Numerical Methods To Solve Initial

Value Problems
An Over View of Runge-Kutta Fehlberg and
Dormand and Prince Methods.

William Mize
Quick Refresher
 We are looking at Ordinary Differential Equations
More specifically Initial Value Problems

Simple Examples:
 Solution of:

 Solution of:
A Problem
 How practical are analytical methods?
Equation:

We chose to find a Numerical solution because

Closed-form is to difficult to evaluate
No close-form solution
Some Quick Ground work
First Start with Taylor Series
Approximations
Then Move onto Runge-Kutta
Methods for Approximations
Lastly onto Runge-Kutta Fehlberg and
Dormand and Prince Methods for
Approximation and keeping control of
error
How these Methods Work
All of the Methods will be using a step
size method.
Error is determined by the size of step,
order, and method used.
When actually calculating these,
almost always done via computer.
Taylor Series Methods(Brief)
 Taylor Series As Follows

Most Basic is Euler’s Method

Higher Order Approximations better Accuracy

But at a cost
What can we do?
Runge-Kutta Methods
Named After Carl Runge and Wilhelm Kutta
What they do?
Do the same Job as Taylor Series Method,
but without the analytic differentiation.
Just like Taylor Series with higher and higher
order methods.
Runge-Kutta Method of Order 4 Well
accepted classically used algorithm.
Runge-Kutta of Order 2
We
  don’t want to take derivatives for approximations
Instead use Taylor series to create Runge-Kutta methods to
approximate solution with just function evaluations.

We Want to Approximate this with

 Find A, B, C

We get:

 Error
Runge-Kutta of Order 4
 

Error of Order
So What's next?
Already Viable Numerical Solution established what's the
next step?
We want to control our Error and Step size at each step.
These methods are called adaptive.
Why?
Cost Less
Keep within Tolerance

Also look for More efficient ways of doing these things.

10 Function Evaluation for RK4 and RK5
Just 6 for RKF4(5)
Runge-Kutta Fehlberg
 

Coefficients are found via Taylor expansions

Next Step to find These Coefficients
Further Deriving
 We assume =1
More and more…
 

So this was way more complicated than I
actually thought it would be.
But it’s all leading some where!
Eventually we want to have all the in terms
of
From there was must figure out our and
 How to find
First Take coefficients from the 5th order equation.
 

Which ultimately leads to

Where we chose = 1/3 and = 3/8

 = 1/3
 = 3/8
Comparison(Problem)
Comparisons of Methods
Dormand and Prince Methods
Visual Comparison of Methods
Conclusion
Taylor’s method uses derivatives to solve ODE
RK uses only a combination of specific function
evaluations instead of derivatives to approximate
solution of the ODE
RKF is beneficial because you can control your step
size so you have your global error within a
predetermined tolerance
RK4 and RK5 uses 10 function evaluations vs RKF
just 6
Runge-Kutta Fehlberg is widely accepted and used
commercially(Matlab, Mathematica, maple, etc)
Sources
Numerical Mathematics and Computing. Sixth
Edition; Ward Cheny, David Kincaid
Low-Order classical Runge-Kutta Formulas with
StepSize Control and their Application to some heat
transfer problems. By Erwin Fehlberg(1969)
A family of embedded Runge-Kutta Formulae. By
Dormand and Prince(1980)