Tournament Ranking: Duality and Efficiency

The feedback arc set problem on tournaments arises in a rich variety of applications, and has been studied extensively in several research fields over the past six decades. It is well known that this problem is $NP$-hard and admits a polynomial-time approximation scheme (PTAS) in general. A tournament $T=(V, A)$ is called cycle Mengerian (CM) if, for every nonnegative integral weight function defined on $A$, the minimum total weight of a feedback arc set is equal to the maximum size of a cycle packing. In 2020 Chen et al. obtained a structural characterization of all CM tournaments; however, their proof is not algorithmic in nature. In this paper we present combinatorial polynomial-time algorithms for finding both minimum feedback arc sets and maximum cycle packings in arc-weighted CM tournaments.