The stabilizers for the action of orthogonal similarity on symmetric matrices and orthogonal $*$-conjugacy on Hermitian matrices

TL;DR

This paper develops algorithms to compute stabilizers of symmetric and Hermitian matrices under orthogonal actions, classifies normal forms, and applies these results to complex geometry.

Contribution

It introduces a recursive algorithm for stabilizers, provides bounds for Hermitian matrix stabilizers, and completes the classification of Hermitian normal forms under orthogonal conjugation.

Findings

Recursive algorithm for stabilizers of symmetric matrices

Lower bounds for stabilizers of Hermitian matrices

Complete classification of Hermitian normal forms under orthogonal conjugation

Abstract

We describe the recursive algorithmic procedure to compute the stabilizers of the group of complex orthogonal matrices with respect to the action of similarity on the set of all symmetric matrices. Futhermore, lower bounds for dimensions of the stabilizers for the action of orthogonal $*$-conjugation on Hermitian matrices are obtained. We also prove a result that completes the classification of normal forms of Hermitian matrices under orthogonal $*$-conjugation. A key step in our proof is to solve a certain block matrix equation with Toeplitz blocks. These results are then applied to provide a theorem on normal forms of the quadratic parts of flat complex points in a real codimension $2$ submanifold in a complex manifold.