Improved Bounding Techniques for Nonconvex MINLP in SHOT

Research output: Chapter in Book/Conference proceedingPublished conference proceedingScientificpeer-review

7 Downloads (Pure)

Abstract

The Supporting Hyperplane Optimization Toolkit (SHOT) is an optimization solver based on polyhedral outer approximation, primarily designed for convex mixed-integer nonlinear programming (MINLP) problems. Nonconvexities are handled through reformulations and heuristic techniques, and thus, the solver is positioned somewhere between a global and heuristic solver for nonconvex problems. This paper introduces two heuristic enhancements to improve SHOT’s nonconvex capabilities, primarily by reducing the optimality gap. The first technique involves solving a convex bounding problem to improve the dual bound, and the second updates an existing primal objective cut heuristic within SHOT to improve the primal bound.
Original languageEnglish
Title of host publicationProceedings of the Stockholm Global Optimization Workshop, STOGO 2025
Pages133-136
Number of pages4
Publication statusPublished - 2 Sept 2025
MoE publication typeA4 Article in a conference publication
EventStockholm Global Optimization Workshop 2025 -
Duration: 2 Sept 20255 Sept 2025
https://sites.google.com/view/stogo25/

Workshop

WorkshopStockholm Global Optimization Workshop 2025
Abbreviated titleSTOGO
Period02/09/2505/09/25
Internet address

Fingerprint

Dive into the research topics of 'Improved Bounding Techniques for Nonconvex MINLP in SHOT'. Together they form a unique fingerprint.

Cite this