A Matrix Positivstellensatz with lifting polynomials
arXiv:1801.04947 · doi:10.1137/18M1203183
Abstract
Given the projections of two semialgebraic sets defined by polynomial matrix inequalities, it is in general difficult to determine whether one is contained in the other. To address this issue we propose a new matrix Positivstellensatz that uses lifting polynomials. Under the classical archimedean condition and some mild natural assumptions, we prove that such a containment holds if and only if the proposed matrix Positivstellensatz is satisfied. The corresponding certificate can be searched for by solving a semidefinite program. An important application is to certify when a spectrahedrop (i.e., the projection of a spectrahedron) is contained in another one.
arXiv comment22 pages, 2 figures
References in corpus (8)
- Lifts of convex sets and cone factorizations
- The convex Positivstellensatz in a free algebra
- The matricial relaxation of a linear matrix inequality
- Spectrahedral Containment and Operator Systems with Finite-Dimensional Realization
- An exact duality theory for semidefinite programming based on sums of squares
- Pure states, positive matrix polynomials and sums of hermitian squares
- Real algebraic geometry for matrices over commutative rings
- Solving moment problems by dimensional extension