English

Quasi-Convex Free Polynomials

Functional Analysis 2012-08-20 v1

Abstract

Let \Rx\Rx denote the ring of polynomials in gg freely non-commuting variables x=(x1,...,xg)x=(x_1,...,x_g). There is a natural involution * on \Rx\Rx determined by xj=xjx_j^*=x_j and (pq)=qp(pq)^*=q^* p^* and a free polynomial p\Rxp\in\Rx is symmetric if it is invariant under this involution. If X=(X1,...,Xg)X=(X_1,...,X_g) is a gg tuple of symmetric n×nn\times n matrices, then the evaluation p(X)p(X) is naturally defined and further p(X)=p(X)p^*(X)=p(X)^*. In particular, if pp is symmetric, then p(X)=p(X)p(X)^*=p(X). The main result of this article says if pp is symmetric, p(0)=0p(0)=0 and for each nn and each symmetric positive definite n×nn\times n matrix AA the set X:Ap(X)0{X:A-p(X)\succ 0} is convex, then pp has degree at most two and is itself convex, or p-p is a hermitian sum of squares.

Keywords

Cite

@article{arxiv.1208.3582,
  title  = {Quasi-Convex Free Polynomials},
  author = {Sriram Balasubramanian and Scott McCullough},
  journal= {arXiv preprint arXiv:1208.3582},
  year   = {2012}
}
R2 v1 2026-06-21T21:52:01.783Z