The Setting
For an integer , the space of real symmetric matrices is the subspace of real matrices equal to their transpose, with denoting that is positive semidefinite.✦ A central question in convex algebraic geometry is whether every closed hyperbolicity cone of a homogeneous hyperbolic polynomial is spectrahedral.
The ambient vector space is , whose points are triples with and , with base point , where is the identity matrix.
The Polynomial and Its Cone
For each , the quadratic matrix is the real matrix defined by (1): (1)✦
The matrix map associated with is the real linear endomorphism of the space of real matrices defined in block form by (2) for , column vectors , and a real matrix , restricting to when and : (2)✦
The polynomial on is the real polynomial defined by the determinantal formula (3), where denotes the adjugate matrix of : (3)✦
At a point , the univariate restriction of relative to is the real univariate polynomial defined by (4): (4)✦
The closed hyperbolicity cone of with respect to is the subset of defined from (4) by (5): (5)✦
The Claim
Theorem 1. The ambient space , the base point , and the polynomial satisfy (6): (6)✦ For every integer and real linear map , (7)✦
In the paper, the result reads: the polynomial (the paper’s (1.4)) is homogeneous of degree and hyperbolic with respect to , with , and its closed hyperbolicity cone is not spectrahedral: for every finite and every real linear map , . The root condition on together with in (6) gives hyperbolicity of with respect to .
What Is Checked
The claims are stated over the ambient space , the base point , the polynomial constructed from and , its univariate restrictions and closed hyperbolicity cone , and arbitrary integers and real linear maps .