A coordinate ascent method for solving semidefinite relaxations of non-convex quadratic integer programs

DOI10.1007/978-3-319-45587-7_10MaRDI QIDQ2835666zbMATH OpenOpenAlexFDO

Authors Christoph Buchheim, Maribel Montenegro, Angelika Wiegele

Publication date 30 November 2016

Published in Lecture Notes in Computer Science (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/1603.02566

zbMATH Keywords

semidefinite programmingcoordinate descent methodnon-convex quadratic integer optimization

Mathematics Subject Classification ID

Quadratic programming (90C20) Semidefinite programming (90C22) Integer programming (90C10)

Abstract: We present a coordinate ascent method for a class of semidefinite programming problems that arise in non-convex quadratic integer optimization. These semidefinite programs are characterized by a small total number of active constraints and by low-rank constraint matrices. We exploit this special structure by solving the dual problem, using a barrier method in combination with a coordinate-wise exact line search. The main ingredient of our algorithm is the computationally cheap update at each iteration and an easy computation of the exact step size. Compared to interior point methods, our approach is much faster in obtaining strong dual bounds. Moreover, no explicit separation and reoptimization is necessary even if the set of primal constraints is large, since in our dual approach this is covered by implicitly considering all primal constraints when selecting the next coordinate.

Recommendations

• SDP-based branch-and-bound for non-convex quadratic integer optimization
• Semidefinite relaxations for non-convex quadratic mixed-integer programming
• A semidefinite programming method for integer convex quadratic minimization
• Semidefinite programming relaxation for nonconvex quadratic programs
• Combining semidefinite and polyhedral relaxations for integer programs

Cites work

• An extension of Karmarkar's projective algorithm for convex quadratic programming
• CSDP, A C library for semidefinite programming
• Improved approximation algorithms for maximum cut and satisfiability problems using semidefinite programming
• On the Shannon capacity of a graph
• Quadratic programming with one negative eigenvalue is NP-hard
• Relaxing nonconvex quadratic functions by multiple adaptive diagonal perturbations
• Semidefinite relaxations for non-convex quadratic mixed-integer programming
• The polynomial solvability of convex quadratic programming

Cited in

(2)

• A branch-and-bound algorithm for solving max-\(k\)-cut problem
• SDP-based branch-and-bound for non-convex quadratic integer optimization

Describes a project that uses

Uses Software

• CSDP

This page was built for publication: A coordinate ascent method for solving semidefinite relaxations of non-convex quadratic integer programs

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2835666)