Lecture 3

Ajay Kumar
[email protected]

August 10, 2023

System of Linear Equations
 Linear equation
• A line in xy-coordinate system can be represented by an
equation ax + by = c, where a, b, c ∈ R and both a, b are not zero.
• A plane in xyz-coordinate system can be represented by an
equation ax + by + cz = d, where a, b, c, d ∈ R and a, b, c all are
not zero.
Note that equation of line is a linear equation in two variable, and
equation of plane is a linear equation in three variable. More
generally, a linear equation in n-variable is of the form

a1 x1 + a2 x2 + · · · + an xn = b,

where ai ∈ R for all i, b ∈ R and atleast one of ai should be nonzero.

In the above equation, if b = 0, then we say that homogeneous linear
equation in n variables.
Examples.
√
• 2 x + 3y − 4z = 5 is not a linear equation.
• sin(x) = y is not a linear equation.
• x2 + y = 0 is not a linear equation.
• 2x1 + 3x4 + 5x3 = 5 is linear equation but not homogeneous.
• x = y is homogeneous linear equation.
A finite set of linear equations is known as linear system of equation.
A general system of linear equations of n-unknowns x1 , . . . , xn can be
written
a11 x1 + a12 x2 + · · · + a1n xn = b1
a21 x1 + a22 x2 + · · · + a2n xn = b2
.. .. .. .. .. (1)
. . . . .
am1 x1 + am2 x2 + · · · + amn xn = bm
If bi = 0 for all i, then the above system is known as system of
homogeneous linear equations.
 Solution of Linear system of equation

• A solution to the system of linear equation (1) in n-unknown is a

sequence of numbers s1 , s2 , . . . , sn such that substitution of xi = si
for all i satisfies (1). This solution can be written as (s1 , s2 , . . . , sn ).
• Solution in two variable:

No solution. One solution. Infinite solutions.

x+y=3 x+y=3 x+y=3
2x + 2y = 8 2x + y = 8 2x + 2y = 6
 Solution in Three variables
• No solution:

• One solution:
• Infinite solutions:
Question
How to solve a linear system in n-variables?
• Write system of linear equation (1) in matrix form Ax = b,where
     
a11 a12 . . . a1n x1 b1
 a21 a22 . . . a2n  x2  b2 
     
A=  .. .. ..  , x =  ..  and b =  ..  .
..     
 . . . .  . .
am1 am2 . . . amn xn bn

• The matrix [A|b] is called augmented matrix associated to linear

system (1).
• We need to define some operations on matrix to solve the linear
system.
 Two Variable

a1 X + b1 Y = c1 (2)
a2 X + b2 Y = c2 (3)

How to find the solution of above system?

Assume a1 6= 0. Multiply (2) equation by a2 /a1 and subtract from (3).
We get
a1 b2 − a2 b1 a1 c2 − a2 c1
0X + Y=
a1 a1
We get a new system of equation

a1 X + b1 Y = c1 (4)
a1 b2 − a2 b1 a1 c2 − a2 c1
0X + Y= (5)
a1 a1
The augmented matrix associated to original system is
" #
a1 b1 : c1
a2 b2 : c2

The augmented matrix associated to a new system is

" #
a1 b1 : c1
0 a1 b2 a−a
1
2 b1
: a1 c2 a−a
1
2 c1

Note that we can obtain the second matrix from the first matrix if we
multiply the first row of first matrix by a2 /a1 and subtract from 2nd row.
 Elementary Row Operations

Let A ∈ Rm×n . We can transformed the matrix A using the following

process:
• Interchange ith and jth rows of A
• Multiply ith row of A by non-zero scalar
• Add a scalar multiple of ith row of A to jth row of A.
These three operations are known as elementary row operations.

Example
" # " #
1 3 5 R2 →R2 −2R1 1 3 5
−−−−−−−→
2 −3 6 0 −9 −4
 Reduced Row Echelon Form

• The first non-zero entry of a non-zero row is called the pivot.

• A matrix A is said to be in row echelon form(REF) if it satisfy the
following:
• The non-zero rows of A precede the zero rows of A.
• Suppose A has r non-zero rows and the pivot of ith row is in ki th
column. Then k1 < k2 < · · · < kr .
• All pivots of A are equal to 1.
• A column containing a pivot called a pivotal column.
• Further, suppose that pivot is the only non zero entry of a pivotal
column of A. Then we say that A is said to be in reduced row
echelon form(RREF).
  
0 1 2 3 0 4
0 0 0 0 0 0
• 
  Not in row echelon form.

0 0 0 1 1 0
0 0 0 0 1 7
 
0 1 2 0 0 4
0 0 0 0 1 2
• 
  Not in row echelon form

0 0 0 1 0 0
0 0 0 0 0 0
 
0 1 2 3 0 4
0 0 0 1 0 0
• 
  Row echelon form

0 0 0 0 1 6
0 0 0 0 0 0
•
 
0 1 2 0 0 4
0 0 0 1 0 0
• 
  Reduced row echelon form

0 0 0 0 1 2
0 0 0 0 0 0
 Convert to Reduced Row Echelon form

1. Find the left most non-zero column of A.

2. Interchange rows, if necessary to make the non zero entry to the
top of column found in step 1.
3. By multiplying a nonzero scalar make the top nonzero entry
found above to 1.
4. Add scalar multiple of top rows to below rows so that all entries
below pivot becomes zero.
5. Now, cover the top row and repeat the steps to the remaining
submatrix until the matrix is in row echelon form.
6. Beginning with the last nonzero row and working upward, add
suitable multiple of each row to the rows above to make zeros
above the pivot.
 Example

     
0 1 2 1 2 1 1 2 1
 R1 ↔R2   R3 →R3 −2R1   R3 →− 13 R3
1 2 1 −−− −→ 0 1 2 −− −−− −−→ 0 1 2 −−−−−−→


2 1 2 2 1 2 0 −3 0
     
1 2 1 1 2 1 1
1 2 1
 R3 →R3 −R2   R3 →− 2 R3 
0 1 2 −− −−− −→ 0 1 2 −−− −−− → 0 1 2(REF)
 

0 1 0 0 0 −2 0 0 1

R2 → R2 − 2R3    
R1 → R1 − R3 1 2 0
 R1 →R1 −2R2 
1 0 0
−−−−−−−−−−−→ 0 1 0 −−−−−−−→ 0 1 0


0 0 1 0 0 1
 Row Equivalent Matrices

Definition
Matrices A and B are said to be row equivalent if either can be
obtained from the other by a sequence of elementary row operations.

Theorem
If the augmented matrices of two system of linear equations are row
equivalent, then both system have the same set of solution.

Theorem
Any matrix is row equivalent to row echelon form matrix.
 Gauss Elimination

Let Ax = b be the given system and let A0 be the REF of the

augmented matrix [A|b].
• Variables corresponding to pivotal columns of A0 are known as
leading variables.
• Variables corresponding to non-pivotal columns of A0 are known
as free variables.
• Identify free variables and assign arbitrary values to them.
• By back substitution, find the solution of the problem.
• Gauss-Jordan elimination: Convert the augmented matrix [A|b]
to RREF instead of REF
 Example
 
1 1 1 1 : 5
Let [A : b] = 2 3 2 4 : 6  .One can check that REF of
 

3 4 3 5 : 11
 
1 1 1 1 : 5
augmented matrix is 0 1 0 2 : −4 .
 

0 0 0 0 : 0
Now, 1st and 2nd columns are pivotal columns. So x1 and x2 are
leading variables, and x3 , x4 are free variable. Now we assign x3 = s
and x4 = t, where s, t ∈ R. Now we get system of equations is

x1 + x2 + x3 + x4 = 5
0x1 + x2 + 0x3 + 2x4 = −4
0x1 + 0x2 + 0x3 + 0x4 = 0
After substituting the values of x3 and x4 in 2nd equation, we get
x2 + 0s + 2t = −4 =⇒ x2 = −4 − 2t.
Now substitute the value of x2 , x3 and x4 in first equation, we get
x1 − 4 − 2t + s + t = 5 =⇒ x1 = 9 − s + t. Thus, we get
{(9 − s + t, −4 − 2t, s, t) : s, t ∈ R}.