Breaking down a numerical integration into two parts
1 vue (au cours des 30 derniers jours)
Afficher commentaires plus anciens
Sorry for this stupic question.
Suppose we have this integral
![](https://www.mathworks.com/matlabcentral/answers/uploaded_files/1636431/image.png)
Of couse this can be solved analytically. But I still want to solve it numerically as:
dx = 0.01;
x = 1:dx:4;
fx = x.^2 + 5*x;
I = trapz(x,fx)
Now, we break down this integral into two part. Analytically, we write it as:
![](https://www.mathworks.com/matlabcentral/answers/uploaded_files/1636436/image.png)
I am confused about the lower limit of the second integral
when solving it numerically. Should it be 2 or
. If we use 2, will not it mean that we are counting one point two times, once in
and once in
?
![](https://www.mathworks.com/matlabcentral/answers/uploaded_files/1636441/image.png)
![](https://www.mathworks.com/matlabcentral/answers/uploaded_files/1636446/image.png)
![](https://www.mathworks.com/matlabcentral/answers/uploaded_files/1636451/image.png)
![](https://www.mathworks.com/matlabcentral/answers/uploaded_files/1636456/image.png)
0 commentaires
Réponse acceptée
Torsten
le 6 Mar 2024
I1 = dx/2 * f(x0) + dx * sum_{i=1}^{n-1} f(xi) + dx/2 * f(xn)
I2 = dx/2 * f(xn) + dx * sum_{i=n+1}^{n+m-1} f(xi) + dx/2 * f(x_{n+m})
->
I1 + I2 = dx/2 * f(x0) + dx * sum_{i=1}^{n+m-1} f(xi) + dx/2 * f(x_{n+m})
Thus since the endpoints in the trapezoidal rule are only counted half, all is fine.
Plus de réponses (0)
Voir également
Catégories
En savoir plus sur Eigenvalue Problems dans Help Center et File Exchange
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!