Energy Minimization Schemes

Once one is able to calculate the energy of the molecule by a suitable program utilizing equations of the sort given in the previous section, then one can worry about the details of the calculation itself. The structure of the molecule will correspond to that geometry where the energy is at a minimum. Therefore if we write the total energy of the molecule as in eqn (9), in principle, all we 

The early work on this problem was carried out by Westheimer (1956, see also Kitaigorodskii, 1960, 1961 Eliel et al., 1965) using hand calculations, and later by Hendrickson (1961, 1962, 1964, 

An improvement on the simple steepest descent method was made by Schleyer (Engler et al., 1973), and utilized the pattern search procedure. In this case the method proceeds as before, but information regarding the direction of motion of the atoms is saved from one iteration to the next, and the correction terms for succeeding motions are summed, and then applied. The advantage of this method is that if a particular atom is moving down a long hill with small slope, while the steepest descent method requires many iterations to move it, the pattern search method accelerates the motion, because the motion is repeatedly in the same direction, and the size of the correction term increases. 

It is very difficult to compare true efficiencies in minimization algorithms. Depending on the sophistication of the program, the machine system used, and other variables difficult to trace, computation times vary widely, even for the same procedure. It would seem that the pattern search method should be considerably faster than the steepest descent method, other things being equal. 

A more sophisticated procedure for locating the energy minima is the Newton-Raphson method, and variants thereof. In this method 

---

Atomic and molecular simulation methods can generally be categorized as either equilibrated or dynamic. Static simulations attempt to determine the structural and thermodynamic properties such as crystal structure, sorption isotherms, and sorbate binding. Structural simulations are often carried out using energy minimization schemes that are similar to molecular mechanics. Elquihbrium prop>erties, on the other hand, are based on thermodynamics and thus rely on statistical mechanics and simulating the system state function. Monte Carlo methods are then used to simulate these systems stochastically. 

Toulouse and Umrigar have recently developed a Monte Carlo energy minimization scheme that allows minimization of Cl, orbital, and Jastrow parameters. The method is based on a linear expansion of the normalized wave function (f>(x p) in terms of the parameters p = ... 

The basic backpiopagation algorithm described above is, in practice, often very slow to converge. Moreover, just as Hopfield nets can sometimes get stuck in undesired spurious attractor states (i.e. local minima see section 10.6.5), so to can multilayer perceptrons get trapped in some undesired local minimum state. This is an unfortunate artifact that plagues all energy (or cost-function) minimization schemes. 

In the literature we may find the procedure for creating localized Hartree-Fock orbitals via an energy minimization based on a Cl procedure employing monoexcitations (see for instance Reference [24]). The scheme starts from a set of given (guess) orbitals and solves iteratively the Hartree-Fock equations via the steps ... 

As discussed in many previous studies of biomolecules, the treatment of electrostatic interactions is an important issue [69, 70, 84], What is less widely appreciated in the QM/MM community, however, is that a balanced treatment of QM-MM electrostatics and MM-MM electrostatics is also an important issue. In many implementations, QM-MM electrostatic interactions are treated without any cut-off, in part because the computational cost is often negligible compared to the QM calculation itself. For MM-MM interactions, however, a cut-off scheme is often used, especially for finite-sphere type of boundary conditions. This imbalanced electrostatic treatment may cause over-polarization of the MM region, as was first discussed in the context of classical simulations with different cut-off values applied to solute-solvent and solvent-solvent interactions [85], For QM/MM simulations with only energy minimizations, the effect of over-polarization may not be large, which is perhaps why the issue has not been emphasized in the past. As MD simulations with QM/MM potential becomes more prevalent, this issue should be emphasized. 

Similar schemes to the above can be used in molecular dynamics simulations in other ensembles such as those at constant temperature or constant pressure (see Frenkel and Smit, and Allen and Tildesley (Further reading)). A molecular dynamics simulation is computationally much more intensive than an energy minimization. Typically with modern computers the real time sampled in a simulation run for large cells is of the order of nanoseconds (106 time steps). Dynamical processes operating on longer time-scales will thus not be revealed. 

A semiempirical predictor of homoaromaticity has been developed based on the interactions between atoms obtained from an energy partitioning scheme (Williams et al., 1988). This technique correlates the energy lowering two-centre interactions of two non-bonded atoms with homoaromaticity. A second part of the predictor is the demonstration of the necessity of including at least a minimal 2x2 configuration interaction (Cl) treatment. This semiempirical predictor has been verified by correctly interpreting the interactions in cycloheptatriene [5], 1,6-... 

Fig. I. (Left) Numbering of vinblastine-type alkaloids in this chapter according to the biogenetic scheme of LeMen and Taylor (134), with equivalent atoms in all synthetic intermediates equally labeled. (Right) Approximation of computer-generated, energy-minimized structure, obtained with the Clark Still MACROMODEL program.

Scheme 2.9 Schematic chemical structures of PTHFM and P3MTHFM. Between squares the dipoles on the one-unit model on the two polymers evaluated using PM3 (converge limit 0.01) for energy minimization and Molecular Mechanic (MM +) force field for molecular dynamic at 300 K. (From ref. [64])...

---