Nonmonotone submodular maximization via a structural continuous greedy algorithm

Moran Feldman, Joseph Naor, Roy Schwartz

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

Consider a suboptimal solution S for a maximization problem. Suppose S's value is small compared to an optimal solution OPT to the problem, yet S is structurally similar to OPT. A natural question in this setting is whether there is a way of improving S based solely on this information. In this paper we introduce the Structural Continuous Greedy Algorithm, answering this question affirmatively in the setting of the Nonmonotone Submodular Maximization Problem. We improve on the best approximation factor known for this problem. In the Nonmonotone Submodular Maximization Problem we are given a non-negative submodular function f, and the objective is to find a subset maximizing f. Our method yields an 0.42-approximation for this problem, improving on the current best approximation factor of 0.41 given by Gharan and Vondrák [5]. On the other hand, Feige et al. [4] showed a lower bound of 0.5 for this problem.

Original languageEnglish
Title of host publicationAutomata, Languages and Programming - 38th International Colloquium, ICALP 2011, Proceedings
Pages342-353
Number of pages12
EditionPART 1
DOIs
StatePublished - 2011
Externally publishedYes
Event38th International Colloquium on Automata, Languages and Programming, ICALP 2011 - Zurich, Switzerland
Duration: 4 Jul 20118 Jul 2011

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
NumberPART 1
Volume6755 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference38th International Colloquium on Automata, Languages and Programming, ICALP 2011
Country/TerritorySwitzerland
CityZurich
Period4/07/118/07/11

Fingerprint

Dive into the research topics of 'Nonmonotone submodular maximization via a structural continuous greedy algorithm'. Together they form a unique fingerprint.

Cite this