Physlib

QuantumInfo.ForMathlib.HermitianMat.CompoundMatrix

Compound matrices on `HermitianMat`

This file lifts the compound-matrix construction from `QuantumInfo.ForMathlib.Majorization` to `HermitianMat`. The compound matrix of a Hermitian matrix is again Hermitian, and inherits the natural spectral and positivity properties of the original.

4 declarations

definition

The kk-th compound Hermitian matrix Ck(A)C_k(A)

Let dd be a finite index set and AA be a d×dd \times d Hermitian matrix over the complex numbers C\mathbb{C}. For any natural number kk, the kk-th compound matrix Ck(A)C_k(A) is a Hermitian matrix whose rows and columns are indexed by subsets SdS \subseteq d with cardinality S=k|S| = k. The entries of this matrix are the k×kk \times k minors of AA. This definition packages the construction as a member of the type of Hermitian matrices `HermitianMat`.

theorem

The eigenvalues of the kk-th compound matrix Ck(A)C_k(A) are the kk-fold products of the eigenvalues of AA

Let dd be a finite index set and AA be a d×dd \times d Hermitian matrix over the complex numbers C\mathbb{C}. For any natural number kk, let Ck(A)C_k(A) denote the kk-th compound Hermitian matrix, whose rows and columns are indexed by the subsets SdS \subseteq d of cardinality kk. There exists a permutation σ\sigma of the set of kk-element subsets of dd such that the eigenvalues of Ck(A)C_k(A) are the products of the eigenvalues of AA over these subsets. That is, for each subset SS with S=k|S|=k, the corresponding eigenvalue of Ck(A)C_k(A) is given by iSλi\prod_{i \in S} \lambda_i, where λi\lambda_i denotes the eigenvalues of the original matrix AA.

theorem

0A    0Ck(A)0 \le A \implies 0 \le C_k(A)

Let dd be a finite index set and AA be a d×dd \times d Hermitian matrix over the complex numbers C\mathbb{C}. If AA is positive semidefinite (denoted by 0A0 \le A in the Loewner partial order), then for any natural number kk, its kk-th compound Hermitian matrix Ck(A)C_k(A) is also positive semidefinite, i.e., 0Ck(A)0 \le C_k(A).

theorem

Ck(BAB)=Ck(B)Ck(A)Ck(B)C_k(B A B^\dagger) = C_k(B) C_k(A) C_k(B)^\dagger

Let dd be a finite index set, AA be a d×dd \times d Hermitian matrix over C\mathbb{C}, and BB be a d×dd \times d complex matrix. For any natural number kk, let Ck()C_k(\cdot) denote the kk-th compound matrix operation. Then the kk-th compound of the congruence transformation BABB A B^\dagger is equal to the congruence transformation of the kk-th compound of AA by the kk-th compound of BB: Ck(BAB)=Ck(B)Ck(A)Ck(B)C_k(B A B^\dagger) = C_k(B) C_k(A) C_k(B)^\dagger where BB^\dagger denotes the conjugate transpose of BB.