Solving Linear Systems with Gaussian Elimination

Lecture 1 1
MATH C192:MATHEMATICS–II
BITS-PILANI HYDERABAD CAMPUS
Presented by
Dr. M.S. Radhakrishnan
Email: msr_bits@yahoo.com
• Solutions of Linear
Systems of Equations
T1 - Ch. 1.6
Bernard Kolman & David R. Hill,
Linear Algebra, Pearson, 8th Ed, 2005
(First Impression, 2009)
2/27/2018 Presented by Dr. M.S. Radhakrishnan BITS, Pilani 2
In this lecture we look at
• Solutions of Linear Systems

of Equations
• Gaussian Elimination method

Linear Systems of Equations
A system of m linear equations in n

unknowns is a collection of m equations,
each of which involves n unknowns (each
only in the first degree). The problem is to
find “solutions” ( = values of unknowns)
that satisfy all the equations simultaneously.

Example
Before discussing the general case, we look at
simple systems, systems of 2 equations in 2
unknowns.
1. 2x + y = 5 2. 2x + y = 6
x + 2y = 4 4x + 2y = 12
3. 2x + y = 6
4x + 2y = 5
Let us look at the first system:
2x + y = 5 ….. (1)
x + 2y = 4 ….. (2)
(1)  2 gives the new equation
4x + 2y = 10 ….. (3)
x + 2y = 4 ….. (2)
(3) – (2) gives 3x = 6 or x = 2
(1) now gives 4 + y = 5 or y = 1
Thus the system has a unique solution
x = 2, y = 1.
We understand from the above, that the
solution will not change if we do any one of
the following:
1. Multiply an equation by a non-zero number
2. Interchange any two equations
3. Add to an equation a non-zero constant
times any other equation.
These are called “elementary Row
operations.” These elementary row
operations are reversible in the sense that if
we apply a finite number of elementary row
operations to a system I and get a system II,
then system I can be obtained by applying
the same number of elementary row
operations to the system II “in the reverse
order.” ( This means if we “multiplied an
equation”, then we should divide in the
second case, etc.)
Equivalent Systems of equations
We say two systems of equations are

equivalent if they have the same solutions.
Thus a given system and a system obtained

from it by applying a finite number of the
above elementary row operations on the
equations will be equivalent.

2. 2x + y = 6
4x + 2y = 12
We note that the second equation is “the same
as” the first equation and so can be omitted.
And the first equation is satisfied by
y = 6 – 2x for any value of x.
Thus (1, 4), (2, 2), (3, 0) … are all solutions.
Thus the given system has an infinite number
of solutions.
3. 2x + y = 6
4x + 2y = 5
We easily see that the two equations are

inconsistent. And thus the above system
has no solution.
We also see the pictorial representation of the

three systems in the next slides.

System 1
Unique Solution

System 2
Infinite number
of Solutions

System 3
No Solution

The situation is the same for any number of
linear equations in any number of unknowns.
This is summarized in the next slide.

The given system is
Consistent (has at Inconsistent (has

least one solution) no solution)
Has a unique Has an infinite

solution number of solutions

a11x  a12 y  b1 With two equations and two variables,
the graphs are lines and the solution (if
a21x  a22 y  b2 there is one) is where the lines
intersect. Let’s look at different
independent possibilities.
dependent inconsistent
consistent consistent
Case 1: The lines Case 2: The lines Case 3: The lines
intersect at a point coincide and there are parallel so there
(a, b), are infinitely many is no solution.
the solution. solutions (all points
on the line).
With two equations and two variables, the graphs were
lines and the solution (if there was one) was where the
lines intersected. Graphs of three variable equations are
planes. Let’s look at different possibilities. Remember the
solution would be where all three planes all intersect.
Planes intersect at a point: consistent with

one solution
Planes intersect in a line: consistent
system called dependent with an infinite
number of solutions
Three parallel planes: no intersection so
system called inconsistent with no solution
No common intersection of all three
planes: inconsistent with no solution
Triangular systems
Consider the system of equations
x + 2y + 3z = 14
y - 2z = - 4
4z = 12
This is called a triangular system (for
obvious reasons). It is very easy to solve
such a system.
From the last (3rd ) equation, we get z = 3.
Now the second equation gives
y - 6 = - 4 or y = 2.
Now the first equation gives
x + 4 + 9 = 13 or x = 1.
So to solve a general system, we apply the
above elementary operations to get an
equivalent triangular system and then solve it.
This is the basis behind Gaussian
Elimination method.
Example
Consider the system of equations
x + y- z= 9
3x +11y + 3z = 21
-2x + 4y - 6z = 40
To “reduce” this system into an equivalent
triangular system, we first get rid of x
from the second and third equations as
follows:
We do not touch the first equation. We
multiply the first equation by the nonzero
constant (-3) and add it to the 2nd equation.
We multiply the first equation by the nonzero
constant 2 and add it to the 3rd equation.
We thus get the equivalent system
x+ y - z= 9
8y + 6z = - 6
6y - 8z = 58
Now we do not touch the first and second
equations. We multiply the second equation
by the nonzero constant (-6/8) and add it to
the 3rd equation.
We thus get the equivalent system
x+ y - z= 9
8y + 6z = - 6
(-100/8)z = 500/8

Thus we have reduced the system into a
triangular system. From the 3rd equation (of
the newest system) we get z = - 5. The
second equation then gives
8y - 30 = - 6 or y = 3.
The first equation then gives
x + 3 +5 = 9 or x = 1.
Hence the given system has the unique solution
x = 1, y = 3, z = -5.
In problems, we do not use such verbose
statements. We use shorthand notation as we
now indicate:
x + y- z= 9
3x +11y + 3z = 21 (Eqn (2) – 3  eqn. (1))
-2x + 4y - 6z = 40 (Eqn (3) +2  eqn. (1))
x+ y - z= 9
 8y + 6z = - 6
6y - 8z = 58 (Eqn (3) - 6/8  eqn. (2))
 x+ y - z= 9
8y + 6z = - 6
(-100/8)z = 500/8
The symbol ‘’ means “equivalent to.”
Also by the side of each equation, we write
what we will be doing in the “next step.”
The system has been triangularized and we
get the solution as before by “back-
substitution”.
We say “back-substitution” as we first look
at the last equation to find z, then go back
to the second to find y and then to the first
to find x.
We use the phrase: The system has been
reduced to an equivalent system in “echelon”
form.
In French, “echelon” means steps and thus
the new system looks like a descending steps;
hence the name “echelon form”.
Linear systems in matrix form
The variables x, y, z need not be written in

the above “elimination process”. We now
redo the same without the variables.
x y z = RHS
 1 1 1 9 
 3 11 3 21
R2 – 3 R1
 
 2 4 6 40 R3 + 2 R1
 1 1 1 9 
0 8 6 6
  
0 6 8 58  R3 – (6/8) R2
1 1 1 9 
  
 0 8 6 6 
0 6 (100 / 8)  500 / 8
Thus the given linear system has been
reduced into echelon form which in
equation form reads as:
x+ y - z= 9
8y + 6z = - 6
(-100/8)z = 500/8
And the solution as before is
x = 1, y = 3, z = -5.
General Linear system
A system of m linear equations in n unknowns
is of the form:
a11 x1  a12 x2  ...  a1n xn  b1
a21 x1  a22 x2  ...  a2 n xn  b2
.
.
am1 x1  am 2 x2  ...  amn xn  bm

A solution of the above system is an n-tuple
of numbers (x1, x2, …, xn) which satisfies all
the above equations.
System in the matrix form
Let
A =  a11 a12 . . a1n  x =  x1  b =  b1 
a  x  b 
 21 a22 . . a2 n   2  2
 .  . .
     
 .  . .
 am1 am 2 . . amn   xn  bm 
It is evident now that the above linear system
can be written in matrix form as:
Ax b
A is (rightly) called the coefficient matrix of
the system. b is the RHS.

Augmented matrix
The matrix  a11 a12 . . a1n b1 
a a . . a2 n 
b2 

A  ( A | b) =
21 22
 . .
 
 . .
 am1 am 2 . . amn bm 
is called the augmented matrix of the
system. Note that it is got by adding (=
augmenting) to the coefficient matrix, one
more column, namely the RHS vector b.
Now to reduce the given linear system to
echelon form, we apply elementary row
operations on the augmented matrix
A
so that the submatrix A (alone) is reduced to
echelon form.

Matrix in Echelon form
A matrix A is said to be in echelon form if
• All zero rows occur below the non-zero
rows.
• Let i < k. If the first non-zero entry ( =
pivot) of row i occurs in column ji and the
first non-zero entry of row k occurs in
column jk, then ji < jk.
• All the entries below the pivot entry (in the
column in which the pivot resides) are zero.
The matrices below are all in echelon form.
1 2 3 4 1 2 3 4 1 2 3 4
0 7  0 7  0
 5 6  5 6
 0 6 7 
0 0 8 9 0 0 8 9 0 0 0 0
     
0 0 0 3 0 0 0 0 0 0 0 0
1 2 3 4
0 5 6 
7

0 0 0 1
 
0 0 0 0
Thus we apply the elementary row
operations to the augmented matrix so
that the first “n” columns form a matrix in
echelon form. Hence the Linear system
we started with has been reduced to an
equivalent system (in the echelon form):

Example
Solve the system by reducing it to echelon
form:
4x  y 4
5x  3 y  z  2
 9x  2 y  z  5

Solution The augmented matrix is
 4 1 0 4 Pivot
 5 3 1 2 
  R2 - (5/4)R1
 9 2 1 5  R3 + (9/4)R1
 4 1 0 4
  0 17 / 4 1 3
 
 0 17 / 4 1 14  R3 + R2
 4 1 0 4
  0 17 / 4 1 3
 
 0 17 / 4 1 14  R3 + R2
x y z = RHS
 4 1 0 4
  0 17 / 4 1 3

 0 0 0 11 
The given system has been reduced to
echelon form.
The last row in equation form reads
0 x + 0 y + 0 z = 11.
which means that the given system is
inconsistent.

Example
form:
2x  y  2z  1
x y 0
x  2 y  6z  3
x  2z  1

 2 1 2 1
 1 1 0 0  R2 – (1/ 2)R1
 
 1 2 6 3 R3 - (1/2)R1
 
 1 0 2 1  R4 - (1/2)R1
2 1 2 1 
 0 1/ 2 1 1/ 2 
  
 0 5 / 2 5 5 / 2  R3 + 5R2
 
 0 1/ 2 3 1/ 2  R4 + R2
 2 1 2 1 
 0 
1/ 2 1 1/ 2 
 
 0 0 0 0  R3  R4
 
 0 0 4 0 
x y z = RHS
2 1 2 1 
0 
1/ 2 1 1/ 2 
 
0 0 4 0 
 
0 0 0 0 
echelon form.
The 3rd row in equation form reads
0 x + 0 y - 4 z = 0. Hence z = 0.
The 2nd Row gives (1/2) y = - (1/2) or y = -1.
The 1st Row gives 2x – 1= 1 or x = 1.

Example
form:
4x  y  4
5x  3 y  z  2
 9x  2 y  z   6

 4 1 0 4
 5 3 1 2 
  R2 - (5/4)R1
 9 2 1 6 R3 + (9/4)R1
 4 1 0 4
 0 17 / 4 1 3
 
 0 17 / 4 1 3 R3 + R2
 4 1 0 4
  0 17 / 4 1 3
 
 0 17 / 4 1 3 R3 + R2
x y z = RHS
 4 1 0 4
  0 17 / 4 
1 3

 0 0 0 0 
echelon form.
Clearly the system is consistent and putting
in equation form it reads:
4x  y  4
 (17 / 4) y  z  3

Hence z is a free variable and we can express
x and y in terms of it.
y = (12/17) + (4/17) z
x = 1– y/4 = 1 – (3/17) - (1/17) z
= (14/17) - (1/17) z
Hence the solution of the given system is:
x = (14/17) - (1/17) z
y = (12/17) + (4/17) z, z is free (i.e. can
take any value.)
Example
form:
4 y  3z  8
2x  z 2
3x  2 y 5

0 4 3 8  R1  R2
 2 0 1 2  (means Interchange
  rows R1 and R2)
 3 2 0 5 
 2 0 1 2 
0 4 3 8 
  
 3 2 0 5  R3 – (3/2) R1
 2 0 1 2 
 0 4 3 8
  
 3 2 0 5  R3 – (3/2)R1
2 0 1 2
 0 4
 3 8 
 0 2 3/ 2 2 R3 – (1/2)R2

 2 0 1 2 
 0 4 3 8

 0 2 3/ 2 2  R3 – (1/2)R2
x y z = RHS
The last row tells
 2 0 1 2 
 0 4 
us that the
 3 8 
system is
inconsistent (=
 0 0 0 2 has no solution).
Example
form:
5 x  5 y  10 z  0
2w  3x  3 y  6 z  2
4w  x  y  2 z  4

 0 5 5 10 0  R1  R2
 2 3 3 6 2  (means
  Interchange rows
 4 1 1  2 4 R1 and R2)
 2 3 3 6 2
 0 5 5 10 0 
  
 4 1 1  2 4 R3 – 2R1
 2 3 3 6 2
  0 5 5 10 0 
 
 4 1 1  2 4 R3 – 2R1
 2 3 3 6 2
  0 5 5 10 0 
 
 0 7 7 14 0  R3 – (7/5)R2

 2 3 3 6 2
 0 5 5 10 0 
  
 0 7 7 14 0  R3 – 7R2
w x y z = RHS
 2 3 3 6 2
  0 5 5 10 0 
 
 0 0 0 0 0 

Thus the system is consistent, the variables y,
z are free and the variables w, x
(corresponding to the pivotal columns) can be
expressed in terms of them.
The second row tells us that
5x + 5y - 10z = 0 or x = - y + 2z
The first row tells us that
2w –3x– 3y + 6z = 2.
Or w = (3/2) x+(3/2) y - 3z + 1
= (3/2) (-y + 2z) +(3/2) y - 3z + 1
= 1
Hence the solution of the given system is:
w = 1
x = 2z – y, y, z free (i.e. can take any
value)
For example, w=1, x=0, y=0, z=0 is a solution.
Gauss-Jordan Elimination
We solved Linear systems by Gauss
Elimination method reducing the system to
Echelon form and then finding the solution
by back-substitution.
Gauss-Jordan method reduces the system to
a special Echelon form so that we can “read
out” the solution.

A matrix is said to be in Row-reduced
Echelon Form (RREF) if
• The first non-zero entry in each row is 1
(Such an entry is called a “pivot”.)
• All the entries above and below the
pivot entry (in the column in which the
pivot resides) are zero.

• Let i < k. If the first non-zero entry of
row i occurs in column ji and the first
non-zero entry of row k occurs in
column jk, then ji < jk.
• The zero rows occur below non-zero rows.

The Gauss-Jordan Method reduces the given
system to Row-reduced Echelon form and
then “reads out” the solution.
To bring a matrix to Row-reduced Echelon

form, we first reduce it to Echelon form and
then make all entries above the pivot entries
(which we remember to make 1) zero,
starting from the rightmost pivot.

Example
Solve the following system by Gauss-
Jordan Elimination method.
x + y- z= 9
3x +11y + 3z = 21
-2x + 4y - 6z = 40

The Augmented matrix is
 1 1 1 9 
 3 11 3 21
  R2 – 3R1
 2 4 6 40 R3 + 2R1
1 1 1 9 

 0 8

6 6 (1/8)  R2

 0 6 8 58 
 1 1 1 9 

 0 1 3/ 4 
3/ 4 

 0 6 8 58  R3 - 6R2
1 1 1 9 
  0 1 3/ 4 3/ 4 
 
 0 0 25/ 2 125/ 2 (-2/25)  R3

 1 1 1 9  R1 + R3
 0 1 3/ 4 3/ 4 R2 – (3/4) R3
  
 0 0 1 5 
 1 1 0 4  R2 – R1
  0 1 0 3 

 0 0 1 5

x y z = RHS
 1 0 0 1
 0 1 0 3 
 
 0 0 1 5
Thus we read out the solution:

x = 1, y = 3, z = -5.

Example
Solve the following system by Gauss-
Jordan Elimination method.
4x + y = 4
5x - 3y + z= 2
- 9x + 2y - z= -6

The Augmented matrix is
 4 1 0 4  (1/4)  R1
 5 3 1 2 
 
 9 2 1 6
 1 1/ 4 0 1 

 5 3 1 2  R2 – 5R1
 
 9 2 1 6 R3 + 9R1
 1 1/ 4 0 1
 
 0 17 / 4 1 3 (-4/17)  R2
 
 0 17 / 4 1 3 
 1 1/ 4 0 1 

 0 1  4/17 12/17 
 
 0 17 / 4 1 3  R3 – (17/4) R2
 1 1/ 4 0 1  R1 – (1/4) R2

 0 1  4/17 12/17 
 
 0 0 0 0 
x y z = RHS
 1 0 1/17 14 /17 

 0 1  4 /17 12 /17 
 
 0 0 0 0 
We now read out the solution as
x = 14/17 - (1/17) z,
y = 12/17 + (4/17) z, z free.

Rank of a matrix
The rank of a m  n matrix A is defined as
the number of non-zero rows in the echelon
form of A.
Thus r  m, the number of rows of A.
Also to each nonzero row, there
corresponds a (pivotal) column of A.
Hence r  n, the number of columns of A.
Thus r  m and r  n.
Consider the system of linear equations
Ax b
We easily see that
The rank s of the augmented matrix A is
either r, the rank of the coefficient matrix A
or r+1 (one more than the rank of the
coefficient matrix).
Summary on the nature of the
solution
Let r be rank of the coefficient matrix and
let s be the rank of the augmented matrix.
Case (i) r < s.
In this case the system is
inconsistent (= has no solution).

Case (ii) r = s.
In this case the system is consistent
(= has a solution).
Subcase (a) r = n, the number of

unknowns.
In this case the system has a unique
solution.

Subcase (b) r < n, the number of
unknowns.
In this case the system has infinite
number of solutions.
The n – r variables corresponding to
the non-pivotal columns are free and
the remaining r variables
(corresponding to the pivotal columns)
can be expressed in terms of them.
Homogeneous Systems
A linear system Ax b is called
homogeneous if the RHS is zero in every

equation; i.e. b is the zero vector.
In this case the system is always consistent
and has always the trivial solution where
every unknown xj is zero.

It will have a non-trivial (= non-zero )
solution if and only if r < n.
Thus a homogeneous system will have

a nontrivial solution if and only if the
number of non-zero rows in the
echelon form of A < number of
unknowns.

Relation between the determinant of
a square matrix A and the
determinant of the row-reduced
echelon form of A
It is easily seen that the determinant of the
row-reduced echelon form of A =
a non-zero multiple of the determinant of A.
Reason
Let A be a n  n square matrix.
I. If B is a matrix got by interchanging
two rows of A, then det (B) = - det(A).
II. If B is a matrix got by multiplying a

row of A by a non-zero scalar c, then
det(B) = c det(A).
III. If B is a matrix got by adding to row i, a
constant times row k, then det(B) =
det(A).
Solve the systems of equations given below
by Gauss-Jordan Elimination method.
1. x1  x2  2 x3  3x4  1 2. x1  2 x2  4 x3  x4  4
2 x1  2 x2  2 x4  1 2 x1  x3  3x4  4
4 x1  x2  x3  x4  1 x1  2 x2  x3 0
x1  2 x2  3x3 1 3x1  x2  x3  5 x4  7
3. 2 x1  x3  x4  x5  2 4. x1  2 x2  x3  2 x4  0
x1  x3  x4  x5  1 2 x1  4 x2  2 x3  4 x4  4
12 x1  2 x2  8 x3  2 x5  12 3x1  6 x2  3x3  6 x4  6
Solutions
 1 1 2 3 1
 2 2 0 2 1 R2 – 2R1
1. 
 4 1 1 1 1 R3 – 4R1
 
 1 2 3 0 1 R4 – R1
 1 1 2 3 1
 
0 4 4 4 1 (1/4) R2
 
 0 5 9 13 3
 
 0 3 1 3 0 
 1 1 2 3 1 
 
0 1 1 1 1/ 4 

  0 5 9 13 3  R3 – 5R2
  R4 – 3R2
 0 3 1 3 0 
 1 1 2 3 1 
 0 1 1 1 1/ 4 
 
 0 0 4 8 7 / 4  (-1/4)R3
 
 0 0 4 0 3/ 4 

 1 1 2 3 1 
 0 1 1 1 1/ 4 

  0 0 1 2 7 /16 
 
 0 0 4 0 3/ 4  R4- 4 R3
 1 1 2 3 1 
 0 1 
1 1 1/ 4 

  0 0 1 2 7 /16 
 
 0 0 0 8 1  (-1/8)R4

 1 1 2 3 1  R1 - 3R4
 0 1 1 1 1/ 4  R2 + R4

  0 0 1 2 7 /16  R3 - 2R4
 
 0 0 0 1 1/8 
 1 1 2 0 5/ 8  R1 - 2 R3
 0 1 1 
0 1/ 8 R2 + R3

  0 0 1 0 3/16 
 
 0 0 0 1 1/ 8 

 1 1 0 0 1/ 4  R1 + R2
 0 1 0 0 1/16 

  0 0 1 0 3/16 
 
 0 0 0 1 1/ 8 
x1 x2 x3 x4 = RHS
1 0 0 0 5 /16 
0 1 0 0 1/16 
 
0 0 1 0 3/16 
 
0 0 0 1 1/ 8 
Hence the system has the unique solution
5 1 3 1
x1  , x2  , x3  , x4  .
16 16 16 8

 1 2 4 1 4
 2 0 1 3 4  R2 – 2R1
2. 
 1 2 1 0 0 R3 – R1
 
 3 1 1 5 7 R4 – 3R1
 1 2 4 1 4
 0 4 9 1 4 
 
 0 4 5 1 4 
(-1/4)R2
 
 0 5 13 8 5

 1 2 4 1 4
 
0 1 9 / 4 1/ 4 1 
 
 0 4 5 1 4  R3 + 4R2
 
 0 5 13 8 5 R4 + 5R2
 1 2 4 1 4
 1 9 / 4 1/ 4 
 

0
0 0 4 2
1
0  (1/4)R3
 
 0 0 7 / 4 37 / 4 0

 1 2 4 1 4
 0 1 9 / 4 1/ 4 
1

  0 0 1 1/ 2 0
  R4 + (7/4) R3
 0 0 7 / 4 37 / 4 0
 1 2 4 1 4
 0 1 9 / 4 1/ 4 1 
 
 0 0 1 1/ 2 0
 
 0 0 0 81/8 0  (-8/81)R4
 1 2 4 1 4 R1 – R4
 0 
1 9 / 4 1/ 4 1  R2 +(1/4) R4

  0 0 1 1/ 2 0  R3 +(1/2) R4
 
 0 0 0 1 0
 1 2 4 0 4  R1 – 4R3
 
 

0
0
1 9/ 4
0 1
0
0
1  R2 -(9/4) R3
0
 
 0 0 0 1 0

 1 2 0 0 4  R1 – 2R2
 0 1 0 0 1

  0 0 1 0 0
 
 0 0 0 1 0
x1 x2 x3 x4 = RHS
1 0 0 0 2
0 0 1  Hence the solution is:
 
0
1
0
0
1 0 0
x1 = 2, x2 = 1,
  x3 = 0, x4 = 0.
0 0 0 1 0
 2 0 1 1 1 2  (1/2) R1
3.  1 0 1 1 1 1 
 12 2 8 0 2 12
 1 0 1/ 2 1/ 2 1/ 2 1 
 1 0 1 
  1 1 1  R2 –R1
 12 2 8 0 2 12 R3 – 12R1
 1 0 1/ 2 1/ 2 1/ 2 1 
  0 0 1/ 2 1/ 2 1/ 2 0 
  R2  R3
 0 2 2 6  4 0
 1 0 1/ 2 1/ 2 1/ 2 1 
 0 2 2
 6  4 0  (1/2) R2

 0 0 1/ 2 1/ 2 1/ 2 0  2 R3
 1 0 1/ 2 1/ 2 1/ 2 1  R1 - (1/2)R3
0 1 1 
  3  2 0 R2 –R3
 0 0 1 1 1 0
 1 0 0 0 0 1
  0 1 0 4  3 0
 
 0 0 1 1 1 0

In equation form this is:
x1  0 x2  0 x3  0 x4  0 x5  1
0 x1  x2  0 x3  4 x4  3x5  0
0 x1  0 x2  x3  x4  x5  0
Thus the solution of the system is:
x1  1, x2   4 x4  3x5 , x3  x4  x5 ,
x4 , x5 free

 1 2 1 2 0 
4.  2 4 2 4 4 R2 – 2R1
 3 6 3 6 6  R3 – 3R1
 1 2 1 2 0 
 0 0 4 8 4 (1/4)R2
  
 0 0 6 12 6 
 1 2 1 2 0
 0 0 1 2 1
  
 0 0 6 12 6 R3 - 6R2
 1 2 1 2 0 R1 + R2
  0 0 1 2 1
 
 0 0 0 0 0
x1 x2 x3 x4 = RHS
 1 2 0 0 1
 0 0 1 2 1
  
 0 0 0 0 0

In equation form this is:
x1  2 x2  0 x3  0 x4  1
0 x1  0 x2  x3  2 x4 1
Thus the solution of the system is:
x1  1  2 x2 , x3 1  2 x4 ,
x2 , x4 free

5. Solve the system
x1  x3  2
x1  x2  2 x3  4
x1  x2  2 x3  4
x1  x2  x3  4
x1  3x2  x3  8

1 0 1 2
1 1 2 4 R2 –R1

1 1 2 4 R3 – R1
 
1 1 1 4 R4 – R1
 1 3 1 8  R5 – R1
 1 0 1 2
 0 1 3 2 

  0 1 1 2  R3 – R2
  R4 – R2
 0 1 2 2
 0 3 0 6  R5 – 3R2
 1 0 1 2
 0 1 3 2 

 0 0 4 0 (-1/4)R3
 
 0 0 1 0
 0 0 9 0 
 1 0 1 2
 0 1 3 2 

  0 0 1 0
  R4 + R3
 0 0 1 0
 0 0 9 0  R5 + 9R3
1 0 1 2  R1 + R3
0 1 3 2  R2 - 3R3

0 0 1 0
 
0 0 0 0
 0 0 0 0 
1 0 0 2
0  Thus the given system is
 1 0 2 consistent and has the
 0

0 1 0

unique solution:
0 0 0 0
x1 = 2, x2 = 2, x3 = 0.
 0 0 0 0 
Solve the following systems of linear
equations:
6. x1  x3  1 7. x1  2 x3  1
2 x1  x2  x3  2 2 x1  x2  2 x3  1
x2  x3  3 x2  2 x3  1
x1  x2  x3  4 x1  x2 1
2 x2  x3  0 x1  x2  4 x3  1
Ans: Inconsistent, Inconsistent

 1 0 1 1
Solu-  2 1 1 2  R2 –2R1
tion of 
 0 1 1 3
Q6.  
 1 1 1 4  R4 – R1
 0 2 1 0 
 1 0 1 1
 0 1 3 0 

  0 1 1 3 R3 – R2
  R4 – R2
 0 1 2 3
 0 2 1 0  R5 – 2R2
1 0 1 1
0 1 3 0 

 0 0 4 3 (-1/4)R3
 
0 0 1 3
 0 0 7 0 
1 0 1 1 
0 1 3 
0 

 0 0 1 3/ 4 
  R4 + R3
0 0 1 3 
 0 0 7 0  R5 + 7R3
 1 0 1 1 
 0 1 3 0  

  0 0 1 3/ 4 
 
 0 0 0 9/4 
 0 0 0 21/ 4 
This is the echelon form of the given

system. The fourth and fifth rows tells us
that the system is inconsistent.

Solve the following systems of linear
equations:
8. x1  x2  x3  6 x4  6 x5  19
x1  7 x4  7 x5  28
2 x2  3x3  18 x4  4 x5  24
9. x1  3x2  x3  x4  7
2 x1  4 x2  x3  6 x4  6
2 x1  x2  x4  0

Answer to Q8: x1  28  7 x4  7 x5
x2  165  57 x4  43x5
x3  118  44 x4  30 x5
x4 , x5 free
Answer to Q9: x1  1  4 x4
x2  2  7 x4
x3  26 x4 , x4 free
2/27/2018
***
Presented by Dr. M.S. Radhakrishnan BITS, Pilani 116

Solving Linear Systems with Gaussian Elimination

Загружено:

Сведения о документе

Исходное описание:

Оригинальное название

Авторское право

Доступные форматы

Поделиться этим документом

Поделиться или встроить документ

Параметры публикации

Этот документ был вам полезен?

Это неприемлемый материал?

Авторское право:

Доступные форматы

Solving Linear Systems with Gaussian Elimination

Загружено:

Авторское право:

Доступные форматы

Lecture 1 1

• Solutions of Linear Systems

2/27/2018 Presented by Dr. M.S. Radhakrishnan BITS, Pilani 3

A system of m linear equations in n

2/27/2018 Presented by Dr. M.S. Radhakrishnan BITS, Pilani 5

We say two systems of equations are

Thus a given system and a system obtained

2/27/2018 Presented by Dr. M.S. Radhakrishnan BITS, Pilani 10

We easily see that the two equations are

We also see the pictorial representation of the

2/27/2018 Presented by Dr. M.S. Radhakrishnan BITS, Pilani 12

2/27/2018 Presented by Dr. M.S. Radhakrishnan BITS, Pilani 13

2/27/2018 Presented by Dr. M.S. Radhakrishnan BITS, Pilani 14

2/27/2018 Presented by Dr. M.S. Radhakrishnan BITS, Pilani 15

2/27/2018 Presented by Dr. M.S. Radhakrishnan BITS, Pilani 16

Consistent (has at Inconsistent (has

Has a unique Has an infinite

2/27/2018 Presented by Dr. M.S. Radhakrishnan BITS, Pilani 17

Planes intersect at a point: consistent with

Consider the system of equations

2/27/2018 Presented by Dr. M.S. Radhakrishnan BITS, Pilani 27

The variables x, y, z need not be written in

2/27/2018 Presented by Dr. M.S. Radhakrishnan BITS, Pilani 35

2/27/2018 Presented by Dr. M.S. Radhakrishnan BITS, Pilani 37

2/27/2018 Presented by Dr. M.S. Radhakrishnan BITS, Pilani 39

2/27/2018 Presented by Dr. M.S. Radhakrishnan BITS, Pilani 42

2/27/2018 Presented by Dr. M.S. Radhakrishnan BITS, Pilani 43

2/27/2018 Presented by Dr. M.S. Radhakrishnan BITS, Pilani 46

2/27/2018 Presented by Dr. M.S. Radhakrishnan BITS, Pilani 47

2/27/2018 Presented by Dr. M.S. Radhakrishnan BITS, Pilani 50

2/27/2018 Presented by Dr. M.S. Radhakrishnan BITS, Pilani 51

2/27/2018 Presented by Dr. M.S. Radhakrishnan BITS, Pilani 54

2/27/2018 Presented by Dr. M.S. Radhakrishnan BITS, Pilani 56

2/27/2018 Presented by Dr. M.S. Radhakrishnan BITS, Pilani 58

2/27/2018 Presented by Dr. M.S. Radhakrishnan BITS, Pilani 60

2/27/2018 Presented by Dr. M.S. Radhakrishnan BITS, Pilani 62

2/27/2018 Presented by Dr. M.S. Radhakrishnan BITS, Pilani 63

2/27/2018 Presented by Dr. M.S. Radhakrishnan BITS, Pilani 66

2/27/2018 Presented by Dr. M.S. Radhakrishnan BITS, Pilani 67

2/27/2018 Presented by Dr. M.S. Radhakrishnan BITS, Pilani 68

To bring a matrix to Row-reduced Echelon

2/27/2018 Presented by Dr. M.S. Radhakrishnan BITS, Pilani 69

2/27/2018 Presented by Dr. M.S. Radhakrishnan BITS, Pilani 70

2/27/2018 Presented by Dr. M.S. Radhakrishnan BITS, Pilani 72

2/27/2018 Presented by Dr. M.S. Radhakrishnan BITS, Pilani 73

Thus we read out the solution:

2/27/2018 Presented by Dr. M.S. Radhakrishnan BITS, Pilani 74

2/27/2018 Presented by Dr. M.S. Radhakrishnan BITS, Pilani 75

2/27/2018 Presented by Dr. M.S. Radhakrishnan BITS, Pilani 79

2/27/2018 Presented by Dr. M.S. Radhakrishnan BITS, Pilani 82

Subcase (a) r = n, the number of

2/27/2018 Presented by Dr. M.S. Radhakrishnan BITS, Pilani 83

A linear system Ax b is called

homogeneous if the RHS is zero in every

2/27/2018 Presented by Dr. M.S. Radhakrishnan BITS, Pilani 85

Thus a homogeneous system will have

2/27/2018 Presented by Dr. M.S. Radhakrishnan BITS, Pilani 86

II. If B is a matrix got by multiplying a

2/27/2018 Presented by Dr. M.S. Radhakrishnan BITS, Pilani 91

2/27/2018 Presented by Dr. M.S. Radhakrishnan BITS, Pilani 92

2/27/2018 Presented by Dr. M.S. Radhakrishnan BITS, Pilani 93

2/27/2018 Presented by Dr. M.S. Radhakrishnan BITS, Pilani 95