**Iterative Solution of Highly Nonlinear Differential Equations Using ***Mathematica*
**Frank Kampas**
Lockheed Martin
Download talk material:
These are just a few of the many interesting and important problems that were
previously excluded from the teaching curriculum as being too difficult but that can now
be solved with *Mathematica* in a simple and lucid manner accessible to students. The
solutions generated by `NDSolve,` *Mathematica'*s
function for numerical solution of ordinary and partial differential equations, are
(interpolating) functions. This unique feature of* Mathematica* enables the
implementation of iterative solution methods for nonlinear boundary value differential
equations in a straightforward fashion. For example, quasi linearization, a form of
Newton's method, can be used to interactively solve nonlinear boundary value
problems that cannot be handled directly by `NDSolve` or by the
shooting method. In this approach, a linearized form of the differential equation is
solved by the shooting method or by superposition, with each successive solution used as
input for the next iteration until the solution has the desired accuracy. Variational
methods can be used to generate good initial functions for the iterative procedure.
Application of this approach to coupled nonlinear differential equations such as those
encountered in semiconductor transport physics will be discussed.
| |