Simple ODE Solvers - Derivation. (pdf file)

A summary of the three methods is given in

Simple ODE Solvers - Summary. (postscript file)

If you are having trouble understanding the mechanics of executing these algorithms look at

Euler's method. This demo contains an annotated implementation of Euler's method. You can run Euler's method one step at a time. Each step is accompanied by a commentary which shows you the computation done during that step.

A spread sheet implementation of Euler's method (postscript file) One good way to be sure you understand the mechanics of, for example, Euler's method is to implement it on a spread sheet. These notes give an example of such an implementation.

Simple ODE solvers. This demo contains a sample implementation of all three methods. You can simultaneously display the results of all three methods with various step sizes to get some first impressions as to how well the methods work.

Simple ODE Solvers - Error Behaviour. (postscript file) These notes give the results of some numerical experiments designed to determine how the error generated by Euler's method, the improved Euler method and the Runge-Kutta method depend on the step size used.

Error Behaviour - A Trivial Example. (postscript file) The initial value problem y'=y, y(0)=1 is so simple that we can easily determine both the exact solution and the approximate solution generated by Euler and his friends. So we can also determine the error generated.

Roundoff error. (postscript file) These notes give the results of a numerical experiment exploring the effect of roundoff error on Euler's method.

Richardson Extrapolation. (postscript file) Automatic step size adjustment for many different algorithms is based on an idea called Richardson extrapolation, that is described in these notes.

Variable Step Size Methods. (pdf file) These notes show how Richardson extrapolation can be used to develop algorithms for generating numerical solutions to ODEs that automatically select the step size used in each step.

A Simple ODE Solver with Automatic Step Size Adjustment. This demo contains an annotated implementation of a very naive variable step size method. You can run it one step at a time. Each step is accompanied by a commentary which shows you the computation done during that step.

How First Order Systems Arise. (postscript file) These notes discuss the basic mechanisms whereby systems of first order ode's arise. One of these mechanisms is a simple trick that can be used to convert any higher order ode into a first order system. It is easy to use this trick to adapt the above algorithms to handle higher order equations.

[ My home page | UBC Math dept ]