Faster Exact and Parameterized Algorithm for Feedback Vertex Set in Tournaments

TL;DR

This paper introduces a faster algorithm for the Feedback Vertex Set problem in tournaments, improving both parameterized and exact exponential algorithms, and proves a new graph partitioning theorem.

Contribution

It presents a novel algorithm with improved running times for Feedback Vertex Set in tournaments and a new graph partitioning theorem.

Findings

Algorithm runs in O(1.618^k + n^{O(1)}) time

Algorithm runs in O(1.46^n) time for exact solutions

New partitioning theorem for undirected graphs

Abstract

A tournament is a directed graph T such that every pair of vertices are connected by an arc. A feedback vertex set is a set S of vertices in T such that T - S is acyclic. In this article we consider the Feedback Vertex Set problem in tournaments. Here input is a tournament T and integer k, and the task is to determine whether T has a feedback vertex set of size at most k. We give a new algorithm for Feedback Vertex Set in Tournaments. The running time of our algorithm is upper bounded by O(1.618^k + n^{O(1)}) and by O(1.46^n). Thus our algorithm simultaneously improves over the fastest known parameterized algorithm for the problem by Dom et al. running in time O(2^kk^{O(1)} + n^{O(1)}), and the fastest known exact exponential time algorithm by Gaspers and Mnich with running time O(1.674^n). On the way to prove our main result we prove a new partitioning theorem for undirected graphs. In particular we show that the vertices of any undirected m-edge graph of maximum degree d can be colored white or black in such a way that for each of the two colors, the number of edges with both endpoints of that color is between m/4-d/2 and m/4+d/4.