $G$ and an integer $d$, either finds an elimination forest of $G$ of depth at most $d$ or concludes that no such forest exists; thus the algorithm decides whether the treedepth of $G$ is at most $d$. The running time is $2^{O(d^2)}\cdot n^{O(1)}$ and the space usage is polynomial in $n$. Further, by allowing randomization, the time and space complexities can be improved to $2^{O(d^2)}\cdot n$ and $d^{O(1)}\cdot n$, respectively. This improves upon the algorithm of Reidl et al. [ICALP 2014], which also has time complexity $2^{O(d^2)}\cdot n$, but uses exponential space.