- #1
ThorX89
- 13
- 0
Hi.
Anyway, learning this sort of makes me feel like I've chosen the wrong school for myself, but I'd like to try and see if I can understand this nevertheless.
See, I'm supposed to learn various numerical methods for solving systems of linear equations, like the Jacobi or the Gauss-Seidel method, and I looked them both up on wikipedia and I understand them up to the point where we get an equation expressing the x vector as a simple function of the same x vector. What I don't get, however, is what would make anyone think that choosing an initial value for the x vector on the right hand side, computing the left hand side from it, and then plugging that to the right hand side again and repeating that over and over again will create a series of x_(k+1)=f(x_k) that will eventually converge.
I mean I could easily derive the
x=f(x) parts
(that is: http://upload.wikimedia.org/math/4/e/f/4ef16455a2e751a078ce9dff8d4d9ad0.png
http://upload.wikimedia.org/math/f/5/1/f5137bf3761d2757a21824fbfc9d7c30.png
in particular for Jacobi or Gauss-Seidel method respectively)
but never would I have thought of deriving an iterative method for getting the values for x from this.
I know it doesn't converge everytime - What I'm asking is what would make anyone think it ever will or in other words: that creating series from the above mentioned equations will lead to the solution.
Thanks if anyone is willing to explain this to me.
Anyway, learning this sort of makes me feel like I've chosen the wrong school for myself, but I'd like to try and see if I can understand this nevertheless.
See, I'm supposed to learn various numerical methods for solving systems of linear equations, like the Jacobi or the Gauss-Seidel method, and I looked them both up on wikipedia and I understand them up to the point where we get an equation expressing the x vector as a simple function of the same x vector. What I don't get, however, is what would make anyone think that choosing an initial value for the x vector on the right hand side, computing the left hand side from it, and then plugging that to the right hand side again and repeating that over and over again will create a series of x_(k+1)=f(x_k) that will eventually converge.
I mean I could easily derive the
x=f(x) parts
(that is: http://upload.wikimedia.org/math/4/e/f/4ef16455a2e751a078ce9dff8d4d9ad0.png
http://upload.wikimedia.org/math/f/5/1/f5137bf3761d2757a21824fbfc9d7c30.png
in particular for Jacobi or Gauss-Seidel method respectively)
but never would I have thought of deriving an iterative method for getting the values for x from this.
I know it doesn't converge everytime - What I'm asking is what would make anyone think it ever will or in other words: that creating series from the above mentioned equations will lead to the solution.
Thanks if anyone is willing to explain this to me.