• Media type: E-Article
  • Title: Forgetfulness Can Make You Faster: An O * (8.097 k ) -time Algorithm for Weighted 3-set k -packing
  • Contributor: Zehavi, Meirav
  • Published: Association for Computing Machinery (ACM), 2023
  • Published in: ACM Transactions on Computation Theory, 15 (2023) 3-4, Seite 1-13
  • Language: English
  • DOI: 10.1145/3599722
  • ISSN: 1942-3454; 1942-3462
  • Keywords: Computational Theory and Mathematics ; Theoretical Computer Science
  • Origination:
  • Footnote:
  • Description: In this article, we study the classic Weighted 3-Set k -Packing problem: given a universe U , a family \( {\mathcal {S}}\) of subsets of size 3 of U , a weight function \(w : {\mathcal {S}} \rightarrow \mathbb {R}\) , \(W \in \mathbb {R}\) , and a parameter \(k \in \mathbb {N}\) , the objective is to decide if there is a subfamily \( {\mathcal {S}}\) ′ ⊆ \( {\mathcal {S}}\) of k disjoint sets and total weight at least W . We present a deterministic parameterized algorithm for this problem that runs in time O * (8.097 k ), where O * hides factors polynomial in the input size. This substantially improves upon the previously best deterministic algorithm for Weighted 3-Set k -Packing , which runs in time O * (12.155 k ) SIDMA [ 18 ], and was also the best deterministic algorithm for the unweighted version of this problem. Our algorithm is based on a novel application of the method of representative sets that might be of independent interest.