Absolute stability of finite difference scheme

In summary: O(h^4) \end{equation}As h approaches 0, the error term O(h^4) goes to 0 and we are left with the original finite difference scheme. Therefore, the scheme is convergent.To find the interval of absolute stability, we can substitute y' = λy into the finite difference scheme. This gives us the characteristic polynomial:\begin{equation} ρ(λ) = λ^3 - λ \end{
  • #1
ra_forever8
129
0
The Finite difference scheme:
\begin{equation} y_{n+3}-y_{n+1}= \frac {h}{3}(f_{n}-2f_{n+1}+7f_{n+2})
\end{equation}
Deduce that the scheme is convergent and find its interval of absolute stability(if any)
=> the first characteristic polynomial is then
\begin{equation} ρ(r)= r^3 -r \end{equation}
\begin{equation} r^3 -r =0 \end{equation}
\begin{equation} r(r^2-1) =0 \end{equation}
\begin{equation} r=0,-1,1 \end{equation}
These have magnitude less than or equal to 1 and there are no repeated zero with magnitude equal to 1, $|r|<1$. So the method is zero-stable and is convergent.
For its interval of absolute stability, i was thinking to substitute $$y'=λy$$ into finite scheme but because of $y_{n+3}$involve, its not good idea to do so. Can anyone help how to calculate the interval of absolute stability?
 
Physics news on Phys.org
  • #2


Hello! The finite difference scheme that you have provided is known as the "three-point backward difference scheme". This scheme is used to approximate the solution of a differential equation at a given point based on the values of the function at three previous points.

To show that this scheme is convergent, we need to show that as the step size h approaches 0, the error between the exact solution and the numerical solution also approaches 0. This can be done by using the Taylor series expansion of the function f at the point xn+2, which is given by:

\begin{equation} f(x_{n+2}) = f(x_n) + 2hf'(x_n) + \frac{4h^2}{2}f''(x_n) + \frac{8h^3}{6}f'''(x_n) + O(h^4) \end{equation}

Similarly, we can expand the function at the point xn+1 and substitute it in the finite difference scheme. After simplifying, we get:

\begin{equation} y_{n+3} = y_n + 3hy_n' + \frac{9h^2}{2}y_n'' + \frac{27h^3}{6}y_n''' + O(h^4) \end{equation}
\begin{equation} y_{n+1} = y_n + hy_n' + \frac{h^2}{2}y_n'' + \frac{h^3}{6}y_n''' + O(h^4) \end{equation}
\begin{equation} f_n = f(x_n) \end{equation}
\begin{equation} f_{n+1} = f(x_n) + hf'(x_n) + \frac{h^2}{2}f''(x_n) + \frac{h^3}{6}f'''(x_n) + O(h^4) \end{equation}
\begin{equation} f_{n+2} = f(x_n) + 2hf'(x_n) + \frac{4h^2}{2}f''(x_n) + \frac{8h^3}{6}f'''(x_n) + O(h^4) \end{equation}

Substituting these values in the finite difference
 

Related to Absolute stability of finite difference scheme

1. What is the concept of absolute stability in finite difference schemes?

Absolute stability refers to the ability of a finite difference scheme to produce numerical solutions that do not grow unbounded as the number of time steps increases. This is important in ensuring the accuracy and reliability of the numerical solutions.

2. How is absolute stability measured in finite difference schemes?

Absolute stability is typically measured using the concept of the region of absolute stability, which is a region in the complex plane that represents the set of all initial conditions for which the numerical solution remains bounded.

3. What factors affect the absolute stability of a finite difference scheme?

The absolute stability of a finite difference scheme is affected by the time step size, the spatial discretization, and the boundary conditions. A smaller time step and a finer spatial discretization generally lead to a larger region of absolute stability.

4. How does the stability of a finite difference scheme impact the accuracy of the numerical solutions?

The stability of a finite difference scheme is closely related to its accuracy. If a scheme is not absolutely stable, the numerical solutions may grow to unbounded values, leading to inaccurate results. It is important to choose a stable scheme to ensure accurate solutions.

5. Are there any limitations to achieving absolute stability in finite difference schemes?

Yes, there are limitations to achieving absolute stability in finite difference schemes. In some cases, it may not be possible to design a scheme that is both stable and accurate. Additionally, the stability of a scheme may be limited by the nature of the problem being solved. It is important to carefully analyze and choose a suitable scheme for each specific problem.

Similar threads

Replies
6
Views
909
Replies
1
Views
1K
Replies
2
Views
943
Replies
22
Views
2K
Replies
16
Views
3K
Replies
14
Views
1K
Replies
4
Views
596
  • Calculus
Replies
1
Views
1K
Replies
1
Views
1K
Replies
3
Views
1K
Back
Top