Você está na página 1de 14

Matrix Algebra

Solution to Simultaneous Algebraic Equations Gauss Elimination


Method:
Consider n simultaneous
a 11 x 1 + a 12 x 2 +, a 13 x 3
a 21 x 1 + a 22 x 2 + a 23 x 3
a 31 x 1 + a 32 x2 + a 33 x 3

equations ,
+ a 1n x n = b1
+ a 2n x n = b2
+ a 3n x n = b3

an1 x 1 + a n2 x 2 + a n3 x 3 + a nn x n = bn
write the given set of equations in matrix form,
a 11
a 21
a 31
.
.
.
.
a n1

a 12
a 22
a 32
.
.
.
.
a n2

a 13
a 23
a 33
.
.
.
.
a n3

.
.
.
.
.
.
.
.

a 1n
a 2n
a 3n

a nn

x1
x2
x3

xn

b1
b2
b3

bn

In Gauss elimination method the variables x2, x n-1, will be


successively eliminated using Row Operations. This step is called forward elimination
. The given matrix will be converted to into an upper triangular matrix, Lower triangular
elements become zeros.
After forward elimination the nth equation (last equation) become simple , it as an
equation with one variable xn , determine xn. Now using (n-1)th equation xn-1 can be
determined. Similarly using (n-2)nd equation xn-2 can be determined. Using (n-3)rd
equation xn-3 can be determined. Continue up to first equation until all the unknowns are
determined. This is called backward substitution.
Forward Elimination
Step 1 : a11 becomes pivot, eliminate x1 from row2 , row3, row4,

Row2 a 21 = a 21 - ( a 21 / a 11) a 11
a 22 = a 22 - ( a 21 / a 11) a 12
a 23 = a 23 - ( a 21 / a 11) a 13

row n etc

a 21 becomes 0
a 22 changes
a 23 changes
1

etc up to a 1n = a 2n - ( a 21 / a 11) a 1n

a 2n changes

b2 = b2 - ( a 21 / a 11) b1 b2 changes
whatever we did to make a 21 = 0 applied the same to other elements
of that row
Row3 a 31 = a 31 - ( a 31 / a 11) a 11 a 31 becomes 0
a 32 = a 32 - ( a 31 / a 11) a 12 a 32 changes
a 33 = a 33 - ( a 31 / a 11) a 13 a 33 changes
etc up to a 3n = a 3n - ( a 31 / a 11) a 1n a 3n changes
b3 = b3 - ( a 31 / a 11) b1 b3 changes
whatever we did to make a 31 = 0 applied the same to other elements
of that row

Row n a n1 = a n1 - ( a n1 / a 11) a 11 a n1 becomes 0


a n2 = a n2 - ( a n1 / a 11) a 12 a n2 changes
a n3 = a n3 - ( a n1 / a 11) a 13 a n3 changes
etc up to a nn = a nn - ( a n1 / a 11) a 1n a nn changes
b3 = b3 - ( a 31 / a 11) b1 b3 changes

whatever we did to make a n1 = 0 applied the same to other elements of that row.

Now, re-write the whole matrix equation. First row remains same, elements of other
rows will be different.
Step2 : : a22 becomes pivot, eliminate x2 from row3 , row4, row5, etc., row n
following the same method
Now, re-write the whole matrix equation. First row , Second row remains same,
elements of other rows will be different
Step3 : : a33 becomes pivot, eliminate x3 from row4 , row5, row6, etc., row n
following the same method
Now, re-write the whole matrix equation. First row , Second row , Third row remains
same, elements of other rows will be different
Continue until the variables
x2, x3, x4 . x n-1 will be
successively eliminated and all the lower triangular elements becomes zero.

Backward substitution.
After forward elimination the nth equation (last equation) become simple , it as an
equation with one variable xn , determine xn. Now , using (n-1)th equation xn-1 can be
determined. Similarly using (n-2)nd equation xn-2 can be determined. Using (n-3)rd
equation xn-3 can be determined. Continue up to first equation until all the unknowns are
determined. The method is best understood by solving problems.

Different Methods used to Solve Set of Simultaneous Equations in FEM.


Method of Matrix Inversion
Gauss elimination method
Cholesky Decomposition Technique
Gauss-Seidal Iteration Technique
Relaxation Method
Numerical examples illustrating Gauss elimination method :

Problem 1. Solve the following set of equation by Gaussian elimination


technique.
5x1 + 3x2 + 2x3 + x4 = 4
4x1 + 3x2 - 3x3 - 2x4 = 5
x1 + 2x2 - 2x3 + 3x4 = 6
-4x1 + 3x2 5x3 + 2x4 = 7
Solution : Write the given equations in Matrix Form

x1

-3

-2

x2

-2

-4

-5

x3

x4

6
7

[ CO ] [ X ] = [ CONS ]
[ a ] [x] = [b]
3

Step 1 : a i j = a i j (a i1 / a 11) a 1j
i = 2, j = 1,2,3,4

b i = b i (a i1 / a 11 ) b 1

Row 2 i = 2 j = 1,2,3,4
4 (4/5) 5 = 4 (4) = 0.0
3 (4/5) 3 = 3 2.4 = 0.6
-3 (4/5)2 = -3 1.6 = -4.6
-2 (4/5) 1 = -2 0.8 = -2.8
Row 3 i = 3 j = 1,2,3,4
1 (1/5)5 = 1 1 = 0.0
2 (1/5)3 = 2 -0 .6 = 1.4
-2 (1/5) 2 = -2 0.4= -2.4

3 (1/5) 1 = 3 0.2 = 2.8

Row 4 : i = 4 j = 1,2,3,4
-4 ( -4 / 5)5 = 4 - 4 = 0
3 ( -4 / 5)3 = 3+2.4 = 5.4
-5 ( -4 /5) 2 = -5 + 1.6 = -3 .4
2 ( -4 /5) 1 = 3 0.2 = 2.8
b i = b i (a i1/ a11) b 1
i = 2 b2 = 5 (4/5)4
= 1.8
i = 3 b3 = 6 (1/5)4
= 5.2
i = 4 b4 = 7 (-4/5)4

= 10.2

The modified matrix equation , after eliminating x1


from 2nd , 3rd and 4th equations.

x1

0 0.6 -4.6 -2.8

x2

1.8

0 1.4 -2.4 2.8 *

x3

= 5.2

0 5.4 -3.4 2.8

x4

10.2

Step 2 : To eliminate x2 from Row 3 and Row 4


a i j = a i j (a i2 / a 22) a 2j b i = b i (a i2 / a 22 ) b 2 i = 3, j = 2,3,4
Row 3 i =3 j = 2,3,4
1.4 (1.4/0.6) 0.6 =1.4 1.4 = 0.0 j = 2
-2.4 (1.4/0.6) (-4.6) = 2.4+10.73 =8.33 j = 3
2.8 (1.4/0.6)(-2.8) = -2.8 6.53 = 9.33 j = 4

Row 4

i=4

j = 2,3,4

5.4 (5.4/0.6)0.6 = 5.4-5.4 = 0.0 j = 2


-3.4 (5.4/0.6)(-4.6) = -3.4+41.4 = 38 j = 3
2.8 (5.4/0.6)2.8 = 2.8 +25.2 = 28 j = 4
b i = b i (a i2/ a22) b 2
i = 3 b3 = b3 (a32/a22)b2
5.2 - (1.4/0.6) 1.8 = 5.2 4.2 = 1
i = 4 b4 = b4 (a42/a22)b2
10.2 - (5.4/0.6)1.8 = 10.2 -16.2 = -6
The modified matrix after step 2 eliminating x2 from 3rd and 4th equations.

x1

0 0.6 -4.6 -2.8

x2

1.8

0 8.33 9.32 *

x3

x4

38 28

Step 3 : To eliminate x3 from Row 4

a i j = a i j (a i3 / a 33) a 3j

1
-6

b i = b i (a i3 / a 33 ) b 3 i = 4, j = 3,4

a43 = a43 (a43/ a33) a33 = 38 (38 / 8.33) 8.33 = 38 38 = 0.0

a44 = a44- (a43/ a33) a34 = 28 - (38 / 8.33) 9.32


= 28 42.52 = -1 4.52
b i = b i (a i3/ a33) b 3 i = 4,
b4 = b4 - (a43/ a33) b3
b4 = -6 - (38 / 8.33) 1 = - 6 - 4.56 = - 10.56
The modified matrix, after step 3 , eliminating x3 from 4th equation.

0 0.6 -4.6 -2.8

x1

x2

1.8

0 8.33 9.32 * x3 =

14.52

x4

1
10.56

Back Substitution:

The modified equations are


5 x1 + 3 x2 + 2 x3 + x4 = 4
0.6 x2 4.6 x3 - 2.8 x4 = 1.8
8.33 x3 + 9.32 x4 = 1
-14.52 x4 = -10.56
x4 = ( -1 0.56/ -1 4.52) = 0.727
8.33 x3 + 9.32 (0.727 ) = 1
x3 = (1 - 6.776) / 8.33 = - 0.693
0.6 x2 4.6( - 0.693) -2.8 (0.727) = 1.8
0.6 x2 = (1.8 - 3.1878 + 2.0356)
x2 = (1.8 - 3.1878 + 2.0356) / 0.6 = 1.079
5x1 + 3 (1.079) + 2 ( - 0.693) + 0.727 = 4
x1 = ( 4 - 3 (1.079) + 2 (0.693) - 0.727 ) / 5 = 0.155
x1= 0.155

x2=1.079

x3= - 0.693

x4= 0.727
6

Prob2 : Solve using gauss elimination method


x1- 2x2 + 6 x3 = 0

ij

2x1 +2x2 + 3x3 = 3

-2

x1

x2

-1

x3

-x 1 + 3x2 = 2

0
3
2
Step 1 :

= a i j (a i1 / a11) a1j I = 2,3 j = 2,3


2- (2/1)* 1 = 0
-1- (-1/1) 1 = 0

2 (2/1) (-2) = 6 3 (2/1) (6) = -9


3 (-1/1) (-2) = 1 0 (-1/1) 6 = 6
bi = bi - (a i1 / a11) b1

3 (2/1) *0 = 3

2 (-1/1) *0 = 2

Modified Matrix After step1


1

-2

x1

-9

x2

x3

Step 2 : a

ij

0
3
2

= a i j (a i2 / a22) a2j i = 3 j = 2,3


1 (1/6)* (6) = 0 6 (1/6) (-9) = 7.5

bi = bi - (a i2 / a22) b2
2 (1/6) 3 = 1.5

Modified Matrix After step2


1

-2

x1

-9

x2

7.5

x3

0
3
1.5

Back Substitution :
7.5 x3 = 1.5 x3 = 1.5 / 7.5 = 0.2

x3 = 0.2

6 x2 9x3 = 3 x2=(3 + 9(0.2))/6 =0.8

x2 = 0.8

x1 2x2 +6x3 = 0 x1 = 2(0.8)6(0.2) = 0.4 x1 = 0.4


Prob 3 :Solve using gauss elimination method
4x1 +6x2 +8x3 = 2 8x1 +4x2 + 6x3 = 4 6x1 + 2x2+4x3= 6
Solution : writing the equations in matrix form
4

Step 1 :

ij

x1

x2

x3

= a i j (a i1 / a11) a1j i = 2,3 j = 2,3

8 - (8/4)* 4 = 0
6 -( 6/4) 4 = 0

4 (8/4) (6) = -8
2 (6/4 (6) = -7

6 -( 8/4)8 = -10
4 (6 /4) 8 = -8

bi = bi - (a i1 / a11) b1
4 (8/4) 2 = 0
6 (6/4) 2 = 3

Modified matrix equation after step 1


4

-8

-10

-7

-8

Step 2 : a

ij

X1

X2

X3

= a i j (a i2 / a22) a2j i = 3 j = 2,3

-7 (-7/-8) (-8) = 0
-8 -(-7 / -8 )( -10) = -8 +70/8 = -8 + 8.75 = 0.75
bi = bi - (a i2 / a22) b2
3 (-7/-8) 0 = 3
Modified matrix equation after step 2 :
4

X1

-8

-10

X2

0.75

x3

0.75 x3 = 3

x3 = 3/0.75 = 4

-8 x2 -10 x3 = 0
4x1+6x2 + 8x3 = 2

-8x2 -10 (4) = 0

x3 = 4
-8x2 = 40 x2 = -5

4x1+ 6(-5) +8(4) = 2

x1 = (2+30 -32 ) / 4 =0
x1 = 0 x2 = -5

x3= 4

Prob 4 : Solve using gauss elimination method


3x1 -3x2 -2x3 = 5
2x1+2x2+3x3= 6
3x1 - 5x2+2x3= 7
3 -3 -2
2 2 3
3 -5 2

5
6
7

Step 1 :
a i j = a i j (a i1 / a11) a1j i = 2,3 j = 2,3
2 - (2/3)3 = 0 2 - (2/3)(-3) = 4 3 -( 2/3)(-2) = 4.33
3 -( 3/3) 3 = 0 -5 (3/3 (-3) = -2 2 (3/3)(-2) = 4
bi = bi - (a i1 / a11) b1
6 (2/3) 5 = 6-10/3 = 2.667
Modified matrix is
3 -3
0
4
0 -2

-2
4.33
4

7 (3/3) 5 = 2

5
2.667
2

Step 2 :
a i j = a i j (a i2 / a22) a2j i = 3 j = 2,3
-2 - (-2/4)4 = 0
4 - (-2/4)(4.333) = 4+ 2.166 = 6.166
bi = bi - (a i2 / a22) b2
2 (-2/4) 2.667 = 2+(2.667/2) = 2+1.333 = 3.333
3 -3
0
4
0 0

-2
4.33
6.166

5
2.667
3.333

6.166 x3 = 3.333 x3 = ( 3.333/6.166) = 0.504


4x2 + 4.333x3 = 2.667
x2 = 0.483 /4 = 0.120

4x2 = 2.667 - 4.333 (0.504)

10

3x1 -3x2 2x3 = 5 3x1 -3(0.120) -2(0.504) = 5


x1 = (5 + 0.360 +1.008) / 3 = 2.122
x1 = 2.122 x2 = 0.120

x3 = 0.504

THUS COMPLETE GAUSSIAN ELIMINATION


TECHNIQUE

11

12

13

14

Você também pode gostar