arXiv++ Combinatorics

Disjointly constrained multilinear programming concerns the problem of maximizing a multilinear function on the product of finitely many disjoint polyhedra. While maximizing a linear function on a polytope (linear programming) is known to be solvable in polynomial time, even bilinear programming is NP-hard. Based on a reformulation of the problem in terms of sum-of-squares polynomials, we study a hierarchy of semidefinite relaxations to the problem. It follows from the general theory that the sequence of optimal values converges asymptotically to the optimal value of the multilinear program. We show that the semidefinite hierarchy converges generically in finitely many steps to the optimal value of the multilinear problem. We outline two applications of the main result. For nondegenerate bimatrix games, a Nash equilibrium can be computed by the sum of squares approach in finitely many steps. Under an additional geometric condition, the NP-complete containment problem for projections of $\mathcal{H}$-polytopes can be decided in finitely many steps.

Spectrahedra are affine sections of the cone of positive semidefinite matrices which form a rich class of convex bodies that properly contains that of polyhedra. While the class of polyhedra is closed under linear projections, the class of spectrahedra is not. In this paper we investigate the problem of deciding containment of projections of polyhedra and spectrahedra based on previous works on containment of spectrahedra. The main concern is to study these containment problems by formulating them as polynomial nonnegativity problems. This allows to state hierarchies of (sufficient) semidefinite conditions by applying (and proving) sophisticated Positivstellensätze. We also extend results on a solitary sufficient condition for containment of spectrahedra coming from the polyhedral situation as well as connections to the theory of (completely) positive linear maps.

Given an $\mathcal{H}$-polytope $P$ and a $\mathcal{V}$-polytope $Q$, the decision problem whether $P$ is contained in $Q$ is co-NP-complete. This hardness remains if $P$ is restricted to be a standard cube and $Q$ is restricted to be the affine image of a cross polytope. While this hardness classification by Freund and Orlin dates back to 1985, for general dimension there seems to be only limited progress on that problem so far. Based on a formulation of the problem in terms of a bilinear feasibility problem, we study sum of squares certificates to decide the containment problem. These certificates can be computed by a semidefinite hierarchy. As a main result, we show that under mild and explicitly known preconditions the semidefinite hierarchy converges in finitely many steps. In particular, if $P$ is contained in a large $\mathcal{V}$-polytope $Q$ (in a well-defined sense), then containment is certified by the first step of the hierarchy.

We study the computational question whether a given polytope or spectrahedron $S_A$ (as given by the positive semidefiniteness region of a linear matrix pencil $A(x)$) is contained in another one $S_B$. First we classify the computational complexity, extending results on the polytope/polytope-case by Gritzmann and Klee to the polytope/spectrahedron-case. For various restricted containment problems, NP-hardness is shown. We then study in detail semidefinite conditions to certify containment, building upon work by Ben-Tal, Nemirovski and Helton, Klep, McCullough. In particular, we discuss variations of a sufficient semidefinite condition to certify containment of a spectrahedron in a spectrahedron. It is shown that these sufficient conditions even provide exact semidefinite characterizations for containment in several important cases, including containment of a spectrahedron in a polyhedron. Moreover, in the case of bounded $S_A$ the criteria will always succeed in certifying containment of some scaled spectrahedron $νS_A$ in $S_B$.