The Setting
Let be a finite index set, and let be the real vector space of complex matrices. Write when is positive semidefinite and when is positive definite, and let denote the conjugate transpose of and denote the functional-calculus square root of .
On , the problem is to equip square complex matrices with a matrix energy and gradient functional together with a variational Hessian form applicable both on the positive definite cone and on positive semidefinite matrices annihilating an orthogonal projection.
Real Trace, Inner Product, and the Sylvester Operator
The real trace of a matrix is the real-valued functional defined by (1)✦ The real Hilbert-Schmidt inner product of two matrices is the real pairing given by (2)✦
The Sylvester operator of a matrix is the matrix map acting on by (3)✦ When is positive definite, is a continuous real-linear automorphism of .✦ The inverse root Sylvester operator of a positive definite matrix is the continuous real-linear inverse of , satisfying for every (4)✦
The Variational Functional and Root Cost
The Sylvester penalty of a pair is the real quantity defined by (5)✦ The variational functional of matrices is the real-valued function given by (6)✦ The root variational cost attached to matrices is the real value defined by (7)✦
The Claim
Definition 1. The matrix energy of a matrix is the real value defined by (8)✦ The energy gradient at is the continuous real-linear functional acting on by (9)✦ The Hessian quadratic form of matrices is the real quantity given by (10)✦ The Hessian bilinear form for is the real pairing defined by (11)✦ The positive definite Hessian of a positive definite matrix and matrices is the real pairing (12)✦ If is positive semidefinite and satisfy , then (13)✦ Whenever is positive definite and , one has (14)✦ If are positive semidefinite with , , and , and satisfy and , then (15)✦
In the paper, the result reads: for positive semidefinite , , , , , and , and for Hermitian and supported on , and ( if ), with standing for and for .
What Is Checked
The matrix energy and Hessian constructions, along with their underlying operators and identities, are stated over a finite index set , with arbitrary matrices in for the real trace, real Hilbert-Schmidt inner product, matrix energy, energy gradient, Sylvester operator, and variational constructions; with for the Sylvester automorphism and for the inverse root Sylvester operator, positive definite Hessian, and coincidence identity; with and for a Sylvester solution; and with satisfying , , and for kernel-projection invariance. For these constructions, the library proves that is additive and real-homogeneous and defines a continuous real-linear functional on , and that for every the operator is additive and real-homogeneous as a real-linear endomorphism of .