Shape, scale, and minimality of matrix ranges

Abstract:

We study containment and uniqueness problems concerning matrix convex sets. First, to what extent is a matrix convex set determined by its first level? Our results in this direction quantify the disparity between two product operations, namely, the product of the smallest matrix convex sets over $K_i \subseteq \mathbb {C}^d$ and the smallest matrix convex set over the product of $K_i$. Second, if a matrix convex set is given as the matrix range of an operator tuple $T$, when is $T$ determined uniquely? We provide counterexamples to results in the literature, showing that a compact tuple meeting a minimality condition need not be determined uniquely, even if its matrix range is a particularly friendly set. Finally, our results may be used to improve dilation scales, such as the norm bound on the dilation of (not necessarily self-adjoint) contractions to commuting normal operators, both concretely and abstractly.