LINEAR ALGEBRA & MATRICES

C.T. Abdallah, M. Ariola, F. Amato

January 21, 2006

Contents
1 Review of Matrix Algebra
1.1 Determinants, Minors, and Cofactors . . . . . . . . . . . . . . . .
1.2 Rank, Trace, and Inverse . . . . . . . . . . . . . . . . . . . . . .
1.3 Elementary Operations and Matrices . . . . . . . . . . . . . . . .

2
4
5
7

2 Vectors and Linear (Vector) Spaces

2.1 Linear Equations . . . . . . . . . . . . . . . . . . . . . . . . . .

9
11

3 Eigenvalues and Eigenvectors

3.1 Jordan Forms . . . . . . . . . . . . . . . . . . . . . . . . . . . .

12
17

4 Inner Products & Norms

4.1 Inner Products . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.2 Norms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

20
20
20

5 Applications of Linear Algebra

5.1 The Cayley-Hamilton Theorem . . . . . .
5.2 Solving State Equations . . . . . . . . . .
5.3 Matrix Decompositions . . . . . . . . . .
5.4 Bilinear Forms and Sign-definite Matrices
5.4.1 Definite Matrices . . . . . . . . .
5.5 Matrix Inversion Lemmas . . . . . . . . .

24
24
24
29
32
32
33

6 MATLAB Commands

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

34

1. Review of Matrix Algebra

LINEAR ALGEBRA & MATRICES

These notes deal with the study of linear Algebra and matrices. Linear Algebra
plays an important role in the subareas of signal processing, control systems, communications, and more broadly in the studies of systems. The notes rely heavily on
[1, 2, 3, 4].

1 Review of Matrix Algebra

We start this chapter by introducing matrices and their algebra. Notationally,
matrices are denoted by capital letters A, M, and are rectangular arrays of elements . Such elements are referred to as scalars and denoted by lowercase letters,
a, b, , etc.. Note that the scalars are not necessarily real or complex constants:
they maybe real or complex numbers, polynomials, or general functions.
Example 1 Consider the matrix

1
2
3
4
5
6
4
3
2 105
1
0.9 7.2 10
0.17 4.96 10

4
0.012338
11.72
2.6
8.7

10
31
22
A=

0
0
1
0
0
0
0
0
0
0
0
30

This is a 5 6 matrix with real entries. On the other hand,

s+1
s+3
P(s) =
s2 + 3s + 2 s2 + 5s + 4

is a 2 2 matrix of polynomials in the variable s. Finally,

1+ j
j
0
=
1 + 3 j 5 2 j
is a 2 3 complex matrix.

A matrix which has m rows and n columns, is said to be m n. A matrix may be

denoted by A = [ai j ] where i = 1, m and j = 1, n.
Example 2 The matrix A in Example 1 has 5 rows and 6 columns. The element
a11 = 1 while a35 = 31.

1. Review of Matrix Algebra

A scalar is a 1 1 matrix. If m = n the matrix is said to be square. If m = 1 the

matrix is a row matrix (or vector). If n = 1 the matrix is an m column matrix (or
vector). If A is square then we define the Trace of A by the sum of its diagonal
elements, i.e.
n

Trace(A) = aii

(1)

i=1

Example 3 Consider P(s) of Example 1, then Trace(P(s)) = s2 + 6s + 5.

Two matrices A and B are equal, written A = B, if and only if A has the same
number of rows and columns as B and if ai j = bi j for all i, j. Two matrices A and
B that have the same numbers of rows and columns may be added or subtracted
element by element, i.e.
C = [ci j ] = A B ci j = ai j bi j i, j

(2)

the multiplication of 2 matrices A and B to obtain C = AB may be performed if and

only if A has the same number of columns as B has rows. In fact,
n

C = AB ci j =

aik bk j

(3)

k=1

the matrix C is m q if A is m n and B is n q. Note that BA may not even be

defined and in general, BA is not equal to AB even when both products are defined.
The matrix of all zeros is the Null matrix, and the square matrix A with aii = 1 and
ai j = 0 for i 6= j is the identity matrix. The identity matrix is denoted by I. Note
that AI = IA = A assuming A is n n as is I.
The following properties of matrix Algebra are easily verified
1. A B = B A
2. A + (B +C) = (A + B) +C
3. (A + B) = A + B for all scalars .
4. A = A
5. A(BC) = (AB)C
6. A(B +C) = AB + AC

1.1 Determinants, Minors, and Cofactors

7. (B +C)A = BA +CA
Let A = [ai j ] be an m n matrix. The transpose of A is denoted by AT = [a ji ]
and is the n m matrix obtained by interchanging columns and rows. The matrix
A is symmetric if A = AT , skew-symmetric if A = AT . Also, it can be seen that
(AB)T = BT AT . In the case where A contains complex elements, we let A be the
conjugate of A whose elements are the conjugates of those of A. Matrices satisfying
A = A T are Hermitian and those satisfying A = A T are skew-Hermitian.

1.1 Determinants, Minors, and Cofactors

In this section we will consider square matrices only. The determinant of a square
matrix A denoted by det(A) or | A | is a scalar-valued function of A and is given as
follows for n = 1, 2, 3.
1. n = 1, then | A |= a11
2. n = 2, then | A |= a11 a22 a12 a21
3. n = 3, then | A |= a11 (a22 a33 a23 a32 ) + a12 (a23 a31 a21 a33 ) + a13 (a21 a32
a22 a31 )
Standard methods apply for calculating determinants. The determinant of a square
matrix is 0, that of an identity matrix is 1, and that of a triangular or diagonal
matrix is the product of all diagonal elements. Each element ai j in an n n square
matrix A has associated with it a minor Mi j obtained as the determinant of the
(n 1) (n 1) matrix resulting form deleting the ith row and the jth column.
From these minors, we can obtain the cofactors given by Ci j = (1)i+ j Mi j .
Example 4 Given

2 4 1
A= 3 0 2
2 0 3

Then, M12 = 5, C12 = 5, M32 = 1 = C32 .

Some useful properties of determinants are listed below:

1. Let A and B be n n matrices, then | AB |=| A | . | B |.
2. | A |=| AT |

1.2 Rank, Trace, and Inverse

3. If A contains a row or a column of zeros, then | A |= 0.

4. If any row (or column) of A is a linear combination of other rows (or columns),
then | A |= 0.
5. If we interchange any 2 rows (or columns) of a matrix A, the determinant of
the resulting matrix is | A |.
6. If we multiply a row (or a column) of a matrix A by a scalar , the determinant of the resulting matrix is | A |.
7. Any multiple of a row (or a column) may be added to any other row (or
column) of a matrix A without changing the value of the determinant.

1.2 Rank, Trace, and Inverse

The rank of an m n matrix A denoted by rA = rank(A) is the size of the largest
nonzero determinant that can be formed from A. Note that rA min{m, n}. If A is
square and if rA = n, then A is nonsingular. If rA is the rank of A and rB is the rank
of B, and if C = AB, then 0 rC min{rA , rB }. The trace of a square matrix A was
defined earlier as Tr(A) = ni=1 aii .
Example 5 Let us look at the rank of:

2 4 1
A= 3 0 2
2 0 3

It is easy to see that rA = 3. Now consider the matrix

2 0
B= 3 0
1 1

whose rank is rB = 2 and form

17 1
A.B = 8 2
1 1

and the rank(C) = 2 {rA , rB }.

1.2 Rank, Trace, and Inverse

If A and B are conformable square matrices, Tr(A + B) = Tr(A) + Tr(B), and

Tr(AB) = Tr(BA). Also, Tr(A) = Tr(AT ).
Example 6 Show that Tr(ABC) = Tr(BT AT CT ). First, write Tr(ABC) = Tr(CAB),
then Tr([CAB]T ) = Tr(CAB), thus proven.

Note that rank(A + B) 6= rank(A) + rank(B) and that Tr(AB) 6= Tr(A) Tr(B).
Example 7 Let

2 4 1
1 0 0
A = 3 0 2 ; B = 0 0 0
2 0 3
0 0 0

and note that rA = 3, rB = 1, while rA+B = 3. Also, note that Tr(AB) = 2 while
Tr(A) Tr(B) = 5 2 = 10.

Next, we define the inverse of a square, nonsingular matrix A as the square

matrix B of the same dimensions such that
AB = BA = I
the inverse is denoted by A1 and may be found by
A1 =

CT
|A|

where C is the matrix of cofactors Ci j of A. Note that for the inverse to exist, | A |
must be nonzero, which is equivalent to saying that A is nonsingular. We also write
CT = Ad joint(A).
The following property holds:
(AB)1 = B1 A1
assuming of course that A and B are compatible and both invertible.

1.3 Elementary Operations and Matrices

1.3 Elementary Operations and Matrices

The basic operations called elementary operations are as follows
1. Interchange any 2 rows (or columns)
2. Multiply any row (or column) by a scalar
3. Multiply any row (or column) by a scalar and add the resulting row (or
column) to any other row (or column)
Note that each of these elementary operations may be represented by a postmultiplication for column operations (or premultiplication for row operations) by elementary matrices which are nonsingular.
Example 8 Let

2 1 0
A= 0 2 1
1 1 1

Then, let use the following row operations:

1. Exchange rows 1 and 3

2. Multiply row 1 by -2 and add it to row 3

3. Multiply row 2 by 1/2 and add it to row 3
4. Multiply row 2 by 1/2
5. Multiply row 3 by 2/5
6. Multiply column 1 by -1 and add it to column 2
7. Multiply column 1 by 1 and add it to column 3
8. Multiply column 2 by -1/2 and add it to column 3
The end result is

1 0 0
B= 0 1 0
0 0 1

1.3 Elementary Operations and Matrices

Example 9 Given the following polynomial matrix

2

s
0
s2
P(s) = 0
1 s+1

then the corresponding row operations are performed:

1. interchange rows 3 and 1.
2. multiply row 1 by s2 and add it to row 3
3. multiply row 2 by s + 1 and add it to row 3.
This corresponds to the following multiplications by the matrices given:
1.

0 0 1
1 s+1
s2 = 0 1 0 P(s)
P1 (s) = 0
1 0 0
s2
0

2.

1
s+1
1 0 0
= 0 1 0 P1 (s)
s2
P2 (s) = 0
2
0 s (s + 1)
s2 0 0
3.

1 s+1
1
0
0
1
0 P(s)
P3 (s) = 0 s2 = 0
0
0
0 s+1 0

In general, for any matrix of rank r we can reduce it via column and row operations to one of the following normal forms

Ir 0
T
(4)
Ir , [Ir | 0], [Ir | 0] , or
0 0
Next, we discuss vector spaces, or as they are sometimes called, linear space..

2. Vectors and Linear (Vector) Spaces

2 Vectors and Linear (Vector) Spaces

In most of our applications, we need to deal with linear real and complex vector
spaces which are defined subsequently.
Definition 1 A real linear vector space (resp. complex linear vector space is a set
V , equipped with 2 binary operations: the addition (+) and the scalar multiplication
(.) such that
1. x + y = y + x, x, y V
2. x + (y + z) = (x + y) + z, x, y, z V
3. There is an element 0V in V such that x + 0V = 0V + x = x, x V
4. For each x V , there exists an element x V such that x + (x) = (x) +
x = 0V
5. For all scalars r1 , r2 R (resp. c1 , c2 C), and each x V , we have r1 .(r2 .x) =
(r1 r2 ).x (resp. c1 .(c2 .x) = (c1 c2 ).x
6. For each r R (resp. c C), and each x1 , x2 V , r.(x1 + x2 ) = r.x1 + r.x2
(resp. c.(x1 + x2 ) = c.x1 + c.x2 )
7. For all scalars r1 , r2 R (resp. c1 , c2 C), and each x V , we have (r1 +
r2 ).x = r1 .x + r2 .x (resp. (c1 + c2 ).x = c1 .x + c2 .x)
8. For each x V , we have 1.x = x where 1 is the unity in R (resp. in C).
Example 10 The following are linear vector spaces with the associated scalar
fields: Rn with R, Cn with C.

Definition 2 A subset M of a vector space V is a subspace if it is a linear vector

space in its own right. One necessary condition for M to be a subspace is that it
contains the zero vector.
Let x1 , . . . xn be some vectors in a linear space X defined over a field F. In these
notes we shall consider either the real field R or the complex field C. The Span of
the vectors x1 , . . . xn over the field F is defined as
n

SpanF {x1 , . . . xn } := {x X : x = ai xi , ai F}.

i=1

2. Vectors and Linear (Vector) Spaces

10

The vectors x1 , . . . xn are said to be linearly independent if

n

ai xi = 0 ai = 0, i = 1, . . . n

i=1

otherwise they are linearly dependent. A set of vectors x1 , . . . xm is a basis of a

linear space X if the vectors are linearly independent and their Span is equal to
X . In this case the linear space X is said to have finite dimension m.
Note that any set of vectors containing the zero vector is linearly dependent.
Also, if {xi } i = 1, , n is linearly dependent, adding a new vector xn+1 will not
make the new set linearly independent. Finally, if the set {xi } is linearly dependent,
then, at least one of the vectors in the set may be written as a linear combination of
the others.
How do we test for linear dependence/independence? Given a set of n vectors
{xi } each having m components. Form the m n matrix A which has the vectors
xi as its columns. Note that if n > m then the set of vectors has to be linearly
independent. Therefore, let us consider the case where n m. Form the n n
matrix G = AT A and check if it is nonsingular, i.e. if | G |6= 0, then {xi } is linearly
independent, otherwise {xi } is linearly dependent.
Geometrically, linear independence may be explained in Rn . Consider the
plane R2 and suppose we have 2 vectors x1 and x2 . If the 2 vectors are linearly
dependent, then x2 = ax1 or both vectors lie along the same direction. If they were
linearly independent, then they form the 2 sides of a parallelogram. Therefore, in
the case of linear dependency, the parallelogram degenerate to a single line. We
can equip a vector space with many functions. One of which is the inner product
which takes two vectors in V to a scalar either in R or in C, the other one is the
norm of a vector which takes a vector in V to a positive value in R.
Given two linear spaces X and Y over the same field F, a function A : X 7
Y is a linear transformation if A (ax + by) = aA (x) + bA (y), for all a, b F.
Let A be a linear transformation A : X 7 X and Y a linear subspace Y X .
The subspace Y is said to be A invariant if
A (y) Y , y Y .
Given a linear transformation, we define the Range or Image of A as the subspace
of Y
Range(A ) := {y Y : y = A (x), x X },
and the Kernel or Null-Space of A as the subspace of X

N (A ) := Ker(A ) := {x X : A (x) = 0}.

Given X = Cm and Y = Cn , a matrix A Cmn , denoted by A = [ai j ; 1 i

m, 1 j n, is an example of linear transformation.

2.1 Linear Equations

11

Example 11 The set {ei } i = 1, , n is linearly independent in Rn . On the other

hand, the 2 vectors, x1T = [1 2]; x2T = [2 4] are linearly dependent in R2 .

Example 12 Let X be the vector space of all vectors x = [x1 xn ]T such that all
components are equal. Then X is spanned by the vector of all 1s. Therefore dim
(X) = 1. On the other hand, if X is the vector space of all polynomials of degree
n 1 or less, a basis is {1,t,t 2 , ,t n1 } which makes dim (X) = n.

2.1 Linear Equations

In many control problems, we have to deal with a set of simultaneous linear algebraic equations
a11 x1 + a12 x2 + + a1n xn = y1

a21 x1 + a22 x2 + + a2n xn = y2

.
..
. = ..

am1 x1 + am2 x2 + + amn xn = ym

(5)

or in matrix notation
Ax = y

(6)

for an m n matrix A, an n 1 vector x and an m vector y. The problem is to find

the solution vector x, given A and y. Three cases might take place:
1. No solution exist
2. A unique solution exists
3. An infinite number of solutions exist
Now, looking back at the linear equation Ax = y, it is obvious that y should be
in R(A) for a solution x to exist, in other words, if we form
W = [A | y]

(7)

then rank(W ) = rank(A) is a necessary condition for the existence of at least one
solution. Now, if z N (A), and if x is any solution to Ax = y, then x + z is also a
solution. Therefore, for a unique solution to exist, we need that N (A) = 0. That
will require that the columns of A form a basis of R(A), i.e. that there will be n of
them and that they will be linearly independent, and of dimension n. Then, the A
matrix is invertible and x = A1 y is the unique solution.

3. Eigenvalues and Eigenvectors

12

3 Eigenvalues and Eigenvectors

Let A be an n m matrix, and denote the corresponding identity matrix by I. Then,
let xi be a nonzero vector in Rn and i be scalar such that
Axi = xi

(8)

Then, i is an eigenvalue of A and xi is the corresponding eigenvector. There will

be n eigenvalues of A (some of which redundant). In order to find the eigenvalues
of A we rewrite the previous equation
(A i I)xi = 0

(9)

Noting that xi can not be the zero vector, and recalling the conditions on the existence of solutions of linear equations, we see that we have to require
det(A i I) =| (A i I) |= 0

(10)

We thus obtain an nth degree polynomial, which when set to zero, gives the characteristic equation
| (A I) | = ( ) = 0

= ( )n + cn1 n1 + + c1 + c0

= (1)n ( 1 )m1 ( 2 )m2 ( p )m p

(11)

Therefore, i is an eigenvalue of A of algebraic multiplicity mi . One can then

shows,
n

Tr(A) =

= (1)n+1cn1

i=1
n

|A| =

i = c0

(12)

i=1

Also, if i is an eigenvalue, then so is i .

How do we determine eigenvectors? We distinguish 2 cases:
1. All eigenvalues are distinct: In this case, we first find Ad j(A I) with as
a parameter. Then, successively substituting each eigenvalue i and selecting
any nonzero column gives all eigenvectors.
Example 13 Let

0
1
0
0
1
A= 0
18 27 10

3. Eigenvalues and Eigenvectors

13

then, solve for

| A I |= 3 10 2 27 18 = 0
which has 3 solutions, 1 = 1, 2 = 3, 3 = 6. Now, let us solve for
the eigenvectors using the suggested method

2
+ 10 + 27
+ 10
1
18
2 + 10

Ad j(A I) =
18
27 18 2
so that for 1 = 1 we can see that column 1 is
x1 = [18 18 18]T
Similarly, x2 = [7 21 63]T and x3 = [1 6 36]T . There is another
method of obtaining the eigenvectors from the definition by actually solving
Axi = i xi

Note that once all eigenvectors are obtained, we can arrange them in an n n
modal matrix M = [x1 x2 xn ]. Note that the eigenvectors are not unique
since if xi is an eigenvector, then so is any xi for any scalar .
2. Some eigenvalues are repeated: In this case, a full set of independent eigenvectors may or may not exist. Suppose i is an eigenvalue with an algebraic
multiplicity mi . Then, the dimension of the Null space of A i I which
is also the number of linearly independent eigenvectors associated with i is
the geometric multiplicity qi of the eigenvalue i . We can distinguish 3 cases
(a) Fully degenerate case qi = mi : In this case there will be qi independent
solutions to (A i I)xi rather than just one.
Example 14 Given the matrix

10/3
0
A=
2/3
2/3

1 1 1/3
4 0
0

1 3 1/3
1 1 11/3

Then, its characteristic equation is

( ) = 4 14 3 + 72 2 160 + 128 = 0

3. Eigenvalues and Eigenvectors

14

There are 4 roots, 1 = 2, 2 = 3 = 4 = 4. We can find the eigenvector associated with 1 as x1 = [1 0 1 1]T . Then, we find

2/3 1 1 1/3
0
0 0 0

(4I A)x =
2/3 1 1 1/3 x
2/3 1 1 1/3

There are an infinite number of solutions, 3 of which are x2 = [1 0 0

2]T ; x3 = [0 1 1 0]T ; x4 = [0 1 0 3]T

Note that now if x1 , xm are eigenvectors for eigenvalue i , then so

is any y = m
i=1 i xi .
(b) Simple degenerate case qi = 1: Here we can find the first eigenvector in
the usual means, then we have mi 1 generalized eigenvectors. These
are obtained as
Ax1 = i x1
(A i )x2 = x1
.
..
. = ..
(A i )xmi

= xmi 1

(13)

Note that all xi found this way are linearly independent.

Example 15

0
1
0
0
0
0
1
0

A=
0
0
0
1
8 20 18 7

its characteristic polynomial is

( ) = 4 + 7 3 + 18 2 + 20 + 8 = 0
then, 1 = 1, 2 = 3 = 4 = 2. The eigenvector of 1 is easily
found to be x1 = [1 1 1 1]T . On the other hand, one eigenvector 0f
2 is found to be x2 = [0.125 0.25 0.5 1]T , then x2 = (A + 2I)x3
leading to x3 = [0.1875 0.25 0.25 0]T also, x3 = (A + 2I)x4 so that
x4 = [0.1875 0.1875 0.125 0]T

3. Eigenvalues and Eigenvectors

15

(c) General case 1 qi mi : Here, a general top-down method should be

used. Here we solve the problem by writing
(A i I)x1 = 0

(A i I)x2 = x1 (A i I)2 x2 = (A i I)x1 = 0

(A i I)x3 = x2 (A i I)2 x3 = (A i I)x2 = x1 6= 0

(A i I)3 x3 = (A i I)x1 = 0

(14)

This approach continues until we reach the index ki of i . This index is

found as the smallest integer such that
rank(A i I)k = n mi

(15)

the index indicates the length of the longest chain of eigenvectors and
generalized eigenvectors associated with i .
Example 16

0
0
A=
0
0

its characteristic polynomial is

0
0
0
0

1
0
0
0

0
1

0
0

( ) = 4 = 0
so 1 = 0 with m1 = 4. Next, form A 1 I = A, and determine that
rank(A 1 I) = 2. There are then 2 eigenvectors and 2 generalized
eigenvectors. The question is whether we have one eigenvector-generalized
eigenvector chain of length 3, or 2 chains of length 2. To check that,
note that n m = 0, and that rank(A 1 I)2 = 0, therefore, the index
is k1 = 2. this then guarantees that one chain has length 2, making the
length of the other chain 2. First, consider (A 1 I)2 x = 0 Any vector satisfies this equation but only 4 vectors are linearly independent.
Let us choose x1 = [1 0 0 0]T . Is this an eigenvector or a generalized eigenvector? It is an eigenvector since (A 1 I)x1 = 0. Similarly,
x2 = [0 1 0 0]T is an eigenvector. On the other hand, x3 = [0 0 1 0]T
is a generalized eigenvector since (A i I)x3 = x1 6= 0. Similarly,
x4 = [0 0 0 1]T is a generalized eigenvector associated with x2 .

3. Eigenvalues and Eigenvectors

16

Example 17 Let

4 1 2
A = 0 2 0 .
0 1 2

The eigenvalues of A are 1 = 4, 2,3 = 2 and the two correspondent eigen

T

T
vectors are x1 = 1 0 0 , x2 = 1 0 1 . Therefore we have three
eigenvalues but only two linearly independent eigenvectors. We call gener
T
alized eigenvector of A the vector x3 = 1 1 1 such that
Ax3 = 2 x3 + x2 .

The vector x3 is special in the sense that, x2 and x3 together span a twodimensional A invariant subspace.

In summary each n n matrix has n eigenvalues and n linearly independent

vectors, either eigenvectors or generalized eigenvectors.
To summarize, given the characteristic polynomial
p( ) := det( I A) = n + a1 n1 + . . . + an ,

(16)

the roots 1 , . . . n of p( ) = 0 are the eigenvalues of A. The set = {1 , . . . n } is

called the spectrum of A. The spectral radius of A is defined as

(A) := max |i |.
i=1,...n

where |.| is the modulus of the argument and thus (A) is the radius of the smallest
circle in the complex plane, centered at the origin, that includes all the eigenvalues.
As described earlier, the non-null vectors xi such that Axi = i xi are the (right)eigenvectors of A. Similarly y 6= 0 is a left-eigenvector of A if y A = y . In
general a matrix A has at least one eigenvector. It is easy to see that if x is an
eigenvector of A, Span{x} is an A invariant subspace.
In general let us suppose that a matrix A has an eigenvalue of multiplicity r
but with only one correspondent eigenvector. Then we can define r 1 generalized
eigenvectors in the following way
Ax1 = x1
Ax2 = x2 + x1
..
.
Axr = xr + xr1 .

3.1 Jordan Forms

17

The eigenvector and the r 1 generalized eigenvectors together span a r-dimensional

A invariant subspace.

3.1 Jordan Forms

Based on the analysis above, we can produce the Jordan form of any n n matrix.
In fact, if we have n different eigenvalues, we can find all eigenvectors xi , such that
Axi = i xi , then let M = [x1 x2 xn ], and = diag(i ). This leads to AM = M.
We know that M 1 exists since all eigenvectors are independent and therefore =
M 1 AM. The Jordan form of A is then . If the eigenvectors are orthonormal,
M 1 = M T . In the case that qi = mi , we proceed the same way to obtain J = . On
the other hand, if qi = 1, then using the eigenvectors and generalized eigenvectors
as columns of M will lead to J = diag[J1 J2 J p ] where each Ji is mi mi and
has the following form

i 1 0 0
0 i 1 0

0 0 i 0

Ji = . .
..
..
..
.
.
. .
.
.
.

0 0 i 1
0 0 0 i
Finally, the general case will have Jordan blocks each of size ki as shown in the
examples below.
Example 18 Let

0
1
0
0
1
A= 0
18 27 10

the eigenvalues are 1 = 1, 2 = 3, 3 = 6. The eigenvectors are

x1 = [1 1 1]T
Similarly, x2 = [1 3 9]T and x3 = [1 6 36]T . Then,

1
1
1
M = 1 3 6
1
9 36

3.1 Jordan Forms

18

then

then = M 1 AM.

1.8
0.9
0.1
M 1 = 1 1.167 0.167
0.2 0.2667 0.067

Example 19 Given the matrix

10/3
0
A=
2/3
2/3
Then, its characteristic equation is

1 1 1/3
4 0
0

1 3 1/3
1 1 11/3

( ) = 4 14 3 + 72 2 160 + 128 = 0
There are 4 roots, 1 = 2, 2 = 3 = 4 = 4. We can find the eigenvector associated
with 1 as x1 = [1 0 1 1]T . Then, we find x2 = [1 0 0 2]T ; x3 [0 1 1 0]T ; x2 =
[0 1 0 3]T Therefore,

1 1 0 0
0
0 1 1

M=
1 0 1 0
1 2 0 3
then

then

1/3
1/6
1
M =
1/3
1/3

2
0
J=
0
0

1/2 1/2 1/6

1/2 1/2 1/6

1/2 1/2 1/6

1/2 1/2 1/6
0
4
0
0

0
0
4
0

0
0

0
4

3.1 Jordan Forms

19

Example 20 Now consider

0
1
0
0
0
0
1
0

A=
0
0
0
1
8 20 18 7

its characteristic polynomial is

( ) = 4 + 7 3 + 18 2 + 20 + 8 = 0
then, 1 = 1, 2 = 3 = 4 = 2. Find eigenvectors as before and form

1 1/8 0.1875 0.1875

1 1/4 1/4 0.1875

M=
1 1/2
1/4
1/8
1
1
0
0

Then,

1 0
0
0
0 2 1
0

J=
0
0 2 1
0
0
0 2

Example 21

0
0
A=
0
0

0
0
0
0

1
0
0
0

its characteristic polynomial is

0
1

0
0

( ) = 4 = 0
so 1 = 0 with m1 = 4. Then,

1
0
M=
0
0

0
0
1
0

0
1
0
0

0
0

0
1

4. Inner Products & Norms

20

therefore making

0
0
J=
0
0

1
0
0
0

0
0
0
0

0
0

1
0

4 Inner Products & Norms

4.1 Inner Products
An inner product is an operation between two vectors of a vector space which will
allow us to define geometric concepts such as orthogonality and Fourier series, etc.
The following defines an inner product.
Definition 3 An inner product defined over a vector space V is a function < ., . >
defined from V to F where F is either R or C such that x, y, z, V
1. < x, y >=< y, x > where the < ., . > denotes the complex conjugate.
2. < x, y + z >=< x, y > + < x, z >
3. < x, y >= < x, y >, F
4. < x, x > 0 where the 0 occurs only for x = 0V
Example 22 The usual dot product in Rn is an inner product, i.e. < x, y >= xT y =
ni=1 xi yi .

4.2 Norms
A norm is a generalization of the ideas of distance and length. As stability theory is
usually concerned with the size of some vectors and matrices, we give here a brief
description of some norms that will be used in these notes. We will consider first
the norms of vectors defined on a vector space X with the associated scalar field of
real numbers R.
Let X be a linear space on a field F. A function k k : X 7 R is called a norm
if it satisfies the following proporties

4.2 Norms

1. kxk 0,

21

x X

2. kxk = 0 x = 0
3. kaxk = |a|kxk,

a F, x X

4. kx + yk kxk + kyk,
as

x, y X

T
Let now X = Cn . For a vector x = x1 . . . xn in Cn the p-norm is defined
kxk p :=

|xi | p

i=1

! 1p

, p 1.

(17)

When p = 1, 2, , we have the three important norms

n

kxk1 := |xi |
i=1

kxk2 :=

|xi |2

i=1

kxk := max |xi |.

i

In a Euclidean space Rn , the 2-norm is the usual Euclidean norm.

Example 23 Consider the vector

1
x = 2
2

Then, k x k1 = 5, k x k2 = 3 and k x k = 2.

Let us now consider the norms for a matrix A Cmn . First let us define the
induced p-norms
kAxk p
kAk p := sup
, p 1.
x6=0 kxk p

4.2 Norms

22

These norms are induced by the p-norms on vectors. For p = 1, 2, , there exists
some explicit formulae for the induced pnorms
m

|ai j |
1 jn

kAk1 := max
kAk2 :=

(maximum absolute column sum)

i=1

p
max (A A) (spectral norm)
n

|ai j |
1im

kAk := max

(maximum absolute row sum).

j=1

Unless otherwise specified, we shall adopt the convention of denoting the 2norm without any subscript. Therefore by kxk and kAk we shall mean respectively
kxk2 and kAk2 .
Another often used matrix norm is the so-called Frobenius norm
s
m n
p
(18)
kAkF := tr (A A) = |ai j |2 .
i=1 j=1

It is possible to show that the Frobenius norm cannot be induced by any vector
norm.
The following Lemma presents some useful results regarding matrix norms.
Lemma 1 Let A Cmn . Then
1. kABk kAkkBk, for any induced norm

(submultiplicative property)

2. Given two unitary matrices U and V of suitable dimensions

kUAV k = kAk

kUAV kF = kAkF

3. (A) kAk, for any induced norm and the Frobenius norm
Given a square nonsingular matrix A, the quantity

(A) := kAk p kA1 k p

(19)

is called the condition number of A with respect to the induced matrix norm k k p .
From Lemma 1, we have

(A) = kAk p kA1 k p kAA1 k p = kIk p = 1.

If (A) is large, we say that A is ill conditioned; if (A) is small (i.e. close to 1),
we say that A is well conditioned. It is possible to prove that, given a matrix A, the
reciprocal of (A) gives a measure of how far A is from a singular matrix.
We now present an important property of norms of vectors in Rn which will be
useful in the sequel.

4.2 Norms

23

Lemma 1 Let k x ka and k x kb be any two norms of a vector x Rn . Then there

exists finite positive constants k1 and k2 such that
k1 k x ka k x kb k2 k x ka x Rn
The two norms in the lemma are said to be equivalent and this particular property will hold for any two norms on Rn .
Example 24 Note the following
1. It can be shown that for x Rn
||x||1

n||x||2

||x|| ||x||1 n||x||

||x||2 n||x||
2. Consider again the vector of ex 23. Then we can check that
||x||1

3||x||2

||x|| ||x||1 3||x||

||x||2 3||x||

Note that a norm may be defined independently from an inner product. Also, we
can define the generalized angle between 2 vectors in Rn using
x.y =< x, y >=k x k . k y k cos

(20)

where is the angle between x and y. Using the inner product, we can define the
orthogonality of 2 vectors by
x.y =< x, y >= 0
which of course means that = (2i + 1) /2.

(21)

5. Applications of Linear Algebra

24

5 Applications of Linear Algebra

5.1 The Cayley-Hamilton Theorem
Note that Ai = A A i times. Then, we have the Cayley-Hamilton theorem
Theorem 1 Let A be an n n matrix with a characteristic equation
| (A I) = ( )

= ( )n + cn1 n1 + + c1 + c0
= 0

(22)

Then,
(A) = (1)n An + cn1 An1 + + c1 A + c0 I
= 0

(23)

Example 25 Use the Cayley-Hamilton theorem to find the inverse of a matrix A

0 = (1)n An + cn1 An1 + + c1 A + c0 I
then,
0 = (1)n An1 + cn1 An2 + + c1 I + c0 A1
A1 =

1
[(1)n An1 + cn1 An2 + + c1 A]
c0

5.2 Solving State Equations

Suppose we want to find a solution to the state-space equation
x = Ax(t) + bu(t); x(0) = x0
and assume initially that we are interested in finding x(t) for u(t) = 0; t 0.
Effectively, we then have
x = Ax(t)

5.2 Solving State Equations

25

Let us then take the Laplace transform,

sX(s) x0 = AX(s)
which gives
X(s) = [sI A]1 x0
which then leads to
x(t) = L 1 [sI A]1 x0
we define the matrix exponential
exp(At) = L 1 [sI A]1
so that
x(t) = exp(At)x0
there are many reasons why that is justified. In fact, it can be shown that the matrix
exponential has the following properties
1.

d
dt exp(At)

= Aexp(At) = exp(At)A

2. expA(t1 + t2 ) = exp(At1 )exp(At2 )

3. exp(A0) = I
4. exp(At) is nonsingular and [exp(At)]1 = exp(At)
i i

5. exp(At) = limk ki=0 A it

The matrix exp(At) is known as the state transition matrix.

Now assume that u(t) is no longer zero. Then, the solution due to both x(0)
and u(t) is given by
x(t) = exp(At)x0 +

Z t
0

eA(t ) bu( )d

Example 26 Consider the system

2 1
x =
x
4 2

1
x(0) =
1

5.2 Solving State Equations

26

To find x(t) let us find exp(At) using 2 methods. First, by evaluating the infinite
series
t2
exp(At) = I + At + A2 +

1 0
2 1
=
+
t +0
0 1
4 2

1 + 2t
t
=
4t
1 2t

(24)

so that

1 + 2t
t
x(t) =
4t
1 2t

1 + 3t
=
6t 1

1
1

next, consider the Laplace transform approach, where we find (sI A)1
s+2 1
1
s2
s2
(sI A) =
4
s2
s2

(25)

s2

so that
exp(At) =

1 + 2t
t
4t
1 2t

(26)

Now consider a third method whereby, we transform A to its Jordan normal form.
The A matrix has a double eigenvalue at zero with an eigenvector x1 = [1 2]T and
a generalized eigenvector at x2 = [0 1]. Therefore, the M matrix is given by

1
0
(27)
2 11
Using T = M we obtain,
J = A = T

AT =

0 1
0 0

so that
x =

0 1
0 0

x;
x(0)
= [1 3]T

(28)

5.2 Solving State Equations

27

we can then solve 2 differential equations

x1 = x2
x2 = 0
so that x2 (t) = 3 and x1 = 3t + 1. Therefore,
x(t) = T 1 x

1 0
3t + 1
=
2 1
3

3t + 1
=
6t 1

(29)

Note that in general

exp(At) = Mexp(Jt)M 1
since
Mexp(Jt)M 1 = Mexp(M 1 AM)M 1
1
= M[I + M 1 AMt + (M 1 AMt)2 + ]M 1
2
1 22
= I + At + A t +
2
= exp(At)
It is extremely simple to calculate exp(Jt) = exp(t) for the case of diagonalizable
matrix A as shown in the following example.
Example 27 Consider the system

1 2
x =
x
2 3

1
x(0) =
1
then there are 2 distinct eigenvalues, and we can find exp(Jt),

1 = 0.2361; 2 = 4.2361

0.8507 0.5257
M=
0.5257 0.8507

5.2 Solving State Equations

28

so that
M

0.8507 0.5257
0.5257 0.8507

Note that M 1 = M T . Then,

0.7237e0.2361t + 0.2764e4.2361t
exp(At) =
0.4472(e4.2361t e0.2361t )

0.4472(e4.2361t e0.2361t )
0.2764e0.2361t + 0.7237e4.2361t

Recall that A matrix A Cnn is said to be Hermitian if A = A , unitary if

A1 = A . A Hermitian
matrix has the property that all its eigenvalues are real. Let
A = a1 a2 an . It is easy to verify that A is unitary if and only if ai a j = i j ,
where i j = 1 if i = j and i j = 0 if i 6= j, i.e. if and only if the columns of A form
an orthonormal basis for Cn .
Let us now consider a nonsingular A partitioned as

A11 A12
A=
A21 A22

(30)

with A11 and A22 square matrices. Suppose that A11 is nonsingular. Then the matrix
:= A22 A21 A1
11 A12
is called the Schur complement of A11 in A. Similarly, if A22 is nonsingular, the
matrix
:= A11 A12 A1 A21

22
is the Schur complement of A22 in A. A useful expression for the inverse of A in
terms of partitioned blocks is

A1
A12 1
1
11
(31)
A =
1
1 A21 A1
11
supposing that all the relevant inverses exist. The following well-established identities are also very useful
(A BD1C)1 = A1 + A1 B(D CA1 B)1CA1
1

(I + AB) A = A(I + BA)

I AB(I + AB)

= (I + AB)

(32)
(33)
(34)

5.3 Matrix Decompositions

29

supposing the existence of all the needed inverses.

If A11 is nonsingular then by multiplying A on the right by

I A1
11 A12
0
I
it is easy to verify that
det(A) = det(A11 ) det().

(35)

Similarly, if A22 is nonsingular, then by multiplying A on the right by

I
0
A1
22 A21 I
we find that

det(A) = det(A22 ) det().

(36)

Consider A Cmn and B Cnm . Using identities (35) and (36), it is easy to
prove that that
det(Im AB) = det(In BA).

5.3 Matrix Decompositions

In this section we shall explore some matrix decompositions that use Unitary matrices.
Theorem 1 (Schur Decomposition) Let A Cnn . Then there exist a Unitary matrix U Cnn so that A = UTU , where T is a triangular matrix whose diagonal
entries are the eigenvalues of A.
Proof 1 We shall prove the theorem by constructing the matrices U and T . Given
a matrix A, we can always find an eigenvector u1 , associated with an eigenvalue

1 , so that Au1 = 1 u1 . Moreover we can assume that ku1 k = 1. Let U1 =

u1 z2 zn be a Unitary matrix. Since the columns of U1 are a basis for
Cn , we have

1
A u1 z2 zn = u1 z2 zn
.
0 A1
Therefore

1
A = U1
U1 .
0 A1

(37)

5.3 Matrix Decompositions

30

Now let us consider A1 C(n1)(n1) . From (37) it is easy to see that the (n 1)
eigenvalues of A1 are also eigenvalues of A. Given 2 an eigenvalue
of A1 with nor-
malized eigenvector u2 , we can construct a Unitary matrix U2 = u2 z2 zn1
C(n1)(n1) so that

A1U2 = U2 2
.
0 A2
Denoting by V2 the Unitary matrix

1 0
V2 =
,
0 U2
we have

1
1
1 0
1 0
= 0 2 .
V2U1 AU1V2 =
0 A1
0 U2
0 U2
0 0 A2

The result can be proven iterating this procedure n times.

Using Theorem 1 it is possible to prove that a Hermitian matrix A Cnn has

a set of n linearly independent eigenvectors. Indeed, since A is Hermitian, from
Theorem 1 we have
A = UTU = UT U T = T .
Since T is triangular, it must be diagonal. Therefore the columns of the matrix U
are the eigenvectors of A.
In Theorem 1, we used the same basis both for the domain and the range of the
matrix. In the following decomposition, we shall show what can be accomplished
when we choose possibly different bases for domain and the range.
Theorem 2 (Singular Value Decomposition) Let A Cmn . Then A can always
be written as A = UV , where U Cmm and V Cnn are Unitary matrices and
= [i j ] Rmn is a diagonal matrix with i j = 0 for i 6= j and ii = i 0.
Proof 2 We shall prove the theorem by induction. First of all, let us suppose,
without any loss of generality that A 6= 0 and m > n. Let now m = p + 1 and n = 1.
In this case we can write

1

A
0
A=
u2 un .. ,
.

5.3 Matrix Decompositions

31

T
with = kAk. Therefore A = UV , with U = A u2 un , = 1 0 0
and V = 1.
We suppose now that the theorem holds for m = p + k and n = k and prove
it for m = p + k + 1 and n = k + 1. Let A C(p+k+1)(k+1) . All the eigenvalues
of A A are real (since A A is Hermitian), nonnegative and at least one is greater
than zero since A 6= 0. Denote by the maximum eigenvalue of A A and by v the
correspondent normalized eigenvector
A Av = v.

Let now = + and u = Av

, so that kuk = 1. Let moreover U1 = u U0

C(p+k+1)(p+k+1) and V1 = v V0 C(k+1)(k+1) be Unitary matrices. We have

A1 = U1 AV1

u AV0
u
.
=
A v V0 =
0 U0 AV0
U0

Since u A = v A A = v , it follows that u AV0 = 0. Now an easy inductive argument completes the proof, noting that U0 AV0 C(p+k)k .
The scalars i are called the singular values of A. They are usually ordered
nonincreasingly 1 2 . . . r 0, with r = min{m, n}. The largest singular
value 1 and the smallest singular value r are denoted by

(A) := 1

(A) := r .

The columns vi of V are called the right singular vectors of A, and the columns ui
of U the left singular vectors. They are related by
Avi = i ui ,

i = 1, . . . r.

The following Lemma shows some of the information we can get from the
singular value decomposition of a matrix A.
Lemma 2 Let A Cmn and consider its singular value decomposition A = UV .
Then
1. The rank k of A equals the number of singular values different from zero
2. Range(A) = Span{u1 , . . . uk }
3. Ker(A) = Span{vk+1 , . . . vn }
4. 1 = kAk

5.4 Bilinear Forms and Sign-definite Matrices

32

5. kAk2F = ri=1 i2
6. Given a square nonsingular matrix A

(A) =

1
(A1 )

5.4 Bilinear Forms and Sign-definite Matrices

The expression < y, Ax >= yT Ax is a bilinear form. When y = x we have a quadratic
form xT Ax. Every matrix A can be written as the sum of a symmetric and skew
symmetric matrices if it is real, and of a Hermitian and skew-Hermitian if it is
complex. In fact, define the symmetric and anti-symmetric parts of A as:
As =

A AT
A + AT
; Aa =
2
2

and the Hermitian and anti-Hermitian parts as:

AH =

A + (A )T
A (A )T
; AAH =
2
2

Then note that < x, Ax >=< x, As x > if A is real, and < x, Ax >=< x, AH x > if A
is complex.
5.4.1 Definite Matrices
Let us consider a Hermitian matrix A. A is said to be positive (semi)definite if
x Ax > () 0, x Cn . We shall indicate a positive (semi)definite matrix by A >
() 0. A Hermitian matrix A is said to negative (semi)definite if (A) is positive
(semi)definite. The following Lemma gives a characterization of definite matrices.
Lemma 3 Let A be a Hermitian matrix. Then
1. A is positive (negative) definite if and only if all its eigenvalues are positive
(negative).
2. A is positive (negative) semidefinite if and only if all its eigenvalues are nonnegative (nonpositive).
Given a real n n matrix Q, then
1. Q is positive-definite, if and only if xT Qx > 0 for all x 6= 0.
2. Q is positive semidefinite if xT Qx 0 for all x.

5.5 Matrix Inversion Lemmas

33

3. Q is indefinite if xT Qx > for some x and xT Qx < 0 for other x.

Q is negative definite (or semidefinite) if Q is positive definite (or semidefinite).
If Q is symmetric then all its eigenvalues are real.
Example 28 Show that if A is symmetric, then A is positive-definite if and only if
all of its eigenvalues (which we know are real) are positive. This then allows us
to test for the sign-definiteness of a matrix A by looking at the eigenvalues of its
symmetric part.

5.5 Matrix Inversion Lemmas

There are various important relations involving inverses of matrices, one of which
is the Matrix Inversion Lemma
1
1
1
1
1
(A1 A2 A1
= A1
4 A3 )
1 + A1 A2 (A4 A3 A1 A2 ) A3 A1

(38)

Example 29 Prove the Matrix Inversion lemma

Solution: Let the LHS be A1 and the RHS be B. The proof proceeds by showing
that
AB = BA = I
so,
1
1
1
1
1
AB = (A1 A2 A1
4 A3 )(A1 + A A2 (A4 A3 A1 A2 ) A3 A1 )
1
1
1
1
= I A2 A1
4 A3 A1 + A2 (A4 A3 A1 A2 ) A3 A1
1
1
1
1
A2 A1
4 A3 A1 A2 (A4 A3 A1 A2 ) A3 A1

1
1
1
1
1
= I A2 A1
+ A3 A1
4 [I A4 (A4 A3 A1 A2 )
1 A2 (A4 A3 A1 A2 ) ]A3 A1

1
1
1
1
= I A2 A1
4 [I (A4 A3 A1 A2 )(A4 A3 A1 A2 ) ]A3 A1

= I

On the other hand,

1
1
1
1
1
BA = (A1
1 + A A2 (A4 A3 A1 A2 ) A3 A1 )(A1 A2 A4 A3 )
1
1
1
1
= I A1
1 A2 A4 A3 + A1 A2 (A4 A3 A1 A2 ) A3
1
1
1
1
A1
1 A2 (A4 A3 A1 A2 ) A3 A1 A2 A4 A3

1
1
1
1
1
1
= I A1
1 A2 [I (A4 A3 A1 A2 ) A4 + (A4 A3 A1 A2 ) A3 A1 A2 )]A4 A3

1
1
1
1
= I A1
1 A2 [I (A4 A3 A1 A2 ) (A4 A3 A1 A2 )]A4 A3

= I

6. MATLAB Commands

34

Now let us consider partition (30) for a Hermitian matrix A

A11 A12
A=
A12 A22

(39)

with A11 and A22 square matrices. The following Theorem gives a characterization
of a positive matrix in terms of its partition (39).
Theorem 3 Let A be a Hermitian matrix partitioned as in (39). Then
1. A is positive definite if and only if A11 and A22 A12 A1
11 A12 are positive
definite

2. A is positive definite if and only if A22 and A11 A12 A1

22 A12 are positive
definite

Proof 3

1. The proof follows from the facts that

A11 A12
=
A12 A22

I A1
11 A12
0
I

A11
0
I A1
11 A12
0 A22 A12 A1
0
I
11 A12

and that

I A1
11 A12
0
I

is a full rank matrix.

2. As in (1), writing A as

A11 A12
=
A12 A22

I A12 A1
A11 A12 A1
0
I A12 A1
22
22 A12
22
.
0
I
0
I
0
A22

6 MATLAB Commands
There are plenty of MATLAB commands dealing with matrices. In this section we
shall give a brief overview of those related with the topics covered in this chapter.
We recall that the (conjugate) transpose of a matrix A is evaluated simply typing A

6. MATLAB Commands

35

and that a polynomial is represented by the row vector of its coefficients ordered in
descending powers.
Q=orth(A) returns a matrix Q whose columns are an orthonormal basis for
the range of a (rectangular) matrix A.
Q=null(A) returns a matrix Q whose columns are an orthonormal basis for
the null space of a (rectangular) matrix A.
det(A) returns the determinant of a square matrix A.
inv(A) returns the inverse (if it exists) of a square matrix A. A warning is
given if the matrix is ill conditioned.
pinv(A) returns the pseudo-inverse of a matrix A.
trace(A) returns the trace of a square matrix A.
rank(A) returns the rank of a matrix A.
cond(A) evaluates the condition number (19) of a square matrix A, using the
matrix spectral norm. In this case (19) becomes

(A) =

(A)
.
(A)

The command condest(A) can be used to get an estimate of the 1-norm condition number.
[V,D]=eig(A) returns a diagonal matrix D whose entries are the eigenvalues
of A and a matrix V whose columns are the normalized eigenvectors of A, such that
A*V=V*D. In general V is singular. The command eigs(A) returns only some
of the eigenvalues, by default the six largest in magnitude.
poly(A) returns a row vector with (n + 1) elements, whose entries are the
coefficients of the characteristic polynomial (16) of A.
norm(A) returns the 2-norm for both vectors and matrices. Anyway there are
some differences in the two cases
If A is a vector, the p-norm defined in (17) can be evaluated typing norm(A,p),
where p can be either a real number or the string inf, to evaluate the
norm.
If A is a matrix, the argument p in the command norm(A,p) can be only
1,2,inf,fro, where the returned norms are respectively the 1, 2, or
Frobenius matrix norms defined in section 4.2.

REFERENCES

36

[U,T]=schur(A) returns an upper quasi-triangular matrix T and a Unitary

matrix U such that A = U*T*U. The matrix T presents the real eigenvalues of A
on the diagonal and the complex eigenvalues in 2-by-2 blocks on the diagonal. The
command [U,T] = rsf2csf(U,T) can be used to get an upper-triangular T
from an upper quasi-triangular T.
[U,S,V]=svd(A) returns a diagonal, in general rectangular, matrix S, whose
entries are the singular values of A, and two Unitary matrices U and V, stacking up
in columns the left and right singular vectors of A, such that A=U*S*V. The
command svds(A) returns only some of the singular values, by default the five
largest.

References
[1] F. R. Gantmacher. The Theory of Matrices. Chelsea Publishing Company, New
York, N.Y., 1959. Two volumes.
[2] R. A. Horn and C. R. Johnson. Matrix Analysis. Cambridge University Press,
1985.
[3] B. Noble and J. W. Daniel. Applied Linear Algebra. Prentice-Hall, Third
edition, 1988.
[4] G. Strang. Introduction to Linear Algebra. Wellesley-Cambridge Press, Cambridge, MA, 3rd edition, 2003.