EUROPT 2025
Abstract Submission

295. Mixed-integer linearity in nonlinear optimization

Invited abstract in session MB-6: Nonsmooth optimization: from continuous to discrete Part I, stream Nonsmooth and nonconvex optimization.

Monday, 10:30-12:30
Room: B100/7013

Authors (first author is the speaker)

1. Alberto De Marchi
University of the Bundeswehr Munich

Abstract

Bringing together nonlinear optimization with polyhedral and integrality constraints enables versatile modelling, but poses significant computational challenges. To address these problems, the talk presents an algorithm that is inspired by proximal-gradient methods but replaces the proximal operator with calls to a generic mixed-integer linear solver. The technique computes feasible iterates based on sequential mixed-integer linearization with trust region safeguard. Convergence to critical, possibly suboptimal, feasible points is established for arbitrary starting points. The theoretical and algorithmic developments are corroborated by numerical applications.

Keywords

Status: accepted


Back to the list of papers