Lecture 2: Linear Algebra: Modern Control Systems

You might also like

You are on page 1of 31

Lecture 2 : Linear Algebra

Modern Control Systems

It Chivorn

Institute of Technology of Cambodia

I4GEE-EA, 2023

IT Chivorn (ITC - GEE Department) Modern Control systems Oct, 10, 2023 1 / 31
Table of Contents

1. Representation of Matrix

2. Determinant

3. Inverse Matrix

4. Matrix Transpose

5. Linear Independence and Matrix Rank

6. Eigenvalues and Eigenvectors

7. Cayley-Hamilton Theorem

IT Chivorn (ITC - GEE Department) Modern Control systems Oct, 10, 2023 2 / 31
Representation of Matrix

Table of Contents

1. Representation of Matrix

2. Determinant

3. Inverse Matrix

4. Matrix Transpose

5. Linear Independence and Matrix Rank

6. Eigenvalues and Eigenvectors

7. Cayley-Hamilton Theorem

IT Chivorn (ITC - GEE Department) Modern Control systems Oct, 10, 2023 3 / 31
Representation of Matrix

Representation of Matrix
···
 
a11 a1m
 .. .. ..  n×m
A= . . .  , or An×m , whereA ∈ R (A ∈ Cn×m )
an1 ··· anm
R stands for a Ring as real number, complex number or rational function etc,
n by m block

A11 · · · A1m
 
 . .. ..  ri ×cj
A =  .. . .  , where Aij ∈ R (A ∈ Cri ×cj )
An1 · · · Anm

Especially, Rn×1 or R1×n matrix is called low or column vector, respectively.


Example:
▶ Many numerical data, e.g., time series data, image data

M = [y(T )| · · · |y(nT )] , y(t) ∈ Rn


▶ Linear equation
Ax = b, A ∈ Rn×m , x ∈ Rm , b ∈ Rn

IT Chivorn (ITC - GEE Department) Modern Control systems Oct, 10, 2023 4 / 31
Representation of Matrix

Representation of Matrix

▶ Inner product (linear functional), exterior product (3 dim.)

< x, y >=: xT y , x, y ∈ Rn
▶ Group operations
▶ Linear Operation
 
m1 0 0
▶ Diagonal matrix D =  0 m2 0  M = diag{mi }, i = 1, ..., p
0 0 m3
 
m11 m12 m13
▶ Upper triangular matrix U =  0 m22 m23 
0 0 m33
 
m11 0 0
▶ Lower triangular matrix L = m21 m22 0 
m31 m32 m33

IT Chivorn (ITC - GEE Department) Modern Control systems Oct, 10, 2023 5 / 31
Determinant

Table of Contents

1. Representation of Matrix

2. Determinant

3. Inverse Matrix

4. Matrix Transpose

5. Linear Independence and Matrix Rank

6. Eigenvalues and Eigenvectors

7. Cayley-Hamilton Theorem

IT Chivorn (ITC - GEE Department) Modern Control systems Oct, 10, 2023 6 / 31
Determinant

Determinant : Definition

Definition
For any i-th low X
|A| := (−1)i+j aij Mij , |A| = det(A)
j

where Mij is a minor of a smaller matrix which is a matrix which is determined by


deleting i-th low and j-th column of matrix A

Application
▶ Calculation of inverse matrix
▶ Change of measure for multivariable integration,i.e., Jacobian.

IT Chivorn (ITC - GEE Department) Modern Control systems Oct, 10, 2023 7 / 31
Determinant

Determinant : Properties

a11 · · · a1n
 
 . .. ..  n×n
Given a square matrix A =  .. . . , A ∈ R
an1 · · · ann
▶ Scalar multiplication of column or row vector, scalar multiplication of matrix

|αA| = αn |A|

ta11 ta12 ta11 a12 a11 a12


= =t
a21 a22 ta21 a22 a21 a22
▶ Determinant of matrix multiplication

|AB| = |A||B|
▶ Determinant of Transpose matrix

|AT | = |A|

IT Chivorn (ITC - GEE Department) Modern Control systems Oct, 10, 2023 8 / 31
Determinant

Determinant : Example 1

▶ Block triangular matrix


A11 ∗
= |A11 ||A22 |
0 A22
▶ Relationship between determinant and eigenvalues
n
Y
|A| = λi
i=1

Where λi (i = 1, 2, ..., n)
▶ If A−1 exists then
1
|A−1 | =
|A|

Example
Compute the determinant of the given matrix
 
  −1 5 2
1 2
A= and B =  1 0 −4
−1 3
4 1 −1

IT Chivorn (ITC - GEE Department) Modern Control systems Oct, 10, 2023 9 / 31
Inverse Matrix

Table of Contents

1. Representation of Matrix

2. Determinant

3. Inverse Matrix

4. Matrix Transpose

5. Linear Independence and Matrix Rank

6. Eigenvalues and Eigenvectors

7. Cayley-Hamilton Theorem

IT Chivorn (ITC - GEE Department) Modern Control systems Oct, 10, 2023 10 / 31
Inverse Matrix

Inverse Matrix : Definition

Definition
For a matrix A ∈ Rn×n , a matrix X which satisfies: AX = XA = In Us called an
inverse matrix A and it is denoted as A−1 .
Inverse formula
···
 
a11 a1n
adj(A)  .. .. .. 
A−1 = , where adj(A) =  . . . 
|A|
an1 ··· ann

aij = (−1)i+j Mji . If |A| ̸= 0, A is called non-singular, and if|A| = 0, is called


singular.
Application
▶ To solve matrix linear equation
▶ Coordinate transformation in orthogonal coordinate space

IT Chivorn (ITC - GEE Department) Modern Control systems Oct, 10, 2023 11 / 31
Inverse Matrix

Inverse Matrix : Properties

▶ If A is nonsingular, then so is A−1 and

(A−1 )−1 = A
▶ If A and B are nonsingular matrices, then AB is nonsignular and

(AB)−1 = B −1 A−1
▶ If A and B are matrices with AB = In then A and B are inverses of each other.

Theorem of Nonsingular Equivalences (TFAE)

1. A is nonsigular
2. Ax = 0 has only the trivial solution
3. A is row equivalent to I
4. The linear system Ax = b has a unique solution for every n × 1 matrix b

IT Chivorn (ITC - GEE Department) Modern Control systems Oct, 10, 2023 12 / 31
Matrix Transpose

Table of Contents

1. Representation of Matrix

2. Determinant

3. Inverse Matrix

4. Matrix Transpose

5. Linear Independence and Matrix Rank

6. Eigenvalues and Eigenvectors

7. Cayley-Hamilton Theorem

IT Chivorn (ITC - GEE Department) Modern Control systems Oct, 10, 2023 13 / 31
Matrix Transpose

Matrix Transpose

··· ···
   
a11 a1m a11 an1
 . .. ..  T  .. .. .. 
A =  .. . .  => A :=  . . . 
an1 ··· anm a1m ··· anm
n×m n×m T m×n m×n
A∈R (A ∈ C ), A ∈ R (A ∈ C )
▶ (AB)T = B T AT
▶ |AT | = |A|
▶ (AT )−1 = (A−1 )T
▶ (cA)T = cAT
Application
▶ Vector representation of inner product
▶ Projection operation total derivative, Definition of dual operator

IT Chivorn (ITC - GEE Department) Modern Control systems Oct, 10, 2023 14 / 31
Linear Independence and Matrix Rank

Table of Contents

1. Representation of Matrix

2. Determinant

3. Inverse Matrix

4. Matrix Transpose

5. Linear Independence and Matrix Rank

6. Eigenvalues and Eigenvectors

7. Cayley-Hamilton Theorem

IT Chivorn (ITC - GEE Department) Modern Control systems Oct, 10, 2023 15 / 31
Linear Independence and Matrix Rank

Linear dependence

Definition
A sequence of vector (v1 , v2 , ..., vk ) from a vector space V is said to be linearly
dependent, if there exist scalars a1 , a2 , ..., ak , not all zero, such that

a1 v1 + a2 v2 + · · · + ak vk = 0

Notice that if not all of scalars are zero, then at one is non-zero, say a1 , in which case
this equation can be written in the form
−a2 −a3 −ak
v1 = v2 + v3 + · · · + vk
a1 a1 a1
Thus,v1 is shown to be a linear combination of the remaining vectors.

Definition
A sequence of vector (v1 , v2 , ..., vk ) from a vector space V , if there are only
ai = 0 (i = 1, 2, ..., k) (The trivial solution), satisfy the equation:

a1 v1 + a2 v2 + · · · + ak vk = 0

Therefor vi are called linearly independent each other.


IT Chivorn (ITC - GEE Department) Modern Control systems Oct, 10, 2023 16 / 31
Linear Independence and Matrix Rank

Criterion for linear independence

Theorem
Let v1 , v2 , ..., vk be n-vector, and let A be the n × n with these vectors as columns.
Then {v1 , v2 , ..., vk } are linearly independent if and only if

v11 v12 ··· v1n


v21 v22 ··· v2n
det(A) = .
.. .. .. .. ̸= 0
. . .
vn1 vn2 ··· vnn

Idea for proof: The linear system Ax = 0 has a unique solution (That is, only the
trivial solution) if and only if det(A) ̸= 0
Example
Show that a1 , a2 , a3 are linearly independent when
     
2 2 −1
a1 = 4 , a2 = 0 , a3 =  2 
0 6 −3

IT Chivorn (ITC - GEE Department) Modern Control systems Oct, 10, 2023 17 / 31
Linear Independence and Matrix Rank

Linear independent : Example 2

Solution
Since we have
2 2 −1
4 0 2 = (−4)0 + (−2)12 = −24 ̸= 0
0 6 −3
it follows that the vector are linearly independent.

Let A be any m × n matrix. Then A consists of n column vectors v1 , v2 , ..., vk which


are m-vectors.
Definition
The rank of A is the maximal number of linearly independent column vectors in A, i.e.
the maximal number of linearly independent vectors among {v1 , v2 , ..., vk }. If A = 0,
then the rank of A is 0.
We write rk(A) for the rank of A. Note that we may compute the rank of any matrix
even square or not

IT Chivorn (ITC - GEE Department) Modern Control systems Oct, 10, 2023 18 / 31
Linear Independence and Matrix Rank

Rank of a matrix : Example 3

Example
Find the rank of the matrix  
1 0 2 1
A = 0 2 4 2
0 2 2 1

Solution
Find the rank of the matrix
   
1 0 2 1 1 0 2 1
A = 0 2 4 2 → 0 2 4 2
0 2 2 1 0 0 −2 1

Since the echelon form has pivots in the first three columns, A has rank rk(A) = 3.
The first three columns of A are linearly independent.

IT Chivorn (ITC - GEE Department) Modern Control systems Oct, 10, 2023 19 / 31
Eigenvalues and Eigenvectors

Table of Contents

1. Representation of Matrix

2. Determinant

3. Inverse Matrix

4. Matrix Transpose

5. Linear Independence and Matrix Rank

6. Eigenvalues and Eigenvectors

7. Cayley-Hamilton Theorem

IT Chivorn (ITC - GEE Department) Modern Control systems Oct, 10, 2023 20 / 31
Eigenvalues and Eigenvectors

Eigenvalues and Eigenvectors

A simple example is that an eigenvector does not change direction in a transformation:

Definition
Let A be a square matrix (or linear transformation). A number is called an eigenvalue
of A if there exists a non-zero vector v such that

Av = λv

where v is an eigenvector and λ is an eigenvalue

IT Chivorn (ITC - GEE Department) Modern Control systems Oct, 10, 2023 21 / 31
Eigenvalues and Eigenvectors

Eigenvalues and Eigenvectors : Example 4

Example
Show the eigenvector of the matrix
   
−6 3 1
A= is v =
4 5 4

Solution
      
−6 3 1 6 1
Av = = =6 = λv
4 5 4 24 4
So, λ = 6 is an eigenvalue which satisfy equation.

IT Chivorn (ITC - GEE Department) Modern Control systems Oct, 10, 2023 22 / 31
Eigenvalues and Eigenvectors

Eigenvalues : Eigenvalues

Consider the equation : Av = λv

⇒ Av = λIv

⇒ Av − λIv = 0
If v is non-zero then we can solve for λ using just the determinant:

⇒ |Av − λIv| = 0

We conclude that λ is an eigenvalue iff below holds. We call that equation


”Characteristic Equation” of A . The eigenspace, the sub space of all eigenvectors
associated with λ, is
Eigenspace=Null(AλI)

IT Chivorn (ITC - GEE Department) Modern Control systems Oct, 10, 2023 23 / 31
Eigenvalues and Eigenvectors

Eigenvalues and Eigenvectors : Example 5

Example
Find all eigenvalues and their eigenspace for
 
3 −2
A=
1 0

Solution

3−λ −2
|A − λI| = = λ2 − 3λ + 2 = (λ − 1)(λ − 2) = 0
1 −λ
Therefore, λ1 = 1, λ2 = 2

IT Chivorn (ITC - GEE Department) Modern Control systems Oct, 10, 2023 24 / 31
Eigenvalues and Eigenvectors

Eigenvalues and Eigenvectors : Example 5

Solution
Eigenvector : (A − λIv) = 0 , v = [v1 v2 ]T
For λ1 = 1     
−2 −2 v1 0
→ =
1 −1 v2 0
T
v= 1 1
For λ1 = 2     
1 −2v1 0
→ =
1 −2v2 0
T
v= 2 1

IT Chivorn (ITC - GEE Department) Modern Control systems Oct, 10, 2023 25 / 31
Cayley-Hamilton Theorem

Table of Contents

1. Representation of Matrix

2. Determinant

3. Inverse Matrix

4. Matrix Transpose

5. Linear Independence and Matrix Rank

6. Eigenvalues and Eigenvectors

7. Cayley-Hamilton Theorem

IT Chivorn (ITC - GEE Department) Modern Control systems Oct, 10, 2023 26 / 31
Cayley-Hamilton Theorem

Cayley-Hamilton Theorem

The solution of linear differential equation: ẋ = Ax ,is

x = eAt x(0)

Theorem
Every matrix A satisfies its own characteristic ( eigenvalue) equation

det(A − λI) = 0

λn + an−1 λn−1 + · · · + a1 λ + a0 = 0
Then,
An + an−1 An−1 + · · · + a1 A + a0 = 0
Where, the determinant P (λ) = det(A − λI) = 0 is called characteristic polynomial of
matrix A .

IT Chivorn (ITC - GEE Department) Modern Control systems Oct, 10, 2023 27 / 31
Cayley-Hamilton Theorem

Cayley-Hamilton Theorem : Calculate eAt

The Cayley-Hamilton theorem is very important in reducing the degree of polynomial


of square matrices. And if a given function is analytic on a region of complex plane
this is much helpful to evaluate the analytic function of square matrices. Here we are
much interested in finding the value of eAt . It is known from calculus that
1 1 1
eλt = 1 +
λt + λt2 + λt3 + · · ·
1! 2! 3!
where λ is any parameter and t is any scalar. if λ is eigenvalue of matrix A, then
1 1 1
eAt = 1 + At + At2 + At3 + · · ·
1! 2! 3!
Assume that a scalar function f (x) is analytic in a region of complex plan. Then in
that region f (x) may be expressed as a polynomial

X
f (x) = ao + a1 x1 + a2 x2 + a3 x3 + · · · = ak xk
k=0

Let A be a square matrix of dimension n, with characteristic polynomial P (λ) and


eigenvalue, then it may be written
f (x) = P (x)Q(x) + R(x)
IT Chivorn (ITC - GEE Department) Modern Control systems Oct, 10, 2023 28 / 31
Cayley-Hamilton Theorem

Cayley-Hamilton Theorem

Where the remainder R(x) is of degree (n − 1) or less.


In particular, for x = λ
n−1
X
f (λk ) = R(λk ) = αk λk
k=0

Example
Find eAt for  
0 1
A=
−1 0

Solution
The characteristic equation is λ2 + 1 = 0, and the eigenvalue are λ1 = +j, λ1 = −j
From equation above:

eλt = αo + α1 λ
(1)
eAt = αo I + α1 A

IT Chivorn (ITC - GEE Department) Modern Control systems Oct, 10, 2023 29 / 31
Cayley-Hamilton Theorem

Cayley-Hamilton Theorem : Example 6

Solution
For λ1 = +j, λ2 = −j

ejt = cos(t) + j sin(t) = α0 + α1 j


(2)
e−jt = cos(t) − j sin(t) = α0 − α1 j

or α0 = cos(t) and α1 = sin(t). Then

eAt = αo I + α1 A
(3)
 
cos(t) sin(t)
=
−sin(t) cos(t)

IT Chivorn (ITC - GEE Department) Modern Control systems Oct, 10, 2023 30 / 31
Cayley-Hamilton Theorem

THANK YOU ...

IT Chivorn (ITC - GEE Department) Modern Control systems Oct, 10, 2023 31 / 31

You might also like