Module 1 Mathematical Preliminaries

Objectives: To understand the basic denitions in set theory and their properties.

Lesson objectives
This module helps the reader in understanding an n-dimensional Euclidean space and the concept of length of a vector by way of norm. understanding the properties of open, closed, compact, dense and connected sets identifying a given set as an open, closed and connected sets. deducing how the sets are mapped using continuous functions

Suggested reading
Mathematical Analysis by Tom M. Apostol, Narosa Publishing House, 1993. Principles of Mathematical Analysis by Walter Rudin, McGraw-Hill International Editions, 1976.

Lecture-2
We begin with a formal denition of real spaces. Denition 2.0.1 Let n > 0 be an integer. An ordered set of n real numbers (x1 , x2 , . . . , xn ) is called an n-dimensional point or a vector with n components. The number xk is called the k th coordinate of the point x or k th component of the vector x. The set of all ndimensional points is called n-dimensional Euclidean space or simply n-space and denoted by I Rn . The real line is an example of a 1-dimensional space and the real plane is a 2-dimensional space. A typical vector in I R2 is denoted by [x1 x2 ] . The algebraic operations on I Rn are carried out using the following rules. 1. For x, y I Rn , x = y if and only if x1 x 2 . = . . xn 2. For x, y I Rn , x2 + y2 x+y = . . . xn + yn x1 + y1 .

y1 y2 . . . yn

NPTEL-Electrical Engineering - Nonlinear Control System

3. For I R and x I Rn ,

x 2 x = . . . . xn 4. 0 = [0 0 0 0] . 5. For x, y I Rn ,
n

x1

x.y =
i=1

xi yi .

6. The length of a vector is captured by the norm of a vector. The norm function ||( )|| : I Rn I R+ satises the following axioms: A1 ||x|| = || ||x|| (positive homogeneity). A2 ||x + y || ||x|| + ||y || (triangle inequality or subadditivity). A3 ||x|| = 0 x = 0 and ||x|| > 0 whenever x = 0 (positive deniteness ). There is class of p-norms dened by ||x||p = (|x1 |p + . . . + |xn |p )1/p , 1 p < and ||x|| = max |xi |.
i 2 2 ||x||2 =< x, x >= x2 1 + x2 + . . . xn

where, the operator < , >: I Rn I Rn I R+ is used to denote the dot product and || ||2 are two dierent p-norms, then there exists two constants c1 , c2 > 0 such that the following holds c1 ||x||1 ||x||2 c2 ||x||1 x I Rn . Lemma 2.0.2 (Cauchy-Schwartz inequality) Let x, y belong to I Rn . Then | < x, y > |
Joint Initiative of IITs and IISc Funded by MHRD

between two vectors in I Rn . All p-norms are equivalent in the sense that if || ||1

||x|| ||y || and | < x, y > | = ||x|| ||y || if and only if the elements x, y are linearly dependent. 5

NPTEL-Electrical Engineering - Nonlinear Control System

Instead of I Rn , we can consider a more general vector space X on which we can dene the notion of distance. This leads to the denition of normed linear space. Denition 2.0.3 A normed linear space is an ordered pair (X, || ||) where X is a linear A1. ||x|| 0 x X ; ||x|| = 0 if only if x = 0, the zero vector in X . A2. ||x|| = ||||x|| x X and I R or C. A3. ||x + y || ||x|| + ||y || x, y X .

vector space and || || : X I R is real-valued function on X such that the following

2.1

Matrix norms

Just as vector norms, one can dene matrix norm. To begin, we proceed with induced norms, dened as follows. Denition 2.1.1 Let || || be a given norm on I Rn . Then for each matrix A I Rnn , the quantity ||A||i dened by ||A||i = ||Ax|| ||Ax|| ||Ax|| = sup = sup x=0,xI Rn ||x|| ||x||=1 ||x|| ||x||1 ||x|| sup

is called the induced matrix norm of A corresponding to the vector norm || ||. The induced norm can be interpreted as the least upper bound of the ratio varies over I Rn . In other words, let C = {y I Rn : y = Ax, x B}. Then ||A||i || ||i : I Rn I Rn [0, ) satises the following axioms: 1. ||A||i 0 A I R n n . 2. ||A||i = || ||A||i A I R n n , I R. 3. ||A + B ||i ||A||i + ||B ||i A, B I R n n . Remark 2.1.2 Corresponding to every vector norm on I Rn there is a corresponding induced norm on I Rnn . The converse is not true as the following example shows. ||A||s = max |aij |.
i,j ||Ax|| ||x||

as x

is the smallest radius that completely covers the set C . The induced norm function

Joint Initiative of IITs and IISc Funded by MHRD

NPTEL-Electrical Engineering - Nonlinear Control System

We drop the subscript i for the induced norm, since we will use only induced norms computed as follows. Let A I Rmn . The one-norm , ||A||1 = max The two-norm , ||A||2 =
1j n

in the rest of the lectures. The norms for p = 1, 2, and the Frobenius norm can be
m

i=1

|aij |.

max [A A] = max [A], where max is the maximum eigen

value of A A 0 and max is the largest singular value of A.

The -norm , ||A|| = max

|aij |.
m i=1 n j =1

The Frobenius-norm , ||A||F = 1 0 0

a2 ij .

. Then ||A||1 = 6, ||A||2 = 4.4425, ||A|| = 5 Example 2.1.3 Consider A = 2 1 2 3 1 1 and ||A||F = 4.5826. Given any induced norm, and matrices A and B , the property ||AB ||p ||A||p ||B ||p ||Ax||p ||A||p ||x||p x.

called the submultiplicative property holds. For any submultiplicative norm

Joint Initiative of IITs and IISc Funded by MHRD

Lecture-3
In this lecture, we will discuss the notion of open set, closed set and compact set. We need the notion of open-ball and interior point to begin with. Let a I Rn and r be a open n-ball of radius r and centered at a and denoted by B (a, r ). The ball B (a, r ) is also termed as the r -neighbourhood of a. S if there is an open n-ball of some radius r > 0 with center a, all of whose points belong We can now dene an open set as follows. Denition 3.0.4 A set S I Rn is called open if all its points are interior points (S = int S ). In I R, the simplest type of open sets are the open intervals of the form (a, b) = {x I R:

given positive number. The set of all points x I Rn such that ||x a|| < r is called an

Let S be a subset of I Rn , and assume that a S . Then a is called an interior point of

to S . The set of all interior points of S is called the interior of S and is denoted by int S .

a < x < b}. Similarly, the open set in I R2 is the open disk {(x1 , x2 ) I R2 : x2 + y 2 < r }. A collection of open sets have the property that arbitrary union of open sets is open intersection of open sets need not be open.
1 1 Example 3.0.5 S = (( n , n ), n = 1, 2, 3, . . .) = {0}, which is not open.

and nite intersection of open sets is open. The following example shows that arbitrary

Denition 3.0.6 A set S in I Rn is said to be closed if its complement I Rn \ S is open. that are common to A and B . A collection of closed sets have the property that a union closed sets is closed. Closed set can be dened in terms of adherent points accumulation point/s, which are dened as follows. 8 The notation A \ B is read as A minus B , consists of all points of A but excluding those

of nite collection of closed sets is closed, and the intersection of an arbitrary collection of

NPTEL-Electrical Engineering - Nonlinear Control System

Denition 3.0.7 Let S be a subset of I Rn and x I Rn , not necessarily in S . Then x is of S . . The set of all adherent points of a set S is called the closure of S and is denoted by S Some points adhere to S because every ball contains points of S distinct from x and this leads to the notion of accumulation point.

said to be adherent to S if for every r > 0 the n-ball B (x, r ) contains at least one point

Denition 3.0.8 If A is a subset of I Rn and if x I Rn , then x is said to be a limit point or accumulation point of A if every neighbourhood of x contains at least one point of A distinct from x. We say x is an accumulation point of S if x adheres to S \ x. Closed set can be

alternatively dened as

A set is closed if and only if it contains all its adherent points. . A set is closed if and only if S = S A set A I Rn is closed if and only if, it contains all its accumulation points. The following Lemma brings out the relation between the closure of a set and the set of its accumulation points. Lemma 3.0.9 Let A denote the set of all accumulation points of a set A in I Rn , and A = A A . the set of all adherent points of A. Then, A Proof: Now if x A , then every neighbourhood of x intersects A (in a point dierent from x). and thus x A A . Therefore x A x / A, then every neighbourhood B (x) contains atleast one element of A, which implies

. since A A . Let x A A . Then x A or x A . If x A, then x A

. Then if x A, then it immediately follows that x A A . If Conversely, if x A 2

B (x) A = = x A . Therefore x A A .

2 b}, while in I R2 they are represented by closed disk {(x1 , x2 ) : I R2 : x2 1 + x2 r }.

On the real line, closed sets are denoted by closed intervals [a, b] = {x R : a x

Example 3.0.10 Let A = (0, 1] I R. Then 0 is the limit point of A and so is every point of [0, 1] a limit point of A. Example 3.0.11 Consider the set A = {1/n : n Z + }, where Z + is the set of positive integers. The only limit point of A is 0.
Joint Initiative of IITs and IISc Funded by MHRD

NPTEL-Electrical Engineering - Nonlinear Control System

Some sets are neither open nor closed, such as [0, 1), which is called half-open or half-closed interval. We next dene a bounded set. Denition 3.0.12 A set S in I Rn is said to be bounded if it lies entirely within an n-ball B (a, r ) for some r > 0 and some a I Rn .

Joint Initiative of IITs and IISc Funded by MHRD

10

Lecture-4
4.1 Supremum and inmum of a set

x S , then b is called an upper bound for S and we say that S is bounded above by b. the upper bounds.

Let S be a set of real numbers. If there is a real number b such that x b for every

Certainly any c b is an upper bound for S , but we are interested in the least among Denition 4.1.1 Let S be a set of real numbers bounded above. A real number b is called a least upper bound for S , denoted by b = sup(S ), if it has the following properties: b is an upper bound for S . No number less than b is an upper bound for S . Naturally, we can dene the greatest lower bound, called the inmum of a set. Denition 4.1.2 Let S be a set of real numbers bounded below. A real number c is called the greatest lower bound for S , denoted by c = inf(S ), if it has the following properties: c is an lower bound for S . No number greater than c is a lower bound for S . Example 4.1.3 Consider S = (1, 3), then sup(S ) = 3 and inf(S ) = 1. Note that the sup and inf of a set need not belong to the set. When they do belong, they are the maximum and minimum elements of the set. Example 4.1.4 Let S = {x : 3x2 10x + 3 < 0}. Now x S satises (x 3)(3x 1) < 0 which implies
1 3

< x < 3. Hence sup(S ) = 3 and inf(S ) = 1/3. 11

NPTEL-Electrical Engineering - Nonlinear Control System

The notion of closed and bounded sets leads to the denition of compact set. We rst need the denition of covering of a set. A collection F of sets is said to be covering of a given set S if S = AF A. The collection F is also said to cover S . If F is a collection of open sets, then F is called an open covering of S . Denition 4.1.5 A set S in I Rn is said to be compact if and only if every open covering of S contains a nite subcover that is, a nite subcollection which also covers S . The Heine-Borel theorem states that every closed and bounded set in I Rn is compact. a boundary point of S I Rn if every n-ball B (x) contains at least one point of S and at We end this lecture with the denition of boundary of a set. A point x I Rn is called

least one point of I Rn \ S . The set of all boundary points of S is called the boundary of S I Rn \ S from which it follows that S is a closed set in I Rn . and is denoted by S = S The notion of dense and connected sets are introduced here. =I Denition 4.1.6 Let S I Rn . Then S is said to be dense in I Rn if S Rn . This is the case if and only if S B (x, r ) = for every x I Rn and r > 0.

Example 4.1.7 The set of rational numbers Q are dense in I R and so are the set of irrationals. We next dene an important concept of connectedness, for which we rst need to say when are two sets are separated. Two subsets A and B of I Rn are said to be separated if B and A B are empty. In other words, no point of A lies in the closure of B both A

and no point of B lies in the closure of A.

Example 4.1.8 Let A = [0, 1] and B = (1, 2). The sets A and B are disjoint but not separated. Example 4.1.9 Let A = (0, 1) and B = (2, 3). The sets A and B are disjoint and separated. Example 4.1.10 Let A = [0, 1] and B = [1, 2). The sets A and B are not disjoint and not separated. We can now dene a connected set. Denition 4.1.11 A set S I Rn is said to be connected if S is not a union of two

non-empty separated sets.

Example 4.1.12 Let S = [0, 1] (1, 2). Since S is a union of two non-empty sets that are not separated, it is connected.

Joint Initiative of IITs and IISc Funded by MHRD

12

Lecture-5
A few solved examples and exercises are presented here. P1. Find the limit point(s) of the set C = {0} (1, 2) I R. Solution : Every point of the interval [1, 2] is a limit point of C. P2. Determine if the following sets in I R are open or closed (or neither). a) All rational numbers Q. b) All numbers of the form
(1)n , (1+(1/n))

n=1,2,. . .

P3. If S and T are subsets of I Rn , prove that (int S ) (int T ) = int (S T ), where int refers to the interior of the set. Solution : We have to show the following inclusions int(S ) int(T ) int(S T ) int(S T ) int(S ) int(T ) to prove the equality. Let x int(S ) int(T ) = x int(S ) and x int(T ). By the denition of interior point, there exists open n-balls BS (x), BT (x) such that S T (recall that nite intersection of open sets is open). Hence x int(S T ). x BS (x) S and x BT (x) T = x Bs BT . But BS BT is open in

B (x) is open in both S and T , which further implies that x int(S ) and x int(T ). Hence x int(S ) int(T ). P4. Give an example of a real-valued continuous function that maps an open set to a closed set. 13

Conversely, let x int(S T ). Then there exists an open n-ball B (x) S T . But

NPTEL-Electrical Engineering - Nonlinear Control System

Solution : Consider the function f : (0, 20) I R dened by f (x) = log(x) + log(20 x). The domain of f , denoted by Df is (0, 20), an open set. The range of f , denoted However, I R \ Rf = (2, +), an open set implying that Rf is a closed set. together with the norm dened as ||x|| = max |x(t)|.
t [a,b]

by Rf , is (, 2] which is not an open set as 2 is not an interior point of (, 2]. P5. The linear space C [a, b] consists of continuous function on the real interval [a, b]

Verify all the axioms of a norm hold for the proposed norm. Solution It is clear that ||x|| 0 and is equal to zero only when the function is identically triangle inequality follows from

equal to zero. Next ||x|| = max |x(t)| = || max |x(t)| = || ||x||. Finally, the ||x + y || = max |x(t) + y (t)| max(|x(t)| + |y (t)|) = max |x(t)| + max |y (t)| = ||x|| + ||y ||.

P6. Prove the following identity ||x + y ||2 + ||x y ||2 = 2||x||2 + 2||y ||2. Solution : Expanding the LOS in terms of inner product, we have ||x + y ||2 + ||x y ||2 = < x + y, x + y > + < x y, x y > = < x, x + y > + < y, x + y > + < x, x y > < y, x y > = < x, x > + < x, y > + < y, x > + < y, y > < x, y > = 2||x||2 + 2||y ||2 + < x, x > + < x, y > < y, x > + < y, y >

This equality is known as the Parallelogram law. P7. Prove that A = (I Rn A). P8. Find the sup and inf of the following set of real numbers. S = {x : (x a)(x b)(x c)(x d) < 0}, where a < b < c < d. P9. Consider the set S = {x I R : x = 2n + 5m , n, m = 1, 2, . . .}. Show that S is neither open nor closed.
Joint Initiative of IITs and IISc Funded by MHRD

14

NPTEL-Electrical Engineering - Nonlinear Control System

Solution : The maximum element of S is achieved when m = n = 1 and equals x = 0.7 S , but a neighbourhood of the form (0.7 , 0.7 + ) does not belong to S since (0.7 + ) is not an element of S . Hence, S is not open. To prove that it is not closed, let m = 2, n = 1 which yields x = 0.54 S . Fix x = 0.54 + , then it is easy to see

that 0.54 + / S since there exists no m and n such that 2n + 5m = 0.54 + . Therefore 0.54 + I R \ S . Now it can be observed that for any > , the interval element 0.54 which is a member of S and hence does not belong to I R \ S implying (0.54 + , 0.54 + + ) is not contained in I R \ S because this interval includes the (0.54 + , 0.54 + + ) S \ 0.54 = .

Joint Initiative of IITs and IISc Funded by MHRD

15