Non-monotone submodular maximization under matroid and knapsack constraints

TL;DR

This paper introduces the first constant-factor approximation algorithms for maximizing non-monotone submodular functions under multiple matroid and knapsack constraints, advancing the theoretical understanding of constrained submodular maximization.

Contribution

It provides novel approximation algorithms for non-monotone submodular maximization with multiple constraints, including matroids and knapsacks, improving previous bounds and covering non-monotone cases.

Findings

First constant-factor approximation for non-monotone submodular maximization with multiple constraints.

Achieves a $({1ackslash}{k+2+{1ackslash}{k}+\epsilon})$-approximation for $k$ matroids.

Provides a $({1ackslash}5-\epsilon)$-approximation for $k$ knapsack constraints.

Abstract

Submodular function maximization is a central problem in combinatorial optimization, generalizing many important problems including Max Cut in directed/undirected graphs and in hypergraphs, certain constraint satisfaction problems, maximum entropy sampling, and maximum facility location problems. Unlike submodular minimization, submodular maximization is NP-hard. For the problem of maximizing a non-monotone submodular function, Feige, Mirrokni, and Vondr\'ak recently developed a $2\over 5$-approximation algorithm \cite{FMV07}, however, their algorithms do not handle side constraints.} In this paper, we give the first constant-factor approximation algorithm for maximizing any non-negative submodular function subject to multiple matroid or knapsack constraints. We emphasize that our results are for {\em non-monotone} submodular functions. In particular, for any constant $k$, we present a $({1\over k+2+{1\over k}+\epsilon})$-approximation for the submodular maximization problem under $k$ matroid constraints, and a $({1\over 5}-\epsilon)$-approximation algorithm for this problem subject to $k$ knapsack constraints ($\epsilon>0$ is any constant). We improve the approximation guarantee of our algorithm to ${1\over k+1+{1\over k-1}+\epsilon}$ for $k\ge 2$ partition matroid constraints. This idea also gives a $({1\over k+\epsilon})$-approximation for maximizing a {\em monotone} submodular function subject to $k\ge 2$ partition matroids, which improves over the previously best known guarantee of $\frac{1}{k+1}$.