Sum-of-squares hierarchies for binary polynomial optimization
Abstract
We consider the sum-of-squares hierarchy of approximations for the problem of minimizing a polynomial over the boolean hypercube . This hierarchy provides for each integer a lower bound on the minimum of , given by the largest scalar for which the polynomial is a sum-of-squares on with degree at most . We analyze the quality of these bounds by estimating the worst-case error in terms of the least roots of the Krawtchouk polynomials. As a consequence, for fixed , we can show that this worst-case error in the regime is of the order as tends to . Our proof combines classical Fourier analysis on with the polynomial kernel technique and existing results on the extremal roots of Krawtchouk polynomials. This link to roots of orthogonal polynomials relies on a connection between the hierarchy of lower bounds and another hierarchy of upper bounds , for which we are also able to establish the same error analysis. Our analysis extends to the minimization of a polynomial over the -ary cube .
Cite
@article{arxiv.2011.04027,
title = {Sum-of-squares hierarchies for binary polynomial optimization},
author = {Lucas Slot and Monique Laurent},
journal= {arXiv preprint arXiv:2011.04027},
year = {2022}
}
Comments
31 pages, 1 figure. Version 3: Changed structure of the paper. Added clarifying statements, shortened several proofs. Fixed minor mistakes/typos