|
||||||||||
|
It could be said with only a slight exaggeration that everything in physics involves differential equations of one kind or another. From Newton's F=ma to Maxwell's equations to Schrodinger's equation, we see that differential equations provide the essential descriptions of the behavior of physical phenomena. In some cases, we can solve these equations analytically but often we can not. Instead, approximation techniques are required and these become the basis of various numerical algorithms for solving the equations. Even when an analytical solution exists, it can still help in understanding the system to use a numerical approach in a computer program to display the functional behavior graphically. Ordinary Differential Equations We look in this and later chapters at numerical solutions of ordinary differential equations (ODE). The methods discussed here apply to first-order ODEs. However, a higher order ODE can always be converted to a first-order ODE by new variable substitution. For example, in (1) let (2) and then equation (1) becomes (3) Thus with the introduction of the v variable, equations (2) & (3) become two coupled first-order differential equations. In general, the problems will reduce to a set of equations as in the form
The goal is to determine x as a function of t. A numerical solution involves starting from an initial point (x0,t0) and using an algorithm to determine a series of points (xn,tn), each calculated based on the previous point(s). The points are separated in time by a fixed increment dt, or tn+1 = tn + dt. ODE problems also can be divided into the following types:
In this chapter we only look at initial value problems. In Chapter 4 : Physics we discuss boundary value problems. (We do not discuss eignevalue problems in this course.) We start with the simplest method of solving ODE problems: the Euler or constant slope method. References & Web Resources
Latest update: Oct. 21, 2005 |
|
||||||||
|