Lower bounds for online bin covering-type problems

János Balogh, Leah Epstein, Asaf Levin

Research output: Contribution to journalArticlepeer-review

Abstract

We study several variants of bin covering and design lower bounds on the asymptotic competitive ratio of online algorithms for these problems. Our main result is for vector covering with d≥ 2 dimensions, for which our new lower bound is d+ 1 , improving over the previously known lower bound of d+12, which was proved more than twenty years ago by Alon et al. Two special cases of vector covering are considered as well. We prove an improved lower bound of approximately 2.8228 for the asymptotic competitive ratio of the bin covering with cardinality constraints problem, and we also study vector covering with small components and show tight bounds of d for it. Finally, we define three models for one-dimensional black and white covering and show that no online algorithms of finite asymptotic competitive ratios can be designed for them.

Original languageEnglish
Pages (from-to)487-497
Number of pages11
JournalJournal of Scheduling
Volume22
Issue number4
DOIs
StatePublished - 15 Aug 2019

Bibliographical note

Funding Information:
J. Balogh was supported by the European Union, co-financed by the European Social Fund (project “Integrated program for training new generation of scientists in the fields of computer science”, No. EFOP-3.6.3-VEKOP-16-2017-00002). L. Epstein and A. Levin were partially supported by a Grant from GIF—the German-Israeli Foundation for Scientific Research and Development (Grant No. I-1366-407.6/2016).

Publisher Copyright:
© 2018, Springer Science+Business Media, LLC, part of Springer Nature.

Keywords

  • Bin covering
  • Competitive analysis
  • Lower bounds

ASJC Scopus subject areas

  • Software
  • Engineering (all)
  • Management Science and Operations Research
  • Artificial Intelligence

Fingerprint

Dive into the research topics of 'Lower bounds for online bin covering-type problems'. Together they form a unique fingerprint.

Cite this