Submodular maximization and its generalization through an intersection cut lens

Research output: Contribution to journalArticlepeer-review

Abstract

We study a mixed-integer set S:={(x,t)∈{0,1}n×R:f(x)≥t} arising in the submodular maximization problem, where f is a submodular function defined over {0,1}n. We use intersection cuts to tighten a polyhedral outer approximation of S. We construct a continuous extension F¯f of f, which is convex and defined over the entire space Rn. We show that the epigraph epi(F¯f) of F¯f is an S-free set, and characterize maximal S-free sets containing epi(F¯f). We propose a hybrid discrete Newton algorithm to compute an intersection cut efficiently and exactly. Our results are generalized to the hypograph or the superlevel set of a submodular-supermodular function over the Boolean hypercube, which is a model for discrete nonconvexity. A consequence of these results is intersection cuts for Boolean multilinear constraints. We evaluate our techniques on max cut, pseudo Boolean maximization, and Bayesian D-optimal design problems within a MIP solver.

Original languageEnglish
JournalMathematical Programming
DOIs
Publication statusAccepted/In press - 1 Jan 2024

Keywords

  • 90C10
  • 90C26
  • 90C57
  • Boolean multilinear functions
  • D-optimal design
  • Intersection cuts
  • MINLP
  • Submodular maximization
  • Submodular-supermodular functions

Fingerprint

Dive into the research topics of 'Submodular maximization and its generalization through an intersection cut lens'. Together they form a unique fingerprint.

Cite this