Variational procedures

Gutdeutsch U, Birkenheuer U and Rdsch N 1998 A strictly variational procedure for cluster embedding based on the extended subspace approach J. Chem. Phys. 109 2056... 

In addition to the mixed results in Table 10-1, the G2 calculation for H2 produces an energy that is lower than the experimental value, in contradiction to the rule that variational procedures reach a least upper bound on the energy. Some new factors are at work, and we must look into the stmcture of the G2 procedure in temis of high-level Gaussian basis sets and electron correlation. 

Such a treatment, while being accurate above T, suffers from the total neglect of the actual form of the potential near the well. It can be the basis for a variational procedure with a parabolic reference [Poliak 1986a]. 

Quadralically Convergent or Second-Order SCF. As mentioned in Section 3.6, the variational procedure can be formulated in terms of an exponential transformation of the MOs, with the (independent) variational parameters contained in an X matrix. Note that the X variables are preferred over the MO coefficients in eq. (3.48) for optimization, since the latter are not independent (the MOs must be orthonormal). The exponential may be written as a series expansion, and the energy expanded in terms of the X variables describing the occupied-virtual mixing of the orbitals. 

The diagonal elements in the sum involving the Hamilton operator are energies of the corresponding deteiminants. The overlap elements between different determinants are zero as they are built from orthogonal MOs (eq. (3.20)). The variational procedure corresponds to setting alt the derivatives of the Lagrange function (4.3) with respect to the at expansion coefficients equal to zero. 

The optimum values of die oq and a coefficients are determined by the variational procedure. The HF wave function constrains both electrons to move in the same bonding orbital. By allowing the doubly excited state to enter the wave function, the electrons can better avoid each other, as the antibonding MO now is also available. The antibonding MO has a nodal plane (where opposite sides of this plane. This left-right correlation is a molecular equivalent of the atomic radial correlation discussed in Section 5.2. 

Consider now making the variational coefficients in front of the inner basis functions constant, i.e. they are no longer parameters to be determined by the variational principle. The Is-orbital is thus described by a fixed linear combination of say six basis functions. Similarly the remaining four basis functions may be contracted into only two functions, for example by fixing the coefficient in front of the inner three functions. In doing this the number of basis functions to be handled by the variational procedure has been reduced from 10 to three. 

The basis functions are normally the same as used in wave mechanics for expanding the HF orbitals, see Chapter 5 for details. Although there is no guarantee that the exponents and contraction coefficients determined by the variational procedure for wave functions are also optimum for DFT orbitals, the difference is presumably small since the electron densities derived by both methods are very similar. ... 

The variational procedure again leads to a matrix equation in the atomic orbital basis which can be written in the form (compare eq. (3.50)). 

Lowdin, P.-O., An elementary iteration-variation procedure for solving the Schrodinger equation."... 

The reader should note that no restrictions were placed on the form of the density expansion Eq. (3.26) in particular there is no limit on the number of terms. As already noted, therefore Eqs. (3.29) are not conventional Kohn-Sham equations. Rather they are an exact one-particle form of the Hohenberg-Kohn variation procedure and use Hohenberg-Kohn potentials in the definition of the... 

Computational strategies can be based on variational procedures using the Dirac-Frenkel time-dependent variational principle (TDVP). Introducing a shorthand notation so that... 

The variational procedure for the states xi can be implemented in many different ways a very general formalism introduces complex parameters z(t). 

The existing SCF procedures are of two types in restricted methods, the MO s, except for the hipest (singly) occupied MO, are filled by two electrons with antiparallel spin, while in unrestricted methods, the variation procedure is performed with individual spin orbitals. In the latter, a total wave function is not an eigenvalue of the spin operator S, which is disadvantageous in many applications because of a necessary annihilation of higher multiplets by the projection operator. Since in practical applications the unrestricted methods have not proved to be remarkably superior, we shall call our attention in this review mainly to the restricted methods. 

In all the other sehemes (UHF, PHF, EHF), the dissociation limit is the correct one corresponding to two neutral hydrogen atoms (2H-) each FSGO-hydrogen atom energy is thus obtained by the simple variational procedure ... 

Physically the variational procedure based on the second entropy may be interpreted like this. If the flux E were increased beyond its optimum value, then the rate of entropy consumption by the subsystem would be increased due to its increased dynamic order by a greater amount than the entropy production of the reservoirs would be increased due to the faster transfer of heat. The converse holds for a less than optimum flux. In both cases the total rate of second entropy production would fall from its maximum value. 

In analogy to using a linear combination of atomic orbitals to form MOs, a variational procedure is used to construct many-electron wavefunctions from a set of N Slater determinants y, i.e. one sets up a N x. N matrix of elements flij = (d>, H d>y) which, upon diagonalization, yields state energies and associated vectors of coefficients a used to define (fi as a linear combination of A,s ... 

The condition that the energy becomes a minimum through a variational procedure, together with the conditions f fa(j>jdvidvj = Sij gives the Hartree equations... 

In Equations 10.1, 10.4, and 10.5, the standard implicit sum on the index (x is assumed. The relativistic Kohn-Sham equations are obtained by minimization of the Equation 10.4 with respect to the orbitals. In contrast to the nonrelativistic case, the variational procedure gives rise to an infinite set of coupled equations (see the summation restrictions in Equation 10.3) that have to be solved in a self-consistent manner ... 

Notice that how the shape function naturally enters this discussion. Because the number of electrons is fixed, the variational procedure for the electron density is actually a variational procedure for the shape function. So it is simpler to restate the equations associated with the variational principle in terms of the shape function. Parr and Bartolotti have done this, and note that because the normalization of the shape function is fixed,... 

Various model-dependent methods for the comparison of two cumulative dissolution data sets have been proposed (21). Usually, these methods involve prior characterization of both profiles by one to three parameters per profile. In some models, these parameters can be interpreted in terms of the kinetics, the shape, and/or the plateau, but in other instances, they have no physical meaning. One issue that requires some attention is that, in cases where more than one parameter is estimated, a multi-variate procedure for the comparison of the parameters must be applied (9,21). 

As Fig. 12 shows, the inner shell electrons of the alkaline ions behave classically like a polarizable spherical charge-density distribution. Therefore it seemed promising to apply a "frozen-core approximation in this case 194>. In this formalism all those orbitals which are not assumed to undergo larger changes in shape are not involved in the variational procedure. The orthogonality requirement is... 

If we apply the variational procedure to hydrogen atoms at ground state, we find that the lowest orbitals in order of increasing energy are... 

The Dirac and the L6vy-Leblond equations establish relationships between the large and the small components of the wavefunctions. If these relationships are to be fulfilled by the functions derived from a variational procedure, the basis sets for the large and for the small components have to be constructed accordingly. In particular, the relation... 

The variational procedure in a many-electron space may be considered as several consecutively executed variational procedures in one-electron spaces (the procedure may be iterative if a self-consistency is required). This means that fulfilment of the one-electron minimax principle is a necessary (but, in general, not sufficient) condition for the fulfilment of a similar principle in a many-electron case. Therefore one should not expect a many-electron generalization of Eq. (7) being valid when, say, parameters L and S are varied. 

Two of the methods proposed appear at first sight to be non-arbitrary. In fact, absolute accuracy has been claimed for one of these (25), a refined two-dimensional variational procedure in which all frequencies but one are used as constraints, while this last one is used as a criterion. In fact, it is not possible to fix two unknowns from one parameter, and the method seems to this reviewer an unusually ingenious and elaborate exercise in self-deception, in which the range of possible solutions of the independent parameter method appears to contract to a point. 

---