Investigation of a new method for numerical integration
Mark A. Pollock and Stanley H. Jury have suggested a new approach to numerical integration which deviates from the usual interpretation that a definite integral is an area of some bounded region in the plane. An investigation of this method for numerical integration has been done.
First we recall some of the classical methods for integration over a finite and infinite intervals, as well as methods for integrating functions with singularities. Then we establish the theory behind the mathematical aspect of the new method without concern for its physical interpretation. The method involves the construction and solution of finite difference equations. The coefficient matrix has another special structure aside from being tri-diagonal which gives us the motivation to derive a more efficient method of solution other than the Thomas algorithm for tri-diagonal systems. During the iterative process the entries in the matrix coefficient are determined by application of simple quadrature rules in each increment.
Lastly, we present several numerical results in Chapter 3 to point out the merits or disadvantages of the algorithm.
Thesis77.P389.pdf
4.78 MB
Adobe PDF
a3e373f2141111ab0365de8d883df7da