LU Decomposition: Solving for A

In summary, the conversation discusses the LU decomposition of a non-singular matrix A and how to show that different decompositions will result in the same upper-triangular matrix. The key is to show that L^-1L' is equal to the identity matrix in order to preserve the upper-triangularity. The use of M-matrices is not necessary and it is sufficient to have L and L' as lower-triangular matrices and U and U' as upper-triangular matrices.
Physics news on Phys.org
  • #2
Suppose we have two different LU decompositions, A = LU and A=L'U'. Because A is non-singular L, U, L' and U' are all non-singular and invertible. This implies that [tex]U = L^{-1}L'U'[/tex]. Now you should be able to show that [tex]I = L^{-1}L'[/tex] in order to preserve upper-triangularity.
 
  • #3
cellotim said:
Suppose we have two different LU decompositions, A = LU and A=L'U'. Because A is non-singular L, U, L' and U' are all non-singular and invertible. This implies that [tex]U = L^{-1}L'U'[/tex].
Aha, very smart!

Now you should be able to show that [tex]I = L^{-1}L'[/tex] in order to preserve upper-triangularity.
Well...from my instructors note the L is defined in a product of M-matrices(see picture) but if that's the case then L and L' have different M's (for we presume L and L' to be different).

If I invert a upper triangular matrix it remains upper triangular so I don't see what you mean by "preserving upper-triangularity".
 
  • #4
The M's are not necessary as far as I can see only that L and L' are lower-triangular and U and U' are upper-triangular, and I mean preserve upper-triangularity of U' to U.
 
  • #5
cellotim said:
The M's are not necessary as far as I can see only that L and L' are lower-triangular and U and U' are upper-triangular, and I mean preserve upper-triangularity of U' to U.

So to preserve upper triangularity of U w.r.t. U' L-1 L' should be I?
 
  • #6
It needs to be diagonal. It remains to prove that that diagonal matrix must be the identity.
 
Last edited:

FAQ: LU Decomposition: Solving for A

What is LU decomposition and why is it useful?

LU decomposition is a method used to solve linear equations by breaking down a matrix into a lower triangular matrix and an upper triangular matrix. It is useful because it simplifies the process of solving large systems of equations, making it more efficient and easier to understand.

How is LU decomposition different from other methods of solving linear equations?

LU decomposition is different from other methods of solving linear equations, such as Gaussian elimination or Cramer's rule, because it breaks down a matrix into two triangular matrices, making it easier to solve and more efficient for larger systems of equations.

Is LU decomposition always possible for any given matrix?

No, LU decomposition is not always possible for any given matrix. The matrix must be non-singular, meaning it has a non-zero determinant, and it must not have any zero pivots in the Gaussian elimination process.

What is the process for using LU decomposition to solve for A?

The process for using LU decomposition to solve for A involves first decomposing the original matrix into a lower triangular matrix and an upper triangular matrix. Then, using back substitution, you can solve for the variables in the lower triangular matrix and then use those values to solve for the variables in the upper triangular matrix.

What are the advantages of using LU decomposition over other methods of solving linear equations?

The advantages of using LU decomposition over other methods of solving linear equations include its efficiency for larger systems of equations, its ability to handle non-square matrices, and the fact that it can be used for repeated solutions with different right-hand sides without needing to repeat the decomposition process.

Similar threads

Replies
6
Views
9K
Replies
5
Views
3K
Replies
5
Views
2K
Replies
2
Views
1K
Replies
1
Views
4K
Replies
2
Views
3K
Back
Top