Linear matrix pencils and noncommutative convexity

TL;DR

This paper surveys the properties of free spectrahedra, convex solution sets of noncommutative polynomial inequalities, highlighting their characterization, detection, and applications in eigenvalue optimization via semidefinite programming.

Contribution

It provides a comprehensive overview of free spectrahedra, including their characterization, detection methods, and the development of a Positivstellensatz for eigenvalue optimization.

Findings

Free spectrahedra are exactly the convex matricial solution sets of noncommutative polynomial inequalities.

A procedure for detecting free spectrahedra and constructing their linear matrix pencils is discussed.

Free spectrahedra admit a Positivstellensatz leading to semidefinite programming formulations.

Abstract

Hermitian linear matrix pencils are ubiquitous in control theory, operator systems, semidefinite optimization, and real algebraic geometry. This survey reviews the fundamental features of the matricial solution set of a linear matrix inequality, the free spectrahedron, from the perspective of free real algebraic geometry. Namely, among matricial solution sets of noncommutative polynomial inequalities, free spectrahedra are precisely the convex ones. Furthermore, a procedure for detecting free spectrahedra and producing their representing linear matrix pencils is discussed. Finally, free spectrahedra admit a perfect Positivstellensatz, leading to a semidefinite programming formulation of eigenvalue optimization over convex matricial sets constrained by noncommutative polynomial inequalities.