Higher Order Approximation Techniques

There are several methods that we can use to increase the order of approximation of the integral in eqn. (8.72). Two of the most common higher order explicit methods are the Adams-Bashforth (AB2) and the Runge-Kutta of second and fourth order. The Adams-Bashforth is a second order method that uses a combination of the past value of the function, as in the explicit method depicted in Fig. 8.19, and an average of the past two values, similar to the Crank-Nicholson method depicted in Fig. 8.21, and written as 

The method increases the order but the stability is compromized due to the extrapolation done by the the linear approximation between the previous times. This stability issue can be improved by adding an extra implicit step using an Adams-Moulton (AM2) as follows 

Runge-Kutta of the second (RK2) and fourth (RK4) order do not use an extrapolation between k — 1 and k to find fell, instead they use future points to do the extrapolation. The methodology is straight forward and can be summarized for the Runge-Kutta of second order (RK2) as [10] 

These Runge-Kutta methods do not require information from the past, and are very versatile if the time steps need to be adjusted as the solution evolves. The stability of the RK2 is similar to the APC2, while the RK4 has less strong conditions for stability [10]. Both are ideal for initial value problems in time or in space, 

For the flow components that are significant in the present geometry, the important components of the stress tensor Oij = —S p + r, can be expressed as, 

---

These conditions are not local (they are integral relations difficult to incorporate in a numerical code). By perturbation techniques one gets local approximations which are partial differential equations on the boundary. Higher order approximations can be obtained at the price of increasing difficulty in the computations. 

An essential factor in the construction of accurate higher order FDTD techniques is the correct stencil manipulation provided by the respective spatial operators. Thus, apart from the most frequently encountered schemes, discussed in the previous chapters, a variety of rigorous approaches have also been developed [24-30]. As an indication of these interesting trends, this section presents three algorithms which, based on different principles, attempt to improve the behavior of higher order spatial sampling and approximation. 

The GGN method is based on a linearization of Fi and the constraints in (2.3) around the current estimate Wk, but yields a quadratic approximation of the optimization problem (rather than just a linear one). Accordingly, the method is superior to both derivative-free and higher-order optimization techniques in this context. Furthermore, it does not converge to minima which are statistically not significant, i.e., minima with large residuals [4, 7]. 

Martinez-Ortiz E, Zoroa N, Molina A, Serna C, Laborda E (2009) Electrochemical digital simulations with an exponentially expanding grid general expressions for higher order approximations to spatial derivatives. The special case of four-point formulas and their application to multipulse techniques in planar and any size spherical electrodes. Electrochim Acta 54 1042-1055... 

Ion exchange (qv see also Chromatography) is an important procedure for the separation and chemical identification of curium and higher elements. This technique is selective and rapid and has been the key to the discovery of the transcurium elements, in that the elution order and approximate peak position for the undiscovered elements were predicted with considerable confidence (9). Thus the first experimental observation of the chemical behavior of a new actinide element has often been its ion-exchange behavior—an observation coincident with its identification. Further exploration of the chemistry of the element often depended on the production of larger amounts by this method. Solvent extraction is another useful method for separating and purifying actinide elements. 

The technique of using the damp ratio hne 0 = cos in Eq. (2-34) is apphed to higher order systems. When we do so, we are implicitly making the assumption that we have chosen the dominant closed-loop pole of a system and that this system can be approximated as a second order underdamped function at sufficiently large times. For this reason, root locus is also referred to as dominant pole design. 

These equations were developed further using similar techniques to those already discussed. The more detailed analysis of liquid structures required to describe the recombination process, than the homogeneous reaction, requires higher-order equations for the liquid structure to be used. This necessarily means that approximations have to be made [286]. 

Numerical solutions of the Schrodinger equation can be obtained within several degrees of approximation, for almost any system, using its exact Hamiltonian. Density functional theory has proven to be one of the most effective techniques, because it provides significantly greater accuracy than Hartree-Fock theory with just a modest increase in computational cost.io> 3-45 The accuracy of DFT method is comparable, and even greater than other much more expensive theoretical methods that also include electron correlation such as second and higher order perturbation theory. 

In the literature a different technique has been widely used to construct effective Hamiltonians, based on the partitioning technique combined with an approximation procedure known as adiabatic elimination for the time-dependent Schrodinger equation (see Ref. 39, p. 1165). In this section we show that the effective Hamiltonian constructed by adiabatic elimination can be recovered from the above construction by choosing the reference of the energy appropriately. Moreover, our stationary formulation allows us to estimate the order of the neglected terms and to improve the approximation to higher orders in a systematic way. 

Higher-order models are rarely applied. In many cases, the true response surface can be sufficiently well approximated by the second-order model. Occasionally, higher-order models can be used when quadratic models are clearly inadequate, for example, when a sigmoid-like relation between the response and a variable is observed (7). Then, either a third-order model, an appropriate transformation, a mechanistic physical model, nonlinear modeling techniques, or neural networks can be applied (1,7). 

---