Scribd
Upload
28 views

LU Decomposition

LU decomposition allows solving systems of linear equations without directly inverting the coefficient matrix A. It decomposes A into the product of a lower triangular matrix L and an upper triangular matrix U. This decomposition can be found by writing out the matrix equation A = LU in component form. Once L and U are determined, the system Ax = b can be solved by solving two successive triangular systems of equations. LU decomposition only needs to be calculated once to solve for multiple right-hand sides.

Uploaded by

Prabhat Ray
Copyright
© Attribution Non-Commercial (BY-NC)
Available Formats
Download as PDF, TXT or read online on Scribd
28 views

LU Decomposition

LU decomposition allows solving systems of linear equations without directly inverting the coefficient matrix A. It decomposes A into the product of a lower triangular matrix L and an upper triangular matrix U. This decomposition can be found by writing out the matrix equation A = LU in component form. Once L and U are determined, the system Ax = b can be solved by solving two successive triangular systems of equations. LU decomposition only needs to be calculated once to solve for multiple right-hand sides.

Uploaded by

Prabhat Ray
Copyright
© Attribution Non-Commercial (BY-NC)
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 2

LU decomposition

Although conceptually simple, finding the solution by calculating A1 can be computationally demanding, especially when N is large. In fact, as we shall now show, it is not necessary to perform the full inversion of A in order to solve the simultaneous equations Ax = b. Rather, we can perform a decomposition of the matrix into the product of a square lower triangular matrix L and a square upper triangular matrix U, which are such that17 A = LU, (1.68) and then use the fact that triangular systems of equations can be solved very simply. We must begin, therefore, by finding the matrices L and U such that (1.68) is satisfied. This may be achieved straightforwardly by writing out (1.68) in component form. For illustration, let us consider the 3 3 case. It is, in fact, always possible, and convenient, to take the diagonal elements of L as unity, so we have A= 100 L21 1 0

L31 L32 1

U11 U12 U13 0 U22 U23 0 0 U33


=

U11 U12 U13 L21U11 L21U12 + U22 L21U13 + U23 L31U11 L31U12 + L32U22 L31U13 + L32U23 + U33
. (1.69) The nine unknown elements of L and U can now be determined by equating the nine elements of (1.69) to those of the 3 3 matrix A. This is done in the particular order illustrated in the example below. Once the matrices L and U have been determined, one can use the decomposition to solve the set of equations Ax = b in the following way. From (1.68), we have LUx = b, but this can be written as two triangular sets of equations

Ly = b and Ux = y, where y is another column matrix to be determined. One may easily solve the first triangular set of equations for y, which is then substituted into the second set. The required solution x is then obtained readily from the second triangular set of equations. We note that, as with direct inversion, once the LU decomposition has been determined, one can solve for various RHS column matrices b1, b2, . . . , with little extra work.

You might also like