English

Simple and Optimal Online Contention Resolution Schemes for $k$-Uniform Matroids

Data Structures and Algorithms 2023-11-28 v2 Discrete Mathematics Computer Science and Game Theory

Abstract

We provide a simple (1O(1k))(1-O(\frac{1}{\sqrt{k}}))-selectable Online Contention Resolution Scheme for kk-uniform matroids against a fixed-order adversary. If AiA_i and GiG_i denote the set of selected elements and the set of realized active elements among the first ii (respectively), our algorithm selects with probability 11k1-\frac{1}{\sqrt{k}} any active element ii such that Ai1+1(11k)E[Gi]+k|A_{i-1}| + 1 \leq (1-\frac{1}{\sqrt{k}})\cdot \mathbb{E}[|G_i|]+\sqrt{k}. This implies a (1O(1k))(1-O(\frac{1}{\sqrt{k}})) prophet inequality against fixed-order adversaries for kk-uniform matroids that is considerably simpler than previous algorithms [Ala14, AKW14, JMZ22]. We also prove that no OCRS can be (1Ω(logkk))(1-\Omega(\sqrt{\frac{\log k}{k}}))-selectable for kk-uniform matroids against an almighty adversary. This guarantee is matched by the (known) simple greedy algorithm that accepts every active element with probability 1Θ(logkk)1-\Theta(\sqrt{\frac{\log k}{k}}) [HKS07].

Keywords

Cite

@article{arxiv.2309.10078,
  title  = {Simple and Optimal Online Contention Resolution Schemes for $k$-Uniform Matroids},
  author = {Atanas Dinev and S. Matthew Weinberg},
  journal= {arXiv preprint arXiv:2309.10078},
  year   = {2023}
}

Comments

26 pages, 15th Innovations in Theoretical Computer Science (ITCS 2024)