You are on page 1of 18

NUMERICAL SOLUTION

OF ORDINARY
DIFFERENTIAL
EQUATIONS
TAYLOR SERIES EXPANSION
EULER / MODIFIED EULER EXPANSION
RUNGE-KUTTA METHODS
NUMERICAL SOLUTION OF ORDINARY DIFFERENTIAL
EQUATIONS
Ordinary Differential Equations is reputed to be one
of the most beautiful parts of mathematics.
It is also an essential tool for modeling many
physical solutions:
Electrical Networks
Mechanical Networks
Bending Beams
Chemical Reactions
Pendulums
The Motion of Rotating Mass Around Another
Body
NUMERICAL SOLUTION OF ORDINARY DIFFERENTIAL
EQUATIONS

The prominence of Ordinary Differential


Equations in Applied Mathematics is due
to the fact that most scientific laws are
more readily expressed in terms of rates of
change.
The solution to a Differential Equation is the
function that satisfies the Differential
Equation and that of the corresponding
initial conditions of the function.
NUMERICAL SOLUTION OF ORDINARY DIFFERENTIAL
EQUATIONS

In solving Differential Equations analytically,


we usually find a general solution
containing arbitrary constants and then
evaluate the general solution to obtain the
values of the arbitrary constants so that
the expression agrees with the initial
conditions.
Unfortunately, the analytical methods are
limited to certain forms of equations only.
NUMERICAL SOLUTION OF ORDINARY DIFFERENTIAL
EQUATIONS

Elementary Differential Equations normally


treat only:
linear equations with constant coefficients
when the degree of the equation is higher
than the first order.
specialized form of linear equations with
variable coefficients known as Euler
Equations or Euler-Cauchy Equations.
NUMERICAL SOLUTION OF ORDINARY DIFFERENTIAL
EQUATIONS

Numerical Methods have no such


limitations to standard forms.
The solution is obtained as a tabulation
of the values of the function at
various values of the independent
variable, and not as a functional
relationship.
NUMERICAL SOLUTION OF ORDINARY DIFFERENTIAL
EQUATIONS

Advantage:
It can solve practically any Differential
Equation; provided the appropriate
method/technique is used.

Disadvantage:
We must recompute the entire table if the
initial conditions are changed!
TAYLOR SERIES EXPANSION
PROBLEM:
dy
y' F ( x , y )
dx
PROCEDURE:
Take the problem & differentiate
repeatedly until it will equal to zero or
until you reach the specified degree
or repetition.
Apply results to the solution for the
nth-order Taylors Series Expansion.
TAYLOR SERIES EXPANSION
SOLUTION:
h2 h3
yi 1 yi hF ( xi , yi ) F' ( xi , yi ) F" ( xi , yi )
2! 3!
h n n 1
... F ( xi , yi )
n!
h2 h3 hn
yi hyi ' yi ' ' yi ' ' ' ...
n'
yi 1 yi
2! 3! n!

Where h = step size


Disadvantage: Requires derivative
computations which are cumbersome.
EULER / MODIFIED EULER
METHOD
As we have discovered in the previous method,
the Taylor Series Expansion may be awkward to apply if
the derivatives become complicated and the error is
difficult to determine.
However, the error in the Taylor Series will be
small if the step size, h, is small. If we make the h small
enough, we will need a few terms of the Expansion to
gain good accuracy.
EULER / MODIFIED EULER
METHOD
The Euler Method follows this idea to the
extreme for the first order Differential Equations it
uses only the first two terms of the Taylor Series. In
simple terms, this is the Taylor Series without the
derivative terms.

PROBLEM:
dy
y' F ( x , y )
dx
EULER / MODIFIED EULER
METHOD
Disadvantage:
It is much less accurate compared to the Taylor
Series Expansion; however, by using small increments
or step sizes, the result is acceptable enough.

SOLUTION:

yi 1 yi hF( xi , yi ) yi hyi '


RUNGE-KUTTA METHOD
The Runge-Kutta Methods are among
the easiest methods in the numerical
solutions of Differential Equations. In most
cases, they are more accurate than the Euler
Method & less tedious than the Taylor Series.
Advantage:
Self-starting feature and consequent
ease of programming.
RUNGE-KUTTA METHOD
Disadvantage:
F(x,y) must be evaluated for several
slightly different values of the x and y in
every step of the solution.
It is more difficult to estimate the per step
error for high order solution than for
solution obtained by some of the multiple
step methods.
RUNGE-KUTTA METHOD
Types
Second-Order Runge-Kutta Method
Third-Order Runge-Kutta Method
Fourth-Order Runge-Kutta MEthod

PROBLEM:
dy
y' F ( x , y )
dx
SECOND-ORDER
RUNGE-KUTTA METHOD

yi 1 yi k1 k 2 yi k1 k 2
1 1 1
2 2 2
k1 hF ( xi , yi )
k 2 hF ( xi h , yi k1 )
THIRD-ORDER
RUNGE-KUTTA METHOD

yi 1 yi k1 4 k 2 k3
1
6
k1 hF ( xi , yi )
h k1
k 2 hF ( xi , yi )
2 2
k3 hF ( xi h , yi k1 2k 2 )
FOURTH-ORDER
RUNGE-KUTTA METHOD
yi 1 yi k1 2k 2 2k3 k4
1
6
k1 hF ( xi , yi )
h k1
k 2 hF ( xi , yi )
2 2
h k2
k3 hF ( xi , yi )
2 2
k4 hF ( xi h , yi k3 )

You might also like