2.

Mathematical review

Nonlinear Control Systems

Ant
onio Pedro Aguiar
pedro@isr.ist.utl.pt
2. Mathematical review

IST-DEEC PhD Course

http://users.isr.ist.utl.pt/%7Epedro/NCS2012/

2011/2012

2. Mathematical review

Mathematical review (Euclidean Space)

Vector and matrix norms
The norm kxk of a vector x Rn is a real valued function with the properties
1. kxk, x Rn , with kxk = 0 if and only if x = 0,
2. kx + yk kxk + kyk, x, y Rn (Triangular inequality)
3. kxk = ||kxk, R and x Rn .
p-norm
1/p
,

kxkp = (|x1 |p + |x2 |p + . . . + |xn |p )

1p<

-norm
kxk = max |xi |
i

The three most commonly used norms are kxk1 , kxk and the Euclidean norm

1/2
kxk2 = |x1 |2 + |x2 |2 + . . . + |xn |2

2. Mathematical review

Mathematical review (Euclidean Space)

Properties

For k.k ,k.k , with 6= there exist positive constants c1 , c2 such that

c1 kxk kxk c2 kxk , x Rn . In particular

kxk2 kxk1

kxk kxk2

n kxk2

n kxk

kxk kxk1 n kxk

H
older inequality

T
x y kxkp kykq ,

1
1
+ = 1, x, y Rn
p
q

2. Mathematical review

Mathematical review (Euclidean Space)

Matrix norms The induced p-norm of A Rmn
kAkp = sup

x6=0

kAxkp
kxkp

sup kAxkp

kxkp =1

In particular
p=1

kAk1 = max
j

m
X

|aij |

i=1

(The maximum by columns)

p=2

1/
2
kAk2 = max AT A
where max (AT A) is the maximum eigenvalue of AT A.
p=

kAk = max
i

n
X

|aij |

j=1

(The maximum by rows)

4

2. Mathematical review

Mathematical review (Euclidean Space)

Some useful properties. Let A Rmn and B Rnl

1.

2.

1
kAk kAk2 mkAk
n

1
kAk1 kAk2 nkAk1
m

3.
kAk2

kAk1 kAk

4.
kABkp kAkp kBkp

2. Mathematical review

Topological concepts in Rn
Convergence of sequences
A sequence of vectors x0 , x1 , ..., xk , ... Rn denoted by {xk }, is said to converge
to a limit vector x if
kxk xk 0
as k , which is equivalent to
>0 N : kxk xk < , kN
A vector x is an accumulation point of a sequence {xk } if there is a subsequence

of {xk } that converges to x. That is, if there is an infinite subset K of the

nonnegative integers such that {xk }kK converges to x.
A bounded sequence {xk } Rn has at least one accumulation point in Rn .
Increasing sequence: {xk } Rn , xk xk+1 , k. If {xk } Rn , xk < xk+1 , k

it is said to be strictly increasing.

Decreasing sequence {xk } Rn , xk xk+1 , k. If {xk } Rn , xk > xk+1 , k

it is said to be strictly decreasing.

An increasing sequence {xk } Rn that is bounded from above converges to a

real number.
Similarly, a decreasing sequence that is bounded from below converges to a real

number.
6

2. Mathematical review

Topological concepts in Rn
Sets Let S Rn
Open set: x S, one can find an -neighborhood of x

N (x, ) = {z Rn : kz xk < }
such that N (x, ) S
Closed set: Every convergent sequence {xk } with elements in S converges to a

point in S.
Bounded set: S is bounded if there is r > 0 such that kxk < r, x S
Compact set: If it is closed and bounded.
Boundary point: A point p is a boundary point of a set S if every neighborhood

of p contains at least one point of S and one point not belonging to S.

The set of all boundary points of S is denoted by S, the interior of a set by S S
and the closure of a set S = S S.

2. Mathematical review

Mathematical review (Euclidean Space)

Continuous functions
A function f : Rn Rm is said to be continuous at a point x if f (xk ) f (x)
whenever xk x. Equivalently, if
>0 >0 : kx yk < kf (x) f (y)k <
Differentiable functions
A function f : R R is said to be differentiable at x if the limit
f 0 (x) = lim

h0

f (x + h) f (x)
h

exists. A function f : Rn Rm is said to be continuously differentiable (denoted as

fi
f C 1 ) at a point x0 if the partial derivatives x
exist and are continuous at x0 for
j
1 i m, 1 j n.

2. Mathematical review

Mathematical review (Euclidean Space)

Gradient vector
For a continuously differentiable function f : Rn R
f (x) =

f
x1

f
x2

f
xn

iT

Jacobian Matrix
For a continuous differentiable function f : Rn Rm
2

J (x) =

f
x

6
=6
4

f1
x1

f1
xn

..
.

..

..
.

7
7
5

fm
x1

fm
xn

Chain rule
Let h(x) = g(f (x)) be a continuously differentiable function at x0 . Then the Jacobian
matrix is given by the chain rule

h
g
f
=

x x=x0
f f =f (x0 ) x x=x0
9

2. Mathematical review

Mathematical review (Euclidean Space)

Mean value theorem
Let f : Rn R be a continuously differentiable function at each point x of an open
set S Rn . Let x and y be two points of S such that the line segment
L(x, y) = {z : z = x + (1 )y, 0 1} belongs to S. Then exists a point z of
L(x, y) such that

f
f (y) f (x) =
(y x)
x
x=z

Implicit Function Theorem

Let f : Rn Rm be continuously differentiable at each point (x, y) of an open set
S Rn Rm . Let (x0 , y0 ) S be a point such that f (x0 , y0 ) = 0 and for which the
(x0 , y0 ) is nonsingular. Then there exist neighborhoods U Rn
Jacobian matrix f
x
of x0 and V Rm of y0 such that for each y V the equation f (x, y) = 0 has a
unique solution x U . Moreover, this solution can be given as x = g(y), where g is
continuously differentiable at y = y0 .

10

2. Mathematical review

Mathematical review
Grownwall-Bellman Inequality
Let : [a, b] R be continuous and : [a, b] R be continuous and nonnegative. If
a continuous function y : [a, b] R satisfies
Zt
y (t) (t) +

atb

(s) y (s) ds,

a

then
Rt

Zt
y (t) (t) +

(s) (s) e

( )d

ds

In particular, if (t) = is a constant, then

Rt

( )d

y (t) ea
If in addition, (t) = > 0 is a constant, then
y (t) e(ta)
11