Lecture Notes: Principles of Mathematical Analysis van abel

A BSTRACT. This is a lecture notes on REAL ANALYSIS, we mainly refers the book of Rudin, Walter PRINCIPLES OF MATHEMATICAL ANALYSIS.

Contents
Chapter 1. Some Acknowledgments 5 7 7 7 8 8 11 13 13 14 15 15 17 17 17 18 18 18 18 23 25

Chapter 2. Real Number System 1. Ordered Sets 2. Fields 3. Ordered Fields 4. The Real Field 5. The Complex Field Chapter 3. Basic Topology 1. Metric Spaces 2. Compactness 3. Perfect Sets 4. Connected Sets Chapter 4. Numerical Sequences and Series 1. Convergent Sequences 2. Subsequences 3. Cauchy Sequences 4. Bounded Sequence and Monotonic Sequences 5. Upper and Lower Limits 6. Series Index Bibliography

CHAPTER 1

Some Acknowledgments

CHAPTER 2

Real Number System

The real number system is an ordered eld with the least-upper-bound property. First let us consider the concept of order, which will make the least upper bound concept meaningful. 1. Ordered Sets Denition 1.1. Let be a set. An order on is a relation, denoted by <, with the following two properties: (1) If and , then one and only one of the statements < , = , < is true. (2) If , , , if < and < , then < . We usually call is an ordered set in this case. The advantage of ordered set is that in which we can compare two elements with the help of the order. Especially we have the denition of bounded . Denition 1.2. Suppose is an ordered set, and . If there exists a such that for every , we say that is bounded above , and call an upper bound of . The most core concept is the lest upper bound of set E. Denition 1.3. Suppose is an ordered set, , and is bounded above. Suppose there exists an with the following properties: (1) is an upper bound of . (2) If < then is not an upper bound of . then is called the least upper bound of or the supremum of , and we write = sup . Denition 1.4. An ordered set is said to have least-upper-bound-property if the following is true: If , is not empty, and is bounded above, then sup exists in . 2. Fields An other aspect of real number system is that it is a eld. Denition 2.1. A eld is a set with two operations, called addition and multiplication , which satisfy the following so-called eld axioms (A), (M) and (D): (A) Axioms for addition (A1) If and , then their sum + is in . (A2) Addition is commutative: + = + for all , . (A3) Addition is associative: ( + ) + = + ( + ) for all , , . (A4) contains an element 0 such that 0 + = for every .
7

2. REAL NUMBER SYSTEM

(A5) To every corresponds an element such that + () = 0. (M) Axioms for multiplication (M1) If and , then their product is in F. (M2) Multiplication is commutative: = for all , . (M3) Multiplication is associative: () = () for all , , . (M4) contains an element 1 = 0 such that 1 = for every . (M5) If and = 0 then their exists an element 1/ such that (1/) = 1. (D) The distributive law ( + ) = + hold for all , , . Remark 2.2. The most basic property of the addition and multiplication of a eld is the cancelation law: (1) if + = + then = . Which implies + = = 0, + = 0 = , and () = . (2) If = 0 and = then = . Which implies analogous properties with addition. With the distribution law we have the following statements, for any , , : (1) 0 = 0. (2) If = 0 and = 0 then = 0. (3) () = () = (). (4) ()() = . 3. Ordered Fields Denition 3.1. An ordered eld is a eld which is also an ordered set, such that (1) + < + if , , and < ; (2) > 0 if , > 0 and > 0. If > 0, we call positive ; if < 0, is negative . Proposition 3.2. The following statements are true in every ordered eld. (1) If > 0 then < 0 and vice versa. (2) If > 0 and < then < . (3) If < 0 and < then < . (4) If = 0 then 2 > 0. In particular, 1 > 0. (5) If 0 < < then 0 < 1/ < 1/. 4. The Real Field The core theorem of this chapter is the existence of real eld. Theorem 4.1. There exists an ordered eld which has the least-upper-bound property. Moreover, contains as an subeld. Remark 4.2. Generally, there are three methods to construct the real number systems. In the book of Landau[Lan51] and Thurston[Thu56], they entirely devoted to number systems. Chapter 1 of Knopps book[Kno28] contains a more leisurely description of how can be

4. THE REAL FIELD

obtained form . Another construction, in which each real number is dened to be an equivalence class of Cauchy sequences of rational numbers, is carried out in Sec. 5 of the book by Hewitt and stromberg[HS65]. The cuts in which Rudins book used were invented by Dedekin. The construction of form by means of Cauchy sequences is due to Cantor. Both Cantor and Dedekind published their constructions in 1872. P ROOF. We will use the cuts method of Dedekin to constructing form . The members of will be certain subsets of , called cuts. setp1. [Construction of cuts] Denition 4.3. A cut is any set with the following three properties (1) is not empty and = . (2) If , and < , then . (3) If , then < for some . Note that (3) simply says that has no largest member, and (2) implies two facts which will be used freely: (1) If and then < . / (2) If and < then . / / setp2. [Dene an order < of Cuts] Dene < to mean: is an proper subset of , ie., and = . Check that this truely denes an order on the cuts. setp3. [ has the least-upper-bound property] Let be a nonempty subset of , and assume that is an upper bound of . Dene to be the union of all . In other words, if and only if for some . We shall prove that and that = sup . Since is not empty, there exists an 0 . This 0 is not empty. Since 0 , is not empty. Next, , Since for every , therefore = . Thus satises property (1) in the denition of cuts. To prove (2) and (3), pick . Then 1 for some 1 . If < , then 1 , hence ; this proves (2). If 1 is so chosen that > , we see that , since 1 , and therefore satises (3). Thus . It is clear that for every . Suppose < . Then there is an and that . Since , for some / . Hence < , and is not an upper bound of . This gives the desired result: = sup . setp4. [Denition of addition] If and we dene + to be the set of all sums + , where and . We dene 0* to be the set of all negative rational numbers. The 0* will play the role of 0. For a xed . Let be the set of all with the following property: There exists > 0 such that . / The will play the role of . Check that the axiom of addition satised. setp5. [Cancelation law] The cancelation law of addition satised, especially we have If , , and < , then + < + . The rst requirement of ordered eld hold. It also follows that > 0* if and only if < 0* . setp6. [Denition of multiplication] We rst conne ourself to + , ie., the set of all with > 0* .

10

2. REAL NUMBER SYSTEM

If , + , we dene to be the set of all such that for some choice of , , , > 0. We dene 1* to be the set of all < 1. Then the axioms of multiplication and distribution law of a eld(+ ) hold. Note that the second requirement of ordered eld holds: If > 0* and > 0* then > 0* . We complete the denition of multiplication by setting 0* = 0* = 0* , and by setting ()() if < 0* , < 0* , [()] if < 0* , > 0* , = [ ()] if > 0* , > 0* . Check that the axioms of multiplication and distribution law of satised. Till now, we have completed the proof that is an ordered eld with the least-upperbound property. setp7. [Contains as a subeld] We associate with each the set * which consists of all such that < . It is clear that * , and satisfy the following relations: (1) * + * = ( + )* , (2) * * = ()* , (3) * < * if and only if < . The above properties says that the map which corresponding each with the cuts * preserves sums, products and order. In other words, The ordered eld is isomorphic to the ordered eld * whose elements are the rational cuts. It is this identication of with * which allows us to regard as a subeld of . This complete the whole theorem. Remark 4.4. Form to , there is an identication, and note that the same phenomenon occurs when the real numbers are regarded as a subeld of the complex eld, and it also occurs at a much more elementary level, when the integers are identied with a certain subset of . In fact, any two ordered elds with the least-upper-bound property are isomorphic. The rst part of Theorem 4.1 therefore characterizes the real eld completely. As a good illustration of the least-upper-bound property we give the following important properties of real number system. Proposition 4.5 (Archimedean property ). If , , and > 0, then there is a positive integer such that > . P ROOF. Consider the set {} and use the least-upper-bound property to derive a =1 contradiction. Proposition 4.6 (Density of in ). If , and < , then there exists a such that < < . P ROOF. Apply the Archimedean property, we have ( ) > 1 for some positive integer . Suppose that there is a rational number / satisfy < / < , then < < , and 1 + < .

5. THE COMPLEX FIELD

11

Thus if we can choose an integer , such that < 1 + < , then we have down. But this is equal to the existence of an integer which satisfy 1 < . Which can be derived from the Archimedean property. 5. The Complex Field Denition 5.1. A complex number is an ordered pair (, ) of real numbers with the following addition and multiplication (, ) + (, ) = ( + , + ),
def def

(, )(, ) = ( , + ).

def

We always write (0, 0) = 0, (1, 0) = 1, (0, 1) = and (, 0) = , Where the has the property 2 = 1 and (, ) = + = (, 0) + (, 0)(0, 1).

CHAPTER 3

Basic Topology
There are some basic properties of real number system, rst it is an uncountable set, while is a countable set. Also it is a topology space, whats more it is a metric space. All these properties will be concerned in this chapter. 1. Metric Spaces Denition 1.1. A set , whose elements we shall call points , is said to be a metric space if with any two points and of there is associated a real number d(, ), called the distance from to , such that (1) d(, ) > 0 if = ; d(, ) = 0; (2) d(, ) = d(, ); (3) d(, ) d(, ) + d(, ), for any . Any function with these properties is called a distance function , or a metric . Example. The most basic and important example is the Euclidean spaces , especially 1 (the real line) and 2 (the complex plane); the distance in is dened by d(, ) = | | = ( )2 , , .
=1

Other examples are the spaces of continuous function C () and square-integrable function spaces L 2 (). Remark 1.2. It is important to observe that every subset of a metric space is a metric space in its own right, with the same distance function. Denition 1.3. we always call (, ) segment and [, ] an interval . If and > 0, then the open (or closed) ball with center at and radius is dened to be the set of all such that | | < (or | | ). We call a set convex if for any , , and any 0 1, we have + (1 ) . Clearly, all balls are convex. Denition 1.4. Let be a metric space. All points and sets mentioned below are understood to be elements and subsets of . (1) A neighborhood of is a ball () centered at and with radius . (2) A point is a limit point of the set if every neighborhood of contains a point of , which is distinct from . All limit points of will be denoted . The closure of is the set = . (3) is closed if every limit point is still in . (4) is open if every point of has a neighborhood , which is contained in . (5) Suppose and . We say that is open related to , if each point to , there is associated a ball () such that whenever () , we have .
13

14

3. BASIC TOPOLOGY

(6) is perfect if is closed and if every point of is a limit point of . (7) is dense in if every point of is a limit point of , or a point of (or both). The following propositions are almost clearly to see. Proposition 1.5. Suppose is a subset of a metric space . (1) every neighborhood of a limit point of contains innitely many points of . (2) A set is open if and only if its complement is closed. (3) open and close sets both closed under nite intersection and union, whereas open sets are closed under any union, and closed sets are closed under any intersection. (4) is a closed set, and it is the smallest closed set that contains . (5) Suppose . A subset of is open relative to if and only if = for some open subset of . The interesting connection with real numbers will be the following Theorem 1.6. Let be a nonempty set of real numbers which is bounded above. Let = sup . Then . Hence if is closed.

2. Compactness Denition 2.1. By an open cover of set in a metric space we mean a collection { } of a open subsets of such that . The basic properties of compactness are stated in the following theorem. Theorem 2.2. Suppose is a metric space, . (1) a subset is compact relative to if and only if it is compact in . (2) Compact subsets of metric spaces are closed. Since metric spaces are Hausdorff space. (3) Closed subsets of compact sets are compact. The most important property of compact sets is that it has the so-called nite intersection property , which stated as a theorem in the following: Theorem 2.3. If { } is a collection of compact subsets of a metric space with the nite intersection property, ie., the intersection of every nite sub-collection of { } is non-empty, then is nonempty too. The most famous case is the intersection of a sequence of non-empty decreasing compact sets will be non-empty. Example. The -cell or -cube is compact. It can be showed with a continuous half-divide of the -cell to derive a contradiction. In , as a special case of metric space. There are some equivalent statement of compactness. Theorem 2.4. If a set in has one of the following three properties, then it has the other two: (1) is closed and bounded. (2) is compact. (3) Every innite subset of has a limit point in E. Corollary 2.5 (Weierstrass). Every bounded innite subset of has limit point in .

4. CONNECTED SETS

15

3. Perfect Sets One fact about perfect set is that any non-empty perfect set in will be uncountable. In fact, if is a countable perfect subset of . Without lost of generality, let = { } . =1 We can construct a sequence of neighborhoods { } as follows. Let 1 be any neighborhood of 1 . Suppose that has been constructed, so that is not empty. It must contains innite point of , since every point of is a limit point of . Thus there is a neighborhood +1 such that +1 , +1 and +1 = . / Put = . Since is closed and bounded, is compact. Clearly, is =1 empty (note that ). But { } has the nite intersection property, this is a contradiction.
1 Example (Cantor set). Let 0 be the interval [0, 1]. Remove the segment ( 3 , 2 ), and let 1 3 denote the union of intervals 2 1 [0, ], [ ]. 3 3 Remove the middle of thirds of these intervals, and 2 be the union of intervals 1 2 3 6 7 8 [0, ], [ , ], [ , ], [ , 1]. 9 9 9 9 9 9 Continuing in this way, we obtain a sequence of compact sets , clearly it has the following properties: (1) 1 2 ; (2) is the union of 2 intervals, each of length 3 ; (3) the removed intervals are ( ) 3 + 1 3 + 2 , , , + . 3 3 The set = , =1

is called the Cantor sets . It can be shown that is a non-empty perfect closed compact set. One of the most interesting properties of the Cantor set is that it provides us with an example of an uncountable set of measure zero. 4. Connected Sets Denition 4.1. Two subsets and of a metric space are side to be separated if both and are empty. A set is said to be connected if is not a union of two nonempty separated sets. The connected subset of the real line have a particularly simple structure. Theorem 4.2. A subset of the real line 1 is connected if and only if it has the following property: If , , and < < , then .

CHAPTER 4

Numerical Sequences and Series

In this chapter, we always assume our sequence is complex, except clearly stated. 1. Convergent Sequences Denition 1.1. A sequence { } in a metric space is said to converge if there is a point such that for every > 0, there is an integer , such that implies that d( , ) < . In this case, we also say that { } is converges to , and is the limit of { }, write

lim = .

If { } is not converge, it is said to diverge . The well known properties about convergent sequences is stated in the following proposition. Proposition 1.2. Let { } be a sequence in a metric space . (1) { } converges to if and only if every neighborhood of contains for all but nitely many . (2) The limit of a convergent sequence is unique. (3) a convergent sequence is bounded. (4) If and if is a limit point of , then there is a sequence { } in such that = lim . (5) The limit operation is commutable with the four arithmetic operation , ie., addition, minus, product and division. (6) If = (1 , 2 , , ) , then { } = (1 , 2 , , ) converges to if and only if lim = for every = 1, 2, , .

2. Subsequences Denition 2.1. Given a sequence { }, consider a sequence of positive integers, such that 1 < 2 < . Then the sequence { } is called a subsequence of { }. It is clear that { } is converges to if and only if every subsequence of { } converges to . A other fact about subsequence is the following theorem. Theorem 2.2. The subsequential limits of a sequence { } in a metric space form a closed subset of . P ROOF. Let * be the set of all subsequential limits of { } and let be a limit point of * . We need to show that * . Choose 1 so that 1 = . (if no such 1 exists, then * has only one point, and there is nothing to prove.) Put = d(, 1 ). Suppose 1 , 2 , , 1 are chosen. Since is a limit point of * , there is an * with d(, ) < 2 . Since * , there is an 1 > 1 such that d(, ) < 2 . Thus d(, ) < 21 ,
17

18

4. NUMERICAL SEQUENCES AND SERIES

for = 1, 2, . This says that { } converges to . Hence * . 3. Cauchy Sequences Denition 3.1. A sequence { } in a metric space is said to be a cauchy sequence if for every > 0 there is an integer such that d( , ) < for any , . Denition 3.2. A metric space in which every Cauchy sequence converges is said to be complete . Proposition 3.3. Let be a metric space. (1) Every convergent sequence is a Cauchy sequence. (2) If is compact, then every Cauchy sequence converge to some point of . (3) is a complete metric space. 4. Bounded Sequence and Monotonic Sequences We already know that every convergent sequence is bounded, but the converse is not true in general. However, if we require that the sequence is monotonic, decrease or increase, the converse will be true. Theorem 4.1. Suppose { } is monotonic. Then { } converges if and only if it is bounded. 5. Upper and Lower Limits Denition 5.1. Let { } be a sequence of real number. Let be the set of numbers (include + and ) such that for some subsequence { }. Dene * = sup , and * = inf . They are called the upper and lower limits of { }, respectively. We also use the notation lim sup = * ,

lim inf = * .

Theorem 5.2. Let { } be a sequence of real numbers. Let and * have the same denition as above. Then * has the following two properties: (1) * . (2) If > * , then there is an integer > 0 such that implies < . Move over, these two properties also dene the upper limits of { }. Analogous result is true for * . Its quite clearly to see that the lim sup and lim inf operation preserve the inequality. 6. Series We need not to constrain ourselves to the real-valued series, if it do, we will explicitly state. Denition 6.1. Given a sequence { }, we use the notation
=

( )

to denote the sum + + 1 + + . With { } we associate a sequence { }, where = For { } we also use the symbolic expression 1 + 2 + 3 +
=1

6. SERIES

19

or, more concisely

=1

Called innite series . The numbers are called partial sums of the series. If { } converges to , we say that the series converges , and write
=1

= .

The number is called the sum of the series. If { } diverges, the series { } is said to diverge . Remark 6.2. It is quite clear that every theorem about sequences can be stated in terms of series, since = 1 , 2, 1 = 1 , and vice versa. Especially the Cauchy criterion for convergent sequence can be restated in the following form: Theorem 6.3. converges if and only if for every > 0 there is an integer such that
=

for every . Particularly, the the term of convergent series must tend to 0 as tends to . Corollary 6.4. If { } converges, then lim = 0. As a consequence of the bounded monotonic theorem (cf. theorem 4.1), we have that Theorem 6.5. A series of nonnegative terms 1converges if and only if its partial sums form a bounded sequence. 6.1. Test Method of Series. To test a series is convergent or divergent, we have the so called comparison test . Theorem 6.6. Suppose { }, { }, { } are series. (1) | | for 0 , where 0 is some xed integer, and if If converges, then converges. (2) If for 0 , and if diverges, then diverges. The comparison test is a very useful one; The most used non-negative series is the geometry series . Theorem 6.7. The series { }, = converges for [0, 1), and diverges other wise. Problem. Suppose 1 2 3 0. Then the series converges if and only if =1 the series 2 2
=0

converges. 1 converges if > 1 and diverges if 1. Use the above result to conclude that Take a thought at the converges or diverges of the following series
=1

, (log )

=3

1 log log log

=3

1 log (log log )2

, .

1nonnegative terms means that every term

> 0.

20

4. NUMERICAL SEQUENCES AND SERIES

There are two derived ways in application of the comparison test, ie., the root test and the ratio test . Theorem 6.8 (Root Test). Given , put = lim sup | |, then (1) if < 1, converges; (2) if > 1, diverges; (3) if = 1, the test gives no information. Theorem 6.9 (Ratio Test). The series , +1 (1) converges if lim sup < 1, +1 (2) diverges if lim sup 1 for all 0 , where 0 is some xed integer. Remark 6.10. The ratio test is frequently easier to apply than the root test, however, the root test has wider scope, since we have Theorem 6.11. For any sequence { } of positive number, +1 +1 lim inf lim inf lim sup lim sup . Neither of the two test is subtle with regard to divergence, and they are both for absolute convergence. 6.2. Power Series. Denition 6.12. Given a sequence { } of complex numbers, the series
=0

is called a power series . The numbers are called the coefcients of the series. The behavior of the series of course depending on the coefcients . In fact, we can show that it will converges if in the inner of a circle, and diverges in the exterior. The radius of the circle called the radius of convergence of the series. Theorem 6.13. Given the power series , put = lim sup

| |,

Then

converges if || < , and diverges if || > . Next we will investigate the series of the form , the following lemma play an important role Lemma 6.14 (Partial Summation Formula ). Given two sequences { }, { }, put =
=0

if 0; Put 1 = 0. Then, if 0 , we have

=

1 =

( +1 ) + 1 .

6. SERIES

21

P ROOF. The results can be derived directly from

=

( 1 ) =

1 +1 . =1

When the { } is monotonic, we have the following theorems Theorem 6.15. Suppose (1) the partial sums of form a bounded sequence; (2) 0 1 2 ; (3) lim = 0. Then converges. Corollary 6.16 (Alternating Series ; Leibnitz). Suppose (1) |1 | |2 | |3 | ; (2) 21 0, 2 0 ( = 1, 2, 3, ),; (3) lim = 0. Then converges. Corollary 6.17. Suppose the radius of convergence of is 1, and suppose 0 1 2 converges at every point on the circle || = 1, except possibly at , lim = 0. Then = 1.

Index

-cell or -cube, 14 addition, 7, 11 Alternating Series, 21 Archimedean property, 10 bounded, 7 bounded above, 7 Cantor sets, 15 cauchy sequence, 18 closed, 13 closure, 13 coefcients, 20 comparison test, 19 complete, 18 complex number, 11 connected, 15 converge, 17 converges, 19 convex, 13 cuts, 9 dense, 14 Density of in , 10 distance, 13 distance function, 13 diverge, 17, 19 eld, 7 nite intersection property, 14 four arithmetic operation, 17 geometry series, 19 innite series, 19 interval, 13 isomorphic, 10 least upper bound, 7 least-upper-bound-property, 7 lest upper bound, 7 limit point, 13 metric, 13 metric space, 13 multiplication, 7, 11 negative, 8 neighborhood, 13
23

open, 13 open (or closed) ball , 13 open cover, 14 open related to , 13 order, 7 ordered eld, 8 ordered set, 7 Partial Summation Formula, 20 partial sums, 19 perfect, 14 points, 13 positive, 8 power series, 20 radius of convergence, 20 ratio test, 20 root test, 20 segment, 13 separated, 15 subsequence, 17 supremum, 7 upper and lower limits, 18

Bibliography
[HS65] [Kno28] [Lan51] [Thu56] E. Hewitt and K. Stromberg, Real and abstract analysis, Springer Publishing Co., Inc, New York, 1965. K. Knopp, Theory and application of innite series, Blackie and Son, Ltd., Glasgow, 1928. E.G.H Landau, Foundations of analysis, Chelsea Publishing Company, New York, 1951. H.A. Thurston, The number system, Blackie and Son, Ltd., London-Glasgow, 1956.

25