Thanks to visit codestin.com
Credit goes to www.scribd.com

Scribd
Upload
116 views13 pages

Stability and Error Analysis in ODEs

The document discusses error, stability, and stiffness in the numerical solution of differential equations. It defines local/truncation error and global error, and how stability depends on the condition of the ODE and growth of errors over multiple steps. Explicit methods like Euler's method require a step size restriction for stability, while implicit methods can be A-stable with no step size limit. Stiff ODEs pose difficulties due to rapidly decaying transients requiring very small step sizes. Choosing a scheme involves convergence rate, applicability, and simplicity.

Uploaded by

Amit Kumar
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
116 views13 pages

Stability and Error Analysis in ODEs

The document discusses error, stability, and stiffness in the numerical solution of differential equations. It defines local/truncation error and global error, and how stability depends on the condition of the ODE and growth of errors over multiple steps. Explicit methods like Euler's method require a step size restriction for stability, while implicit methods can be A-stable with no step size limit. Stiff ODEs pose difficulties due to rapidly decaying transients requiring very small step sizes. Choosing a scheme involves convergence rate, applicability, and simplicity.

Uploaded by

Amit Kumar
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 13

Error and stability analysis of DE

U. Saravanan
Error
 Local or Truncation error
 Global error

In one step methods if the truncation error is


O(hk+1) global error is O(hk)
Ill conditioned ODE
 If the errors at one step of the computation
gets magnified irrespective of the method
used, such ODEs are called ill-conditioned

 Eg: If y’ = 3y – t2 and y(0) = 2/27 + d


then y = de3t + t2/3 + 2t/9 + 2/27
Exponential term is called the parasitic
solution
Stability of the numerical
procedure
 A numerical method is stable if the errors incurred at
one stage of the process do not tend to be magnified
at later stages for ODE that are not ill-conditioned

 This usually involves investigation of the error for the


problem y’ = ay

 If the method is unstable for the model equation, the


method is considered to be unstable in general
Stability of the numerical
procedure
 A-Stable: A method is A-stable if any solution
produced when the method is applied to the
problem y’ = ay (with a < 0) tends to zero as x
tends to infinity

 Weakly stable: A method is called weakly stable


if it yields unstable solutions when a < 0
Stability

 Consider the explicit method to solve the ODE:


y’ = -4y with IC: y(0) = 1
Stability – Euler method
Let y1 = y0 + hf0

Y Y

x x
h = 0.5 h = 0.05
Stability – Euler method
Let y1 = y0 + hf0 and y’ = 4y

Y Y

x
x
h = 0.01 h = 0.1
Stability

A strongly stable method is stable for y’ = ay


regardless of the sign of ‘a’

 Euler’s method requires: -2 < ah < 0


 Implicit 2 step method: -Inf < ah < 0
Stability – Euler method

 Consider Euler method:


yi+1 = yi+hf(xi,yi)
and the ODE: y’ = -4y with IC: y(0) = 1
Then h < 0.5 for stability
Stability – Euler method

Y Y

x x
h = 0.5 h = 0.05
Stiff ODE
 An ODE in which there is a rapidly decaying
transient solution causing difficulties for numerical
solution, requiring an extremely small step size in
order to obtain an accurate solution is called a stiff
ODE.

 Eg: If u’ = 98u + 198v, v’ = -(99u + 199v) and u(0)


= 1, v(0) = 0, then u = 2e-t – e-100t, v = -e-t + e-100t
Criteria for choosing a scheme

 Faster convergence – Less error in the


solution for the same number of function
evaluations
 Convergence for a broad class of ODE
 No restriction on the step size, h
 Simpler the better – Explicit vs. implicit
schemes

You might also like