English

Maximization of Approximately Submodular Functions

Data Structures and Algorithms 2024-11-19 v1 Computational Complexity

Abstract

We study the problem of maximizing a function that is approximately submodular under a cardinality constraint. Approximate submodularity implicitly appears in a wide range of applications as in many cases errors in evaluation of a submodular function break submodularity. Say that FF is ε\varepsilon-approximately submodular if there exists a submodular function ff such that (1ε)f(S)F(S)(1+ε)f(S)(1-\varepsilon)f(S) \leq F(S)\leq (1+\varepsilon)f(S) for all subsets SS. We are interested in characterizing the query-complexity of maximizing FF subject to a cardinality constraint kk as a function of the error level ε>0\varepsilon>0. We provide both lower and upper bounds: for ε>n1/2\varepsilon>n^{-1/2} we show an exponential query-complexity lower bound. In contrast, when ε<1/k\varepsilon< {1}/{k} or under a stronger bounded curvature assumption, we give constant approximation algorithms.

Keywords

Cite

@article{arxiv.2411.10949,
  title  = {Maximization of Approximately Submodular Functions},
  author = {Thibaut Horel and Yaron Singer},
  journal= {arXiv preprint arXiv:2411.10949},
  year   = {2024}
}

Comments

12 pages

R2 v1 2026-06-28T20:02:32.970Z