…GAUSSIAN

ELIMINATION…

VIVIANA MARCELA BAYONA CARDENAS

Linear equations…
A set of n linear equations in n variables, xi
a11 x1  a12 x2  a1n xn  b1
a21x1  a22 x2  a2n xn  b2
 
an1 x1  an2 x2  ann xn  bn

can be written in matrix form, Ax = b:

 a11 a12  a1n x1  b1 

a21 a22 a2n x2  b2 
  
   
  
 x  b 
an1 an2
 ann  n   n 
…Linear equations
e.g. the 2  2 system of linear equations
2x  y  5
xy  3
can be written in matrix form, Ax = b:
2 1x 5
 
1 1
 y 3
    
To solve, we premultiply the equation by A–1, A–
1
Ax = A–1b, giving x = A–1b: 1 1
2 1 2 1x 2 1 5

   
  1 1 

1 1 
1 1  
y  3

x 1 1 15 2
y  1 2 3 1
  1     
Matrix equations…
We can use this method for matrix equations of the
form AX = B: e.g. find the matrix X if
5 4 1 1 4
 X  0 1 2
2 2
 
 
We can get rid of the matrix to the left of X by
premultiplying each side of the equation by the

inverse 5 415 4 5 411 1 4
   X  2 2 0
 2

2 2  
2 2     1 

5 411 1 4 1 2 41 1 4

Thus X  2 2 0 1 2  2 5 0 1 2
    2   
 1 3 0
 1 3.5 1
 
…Matrix equations
Similarly for 4  1 e.g.
5XA = B: 1
find the matrix X if
X2 2 0 1 
  4 2 

We can remove matrix


 A
1
 by1postmultiplying
1 1
 each
 5 4 5 4  inverse, 5 4
0 1 2A–12
side of theXequation

2
 2 
2
 2
by 
 4 2 
its 

1 1 1
 1 11   2 4.5
5 4 2 4
X  0 1 2 2  0 1  2 5  1 2.5 
Thus
 4 2   4 2 2   2 3 
Solving n  n linear equations
We can solve a general system of n  n linear equations by
writing it in matrix form, AX = B, and premultiplying the
matrix equation by the inverse of the matrix of
coefficients, A–1.
i.e. if AX = B
then A–1AX = A–1B,
but A–1A = I
so X = A–1B
However, this method is almost never used because it is
inefficient and prone to rounding errors.
We shall investigate Gaussian elimination, a more
efficient and accurate method for solving linear equations.
Visualising n  n linear equations
2 variables • 3 variables
• Solution is at the
Solution is at the intersection of 3 planes
intersection of 2
lines
Gaussian elimination…
Consider a system of 3  3 linear equations in matrix
form, Ax = b:
a11 a12 a13 x1  b1 
a21 a22 a23 x2  b2 

a31 a32 a33 

x 3 
 
b3 

To make book-keeping simpler, we represent the
system by an augmented matrix:

a11 a12 a13 b1 
a21 a22 a23 b2 

a31 a32 a33 b3 

…Gaussian elimination…
We can zero the first column by subtracting a21/a11 times
the first row from the second row, and subtracting a31/a11
times the first row from the third row (primes indicate
changed values)
a11 a12 a13 b1 
 0 a22 a23
 b2

 0 a32 a33 b3

Similarly, we can now zero the second column by
subtracting a’32/a’22 times the first row from the third row

(double primes indicate changed values), forming an
upper triangular matrix:
a11 a12 a13 b1 
 0 a22  b2
a23

 0 0 a33 b3

…Gaussian elimination…
The last row represents an equation in a single variable
a11 a12 a13 b1 
 0 a
22  b2
a23

 0 0 a33 b3

a”33 x3 = b”3
which can be solved as x3 = b”3 / a”33
The second row represents an equation in two variables

a’22 x2 + a’23 x3 = b’2
Since the variable x3 has already been found in the previous
step, x2 can be solved as
x2 = (b’2 – a’23 x3) / a’22
…Gaussian elimination
The first row represents an equation in three variables
a11 x1 + a12 x2 + a13 x3 = b1
Since the variables x2 and x3 have already been found
in the previous steps, x1 can be solved as
x1 = (b1 –  a12 x2 – a13 x3) / a11
This process of solving an upper triangular matrix
equation is called back substitution.
Gaussian elimination: example 1…
Solve the system of equations
x yz  4
x  y  z  2
2x  8y  z  19

Represent the system as an augmented matrix:

 1 1 1 4 
1 1 1 2
2 8 1 19 
We introduce a lower triangular matrix, L, to record
row multipliers in the • positions
 1 0 0
 1 0
  1
…Gaussian elimination: example 1…
Calculate the row multipliers and record them in the L
matrix, l21 = a21/a11 and l31 = a31/a11
1 0 0
L  1 1 0
2  1
Zero the first column by subtracting l21 = a21/a11 times
the first row from the second row, and subtracting

l31 = a31/a11 times the first row from the third row
1 1 1 4  1 1 1 4 
1 1 1 2 0 2 2 6
2 8 1 19  0 6 1 11 
…Gaussian elimination: example 1…
Calculate the row multiplier and record it in the L
matrix, l32 = a’32/a’22
1 0 0
L  1 1 0
2 3 1
Zero the second column by subtracting l32 = a’32/a’22
times the second row from the third row, forming an

upper triangular matrix
1 1 1 4  1 1 1 4 
0 2 2 6 0 2 2 6
0 6 1 11  0 0 7 7
…Gaussian elimination: example 1
 Back substitute
1 1 1 4 
0 2 2 6
0 0 7 7
 The last row represents an equation in a single variable, a”33 z = b”3,
which can be solved as z = b”3 / a”33 = –7 / –7 = 1
 row represents an equation in two variables,
 The second
a’22 y + a’23 z = b’2, which can be solved as
y = (b’2 – a’23 z) / a’22 = (–6 + 2) / (–2) = 2
 The first row represents an equation in three variables
a11 x + a12 y + a13 z = b1, which can be solved as
x1 = (b1 –  a12 y – a13 z) / a11 = (4 – 2 – 1) / 1 = 1
 The solution is thus x = 1, y = 2, z = 1.
Gaussian elimination: example 2…
Solve the system of equations
2x  5y  3z  7
x  2.5y 1.5z  7.5
2x  7y  4.5z  12
Represent the system as an augmented matrix:

 2 5 3 7 
1 2.5 1.5 7.5
2 7 4.5 12 


…Gaussian elimination: example 2…
Calculate the row multipliers and record them in the L
matrix, l21 = a21/a11 and l31 = a31/a11
 1 0 0
L  0.5 1 0
 1  1

Zero the first column by subtracting l21 = a21/a11 times

the
 first row from the second row, and subtracting
l31 = a31/a11 times the first row from the third row
2 5 3 7  2 5 3 7 
1 2.5 1.5 7.5 0 5 3 11
2 7 4.5 12  0 2 1.5 5 
…Gaussian elimination: example 2…
Calculate the row multiplier and record it in the L
matrix, l32 = a’32/a’22
 1 0 0
L  0.5 1 0
 1 0.4 1

Zero the second column by subtracting l32 = a’32/a’22

times the second row from the third row, forming an

upper triangular
2 5matrix
3 7  2 5 3 7 
0 5 3 11 0 5 3 11
0 2 1.5 5  0 0 0.3 0.6 
…Gaussian elimination: example 2
Back substitute 2 5 3 7 
0 5 3 11
0 0 0.3 0.6 

The last row can be solved as z = b”3 / a”33 = 0.6 / 0.3 = 2

The second
 row can be solved as
y = (b’2 – a’23 z) / a’22 = (–11 + 6) / (–5) = 1
The first row can be solved as
x = (b1 – a12 y – a13 z) / a11 
= (7 – 5 – 6) / 2 = –2
The solution is thus x = –2, y = 1, z = 2.
Gaussian elimination: example 3…
Solve the system of equations
0.0001x  0.0001y 1.99z  10
2x  2.001y  z  1
4x  3y  2.982z  1

Work to 4 significant figures and give the answer to 3.

Check the answer by substituting into the original

equation.
Represent the system as an augmented matrix:
0.0001 0.0001 1.99 10
 2 2.001 1 1 
 4 3 9.282 1 
…Gaussian elimination: example 3…
Calculate the row multipliers and record them in the L
matrix, l21 = a21/a11 and l31 = a31/a11
 1 0 0
L  20, 000 1 0

40, 000  1 
Zero the first column by subtracting l21 times the first row
from the second row, and subtracting l31 times the first row
from the third

0.0001 0.0001 1.99 10
 2 2.001 1 1 
 4 3 9.282 1 
0.0001 0.0001 1.99 10 
 0 0.001 39, 800 200, 000

 0 1 79, 600 400, 000

…Gaussian

elimination: example 3…
Calculate the row multiplier and record it in the L
matrix, l32 = a’32/a’22
 1 0 0
L  20, 000 1 0

40, 000 1, 000 1

Zero the second column by subtracting l32 = a’32/a’22
times the second row from the third row, forming an
  matrix
upper triangular 
0.0001 0.0001 1.99 10
 0 0.001 39, 800 200, 000

 0 1 79, 600 400, 000

0.0001 0.0001 1.99 10 
 0 0.001 39, 800 200, 000 

 0 0 39, 880, 000 200, 400, 000

…Gaussian elimination: example 3… 
0.0001 0.0001 1.99 10
Back substitute  0 0.001 39, 800 200, 000 

 0 0 39, 880, 000 200, 400, 000


The last row can be solved as z = –39,880,000 / –

200,400,000 = 5.025
 row can be solved as
The second
y = (–200,000 + 199,995) / (0.001) = –5,000
The first row can be solved as
x = (b1 – a12 y – a13 z) / a11 
= (10 + 0.5 – 9.99975) / 0.0001 = 5,003
The computed solution is thus
x = 5,003, y = –5,000, z = 5.03
…Gaussian elimination: example 3
Check the computed solution against the original
equations 0.0001 0.0001 1.99 x 10
 2 2.001 1 y 1 
 4 3 9.282
z 
 1 
0.0001 0.0001 1.99 5, 000   10 
 2 2.001 1 5, 000  0.03 
 4 3 9.282
 5.03  5, 015
Numerical rounding has caused these errors.
Working to more significant figures gives
x = –7.939 c.f. 5,000

y = 5.922 c.f. –5,000
z = 5.025 c.f. 5.03
Geometric interpretation
Ill-conditioning, as seen in example 3, can be
interpreted geometrically:
The shaded red area represents uncertainty
in the lines’ positions due to rounding errors.
The solid red area represents
where the intersection may be
found due to these errors.
The closer to parallel the
lines are, the larger the
possible error.
Geometric interpretation: 1 solution
Gaussian elimination
to the form
   
0   
0 0  
where none of the
diagonal elements are
zero
 We get the usual form
of the augmented
matrix and can fine a
unique solution by
back substitution
Geometric interpretation: 0 solution
Gaussian elimination
to the form
   
0   
0 0 0 
where b3 is not zero
We end up with
 0 z = non-zero
i.e. no solution can
exist because at least 2
planes, or 3 lines of
intersection, are
parallel
Geometric

interpretation: ∞ solution
Gaussian elimination
to the form
   
0   
0 0 0 0
i.e. b3 is zero
We end up with
 0 z = zero
i.e. an infinite number
of solutions exist
because this is always
true for any value of z
REFERENCE
http
://es.wikipedia.org/wiki/Eliminaci%C3%B3n_de_Gaus
s-Jordan

http://www.monografias.com/trabajos72/resolucion-si
stemas-metodo-gauss-jordan/resolucion-sistemas-me
todo-gauss-jordan.shtml
 
http://cbi.azc.uam.mx/archivos/varios/ProblemarioW
.pdf