English

Cardinality constrained submodular maximization for random streams

Data Structures and Algorithms 2021-11-16 v1

Abstract

We consider the problem of maximizing submodular functions in single-pass streaming and secretaries-with-shortlists models, both with random arrival order. For cardinality constrained monotone functions, Agrawal, Shadravan, and Stein gave a single-pass (11/eε)(1-1/e-\varepsilon)-approximation algorithm using only linear memory, but their exponential dependence on ε\varepsilon makes it impractical even for ε=0.1\varepsilon=0.1. We simplify both the algorithm and the analysis, obtaining an exponential improvement in the ε\varepsilon-dependence (in particular, O(k/ε)O(k/\varepsilon) memory). Extending these techniques, we also give a simple (1/eε)(1/e-\varepsilon)-approximation for non-monotone functions in O(k/ε)O(k/\varepsilon) memory. For the monotone case, we also give a corresponding unconditional hardness barrier of 11/e+ε1-1/e+\varepsilon for single-pass algorithms in randomly ordered streams, even assuming unlimited computation. Finally, we show that the algorithms are simple to implement and work well on real world datasets.

Keywords

Cite

@article{arxiv.2111.07217,
  title  = {Cardinality constrained submodular maximization for random streams},
  author = {Paul Liu and Aviad Rubinstein and Jan Vondrak and Junyao Zhao},
  journal= {arXiv preprint arXiv:2111.07217},
  year   = {2021}
}

Comments

To appear in NeurIPS 2021

R2 v1 2026-06-24T07:37:30.117Z