Combining Gradient Information and Primitive Directions for High-Performance Mixed-Integer Optimization

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Lapucci, Matteo, Liuzzi, Giampaolo, Lucidi, Stefano, Mansueto, Pierluigi
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866917206419308544
author Lapucci, Matteo
Liuzzi, Giampaolo
Lucidi, Stefano
Mansueto, Pierluigi
author_facet Lapucci, Matteo
Liuzzi, Giampaolo
Lucidi, Stefano
Mansueto, Pierluigi
contents In this paper we consider bound-constrained mixed-integer optimization problems where the objective function is differentiable w.r.t.\ the continuous variables for every configuration of the integer variables. We mainly suggest to exploit derivative information when possible in these scenarios: concretely, we propose an algorithmic framework that carries out local optimization steps, alternating searches along gradient-based and primitive directions. The algorithm is shown to match the convergence properties of a derivative-free counterpart. Most importantly, the results of thorough computational experiments show that the proposed method clearly outperforms not only the derivative-free approach but also the main alternatives available from the literature to be used in the considered setting, both in terms of efficiency and effectiveness.
format Preprint
id arxiv_https___arxiv_org_abs_2407_14416
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Combining Gradient Information and Primitive Directions for High-Performance Mixed-Integer Optimization
Lapucci, Matteo
Liuzzi, Giampaolo
Lucidi, Stefano
Mansueto, Pierluigi
Optimization and Control
90C11, 90C26, 90C30
In this paper we consider bound-constrained mixed-integer optimization problems where the objective function is differentiable w.r.t.\ the continuous variables for every configuration of the integer variables. We mainly suggest to exploit derivative information when possible in these scenarios: concretely, we propose an algorithmic framework that carries out local optimization steps, alternating searches along gradient-based and primitive directions. The algorithm is shown to match the convergence properties of a derivative-free counterpart. Most importantly, the results of thorough computational experiments show that the proposed method clearly outperforms not only the derivative-free approach but also the main alternatives available from the literature to be used in the considered setting, both in terms of efficiency and effectiveness.
title Combining Gradient Information and Primitive Directions for High-Performance Mixed-Integer Optimization
topic Optimization and Control
90C11, 90C26, 90C30
url https://arxiv.org/abs/2407.14416