Linear systems, partial solution

In principle, the task of solving a linear algebraic systems seems trivial, as with Gauss elimination a solution method exists which allows one to solve a problem of dimension N (i.e. N equations with N unknowns) at a cost of O(N ) elementary operations [85]. Such solution methods which, apart from roundoff errors and machine accuracy, produce an exact solution of an equation system after a predetermined number of operations, are called direct solvers. However, for problems related to the solution of partial differential equations, direct solvers are usually very inefficient Methods such as Gauss elimination do not exploit a special feature of the coefficient matrices of the corresponding linear systems, namely that most of the entries are zero. Such sparse matrices are characteristic of problems originating from the discretization of partial or ordinary differential equations. As an example, consider the discretization of the one-dimensional Poisson equation... 

Summarizing, we conclude that the problem of construction of P(l,3)-invariant ansatzes reduces to finding solutions of linear systems of first-order partial differential equations that are integrated by rather standard methods of the general theory of partial differential equations. 

Reprint H is a late paper written as a festschrift paper in honor of Davidson s retirement. It combines the continuous mixture techniques with the model reduction method given in Chapter 2, the section entitled Scaling and Partial Solution in Linear Systems. A closely similar paper was requested... 

Steady state mass or heat transfer in solids and current distribution in electrochemical systems involve solving elliptic partial differential equations. The method of lines has not been used for elliptic partial differential equations to our knowledge. Schiesser and Silebi (1997)[1] added a time derivative to the steady state elliptic partial differential equation and applied finite differences in both x and y directions and then arrived at the steady state solution by waiting for the process to reach steady state. [2] When finite differences are applied only in the x direction, we arrive at a system of second order ordinary differential equations in y. Unfortunately, this is a coupled system of boundary value problems in y (boundary conditions defined at y = 0 and y = 1) and, hence, initial value problem solvers cannot be used to solve these boundary value problems directly. In this chapter, we introduce two methods to solve this system of boundary value problems. Both linear and nonlinear elliptic partial differential equations will be discussed in this chapter. We will present semianalytical solutions for linear elliptic partial differential equations and numerical solutions for nonlinear elliptic partial differential equations based on method of lines. 

The variational equations (12) are a system of four linear differential equations with time dependent coefficients. If the solution x t) is T-periodic, then the partial derivatives are also T-periodic. In this latter case the system of variational equations is a linear system with periodic coefficients. The theory related to the study of such systems is the Flo-quet theory and some elements of it will be presented in the following sections. 

The fourth and final step in the stability analysis is the reduction of the linear system of partial differential equations to a system of ordinary linear differential equations, the solution of which, subject to the appropriate boundary conditions, yields the eigenfunction

associated complex wave velocity c. 

Network calculations are tasks which occur in many areas of engineering technology. Partial solutions should be chosen only, if they may be incorporated into globally valid concepts. Reasonably chosen intermediate steps of the calculation process may stiU be used. For direction observations, i.e., in traverse networks, it is advisable to analyse the directions first in order to get a suitable linear system of interconnected directions as described. In the next step an extension to the global network can be made in order to get coordinate values. Generating the topology of the network can be done implicitly using only the observational data. 

This is a time-dependent, linear (hyperbolic) partial differential equation. The steady state solutions of this equation (defined by 9 /= 0) are of the form g(z, f) = (p z) where cp satisfies Cf(p z) = 0. These are just the first integrals of the ordinary differential equation system z = /(z). 

Here, Bn is aa N x N tridiagonal matrix, and In denotes the V X V identity matrix (see Section II.D). This example demonstrates how the finite difference method reduces the solution of partial differential equations to the solution of linear systems [Eq. (1)] by replacing derivatives by divided differences. The matrices of the resulting linear systems are sparse and well structured. 

In many applications, linear systems have block-band structure. In particular, the numerical solution of partial differential equations is frequently reduced to solving block tridiagonal systems (see Section II. A). For such systems, block triangular factorizations of A and block Gaussian elimination are effective. Such systems can be also solved in the same way as usual banded systems with scalar coefficients. This would save flops against dense systems of the same size, but the algorithms exploiting the block structure are usually far more effective. 

Special linear systems arise from the Poisson equation, d uldx + d uldy = f x, y) on a rectangle, 0 Laplace equation of Section II.A is a special case where fix, y) = 0.] If finite differences with N points per variable replace the partial derivatives, the resulting linear system has equations. Such systems can be solved in 0(N log N) flops with small overhead by special methods using fast Fourier transform (FFT) versus an order of AC flops, which would be required by Gaussian elimination for that special system. Storage space also decreases from 2N to units. Similar saving of time and space from O(N ) flops, 2N space units to 0(N log N) flops and space units is due to the application of FFT to the solution of Poisson equations on a three-dimensional box. 

Figure 6.92 shows the analytical solution of problem of optimal tank, the system of two non-linear equations (partial derivatives of the function S with respect to arguments R and H) that returns 12 roots one of which (R= 0.87776 and L = 0.78512) is the coordinate of the deepest place in Fig. 6.91. 

Abstract Chapter 5 provides an examination of the numerical solutions of the dyeing models that can be applied to different conditions. Numerical simulation of the system involves the use of Matlab software to solve systems of highly non-linear simultaneous coupled partial differential equations. The finite difference and finite element methods are introduced The partition of the fibrous assembly geometry into small units of a simple shape, or mesh, is examined. Polygonal shapes used to define the element are briefly described. The defined geometries, boundary conditions, and mesh of the system enable solutions to the equations of flow or mass transfer models. 

With Gaussian elimination and partial pivoting, we have a method for solving linear systems that either finds a solution or fails under conditions in which no unique solution exists. In this section, we consider at more depth the question of when a linear system possesses a real solution (existence) and if so, whether there is exactly one (uniqueness). These questions are vitally important, for linear algebra is the basis upon which we build algorithms for solving nonlinear equations, ordinary and partial differential equations, and many other tasks. 

An alternative procedure is the dynamic programming method of Bellman (1957) which is based on the principle of optimality and the imbedding approach. The principle of optimality yields the Hamilton-Jacobi partial differential equation, whose solution results in an optimal control policy. Euler-Lagrange and Pontrya-gin s equations are applicable to systems with non-linear, time-varying state equations and non-quadratic, time varying performance criteria. The Hamilton-Jacobi equation is usually solved for the important and special case of the linear time-invariant plant with quadratic performance criterion (called the performance index), which takes the form of the matrix Riccati (1724) equation. This produces an optimal control law as a linear function of the state vector components which is always stable, providing the system is controllable. 

---