Course Title: Numerical Analysis.
Course code: MATH_601.
.
Presented By :
Manahil
(2021_GCBM_371). Presented To :
Laiba Kareem
Lecturer Abdul Basit
(2021_GCBM_381).
Hina Batool
(2021_GCBM_412).
DEPARTMENT OF MATHEMATICS
Overview
Method to sove LINEAR SYSTEM
DIRECT Method
ITERATIVE Method
INVERSE based method,
CRAMER’s Rule Gauss Jordan Rule LU Factorization Method
LU FACTORIZATION METHOD
v Categories of LU FACTORIZATION method :
• Doolittle’S Method
• LDLT Method
• Cholesky ‘S Method
• Crout ‘S Method
Introduction to LU Factorization Method :
Ø LU Factorization was introduced by mathematician Tadeusz Banchiewicz
in 1938.In NUMERICAL ANALYSIS and LINEAR ALGEBRA , LU
FACTORIZATION of a matrix is the factorization of a given square matric into
two triangular matrices , ONE upper triangular matrix and another is lower
triangular matrix gives the original matrix.
Assume
[A] =Original matrix.
[L]=Lower triangular matrix.
[U]=Upper triangular matrix.
[A]=[L][U]
LU Factorization
Conditions for LU Factorization :
The steps for solving the linear system by involving an LU
Factorization method are as fellows :
1) Determine the factorization of a matrix A as A=LU ,
where L is lower triangular matrix and U is an upper
triangular matrix. So that , AX=B implies LUX=B.
2) Assume the vector Y=UX and solve the lower triangular
system LY=B for Y.
3) Using Y, solve the upper triangular system UX=Y for X.
LU Factorization
Method # 01
• The Doolittle’s Method :
Consider a 3×3 matrix A to express it as A=LU , where L and U are
the lower and upper triangular matrix respectively .
�11 �12 �13 �11 0 0 �11 �12 �13
�21 �22 �23 = �21 �22 0 0 �22 �23
�31 �23 �33 �31 �32 �33 0 0 �33
To determine the values of ��� and ��� , According to the Doolittle’s
approach set lii=1.
A=LU
�11 �12 �13 1 0 0 �11 �12 �13
�21 �22 �23 = �21 1 0 0 �22 �23
�31 �23 �33 �31 �32 1 0 0 �33
�11 �12 �13 �11 �12 �13
�21 �22 �23 = �21�11 �21�12 + �22 �21�13 + �23
�31 �23 �33 �31�11 �31�12 + �32�22 �33 + �31�13 + �32�23
Equating the crossPonding elements (Row_wise )on both sides and rearranging the required
values lij and uij as below;
• u11=a11 , u12=a12 , u13=a13
• l21=a21/u11 , l31=a31/u11 , l32=1/u22[a32-l31u12]
• u22=a22-l21u12 , u23=a23-l21u13 , u33=a33-[l31u13+l32u23]
Ø For n×n matrix :
Generalizing this procedure for n×n matrix A=[aij]
By factorizing A ,
A=LU
That gives the following formula for the Entries
of L and U :
• lii=1 i=1,2,3........n.
�−1
• uij=aij- �=1
���. ��� i=1,2....n , j=i,i+1,....n (for all i≤j).
• uij=0 (for all i>j)
�−1
• lij=1/uij(aij- �=1
���. ���) j=1,2,3.....n ,
i=j+1,j+2,....,n (for all i>j).
• lij=0 (for all i<j).
Related Example:
Solve the following linear system Ax=B by using the
DOOLITTLE method
1.7x1+2.3x2-1.5x3=2.35
1.1x1+1.6x2-1.9x3=-0.94
2.7x1 +-2.2x2+1.5x3=2.70 .
Solution:
By using the LU Factorization method
1.7 2.3 −1.5 x1 2.35
A= 1.1 1.6 −1.9 , X= x2 , B= −0.94
2.7 −2.2 1.5 x3 2.70
By Doolittle’s method Factorize A matrix ,
1 0 0 �11 �12 �13
A=LU= �21 1 0 0 �22 �23
�31 �32 1 0 0 �33
SINCE Diagonal entries in the lower triangle is replaced by 1.
1.7 2.3 −1.5 �11 �12 �13
1.1 1.6 −1.9 = �21�11 �21�12 + �22 �21�13 + �23
2.7 −2.2 1.5 �31�11 �31�12 + �32�22 �33 + �31�13 + �32�23
Equating the Components of First Rows on both sides,
• u11=1.7 , u12=2.3 , u13=-1.5.
Equating the Components of second Rows on both sides,
• l21=0.6471 , u22=0.1117 , u23=-0.9294
Equating the Components of third Rows on both sides,
• l31=1.5882 , l32=-52.4129 , u33=-44.8276
Thus,
1 0 0 1.7 2.3 -1.5
L= 0.6471 1 0 , U= 0 0.1117 -0.9294
1.5882 -52.4129 1 0 0 -44.8276
Therefore,
AX=B
LUX=B
Step:01
Assume vector Y=[y1 y2 y3]T such that UX=Y,
LY=B , Y=?
1 0 0 y1 2.35
0.6471 1 0 y2 = −0.94
1.5882 −52.4129 1 y3 2.70
Now, solve this system By Forward substitution :
• �1 =2.35
• 0.6471�1+�2=-0.94 or �2 = -2.4607
• 1.5882�1+(-52.4129)�2+�3 =2.70 or �3 =-130.0047
so,
Y=[2.35 -2.4607 -130.0047]T
Step:02
Putting the value of Y=[2.35 -2.4607 -130.0047]T
in UX=Y .
Now, Solve this System By Backward substitution ,
• -44.827x3 =-130.0047 or x3 =2.900
• 0.1117x2 +(-0.9294)x3=-2.4609 or x2 =1.099
• 1.7x1 +2.3x2 +(-1.5)x3 =2.35 or x1=-2.101
Thus ,
The Required solution vector is given by;
-2.101
X= 1.099 .
2.900
