ON THE RELAXATION COMPLEXITY OF NONCONVEX QUADRATIC GLOBAL OPTIMIZATION

  • Tongli Zhang
  • , Yong Xia*
  • *Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

Abstract

We study the relaxation complexity for nonconvex quadratic global optimization, which is defined as the number of convex relaxation subproblems to be solved. The relaxation complexity for quadratic programming with fixed nonconvex-rank is known to be a polynomial function of the dimension. In this paper, we show that the relaxation complexity for nonconvex quadratic optimization with convex quadratic constraints may not depend on the dimension, as long as the objective function has a fixed nonconvex rank.

Original languageEnglish
Article number19
JournalCommunications in Optimization Theory
Volume2024
DOIs
StatePublished - 2024

Keywords

  • Dikin ellipsoid
  • Global optimization
  • Löwner-John ellipsoid
  • Quadratic constrained quadratic optimization

Fingerprint

Dive into the research topics of 'ON THE RELAXATION COMPLEXITY OF NONCONVEX QUADRATIC GLOBAL OPTIMIZATION'. Together they form a unique fingerprint.

Cite this