Adaptive Submodular Optimization under Matroid Constraints

Abstract

Many important problems in discrete optimization require maximization of a monotonic submodular function subject to matroid constraints. For these problems, a simple greedy algorithm is guaranteed to obtain near-optimal solutions. In this article, we extend this classic result to a general class of adaptive optimization problems under partial observability, where each choice can depend on observations resulting from past choices. Specifically, we prove that a natural adaptive greedy algorithm provides a $1/(p+1)$ approximation for the problem of maximizing an adaptive monotone submodular function subject to $p$ matroid constraints, and more generally over arbitrary $p$-independence systems. We illustrate the usefulness of our result on a complex adaptive match-making application.