Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
1
System of Linear Equations
When mathematics is used to solve a problem it often
becomes necessary to find a solution to a so-called system
of linear equations. Historically, linear algebra developed
from studying method for solving such equations. This
module introduces methods for solving system of linear
equations and includes examples of problems that reduce
to solving such equations. The techniques of this module
will be used throughout the remainder of the course.
Example: 𝑥 + 3𝑦 = 9.
The graph of this equation is a straight line in the 𝑥 − 𝑦
plane.
−2𝑥 + 𝑦 = −4
A pair of values of 𝑥 and 𝑦 that satisfy both the equations is
called a solution. It can be seen by substitution that
𝑥 = 3, 𝑦 = 2 is a solution to this system. A solution to such a
system will be a point at which the graphs of the two
equations intersect. The following examples illustrate that
three possibilities can arise for such system of equations.
There can be a unique solution, no solution, or many
solutions.
Unique Solution
𝑥 + 3𝑦 = 9
−2𝑥 + 𝑦 = −4
Unique solution, 𝑥 = 3, 𝑦 = 2.
x+3y=9
-2x+y=-4
(3,2)
No Solution
−2𝑥 + 𝑦 = 3
−4𝑥 + 2𝑦 = 2
Lines are parallel. No point of intersection.
-4x+2y=2
-2x+y=3
Many Solutions
4𝑥 − 2𝑦 = 6
6𝑥 − 3𝑦 = 9
Both equations have the same graph. Any point of the
graph is a solution.
4x-2y=6
6x-3y=9
Our aim in this module is to analyze larger system of
linear equation. The following is an example of a system of
three linear equations.
𝑥1 + 𝑥2 + 𝑥3 = 2
2𝑥1 + 3𝑥2 + 𝑥3 = 3
𝑥1 − 𝑥2 − 2𝑥3 = −6
C
Three planes A, B and C intersects at a single point P.
P corresponds to a unique solution.
A
A Q
C
Three planes A, B and C intersects in a line PQ.
Any point on the line is a solution
𝑎𝑚1 𝑥1 + 𝑎𝑚 2 𝑥2 + ⋯ + 𝑎𝑚𝑛 𝑥𝑛 = 𝑏𝑚
The matrix representation of the above system of linear
equations is
𝑥1 𝑏1
𝑥2 𝑏2
𝑎11 𝑎12 … . . 𝑎1𝑛
. .
𝑎21 𝑎22 … . . 𝑎2𝑛
. =
𝑎𝑚1 𝑎𝑚2 … . . 𝑎𝑚𝑛 .
𝑚×𝑛 . .
𝑥𝑛 𝑛×1 𝑏𝑚 𝑚×1
i.e., 𝐴𝑋 = 𝐵
1 7 0 8
1 0 0 2
0 1 0 3
0 0 1 4
0 0 1 2
0 1 0 3
0 0 0 0
Leading 1 in row 3 not to Non zero element
the right of leading 1 in above leading 1 in row
row 2. 2
𝑥1 − 2𝑥2 + 4𝑥3 = 12
2𝑥1 − 𝑥2 + 5𝑥3 = 18
Step 1:
Start with the augmented matrix and use the first row
to create zeros in the first column (This corresponds to
using the first equation to eliminate 𝑥1 from the second and
third equations).
1 2 4 12 1 2 4 12
2 1 5 18 R 2 2 R1 0 3 3 6
1 3 3 8 R3 R1 0 1 1 4
Step 2:
1
Next, multiply row 2 by to make the (3,3) element
3
1. (This corresponds to making the coefficient of 𝑥2 in the
second equation 1.)
1 2 4 12
0 1 1 2
1
R2 0 1 1 4
3
Step 3:
Create zeros in the second column as follows. (This
corresponds to using the second equation to eliminate 𝑥2
from the first and third equations.)
1 0 2 8
R1 (2) R 2 0 1 1 2
R3 (1) R 2 0 0 2 6
Step 4:
1
Multiply row 3 by . (This corresponds to making
2
the coefficient of in the third equation 1.)
1 0 2 8
0 1 1 2
1
R3 0 0 1 3
2
1 0 0 2
R1 2 R3 0 1 0 1
R 2 R3 0 0 1 3
𝑥2 = 1
𝑥3 = 3
The solution is 𝑥1 = 2, 𝑥2 = 1, 𝑥3 = 3.
Problem 2: Solve, if possible, the system of equations.
2𝑥3 − 2𝑥4 = 2
3𝑥1 + 3𝑥2 − 3𝑥3 + 9𝑥4 = 12
Step 1:
We interchange rows, if necessary, to bring a nonzero
element to the top of the first nonzero column. This
nonzero element is called a pivot.
Pivot
3 3 3 9 12
0 0 2 2 2
R1 R 2
4 4 2 11 12
Step 2:
Create a 1 in the pivot location by multiplying the
1
pivot row by
𝑝𝑖𝑣𝑜𝑡
1 1 1 3 4
0 0 2 2 2
1
R1 4 4 2 11 12
3
Step 3:
Create zeros elsewhere in the pivot column by
adding suitable multiple of the pivot row to all other rows of
the matrix
1 1 1 3 4
0 0 2 2 2
R3 (4) R1
0 0 2 1 4
Step 4:
Cover the pivot row and all rows above it. Repeat
steps 1 and 2 for the remaining sub Matrix. Repeat step 3
for the whole matrix. Continue until the reduced echelon
from is reached. Pivot
1 1 1 3 4
0 0 2 2 2
0 0 2 1 4
1 1 1 3 4
0 0 1 1 1
1
R2 0 0 2 1 4
2
1 1 0 2 5
R1 R 2 0 0 1 1 1
R3 (2) R 2 0 0 0 1 6
Pivot
1 1 0 0 17
R1 (2) R3 0 0 1 0 5
R 2 R3 0 0 0 1 6
𝑥1 + 𝑥2 = 17
𝑥3 = −5
𝑥4 = −6
3𝑥1 − 5𝑥2 − 𝑥3 = 7
Solution: Start with the augmented matrix and follow the
Gauss-Jordan algorithm. Pivots and leading 1s are circled
3 3 3 9 1 1 1 3
2 1 4 7 2 1 4 7
1
3 5 1 7 R1 3 5 1 7
3
1 1 1 3 1 0 3 4
0 1 2 1 R1 R2 0 1 2 1
R2 (2)R1
R3 (3)R1 0 2 4 2 R3 2R2 0 0 0 0
𝑥2 + 2𝑥3 = 1
x2 2x3 1
r 1 gives x1 1, x2 1, x3 1
𝑥1 − 𝑥2 + 2𝑥3 = 3
2𝑥1 − 2𝑥2 + 5𝑥3 = 4
𝑥1 + 2𝑥2 − 𝑥3 = −3
2𝑥2 + 2𝑥3 = 1
1 1 2 3 1 1 2 3
2 2 5 4 0 0 1 2
R 2 (2) R1
1 2 1 3 0 3 3 6
R3 (1) R1
0 2 2 1 0 2 2 1
1 1 2 3
0 3 3 6
R 2 R3 0 0 1 2
0 2 2 1
1 1 2 3
0 1 1 2
1
R2 0 0 1 2
3
0 2 2 1
1 0 11
0 1 1 2
R1 R 2
0 0 1 2
R 4 (2) R 2
0 0 4 5
1 0 0 3
R1 (1) R3 0 1 0 4
R 2 R3 0 0 1 2
R 4 (4) R3 0 0 0 13
1 0 0 3
0 1 0 4
1
R4 0 0 1 2
13
0 0 0 1
This matrix is still not in reduced echelon form; zeros still
have to be created above the 1 in the last row. However, in
such a situation, when the last nonzero row of a matrix is of
the form (0 0 . . . 0 1), there is no need to proceed further.
They system has no solution. To see this, let us write down
the equation that corresponds to the last row of the matrix.
(a) 𝑥1 + 3𝑥2 = 7
2𝑥1 − 5𝑥2 = −3
𝑥1 + 3𝑥2 = 6
4𝑥2 + 3𝑥3 = 0
𝑥1 − 𝑥3 = 6
8 7 5 1
(b) 4 6 2 4
9 3 7 6
0 2 4
(c) 5 7 3
6 0 8
1 2 1 6
(d) 0 1 4 5
0 0 1 2
1 2 3 1
1 1 7 1 R 2 R1
(b)
2 4 5 3 R3 2 R1
1 0 4 3
0 1 3 2 R1 4 R3
(c)
0 0 1 5 R 2 3R3
𝑥1 − 𝑥2 = −4
2𝑥1 − 𝑥2 + 2𝑥3 = −1
1 2 5 6
(b) 0 1 3 7
1 0 0
(c) 0 1 0
0 0 1
1 0 0 3 2
(d) 0 2 0 6 1
0 0 1 2 3
6. Determine the following matrices are in reduced echelon
form. If a matrix is not in reduced echelon form, give a
reason.
1 0 3 2
(a) 0 0 1 8
0 1 4 9
1 5 0 2 0
0 0 1 9 0
(b)
0 0 0 0 1
0 0 0 0 0
1 0 2 0 3
0 0 0 0 0
(c)
0 1 2 0 7
0 0 0 1 3
1 0 0 5 3
(d) 0 0 1 0 3
0 1 2 3 7
1 5 3 0 7
(e) 0 0 0 1 4
0 0 0 0 0
(a) 𝑥1 +2𝑥2 − 𝑥3 − 𝑥4 = 0
𝑥1 + 2𝑥2 + 𝑥4 = 4
(c) 𝑥1 − 𝑥2 − 2𝑥3 = 7
2𝑥1 −2𝑥2 + 2𝑥3 − 4𝑥4 = 12
−𝑥1 + 𝑥2 − 𝑥3 + 2𝑥4 = −4
Answers:
1 3 5 5 2 4
1 3 2 2 4 0 4 3
1. (a) 2 5 (b) (c)
1 3 0 1 0 1
(c) −2𝑥2 = 4
5𝑥1 + 7𝑥2 = −3
6𝑥1 =8
(d) 𝑥1 + 2𝑥2 − 𝑥3 = 6
𝑥2 + 4𝑥3 = 5
𝑥3 = −2
1 3 2 0 1 2 3 1 1 0 0 23
3. (a) 1 2 3 6 (b) 0 3 10 0 (c) 0 1 0 17
8 3 2 5 0 8 1 1 0 0 1 5
4. (a) 𝑥1 = 3, 𝑥2 = 0, 𝑥3 = 2 (b) 𝑥1 = 0, 𝑥2 = 4, 𝑥3 = 2
(c) 𝑥1 = 2, 𝑥2 = 3, 𝑥3 = −1
5. (a) Yes (b) No (c) Yes (d) No
(c) No solution
1.2
Vector Space ℝ𝑛
Introduction to vectors
The locations of points in a plane are usually discussed in
terms of a coordinate system. For example, in Figure below,
the location of each point in the plane can be described
using a rectangular coordinate system. The point 𝐴 is the
point (5,3).
EXAMPLE 1:
The direction will depend upon the sign of the scalar. The
general result is as follows. Let 𝑢 be a vector and 𝑐 a scalar.
The direction of 𝑐𝑢 will be the same as the direction of 𝑢 if
𝑐 > 0, the opposite direction to 𝑢 if 𝑐 < 0. The length of 𝑐𝑢 is
𝑐 times the length of 𝑢. See the figure below.
Zero vector:
The vector (0,0, … ,0) having 𝑛 zero components, is called the
zero vector of ℝ𝒏 and is denoted by 0. For example, (0,0,0)
is zero vector of ℝ3 . We shall find that zero vectors play a
central role in the development of vector spaces.
Negative Vector:
THEOREM:
Let 𝒖, 𝒗 and 𝒘 be vectors in ℝ𝑛 and let 𝑐 and 𝑑 be scalars.
(a) 𝒖 + 𝒗 = 𝒗 + 𝒖
(b) 𝒖 + (𝒗 + 𝒘) = (𝒖 + 𝒗) + 𝒘
(c) 𝒖 + 𝟎 = 𝟎 + 𝒖 = 𝒖
(d) 𝒖 + (−𝒖) = 𝟎
(e) 𝒄 (𝒖 + 𝒗) = 𝒄𝒖 + 𝒄𝒗
(f) (𝒄 + 𝒅) 𝒖 = 𝒄𝒖 + 𝒅𝒖
(g) 𝒄 (𝒅𝒖) = (𝒄𝒅)𝒖
(h) 𝟏𝒖 = 𝒖
These results are verified by writing the vectors in terms of
components and using the definitions of vector addition
and scalar multiplication, and the properties of real
numbers. We give the proofs of (a) and (e).
𝒖 + 𝒗 = 𝒗 + 𝒖:
Let 𝑢 = 𝑢1 , … . . 𝑢𝑛 and 𝑣 = 𝑣1 , … . . 𝑣𝑛 Then
𝑢 + 𝑣 = 𝑢1 , … . . 𝑢𝑛 + 𝑣1 , … . . 𝑣𝑛
= 𝑢1 + 𝑣1 , … 𝑢𝑛 + 𝑣𝑛
= 𝑣1 + 𝑢1 , … 𝑣𝑛 + 𝑢𝑛
=𝑣+𝑢
𝒄(𝒖 + 𝒗) = 𝒄𝒖 + 𝒄𝒗:
𝑐 𝒖 + 𝒗 = 𝑐 𝑢1 , … . . 𝑢𝑛 + 𝑣1 , … . . 𝑣𝑛
= 𝑐( 𝑢1 + 𝑣1 , … 𝑢𝑛 + 𝑣𝑛 )
= 𝑐 𝑢1 + 𝑣1 , … , 𝑐 𝑢𝑛 + 𝑣𝑛
𝑎 + 𝑏 ∈ 𝑉 𝐹 , 𝛼𝑎 ∈ 𝑉 𝐹 .
2) Commutativity:
𝑎+𝑏 =𝑏+𝑎
3) Associativity:
∃ 0 ∈ 𝑉 𝐹 such that 𝑎 + 0 = 0 + 𝑎 = 𝑎
5) Existence of additive inverse:
∃ − 𝑎 ∈ 𝑉 𝐹 such that 𝑎 + −𝑎 = −𝑎 + 𝑎 = 0
6) Distributive Laws:
𝛼 𝑎 + 𝑏 = 𝛼𝑎 + 𝛼𝑏
7) 1𝐹 𝑎 = 𝑎
8) 𝛼𝛽 𝑎 = 𝛼 𝛽𝑎 .
Note: Throughout this course we use the field of real
numbers as scalar set. We may refer a vector space simply
Example:
1) The set of all 2×2 matrices with entries real numbers is
a vector space over the field of real numbers under usual
addition and scalar multiplication of matrices.
2) The set of all functions having real numbers as their
domain is a vector space over the field of real numbers
under the following operations.
𝒇+𝒈 𝒙 =𝒇 𝒙 +𝒈 𝒙
𝒄𝒇 𝒙 = 𝒄. 𝒇 𝒙
For all functions 𝑓 and 𝑔 and scalar 𝑐.
We now give a theorem that contains useful properties of
vectors. These are properties that were immediately
apparent for ℝ𝑛 and were taken almost for granted. They
are not, however, so apparent for all vector spaces.
Theorem: Let „V‟ be a vector space, 𝑣 a vector in 𝑉, 0 the
zero vector of 𝑉, ‘𝑐’ a scalar, and 0 the zero scalar. Then
(a) 0𝑣 = 0
(b) 𝑐𝑣 = 0
(c) −1 𝑣 = −𝑣
(d) If 𝑐𝑣 = 0, then either 𝑐 = 0 or 𝑣 = 0
Proof: (a) 0𝑣 + 0𝑣 = 0 + 0 𝑣 = 0𝑣
Add the negative of 0𝑣 namely −0𝑣 to both sides of this
equation.
0𝑣 + 0𝑣 + −0𝑣 = 0𝑣 + −0𝑣
⇒ 0𝑣 + 0𝑣 + −0𝑣 =0
⇒ 0𝑣 + 0 = 0
⇒ 0𝑣 = 0
(b) 𝑐0 = 𝑐 0 + 0
⇒ 𝑐0 + 𝑐0
⇒ 𝑐0 = 0.
(c) −1 𝑣 + 𝑣 = −1 𝑣 + 1𝑣
= −1 + 1 𝑣
= 0𝑣 = 0
Thus (-1) 𝑣 is the additive inverse of 𝑣.
i.e (-1) 𝑣 = −𝑣
(d) Assume that 𝑐 ≠ 0𝐹 .
𝑐𝑣 = 0 ⇒ 𝑐 −1 𝑐𝑣 = 𝑐 −1 0
⇒ 𝑐 −1 𝑐𝑣 = 0
⇒ 1𝐹 𝑣 = 0
⇒ 𝑣 = 0.
SUBSPACES
Definition: Let 𝑉(𝐹) be a vector space. A non-empty subset
𝑈 ⊆ 𝑉 which is also a vector space under the inherited
operations of 𝑉 is called a vector subspace of 𝑉.
So 0 = 𝑣 − 𝑣 ∈ 𝑈
All other properties hold as they hold in 𝑉.
Therefore ‘𝑈’ is a vector space with the same binary
operations as on 𝑉 and hence 𝑈 is a subspace of 𝑉.
Theorem: Let X⊆ 𝑉, 𝑌 ⊆ 𝑉 be vector subspaces of a vector
space 𝑉(𝐹). Then their intersection 𝑋 ∩ 𝑌 is also a vector
subspace of 𝑉.
Proof: Follows from the above Theorem.
Problem 1: Let 𝑢 = −1,4,3,7 and 𝑣 = −2, −3,1,0 be vectors
in ℝ4 . Find 𝑢 + 𝑣 and 3𝑢.
Solution: We get
= 20,7, −31 .
Problem 3: Let ℂ denote the complex numbers and ℝ
denote the real numbers. Is ℂ a vector space over ℝ under
ordinary addition and multiplication? Is ℝ a vector space
over ℂ?
Suppose that 𝑈1 ⊈ 𝑈2 on 𝑈2 ⊈ 𝑈1
So ∃ 𝑢1 ∈ 𝑈1 and 𝑈2 ∈ 𝑈1 consider 𝑢1 + 𝑢2
Then 𝑢1 + 𝑢2 cannot be in 𝑈1 and 𝑈2 .
(If 𝑢1 + 𝑢2 ∈ 𝑈1 , then 𝑢1 + 𝑢2 − 𝑢1 = 𝑢2 ∈ 𝑈1 )
= 𝑎1 − 𝑏1 − 3𝜆1 + 𝛼 𝑎2 − 𝑏2 − 3𝜆2
= 0 + 𝛼 0 = 0.
Problem 6: Prove that 𝑋 = 𝑎, 2𝑎 − 3𝑏, 5𝑏, 𝑎 + 2𝑏, 𝑎 : 𝑎, 𝑏 ∈ 𝑅
is a vector subspace of ℝ5 .
Solution:
(b) 2𝑢 + 3𝑣 − 𝑤
(c) −3𝑢 + 4𝑣 − 2𝑤
2. Prove that the set 𝐶 𝑛 with the operations of addition and
scalar multiplication defined as follow is a vector space
𝑢1 , … , 𝑢𝑛 + 𝑣1 , … , 𝑣𝑛 = 𝑢1 + 𝑣1 , … , 𝑢𝑛 + 𝑣𝑛
𝑐 𝑢1 , … , 𝑢𝑛 = 𝑐𝑢1 , … , 𝑐𝑢𝑛
(a) 𝑎, 𝑏, 0
(b) 𝑎, 2𝑎, 𝑏
(c) 𝑎, 𝑎 + 𝑏, 3𝑎
(a) 𝑎 + 𝑏 + 𝑐 = 0
(b) 𝑎𝑏 = 0
(c) 𝑎𝑏 = 𝑎𝑐.
(a) (𝑎, 𝑎 + 1, 𝑏)
(b) 𝑎, 𝑏, 𝑎 + 𝑏 − 4 .
10. Let 𝑈 be the set of all vectors of the form 𝑎, 𝑏, 𝑐 and 𝑉
be the set of all vectors of the form 𝑎, 𝑎 + 𝑏, 𝑐 . Show that 𝑈
and 𝑉 are the same set. Is this set a subspace of ℝ3 ?
Answers
1. (a) 13, −1 (b) 17, −4 (c) 19, −20
2. (a) 7 − 𝑖, 4 + 7𝑖 (b) 4 − 7𝑖, 17 + 6𝑖
3. 𝑊 is not a subspace.
6. Not a subspace
𝐶1 + 2𝐶3 = −1
2𝐶1 + 𝐶2 + 3𝐶3 = 1
3𝐶1 + 4 𝐶2 + 6𝐶3 = 5
It can be shown that this system of equations has the
unique solution.
𝐶1 = 1, 𝐶2 = 2, 𝐶3 = −1.
2𝐶1 + 𝐶2 + 3𝐶3 = 5
3𝐶1 + 4 𝐶2 + 2𝐶3 = 5
This system of equations has many solutions,
𝐶1 = −2𝑟 + 3, 𝐶2 = 𝑟 − 1, 𝐶3 = 𝑟
𝐶1 − 𝐶2 + 𝐶3 = 3
2𝐶1 − 𝐶2 + 4𝐶3 = −4
3𝐶1 − 2 𝐶2 + 5𝐶3 = 6
This system has no solution. Thus (3, −4, −6) is not a linear
combination of the vectors
Thus, 𝐶1 + 𝐶3 = 𝑥
2𝐶1 + 𝐶2 + 𝐶3 = 𝑦
−𝐶2 + 2𝐶3 = 𝑧
𝐶1 = 3𝑥 − 𝑦 − 𝑧,
𝐶2 = −4𝑥 + 2𝑦 + 𝑧,
𝐶3 = −2𝑥 + 𝑦 + 𝑧.
𝑣 = 𝑎 𝑣1 + 𝑏 𝑣2
we can write
𝑎 𝑏
𝑣= 𝑘1 𝑣1 + 𝑘2 𝑣2
𝑘1 𝑘2
= (𝑎 − 3𝑏, 2𝑎 + 𝑏, 2𝑏)
5𝑝 + 𝑞 = 2𝑎 + 𝑏
2𝑝 − 2𝑞 = 2𝑏.
𝑎+𝑏 𝑎−2𝑏
This system of eqs has unique solution 𝑝 = ,𝑞 = .
3 3
(a)𝑓 𝑥 = 3𝑥 2 + 2𝑥 + 9; 𝑔 𝑥 = 𝑥 2 + 1, 𝑥 = 𝑥 + 3
(b) 𝑓 𝑥 = 𝑥 2 + 4𝑥 + 5; 𝑔 𝑥 = 𝑥 2 + 𝑥 − 1, 𝑥 = 𝑥 2 + 2𝑥 + 1
Answers
2. (a) −3,3,7 = 2 1, −1,2 − 2,1,0 + 3 −1,2,1
(b) 0,10,8 = 2 − 𝑐 −1,2,3 + 2 − 2𝑐 1,3,1 + 𝑐 1,8,5 ,
whether c is any real number
Assume that 𝑐1 ≠ 0.
The above identity can be rewritten as
−𝑐2 −𝑐𝑚
𝑣1 = 𝑣2 + ⋯ + 𝑣
𝑐1 𝑐𝑚 𝑚
𝑣1 = 𝑣2 𝑑2 , … , 𝑑𝑚 𝑣𝑚 .
𝑐1 0 + 𝑐2 𝑣2 + ⋯ + 𝑐𝑛 𝑣𝑛 = 0.
𝐶1 − 2𝐶2 + 8𝐶3 = 0
2𝐶1 + 𝐶2 + 6𝐶3 = 0
3𝐶1 + 𝐶2 + 10𝐶3 = 0
3 𝐶1 + 3 𝐶2 + 𝐶3 = 0
−2 𝐶1 − 𝐶2 = 0
2 𝐶1 + 4 𝐶2 + 5 𝐶3 = 0
⇒ 𝑎 + 𝑏 𝑣1 + 𝑎 − 𝑏 𝑣2 = 0
(a) {𝑣1 , 𝑣1 + 𝑣2 , 𝑣3 }
and since
𝛼1 , … , 𝛼𝑘 , 𝛽1 , … , 𝛽𝑚
is also an independent set, each 𝑥𝑖 = 0 and each 𝑦𝑗 = 0 .
Thus 𝛼1 , … , 𝛼𝑘 , 𝛽1 , … , 𝛽𝑚 , 𝛾1 , … , 𝛾𝑛
is a basis for 𝑊1 + 𝑊2 . Finally
dim 𝑊1 + dim 𝑊2 = 𝑘 + 𝑚 + 𝑘 + 𝑛
=𝑘+ 𝑚+𝑘+𝑛
= dim 𝑊1 ∩ 𝑊2 + dim 𝑊1 + 𝑊2 .
Problem 1: Show that the set 1,0, −1 , 1,1,1 , 1,2,4 is a
basis for ℝ3 .
𝑎2 + 2𝑎3 = 𝑥2
−𝑎1 + 𝑎2 + 4𝑎3 = 𝑥3
This system of equations has the solution
𝑎1 = 2𝑥1 − 3𝑥2 + 𝑥3
𝑎3 = 𝑥1 − 2𝑥2 + 𝑥3
Thus the set spans the space. We now show that the set is
linearly independent.
𝑏2 + 2𝑏3 = 0
−𝑏1 + 𝑏2 + 4𝑏3 = 0
This system has the unique solution 𝑏1 = 0, 𝑏2 = 0, and 𝑏3 =
0. Thus the set is linearly independent.
−𝑐1 + 4𝑐3 = 0
(a) The vectors (1, 2), (-1, 3), (5, 2) are linearly dependent in
ℝ 2.
(b) The vectors (1, 0, 0), (0, 2, 0), (1, 2, 0) span ℝ3.
(a) {(3, 1), (2, 1)} (b) {(1, −3), (−2, 6)}
𝛽 = 𝑦1 𝛼1 + ⋯ + 𝑦𝑛 𝛼𝑛
be in V and let c be any scalar. Now
𝑐𝛼 + 𝛽 = 𝑐𝑥1 + 𝑦1 𝛼1 + ⋯ + 𝑐𝑥𝑛 + 𝑦𝑛 𝛼𝑛
and so by definition
𝑇 𝑐𝛼 + 𝛽 = 𝑐𝑥1 + 𝑦1 𝛽1 + ⋯ + 𝑐𝑥𝑛 + 𝑦𝑛 𝛽𝑛
on the other hand,
𝑛 𝑛
𝑐 𝑇(𝛼) + 𝑇(𝛽) = 𝑐 𝑥𝑖 𝛽𝑖 + 𝑦𝑖 𝛽𝑖
𝑖=1 𝑖=1
𝑛
= (𝑐𝑥𝑖 + 𝑦𝑖 ) 𝛽𝑖
𝑖=1
𝑈(𝛼) = 𝑈 𝑥𝑖 𝛼𝑖
𝑖=1
𝑛
= 𝑥𝑖 (𝑈(𝛼𝑖 ))
𝑖=1
𝑛
= 𝑥𝑖 𝛽𝑖
𝑖=1
𝑇 0𝑈 = 𝑇 0 𝑢 = 0 𝑇 𝑢 = 0𝑣 = 0𝑉 .
DEFINITION: Let 𝑇: 𝑈 → 𝑉 be a linear transformation. The
set of vectors in 𝑈 that are mapped into the zero vector of 𝑉
is called the kernel of 𝑇. The kernel is denoted by 𝑘𝑒𝑟 𝑇 .
The set of vectors in 𝑉 that are images of vectors in 𝑈
is called the range of 𝑇. The range is denoted by 𝑟𝑎𝑛𝑔𝑒 𝑇 .
We illustrate these sets in the following figure.
U V
O
Kernel
𝑇→
All vectors in 𝑈 that are mapped into 0.
U V
range
𝑇→
All vectors in 𝑉 that are images of vectors in 𝑈.
(a) T(x, y) = y2
(b) T(x, y, z) = (x + y, z) of 𝑹3 → 𝑹2
(b) The kernel is the set {(r,-r,0)} and the range is R2. Dim
ker(T) =1, dim range(T)=2, and dim domain(T)= 3, so dim
ker(T)+ dim range(T) = dim domain(T).
(c) The kernel is the zero vector and the range is the set
{(3a,a-b,b)}. Dim ker(T) =0, dim range(T)= 2, so dim ker(T)+
dim range(T) = dim domain(T).
a1
The column vector 𝑢𝑩 =
is called the coordinate
an
vector of u relative to this basis. The scalars 𝑎1 , . . . . , 𝑎𝑛
are called the coordinates of 𝒖 relative to this basis.
Note: We will use a column vector from coordinate vectors
rather than row vectors. The theory develops most
smoothly with this convention.
Example: Find the coordinate vector of 𝒖 = (4,5) relative to
the following bases 𝑩 and 𝑩′ of 𝑅 2 :
(a) The standard basis, 𝐵 = 1,0 , 0,1 and
(b) 𝐵′ = { 2,1 , −1,1 }.
Solution:
(a) By observation, we see that
(4, 5) = 4(1, 0) + 5(0, 1)
4
Thus 𝒖𝐵 = . The given representation of u is, in
5
fact, relative to the standard basis.
(b) Let us now find the coordinate vector of u relative
to 𝐵′, a basis that is not the standard basis. Let
(4, 5) = 𝑎1 (2, 1) + 𝑎2 (-1, 1)
Thus
2a1 a2 4
a1 a2 5
This system has the unique solution
𝑎1 = 3 , 𝑎2 = 2
3
Thus 𝑢𝐵 ′ =
2
Definition: Let 𝑩 = 𝑢1 , . . . , 𝑢𝑛 and 𝑩′ = 𝑢′1 , . . . , 𝑢′ 𝑛 be
bases for a vector space 𝑈. Let the coordinate vectors of
𝑢1 , . . . , 𝑢𝑛 relative to the basis 𝑩′ = 𝑢′1 , . . . , 𝑢′ 𝑛 be
(𝑢1 )𝐵 ′ , . . . . , (𝑢𝑛 )𝐵 ′ . The matrix 𝑃, having these vectors as
columns , plays a central role in our discussion. It is called
the transition matrix from the basis 𝑩 to the basis 𝑩′.
Transition matrix 𝑷 = [(𝑢1 )𝐵 ′ , . . . . , (𝑢𝑛 )𝐵 ′ ].
𝒖𝐵 ′ = 𝑃𝒖𝐵
= [ 𝑢1 𝐵′ , . . . . , 𝑢𝑛 𝐵 ′ ]𝒖𝐵
1 3
The coordinate vectors of (1,2) and (3, −1) are and
2 1
.
The transition matrix 𝑃 is thus
1 3
P
2 1
(Observer that the columns of 𝑃 are the vectors of the
basis.) We get
1 3 3 15
𝒖𝐵 ′
2 1 4 2
f x x 3
𝑢 . . . . . . . . . . 𝑇 . . . . . . . . . . .𝑇(𝑢)
Coordinate
Mapping
......... .........
a A b
Figure
𝑇 𝑢1 = 𝑐11 𝑣1 + . . . + 𝑐1𝑚 𝑣𝑚
𝑇 𝑢2 = 𝑐21 𝑣1 + . . . + 𝑐2𝑚 𝑣𝑚
.
.
.
𝑇 𝑢𝑛 = 𝑐𝑛1 𝑣1 + . . . + 𝑐𝑛𝑚 𝑣𝑚
Thus
𝐵 = 𝑃−1 𝐴𝑃
𝐴′ = 𝑃−1 𝐴𝑃
Proof: Consider a vector 𝑢 in 𝑈. Let its coordinate vector
relative to B and 𝐵′ be 𝑎 and 𝑎′. The coordinate vectors of
𝑇(𝑢) are 𝐴𝑎 and 𝐴′𝑎′. Since P is the transition matrix from 𝐵′
to 𝐵, we know that
𝑃−1 𝐴𝑎 = 𝐴′𝑎′
1 1 3 3 2 2
L , L , L .
4 4 1 1 6 6
2, 6
1, 4
3, 1
3, 1
1, 4
2, 6
1 4 3 1 2 6
L , L , L .
4 1 1 3 6 2
2, 6
1, 4
3, 1
4, 1
1, 3
6, 2
3. Rotation:
u1 u1 cos sin u1
2
2
L : R R , L A u
u 2 u 2 sin cos 2
For example, as 2,
cos
A
2 0
sin 1
cos 1
2
0 .
sin 2 2
Thus, the rotation for the triangle with vertices 0, 0, 1, 0, 1, 1
is
and
1 0 1 1 1
L 1 1 .
1 1 0
0, 1
1, 1
1, 1
0, 0 1, 0
Problem 1: Consider the linear transformation T : R3 R2
defined as follows on basis vectors of R 3 . Find T 1, 2,3.
8, 3
Problem 2: Let 𝑇: 𝑈 → 𝑉 be a linear transformation. T is
defined relative to bases 𝐵 = {𝑢1 , 𝑢2 , 𝑢3 } and 𝐵′ = 𝑣1 , 𝑣2 of 𝑈
and 𝑉 as follows
𝑇 𝑢1 = 2𝑣1 − 𝑣2
𝑇 𝑢2 = 3𝑣1 + 2𝑣2
𝑇 𝑢3 = 𝑣1 − 4𝑣2
Find the matrix representation of 𝑇 with respect to these
bases and use this matrix to determine the image of the
vector 𝑢 = 3𝑢1 + 2𝑢2 − 𝑢3 .
2 3 1
A
1 2 4
Let us now find the image of the vector 𝑢 = 3𝑢1 + 2𝑢2 − 𝑢3
using this matrix.
3
The coordinate vector of 𝑢 is a 2 .
1
We get
3
2 3 1 11
Aa 2 5
1 2 4 1
11
𝑇(𝑢) has coordinate vector . Thus 𝑇 𝑢 = 11𝑣1 + 5𝑣2 .
5
Problem 3: Consider the linear transformation 𝑇: 𝑅 3 → 𝑅 2 ,
defined by 𝑇 𝑥, 𝑦, 𝑧 = (𝑥 + 𝑦, 2𝑧). Find the matrix of 𝑇 with
respect to the bases {𝑢1 , 𝑢2 , 𝑢3 } and {𝑢′1 , 𝑢′2 } of 𝑅 3 and 𝑅 2 ,
where
𝑢1 = 1,1,0 , 𝑢2 = 0,1,4 , 𝑢3 = 1,2,3 𝑎𝑛𝑑 𝑢′1 = 1,0 , 𝑢′2 = (0,2).
2 0
A
1 1
Therefore
1
2 1 2 0 2 1
A ' P 1 AP 1 1 3 1
3 1
1 1 2 0 3 2
3 2 1 1 10 6
Exercise
1) Let 𝑇: 𝑈 → 𝑉 be a linear transformation. Let 𝑇 be
defined relative to bases {𝑢1 , 𝑢2 , 𝑢3 } and {𝑣1 , 𝑣2 , 𝑣3 } of 𝑈
and 𝑉 as follows:
𝑇 𝑢1 = 𝑣1 + 𝑣2 + 𝑣3
𝑇 𝑢2 = 3𝑣1 −2𝑣2
𝑇 𝑢3 = 𝑣1 + 2𝑣2 − 𝑣3 .
Find the matrix of 𝑇 with respect to these bases. Use
this matrix to find the image of the vector
𝑢 = 3𝑢1 + 2𝑢2 − 5𝑢3 .
1 3 1
1) 1 2 2 , 4v1 11v2 8v3
1 0 1
1 0 0
2) (a) 0 2 0 ,( 1,10,6)
0 0 3
1 0 0
(b)
0 0 0 ,( 1,0,0)
0 0 0
2 2 1
3) ,(7,3)
1 3 2
0 0 0
4) 4 0 0 ,6 x 2
0 1 0
0 1 1
5) (a) 0 1 1
0 0 0
0 1
(b) 1 0
0 1
2 0 2 2
6) , 0 1
1 1
8) No
1.8
The Inverse of a Matrix
In this module we introduce the concept of the matrix
inverse. We will see how an inverse can be used to solve
certain systems of linear equations, and we will see an
application of matrix inverse in cryptography, the study of
codes.
We motivate the idea of the inverse of a matrix by looking
at the multiplicative inverse of a real number. If number 𝑏
is the inverse of 𝑎, then
𝑎𝑏 = 1 and 𝑏𝑎 = 1
1
For example, is the inverse of 4 and we have
4
1 1
4 = 4=1
4 4
and
−2 1 1 2 1 0
𝐵𝐴 = 3
−
1 = = 𝐼2
2 2
3 4 0 1
𝐶𝐴 𝐵 = 𝐶
𝐼𝑛 𝐵 = 𝐶
𝐵=𝐶
𝐴−1 = 𝑋1 𝑋2 … 𝑋𝑛 and 𝐼𝑛 = 𝐶1 𝐶2 … 𝐶𝑛
Thus
𝐴: 𝐼𝑛 ≈ ⋯ ≈ 𝐼𝑛 : 𝐴−1
Let 𝐴 be an 𝑛 × 𝑛 matrix.
1. Adjoin the identity 𝑛 × 𝑛 matrix 𝐼𝑛 to 𝐴 to form the matrix
𝐴: 𝐼𝑛 .
2. Compute the reduced echelon form of 𝐴: 𝐼𝑛 .
If the reduced echelon form is of the type 𝐼𝑛 : 𝐵 , then 𝐵 is
the inverse of 𝐴.
If the reduced echelon form is not of the type 𝐼𝑛 : 𝐵 , in
that the first 𝑛 × 𝑛 submatrix is not 𝐼𝑛 , then 𝐴 has no
inverse.
𝐴𝑋 = 𝐴 𝐴−1 𝐵 = 𝐴𝐴−1 𝐵 = 𝐼𝑛 𝐵 = 𝐵
𝐼𝑛 𝑋1 = 𝐴−1 𝐵
𝑋1 = 𝐴−1 𝐵
1. 𝐴−1 −1 =𝐴
1
2. 𝑐𝐴 −1 = 𝐴−1
𝑐
3. 𝐴𝐵 −1 = 𝐵 −1 𝐴−1
4. 𝐴𝑛 −1
= 𝐴−1 𝑛
5. 𝐴𝑡 −1 = 𝐴−1 𝑡
𝐴𝐴−1 = 𝐴−1 𝐴 = 𝐼𝑛
= 𝐴𝐼𝑛 𝐴−1
= 𝐴𝐴−1
= 𝐼𝑛
PREPARE TO ATTACK
and the encoding matrix be
−3 −3 −4
0 1 1
4 3 4
We assign a number to each letter of the alphabet. For
convenience, let us associate each letter with its position in
the alphabet: 𝐴 is 1, 𝐵 is 2, and so on. Let a space between
words be denoted by the number 27. Thus the message
becomes
P R E P A R E * T O * A T T A C K
16 18 5 16 1 18 5 27 20 15 27 1 20 20 1 3 11
P R E P A R E * T O * A T T A C K
Problem 1: Determine the inverse of the matrix
1 −1 −2
𝐴= 2 −3 −5
−1 3 5
Solution: Applying the method of Gauss-Jordon
elimination, we get
1 −1 −2 1 0 0
𝐴: 𝐼3 = 2 −3 −5 0 1 0
−1 3 5 0 0 1
≈ 1 −1 −2 1 0 0
𝑅2 + −2 𝑅1 0 −1 −1 −2 1 0
𝑅3 + 𝑅1 0 2 3 1 0 1
≈ 1 −1 −2 1 0 0
−1 𝑅2 0 1 1 2 −1 0
0 2 3 1 0 1
≈ 1 0 −1 3 −1 0
𝑅1 + 𝑅2 0 1 1 2 −1 0
𝑅3 + −2 𝑅2 0 0 1 −3 2 1
≈ 1 0 0 0 1 1
𝑅1 + 𝑅3 0 1 0 5 −3 −1
𝑅2 + −1 𝑅3 0 0 1 −3 2 1
Thus
0 1 1
𝐴−1 = 5 −3 −1
−3 2 1
Problem 2: Determine the inverse of the following matrix, if
it exists.
1 1 5
𝐴= 1 2 7
2 −1 4
Solution: Applying the method of Gauss-Jordon
elimination, we get
1 1 5 1 0 0
𝐴: 𝐼3 = 1 2 7 0 1 0
2 −1 4 0 0 1
≈ 1 1 5 1 0 0
𝑅2 + −1 𝑅1 0 1 2 −1 1 0
𝑅3 + −2 𝑅1 0 −3 −6 −2 0 1
≈ 1 0 3 2 −1 0
𝑅1 + −1 𝑅2 0 1 2 −1 1 0
𝑅3 + 3𝑅2 0 0 0 −5 3 1
There is no need to proceed further. The reduced echelon
form cannot have a one in the 3,3 location. The reduced
echelon form cannot be of the form 𝐼𝑛 : 𝐵 . Thus 𝐴−1 does
not exist.
Problem 3: Solve the system of equations
𝑥1 − 𝑥2 − 2𝑥3 = 1
2𝑥1 − 3𝑥2 − 5𝑥3 = 3
x1 2 x2 x3 2
x1 x2 2 x3 0
a.
x1 x2 x3 1
x1 2 x2 3 x3 1
2 x1 5 x2 3 x3 3
b.
x1 8 x3 15
x1 x2 5
x1 x3 2
c.
6 x1 2 x2 3 x3 1
2.
1 1 1
5 5 5 0
1 1 0
1
5 5 5
a. 1
1 1
0
5 5 5
1 1 1
0
5 5 5
1 0 1 1
0 1 3 4
b.
1 0 1 2
3 0 0 1
3.
1 3 5 7
𝑥1 9 9 9 2 9
3 3 3
a. 𝑥2 = 0 − 0 =
9 9 9
𝑥3 2 3 1 1 5
− − −
9 9 9 9
𝑥1 −40 16 9 1 143
b. 𝑥2 = 13 −5 −3 3 = −47
𝑥3 5 −2 −1 15 −16
𝑥1 2 3 1 5 5
c. 2 = 3 3 1 −2 = 10
𝑥
𝑥3 2 4 1 1 3
𝒗𝑗 = 𝑏𝑗 1 , 𝑏𝑗 2 , … , 𝑏𝑗𝑛
Answers
1.
a. 2
b. 1
c. 2
2.
1 0 0
a. 0 1 0 . Basis 1,0,0 , 0,1,0 and 0,0,1 . Rank 3.
0 0 1
1 −3 2
b. 0 0 0 . Basis 1, −3,2 . Rank 1.
0 0 0
3.
a. The given vectors are linearly independent so they
are a basis for the space they span, which is 𝑹3 .
1 −1 3 1 0 1
b. 1 0 1 ≈ ⋯ ≈ 0 1 −2 . Vectors 1,0,1 and
−2 1 −4 0 0 0
0,1, −2 are a basis for the space spanned by the
given vectors.
1 2 −1 1 0 0
4. 𝐴 = 0 1 3 ≈ ⋯ ≈ 0 1 0 . Vectors 1,0,0 , 0,1,0
1 4 6 0 0 1
and 0,0,1 are a basis for the row space of 𝐴.
1.10
Eigenvalues and Eigenvectors
𝐴 𝑐𝑋1 = 𝜆 𝑐𝑋1 .
Therefore 𝑐𝑋1 is a vector in the eigenspace of 𝜆 . The
eigenspace is closed under scalar multiplication.
n1
a an2 ann
= 1 2 n
𝑎1
=− = 𝑎11 + 𝑎22 + ⋯ + 𝑎𝑛𝑛 (Using 4)
𝑎0
= trace of 𝐴.
Product of the eigenvalues = product of the roots
= 1 , 2 ,n
𝑎 (−1)𝑛 𝑎 𝑛
= (−1)𝑛 𝑛 =
𝑎0 (−1)𝑛
= 𝑎𝑛
= 𝐴 (from 2 )
Property 3: The eigen values of 𝐴 and its transpose 𝐴𝑇 are
the same.
𝐴 − 𝜆𝐼 = (𝐴 − 𝜆𝐼)𝑇 = 𝐴𝑇 − (𝜆𝐼)𝑇 = 𝐴𝑇 − 𝜆𝐼
Hence the result.
𝐼𝑋 = 𝜆𝐴−1 𝑋
1
∴ 𝐴−1 𝑋 = 𝑋.
𝜆
1
∴ is an eigen value of 𝐴−1 .
𝜆
Then 𝐴𝑋 = 𝜆𝑋.
Now, 𝐴2 𝑋 = 𝐴𝐴 𝑋 = 𝐴(𝐴𝑋)
= 𝜆2 𝑋.
∴ 𝜆2 is an eigen value of 𝐴2 .
𝐴 − 𝜆𝐼 adj 𝐴 − 𝜆𝐼 = 𝐴 − 𝜆𝐼 𝐼.
i.e.,
𝐴 − 𝜆𝐼 𝐵𝑛 −1 𝜆𝑛−1 + ⋯ + 𝐵0 = −1 𝑛
𝜆𝑛 + 𝑎𝑛−1 𝜆𝑛−1 + ⋯ + 𝑎0 𝐼.
…...……………………..
𝐴𝐵1 − 𝐵0 = −1 𝑛 𝑎1 𝐼.
𝐴𝐵0 = −1 𝑛 𝑎0 𝐼.
0 = −1 𝑛 𝐴𝑛 + 𝑎𝑛−1 𝐴𝑛−1 + ⋯ + 𝑎1 𝐴 + 𝑎0 𝐼
⇒ −1 𝑛 𝐴𝑛 + 𝑎𝑛 −1 𝐴𝑛−1 + ⋯ + 𝑎1 𝐴 + 𝑎0 𝐼 = 0
i.e −1 𝑛 𝐴𝑛 + 𝑎𝑛 −1 𝐴𝑛−1 + ⋯ + 𝑎1 𝐴 + 𝑎0 𝐼 = 0
⇒ 𝐴𝑛 + 𝑎𝑛−1 𝐴𝑛 −1 + ⋯ + 𝑎1 𝐴 + 𝑎0 𝐼 = 0
Applications
Eigenvalues and eigenvectors are extremely important tools
in applying mathematics. They are used in many branches
of engineering and the natural and social sciences. Now we
see use of eigenvectors in long term behavior of the
population distribution.
(from) (to)
City Suburb
= 57.1 million
Suburban population in 2001 = People who moved in from
the city +People who stayed
from 2000
= 142.9 million.
𝑋2 = 𝑃𝑋1
𝜆2 + 𝜆 − 2 = 0
𝜆−2 𝜆+1 =0
𝜆 = 2 or −1.
3𝑥1 + 3𝑥2 = 0
= 1 − 𝜆 𝜆 − 10 𝜆 − 1 = − 𝜆 − 10 𝜆 − 1 2
𝑐𝐴𝑋𝑖 = 𝑐𝜆𝑖 𝑋𝑖 .
ANSWERS
1)
a. Characteristic polynomial −𝜆3 + 2𝜆2 + 𝜆 − 2 ;
eigenvalues 1,-1,2; corresponding eigenspaces
1 1 0
r 0 , s 1 , t 1
1 1 1
b. Characteristic polynomial 1 − 𝜆 2 (3 − 𝜆) ; eigen
1
values 1,3; corresponding eigenspaces r 1 ,
2
0
s 1
1
c. Characteristic polynomial 1 − 𝜆 2 − 𝜆 8 − 𝜆 ;
eigenvalues 1,2,8; corresponding eigenspaces
1 0 1
r 1 , s 1 , t 0
1 1 1
2) Characteristic polynomial 2 − 𝜆 2 − 𝜆 4 − 𝜆 (6 − 𝜆) ;
eigenvalues 2,4,6; corresponding eigenspaces
1 0 0 1
1 0 1 0
r s , t , p
0 1 0 0
0 1 1 1
3 1 −1
1
8) 1 3 1 .
4
−1 1 3
9)𝜆3 − 5𝜆2 + 7𝜆 − 3 = 0, 𝐼
10)625𝐼