Deterministic guarantees for Burer-Monteiro factorizations of smooth semidefinite programs
Abstract
We consider semidefinite programs (SDPs) with equality constraints. The variable to be optimized is a positive semidefinite matrix of size . Following the Burer--Monteiro approach, we optimize a factor of size instead, such that . This ensures positive semidefiniteness at no cost and can reduce the dimension of the problem if is small, but results in a non-convex optimization problem with a quadratic cost function and quadratic equality constraints in . In this paper, we show that if the set of constraints on regularly defines a smooth manifold, then, despite non-convexity, first- and second-order necessary optimality conditions are also sufficient, provided is large enough. For smaller values of , we show a similar result holds for almost all (linear) cost functions. Under those conditions, a global optimum maps to a global optimum of the SDP. We deduce old and new consequences for SDP relaxations of the generalized eigenvector problem, the trust-region subproblem and quadratic optimization over several spheres, as well as for the Max-Cut and Orthogonal-Cut SDPs which are common relaxations in stochastic block modeling and synchronization of rotations.
Cite
@article{arxiv.1804.02008,
title = {Deterministic guarantees for Burer-Monteiro factorizations of smooth semidefinite programs},
author = {Nicolas Boumal and Vladislav Voroninski and Afonso S. Bandeira},
journal= {arXiv preprint arXiv:1804.02008},
year = {2019}
}
Comments
28 pages, Communications on Pure and Applied Mathematics: https://onlinelibrary.wiley.com/doi/abs/10.1002/cpa.21830