You are on page 1of 10

LU-Decomposition

Another way of solving a system of


equations is by using a factorization
technique for matrices called LU
decompostion. This factorization is
involves two matrices, one lower
triangular matrix and one upper
triangular matrix.
LU-Decomposition

A m×n matrix is said to have a LU-Decomposition if there


exists matrices L and U with the following
properties:
1. A square matrix B (bij) is said to be lower Triangular bij
= 0 for all i<j.
2. A square matrix B (bij) is said to be unit lower Triangular
if it is lower triangular and each bii = 1.
3. A square matrix B (bij) is said to be upper Triangular if
bij = 0 for all i>j.
Suppose we want to solve a  m×n  system AX = b.
If we can find a LU-decomposition for A , then to solve
AX =b, it is enough to solve the systems
                
Thus the system LY = b can be solved by the method of
forward substitution and the system UX = Y
can be solved by the method of backward substitution.
To illustrate, with the following example:
How to find an
LU-Decomposition
1. Row swapping is not allowed. If you swap rows, then an
LU decomposition will not exist.
2. It is not necessary to get leading ones on the main
diagonal when using Gaussian Elimination. In some
matrices, however, it is recommended to get leading ones
to use nice row operations.
3. When using Gaussian Elimination to find such an LU
decomposition, record all row operations involved. The
row operations will help find the lower triangular matrix
using the identity matrix.
4. An LU decomposition is not unique. There can be more
than one such LU decomposition for a matrix.
Steps to solve a system of equations
by the LU-Decomposition

1. Get the lower triangular matrix L and upper


triangular matrix U.
2. Solve Ly = b (to find y).
3. The result of the previous step is saved in a new
array named "y".
4. Make Ux = y (to find x).
5. The result of the previous step is stored in a new
array called "x", which provides the values for the
unknowns of the equation.
Example. Solve the following system using an
LU-Decomposition.
2 8 1 1
4 13 3  1
A 
 2  5  3 3 
 
  6  18  1 1 
1. Set up the equation A = LU.

2 8 1 1 1 0 0 0  2 8 1 1
4 13 3  1 2 1 0 0 0  3 1  3
A    
 2  5  3 3    1  1 1 0  0 0  1 1 
    
  6  18  1 1   3  2  4 1 0 0 0 2
2. This will yield the equation:
 (LU)x = b.

1 0 0 0  2 8 1 1   x1   1 
2
 1 0 0 0  3 1  3  x2   0 

 1 1 1 0 0 0  1 1   x3   2
     
 3  2  4 1  0 0 0 2   x4   1 

L U b
and Ly = b

1 0 0 0  y1   1 
2
 1 0 0  y2   0 

 1 1 1 0   y3    2 
    
 3  2  4 1  y4   1 

Now solving for y gives the following values:

y1 1 y1  1
2 y1  y2  0  y2   2
 y1  y2  y3 6 y3   3
 3 y1  2 y2  4 y3  y4 1 y4   12
4. Take the values for y and solve the equation Ux=y for x. This
will give the solution to the system Ax = b.

2 8 1 1   x1   1 
0  3 1  3  x    2 
  2    
0 0  1 1   x3    3 
    
0 0 0 2   x4   12

Therefore, the solution to the system is x1 = -53/3, x2 = 17/3 , x3=-3


and x4 = -6.
2 x1  8 x2  x3  x4  1 x1   53 3
 3 x2  x3  3 x4   2  x2  17 3
 x3  x4   3 x3   3
2 x4  12 x4   6
BIBLIOGRAPHY

 http://www.monografias.com/trabajos45/descomposi
cion-lu/descomposicion-lu.shtml

You might also like