C*-Algebra C(ex(K)) and Spectrahedral Structures

• C*-algebra C(ex(K)) is a function space defined on the extreme boundary of compact convex sets, serving as the minimal carrier of function theory on spectrahedra.
• It links spectrahedral representation size bounds, matrix convexity, and noncommutative boundaries, offering a framework for operator theory and convex algebraic analysis.
• Analytical tools like Lasserre relaxations and extreme point decomposition provide practical methods to study and compute the functional properties encoded in C(ex(K)).

The C*-algebra $C(\mathrm{ex}(K))$ arises naturally in the study of compact convex sets $K$ via their extreme points, especially within the context of spectrahedral geometry and convex algebraic analysis. The interplay between the structure of $K$, the spectrahedral representations, and the functional analysis of the associated C*-algebra is central in operator theory and noncommutative convexity. This article details the key aspects of $C(\mathrm{ex}(K))$ in relation to compact spectrahedra, matrix convexity, extremal representations, and boundary phenomena.

1. Compact Spectrahedra and Extreme Points

A spectrahedron in $\mathbb R^n$ is defined by a linear matrix inequality:

$$S = \{x \in \mathbb R^n : A_0 + x_1 A_1 + \dots + x_n A_n \succeq 0 \}$$

where the $A_i \in \mathrm{Sym}_r(\mathbb R)$ are real symmetric matrices, and $\succeq 0$ denotes positive semidefiniteness. If $S$ is compact, the extreme points $\mathrm{ex}(S)$ play an essential role in its convex geometry. The C*-algebra $C(\mathrm{ex}(S))$ consists of all continuous complex-valued functions on $\mathrm{ex}(S)$, serving as a foundational object for function theory on the boundary of spectrahedra. In matrix convex settings, the situation becomes noncommutative, and the structure of extreme points generalizes.

2. Matrix Convexity, Free Spectrahedra, and Noncommutative Boundaries

Matrix convex sets are dimension-free analogs of convex sets, closed under matrix convex combinations across arbitrary sizes. For free spectrahedra $\mathcal D_A$ (solution sets to monic linear pencils $L_A(x)$ in all sizes), the matrix convex hull is generated by matrix convex combinations of tuples $X$ in $\mathrm{Sym}_n(\mathbb R)^g$. In this framework, there are three notable extreme point notions: ordinary, matrix, and absolute (Arveson) extreme points. The central result is that every compact free spectrahedron is the matrix convex hull of its absolute extreme points, with the absolute boundary being the minimal irreducible spanning set (Evert et al., 2018).

The C*-algebra $C(\mathrm{ex}(K))$ encodes the function theory on the Arveson boundary, generalizing the classical Choquet boundary in commutative convexity. In noncommutative convexity, Arveson extreme points are irreducible tuples not admitting nontrivial one-step dilations, matching the operator-system boundary framework.

3. Lower Bounds on Spectrahedral Matrix Sizes and Polynomial Boundaries

A fundamental problem concerns the minimal possible matrix size $r$ of an LMI representation for a convex body $K \subset \mathbb R^n$. For the $n$-dimensional Euclidean unit ball or any compact convex set defined by a quadratic polynomial, the matrix size lower bound is $r \geq n/2$, and in certain dyadic cases $r = n$ is sharp (Kummer, 2015). For convex regions in $\mathbb R^3$ with smooth cubic algebraic boundaries, no $4 \times 4$ LMI suffices; $r \geq 5$ is necessary, dictated here by singularity in the boundary surface. These constraints reflect deep geometric obstructions not captured solely by algebraic degree, and impact the possible function theory encoded in $C(\mathrm{ex}(K))$, since the structure of the extreme boundary depends on the spectrahedral representation.

4. Lasserre Relaxations, Semidefinite Representability, and Algebraic Shadows

Semidefinite representations of compact convex basic closed semialgebraic sets are achieved via Lasserre (moment-SDP) relaxations, subject to the Archimedean condition and strict quasiconcavity or SOS-concavity of the defining polynomials (Schweighofer et al., 2017). When these conditions are met, the Lasserre hierarchy becomes exact at finite level, furnishing the set as the projection onto $x$-coordinates of a spectrahedron. The C*-algebra $C(\mathrm{ex}(K))$ then captures functional properties of the extreme points of $K$, which correspond to atomic structures in the lifted spectrahedral shadow.

The distinction between the classical Helton–Nie covering method and the refined Kriel–Schweighofer Lasserre relaxation places emphasis on global extremal structures in $K$, potentially affecting the spectrum and functional calculus in $C(\mathrm{ex}(K))$.

5. Structure and Span of Extremal Boundaries

For any compact matrix convex set defined by an LMI, every element is a matrix convex combination of its absolute extreme points. The minimality of the absolute boundary ensures that $C(\mathrm{ex}(K))$ is the unique smallest irreducible generating C*-algebra for the function theory of $K$ (Evert et al., 2018, Evert, 2023). Explicitly, each tuple $X$ can be decomposed as a compression of a direct sum of absolute extremes, with optimal Carathéodory-type bounds on matrix sizes. This decomposition is algorithmic via iterated maximal one-step dilations, systematically reducing the dimension of the dilation subspace.

In the context of projections of spectrahedra ("spectrahedrops") or generalized free spectrahedra with compact operator tuples, the span by free extreme points persists (Evert, 2023). Thus, $C(\mathrm{ex}(K))$ functions as the algebraic carrier of minimal representation and atomic decomposition for the compact convex (or matrix convex) set $K$.

6. Orbitopes, Polar Structures, and Exposed Faces

In representation-theoretic settings, orbitopes and polar orbitopes arising from compact Lie group actions possess spectrahedral structures with explicit LMI representations (Kobert et al., 2020, Kobert, 2016). The extreme points structure is controlled by highest-weight theory and momentum polytope geometry, ensuring all faces are exposed and accessible via supporting matrix inequalities. In these contexts, $C(\mathrm{ex}(K))$ is determined by the symmetries and orbit-geometry, and often reflects the underlying representation algebra via its fundamental weights.

Furthermore, in doubly spectrahedral cases—where both the orbitope and its polar admit LMI descriptions—the dual function theory on $C(\mathrm{ex}(K))$ and $C(\mathrm{ex}(K)^o)$ is tractable and rich in families of Ky Fan balls, nuclear-norm balls, and related operator-norm structures.

7. Faces of Gram Spectrahedra, Singularities, and Smoothness

Gram spectrahedra parametrize all sum-of-squares representations of a fixed form $f$. The dimension of a face of the Gram spectrahedron is combinatorially maximized in singular (monomial) cases and improves for smooth (base-point–free) forms (Vill, 2020). The subspaces $U$ achieving maximal face dimensions correspond to singularities, affecting the boundary and thus the functional structure of $C(\mathrm{ex}(K))$. The presence or absence of base-points in $U$ induces sharp transitions in the algebraic topology of extremal boundaries.

This suggests that in spectrahedra with smooth boundaries, the algebra $C(\mathrm{ex}(K))$ may admit more favorable analytic properties (e.g., lower-dimensional facial decompositions), whereas singular boundaries lead to more complicated or higher-dimensional $C^*$-algebras.

---

In summary, the C*-algebra $C(\mathrm{ex}(K))$—whether in classical, matrix convex, or representation-theoretic contexts—embodies the function theory of the extremal boundary of a compact convex or spectrahedral set. Its structure is intricately linked to spectrahedral representation size bounds, boundary regularity, moment-SDP relaxations, matrix convexity theory, and group-theoretic symmetries. It is the canonical carrier of the minimal extremal function theory, encoding irreducible atomic data, exposed face decomposition, and duality phenomena intrinsic to spectrahedral and algebraic convex geometry.