Matrix anti-hermitian

Exercise. Prove the following lemma If H is a positive semi-definite Hermitian matrix, and F anti-Hermitian then the eigenvalues of A = H + F have nonnegative real parts. Moreover, if the real part is zero the corresponding eigenvector is an eigenvector of H and F separately. Use this lemma to show that (5.12) is the solution of (5.10). 

Since any operator can be written as the sum of Hermitian and anti-Hermitian operators, we can restrict our discussion to these two types only. Further, any operator can be written as a linear combination of irreducible symmetry operators, so we can restrict ourselves to irreducible tensor operators. An operator matrix 0(r, K) that transforms according to the symmetry (T, K) obeys the relationship... 

A full proof of this can be found in Taylor [15]. The factor p is +1 if the operator is Hermitian and —1 if it is anti-Hermitian. We can see an immediate complication relative to our earlier formula Eq. 5.11 in that a full representation matrix for irrep T is required. This is considered in more detail below. Additional redundancies in the P2 list that arise from the form of particular operators axe also treated by Taylor. 

We can now identify the unitary matrix U = exp(-T). It is a general property of unitary matrices that they can be written as the exponential of an anti-Hermitian matrix First it is immediately clear that exp (-T) is a unitary matrix, when T is anti-Hermitian. Secondly it is possible to show that all... 

Show that a unitary matrix U can always be written in the form U = exp(T), where T is an anti-Hermitian matrix. 

If the elements a are complex numbers of the form a- -ib (z=-v/ l) tb complex conjugates, such as a—ib, are denoted by % and the matrix A=[a(/] is called the complex conjugate of A. If A =A, the matrix A is square and is unchanged by the operations of transposition and taking complex conjugates it is called a Hermitian matrix. If A =—A, the square matrix A is called a skew (or anti-) Hermitian matrix. 

Some types of matrices that appear frequently are Hermitian matrices, for which A = A anti-Hermitian matrices, for which A = — A and unitary matrices, for which = U The MCSCF methods diseussed in most detail in this review will involve only operations of real matrices. In this case these matrix types reduce to symmetric matrices, for which A = A , antisymmetric matrices, for which A = — A, and orthogonal matrices, for which U = U. A particular type of orthogonal matrix is called a rotation matrix and satisfies the relation Det(R)= -I-1, where Det(R) is the usual definition of a determi-... 

This shows that, just as a general matrix, the solution vector Xt(to) may be split into Hermitian and anti-Hermitian parts... 

From the above observations we can draw the following conclusions The gradient vector appearing on the right hand side of the first-order response equation (223) generally has positive hermiticity. From (244) and (245) it is then clear that the Hessian and the matrix 5 should be multiplied with Hermi-tian and anti-Hermitian trial vectors, respectively. In the static case (O — O all trial vectors should consequently be restricted to have the Hermitian structure. Furthermore, tried vectors should have the time reversal symmetry of the property gradient since both 4 conserves time reversal symmetry. Then,... 

Clearly, this relationship holds also for linear combinations of determinants. Note that this representation of unitary spin-orbital rotations and unitarily transformed states is independent, in the sense that the numerical parameters kpq may be chosen freely in accordance with the simple requirement that K constitutes an anti-Hermitian matrix. 

Since C is an anti-hermitian matrix ("C+ = — C) the second derivative can be easily expressed from the first and second derivatives ... 

We start from a one-determinant wavefunction Wg of spin-orbitals assumed to be orthonormal, and consider the fluctuation of these orbitals under the influence of a perturbation operator (12.1.8) or, equivalently, a single Fourier component (12.1.14). The variation is conveniently introduced by considering i >—as in Section 8.2, where A is an anti-Hermitian matrix that describes the admixture of virtual orbitals tpm into the occupied set (V i). On expanding the exponential and taking the distinct elements A (r>s) as the parameters the variation yields, up to second order,... 

Every unitary matrix U can be represented by an exponential function of an anti-Hermitian matrix... 

We first show that, for any unitary matrix U, we can always find an anti-Hermitian matrix X such that (3.1.9) is satisfied. For this purpose, we recall that the spectral theorem states that any unitary matrix can be diagonalized as... 

Since iV8V is anti-Hermitian, we have shown that any unitary matrix can be written in the exponential form (3.1.9). 

We have now satisfied the first of the three requirements for the unitary parametrization stated in the introduction to Section 3.1. To satisfy the second requirement, we note that, for any anti-Hermitian matrix X, the exponential exp(X) is always unitary since, from the relation X = —X, it follows that... 

Finally, we note that the third requirement for the parametrization is also satisfied since anti-Hermitian matrices are trivially represented by a set of independent parameters. We may, for instance, take the matrix elements at the diagonal and below the diagonal as the independent ones and generate the remaining elements of the matrix from the anti-Hermitian condition = —Xgp. Note that the diagonal elements of an anti-Hermitian matrix are pure imaginary and that the off-diagonal elements are complex. 

The exponential parametrization of a unitary matrix in (3.1.9) is a general one, applicable under all circumstances. We shall now consider more special forms of unitary matrices. We begin by writing the anti-Hermitian matrix X in the form... 

The expansion of the exponential matrix (3.1.2) is rapidly convergent and may often be used for the evaluation of unitary matrices, especially if the anti-Hermitian matrix X has a small norm and high accuracy is not required. An alternative strategy is to diagonalize X ... 

According to the discussion in Section 3.1, the unitary matrix U may be written in terms of an anti-Hermitian matrix K as... 

In Exercise 3.6, the proof given for (3.4.14) is valid only for matrices belonging to the domain where the logarithmic and exponential matrix functions are inverse functions. Finally, (3.4.15) - which follows from (3.4.12) and (3.4.13) - shows that the logarithmic function maps unitary matrices onto anti-Hermitian matrices as expected from the fact that flie exponential function maps anti-Hermitian matrices onto unitary matrices. 

Determine closed-form expressions for exp(ic) for the following anti-Hermitian matrices k 1. a general 2x2 anti-Hermitian matrix... 

---