LA-W6 Lineer Alg

Download as pdf or txt
Download as pdf or txt
You are on page 1of 11

Linear Algebra and Applications

MAT281E – CRN: 21534

Week #6

Linear Algebra and Applications 1


Note: Result: 1 1 2 4
If a matrix A, has zero nullspace, this If a matrix A, has dimensions 𝑛 > 𝑚, 𝐴= 1 2 2 5
means that all the columns of A are that is, if 𝐴𝒙 has more unknowns than
independent. equations there must be at least one
1 3 2 6
(All columns have pivots and there are free column. p p
no free columns.) f f
Then, 𝐴𝒙 = 𝟎 has at least one non-zero
solution. Two pivot columns: Rank2
There are two special solutions for
The dimension of a nullspace is the
number of free variables.
𝐴𝒙 = 𝟎.

Rank of a matrix: Rank1 matrix :

The rank of A is the number of pivots. 1 3 10 𝑅 1 3 10


2 6 20 ՜ 0 0 0
𝑚 × 𝑛 is the dimension of a matrix, but
not the true size of a linear system. 3 9 30 0 0 0
The true size of A is given by its rank. Every row is a multiple of pivot row.
EXAMPLE:
1 1 2 4 Find the special 𝑥1 + 2𝑥3 + 3𝑥4 = 0 −2 −3
𝐴= 1 2 2 5 solutions for 𝐴𝒙 = 𝟎. 0 + 𝑥 −1
1 3 2 6 𝑥2 + 𝑥4 = 0 𝒙 = 𝑥3 4 0
1
0 1
𝑥1 = −2𝑥3 − 3𝑥4
1 0 2 3
𝐴՜𝑈՜𝑅= 0 1 0 1 𝑥2 = −𝑥4 𝑥3 = 1; 𝑥4 = 0 𝑥3 = 0; 𝑥4 = 1
0 0 0 0 𝑥1 −2𝑥3 − 3𝑥4 𝒔𝟏 = (−2,0,1,0) 𝒔𝟐 = (−3, −1,0,1)
2 pivot columns and two free columns. 𝑥2 −𝑥4 Special solutions
𝑥3 = 𝑥3
We will have two special solutions for 𝑥4 𝑥4
𝑅𝒙 = 𝟎.

First find the Nullspace ; 𝑅𝒙 = 𝟎

𝑥
1 0 2 3 𝑥1
0 1 0 1 𝑥2 = 𝟎
3
0 0 0 0 𝑥
4
RANK-1 Another definition of rank :
Matrices of Rank-1 have only one
pivot. The rank is the dimension of the
column space (or the row space).
1 3 10 𝑅 1 3 10
2 6 20 ՜ 0 0 0
3 9 30 0 0 0 𝑑𝑖𝑚𝐶 𝐴 = 𝑟𝑎𝑛𝑘 = 𝑑𝑖𝑚𝐶(𝐴𝑇 )
𝑑𝑖𝑚𝑒𝑛𝑠𝑖𝑜𝑛 𝑜𝑓 𝑡ℎ𝑒
𝑟𝑜𝑤 𝑠𝑝𝑎𝑐𝑒
A can also be written as;

1 1 3 10
𝐴= 2 1 3 10 = 2 6 20
3 3 9 30

NOTE: Every Rank-1 matrix can be


written as 𝐴 = 𝑢𝑣 𝑇 where 𝑢 and 𝑣 are
Column matrices.
The complete solution to 𝐴𝒙 = 𝒃 Apply the usual elimination steps
In the last section 𝒃 = 𝟎.
𝐴 𝒃 ՜ Subtract Row1 from
Now, 𝒃 ≠ 𝟎. Row3 then, subtract
So, 𝐴𝒙 = 𝒃 ՜ 𝑅𝒙 = 𝒅 row2 from row3

Let’s use the augmented matrix to 1 3 0 2 1


Show this. Consider, ՜ 0 0 1 4 6 = 𝑅 𝒅
0 0 0 0 0
𝑥
1 3 0 2 𝑥1 1 For general 𝒃 what we did is;
2
0 0 1 4 𝑥 = 6
3
1 3 1 6 𝑥 ด
7 1 3 0 2 𝑏1 1 3 0 2 𝑏1
𝐴
ถ4
𝒃
𝒙 𝐴 𝒃 = 0 0 1 4 𝑏2 ՜ 0 0 1 4 𝑏2 = 𝑅 𝒅
This can be represented by the 1 3 1 6 𝑏3 0 0 0 0 𝑏3 − 𝑏1 − 𝑏2
augmented matrix;
1 3 0 2 1 Here, 𝑏3 − 𝑏1 − 𝑏2 = 0.
0 0 1 4 6 = 𝐴 𝒃 So equations can be solved for known
1 3 1 6 7 𝑏1 and 𝑏2 .
Special , particular and complete 𝑥1 = 1 − 3𝑥2 − 2𝑥4 We find the particular solution by
solutions for 𝑅𝒙 = 𝒅 setting 𝑥2 = 𝑥4 = 0;
𝑥3 = 6 − 4𝑥4
𝒙𝒑 = (1,0,6,0).
1 3 0 2 1 𝑥1 and 𝑥3 are our pivot variables
0 0 1 4 6 = 𝑅 𝒅 (basic variables). So we write
Method of finding the particular solution
0 0 0 0 0 the pivot variables in terms of free can be summarized as set
variables (𝑥2 , 𝑥4 ). 𝑓𝑟𝑒𝑒 𝑣𝑎𝑟𝑖𝑎𝑏𝑙𝑒𝑠 = 0, then get pivot
𝑅𝒙 = 𝒅; variables from 𝒅.
𝑥1 1 − 3𝑥2 − 2𝑥4
𝑥 On the other hand we have two special
1 3 0 2 1 𝑥1 1 𝑥2 𝑥2
2 𝑥3 = solutions
0 0 1 4 6 𝑥 = 6 6 − 4𝑥4
3
0 0 0 0 0 𝑥 0 𝑥4 𝑥4
4 𝒔𝟏 = −3,1,0,0
𝒔𝟐 = (−2,0, −4,1)
1 −3 −2
𝑥1 + 3𝑥2 + 2𝑥4 = 1
𝒙 = 0 + 𝑥2 1 + 𝑥4 0
𝑥3 + 4𝑥4 = 6 6 0 −4
0 0 1

Complete solution for 𝒙.


Therefore;
EXAMPLE: Given,
𝐶𝑜𝑚𝑝𝑙𝑒𝑡𝑒 𝑠𝑜𝑙𝑢𝑡𝑖𝑜𝑛 = 𝑝𝑎𝑟𝑡𝑖𝑐𝑢𝑙𝑎𝑟 𝑠𝑜𝑙𝑢𝑡𝑖𝑜𝑛
+𝑠𝑝𝑒𝑐𝑖𝑎𝑙 𝑠𝑜𝑙𝑢𝑡𝑖𝑜𝑛𝑠. 1 1 𝑏1
𝒙 = 𝒙𝑝 + 𝒙𝑛 𝐴= 1 2 and 𝑏 = 𝑏2
−2 −3 𝑏3
Particular solution solves
find the condition on (𝑏1 , 𝑏2 , 𝑏3 ) for
𝑅𝒙𝒑 = 𝒅 𝐴𝒙 = 𝒃 to be solvable.
Discuss the complete solution.
The 𝑛 − 𝑟 special solutions solve
𝑅𝒙𝒏 = 𝟎 First remind that, 𝐴𝒙 = 𝒃 is solvable ⇔ 𝑏 is in the 𝐶 𝐴 .
Next, use the augmented matrix 𝐴 𝒃
Note: If a square invertible matrix to reach 𝑅 𝒅 by using elimination steps.
has dimensions 𝑚 = 𝑛 = 𝑟,
Then we have 𝑛 − 𝑟 = 0 special solutions.

Only solution is the particular solution

𝒙𝒑 = 𝐴−1 𝒃
The last line is 0 = 0 provided that
𝑏3 + 𝑏2 + 𝑏1 = 0
This means that 𝐴𝒙 = 𝒃 is solvable
When 𝑏3 + 𝑏2 + 𝑏1 = 0.

This is called ‘full column rank’ (𝑛 = 𝑟)


Full Row Rank

EXAMPLE:
Therefore, the complete solution,
2 −3
𝒙 = 1 + 𝑥3 2
0 1

You might also like