Abstract
In this work, we study the set of strictly accretive matrices, that is, the set of matrices with positive definite Hermitian part, and show that the set can be interpreted as a smooth manifold. Using the recently proposed symmetric polar decomposition for sectorial matrices, we show that this manifold is diffeomorphic to a direct product of the manifold of (Hermitian) positive definite matrices and the manifold of strictly accretive unitary matrices. Utilizing this decomposition, we introduce a family of Finsler metrics on the manifold and characterize their geodesics and geodesic distances. Finally, we apply the geodesic distance to a matrix approximation problem and also give some comments on the relation between the introduced geometry and the geometric mean of strictly accretive matrices as defined by Drury [1].
Disclosure statement
No potential conflict of interest was reported by the author(s).
Notes
1 Note that this is equivalent to the Toeplitz decomposition since if , then
and
, see [Citation8, p. 7].
2 The naming used here is the same as in [Citation9, p. 281], in contrast to [Citation10].
3 In [Citation24], these were called canonical angles.
4 To see this, note that a matrix A is normal if and only if H(A) and commute, see, e.g. [Citation15, Thm. 9.1].
5 Note that the latter has unfortunately also been termed dissipative in the literature, see [Citation9, p. 279].
6 In [Citation7], the convention is that the Finsler structure is squared compared to the one in [Citation34]. We follow the convention of the latter.
7 Note that functions Ψ fulfilling i)-v) are not necessarily symmetric gauge functions. As an example, consider Ψ(x, y) = (x2 + 3xy + y2)2 [Citation7, Rem. 6]; this is not a symmetric gauge function since in general Ψ(x, y) ≠ Ψ(x, − y). Conversely, symmetric gauge functions do not in general fulfil (i)–(v). As an example, consider Φ(x, y) = max{|x|, |y|} [Citation12, p. 138]; at any point such that x>y, ∂yΦ = 0 and hence condition (iv) is not fulfilled for this function.
8 To see this, take and
in [Citation12, Thm. 9.H.1.f], and use the fact that the eigenvalues of
are invariant under the similarity transform
.