0% found this document useful (0 votes)
13 views41 pages

Real Analysis

The document discusses mathematical concepts related to real numbers, including upper and lower bounds, completeness properties, and sequences. It covers definitions and theorems such as the Least Upper Bound Axiom, properties of bounded and unbounded sets, and the characteristics of Cauchy sequences. Additionally, it touches on metric spaces and convergence, emphasizing the completeness of real numbers and the existence of limits in various contexts.
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
13 views41 pages

Real Analysis

The document discusses mathematical concepts related to real numbers, including upper and lower bounds, completeness properties, and sequences. It covers definitions and theorems such as the Least Upper Bound Axiom, properties of bounded and unbounded sets, and the characteristics of Cauchy sequences. Additionally, it touches on metric spaces and convergence, emphasizing the completeness of real numbers and the existence of limits in various contexts.
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 41

equI It is a mappingfrom N to R

I
Ex fin R
5

o i
C
is Is d.es

References RadingPrinciples
Analysis 18 2 byTao
M Kumaresan Ajithbook
Kapostol book

Calculus RealAnalysis

Q Howdo we write Realnot


Q Howdowe write natural nos first
Piano'stxion we declare N hi2,3 3
Completeness property in noSystem

Let A be a subset of R M ER we say m is an upperbound of set A if


atA hem is satisfied

Similarly m is a lowerboung of set A if utA n m is satisfied

2101 Check Number notes


system

α is an upper bound of A if nfA In α similardef'sforlower


bound
α is not an ab of A if ntA s.tn α

AECR is bounded about if 7 GR s t α is an ab of A αER NEA n α


not bounded above if ER α is not an ab of A αtRf nEA nsa

If A is boundedabove bounded below then A isbounded


4 Let A B IR AEB then i if B is bounded then A is bounded converse

Finitesets arebounded

Least Upper Bound lab αᵗ s lab of AER A if


i α is an ab of A
iil if β is an ab of A then
β
Greatest lower bound geb α is geb of AER A if
i α is a eb of A
it if β is a lb ofA then α β
If his nonempty
if A is bounded above LUB Axiomof IR
then hub of A exists in 12

pf AER A is bounded above


LER is the lab of A
if i α is an ub
ill if β is an ub then α eβ If β α then β isnot an ubofA
FREA Rsp
if a club Al then
7yfA st
you
Ex A 0.1 lab A 1
If α is an ab of E αfE α lab E Max E

If α hub Al β lab B then lab AUB Max α β 8


Proof to prove 8 is labAUB 8 is ub of AVB
if 8 is an ab then 8128
givenany element REAUB Rt A or a B
α or n
a β
n x d or n βd
K 8 8 is ub of AUB
if d is an ub 6 8 max a
β
suppose 8 CJ wot let f α a 8k α
Sinie Subtal kt A s t t's no α

at A AUB s t ns8 contradiction

i 8128

prove that N is not boundedabove


Proof let us consider it is bounded above then by LVB Axiom αtR
st α lab N
since tablet Fn FN sit α 1 Cn
α cuel
Sinie UFM Rtlf N
i 7 Milf ns.t uti α lab N
i α cannot be lublix contradiation

i Nis not bounded above


MEN sit Archimedian Property
let
niy ER a 0 then
nnsy
Intel In
Negation yfR
nsositnne n.la

ie is an ab of
which is nottrue henie Archimedian
property is
true
Given any two real no s there is a rational no betweenthem

n.yfR hey afQ sit neacy


If Given niytR nay we need to find a
Iata set
neacy pigez 9 0
nsf y quopegy
for p to exists in between
qn gy gy qn 1
gly a 21
i.e we have to g s't
choose
gly n 1
By archimidean's principle we can always findsuch g
Now we ran take 1 or
p qn p gy
i a
Iqexists for given a
y
Given any two rational no s we can find an irrational no in betweenthem

7pfQ sit
niyfQ nay ncpaq
If Start with a V2
y f since
they are real
no's a rational no a

betweenthem RCA
n C
y V2
acatricy
since ath is irrational the statement holds

If α lab a then so FactA st α Ecae


We can also write nfN Fanta s t α I can

labATB labA lab B ATB fatb ata bfB


net lab Atp 8 lab a α lun B1 p
A 11101 B 0,11
afA DEB aEα b β At B f1 I
ath p
i β is an ab of AIB AIB atB BAtb
let w be an ub of ATB then
atb w a w b Hata w b α w b btB
w α 2 β WE β β is lab A B
21101

Sequences f NTR fin an

Triangle Inequality lab la Ibl


Prof Exerrice la bibl la la b tlb
la bl 191 1611
A sequence is bounded if 7M sit ntN Ian M Everlise find M interms
A of tb or ab
sequente is unbounded if MSO IntN sit an M

convergence of a sequence an I SO n.TN sit n no Ian l CE


an converges etR sit an I

If an converged an is bounded
Bounded sequente need not converge

Monotonic sequence Increasing or Decreasing sequence

antant
nfN a ant
I
Montonic sequence bounded converges

Intreating bd above converges


proof ext
Decreasing bd below converges

IIII II tunitions

Every bounded sequence has a converging sequente


Every sequente has monotone sequente True

Prof 1 s In Akon andan IN in

S
Infinity
If 5 0 the sequence is an increasing one ksn ancan
i a montoni subsequence can be found

since S is finite let no max S f N


consider not 5 k not anotkan
let n k
nits 7 neon Can can
sequente

245 ng n anacans found

let smallest in 5 be no
since no chicnic
smallest in si no be ni
ano an an
smallest in sifno.nl be no
i a montonic subsequenie is found

Cauchy sequence in R SO 7NotN n.msNo an am a E


an

Every bounded sequente has a converging subsequence

Every Cauchy sequence has a converging subsequence

4 Cauchy sequences are bounded


converging sequentes are Cauchy in IR
Cauchy sequences are converging in IR completeness of IR
Let canI be touchy

1am converges to say

I
a a
claim an Ian l Ian ane ane e an an
Jan anic anic l

we can choose Mic sit


nk No in ranchy deft h
An No 1in limitdef

we can choose his sit


an anal a
lann l Elz
a fan Ann tank l C E
fan 1 CE

let X be a non empty set d Xxx R st i dfn.gl 20 n.ge


it ding D n
y
iii dfn.nl dfy.nl
iv d him dn2 d2g
then d is said to be a metric on x d is said to be a metric space
If d satisfies everything excluding it then it is called pseudometric

usual metric
Examples i R d min n.gl
ii R dcn.nl In 917 where

d Xx X R defined by dfu.gl 9
f
be l
any non empty set
n
O y
this x d disirete metric
lit anti yn be binary sequence
hi yil
dflnnl.hn a

t let x idi X2dz be two metrit spares then for set Xix we can define

d fungi a gal d ning dalm.gr sit xxx d is a


mafia
we canalsodefine d cnn.mil cnn.nl
FI dzfm.fi
Open Unitballs Bfa11 ntx dinia 1
Closed unitballs B a I atx dlu.at 1
Unit Sphere Sca fntxidln.at 1

for R Adieirete
B 0,1 Gntr d note 1 0

B 0 I at R d no 1 IR

5107 IR 403
translation invariant afn y d nth
yth
Scaling faitor won't work in metri spatel

IR Xp d m d 91.41 1
t e 1m na t ly y I
B go
fade
deminal 4141,91 1 FETT I

chalk for 0
convergenie
p
pre
IF a fans is well defined it fful flu M 2
M n fful flail
let D fall Fu
then D setofdiff funis contfun s

We canalso define funi from one metric space to another

f IR IR f is continuous at atR so 78 0 In a as fin fake

An a ffnn ffa
Write above def in terms of metric spaces

03101

iii I f ni un
convergente 1 70 THEN sit n N fun a E
Mn in
Ex In disinite metri spare R dais the sequente converges to n if lunt is
eventually constant ie FNEN.tn N an n

limit is unique
Proof let us assume limit is not unique i e n limits ofthe
y are sequence

dfu.nl d n Maxim
mn d kmaxlmNd Y CE

n Eso
y
Tiny feso deniy 0 y
Are converging sequences bounded In metri spares
Does every bounded sequence have a converging subsequence

Observation There is no increasing deisealing sequences in metric spaces So we


can't use the proofs discussed in lase of IR

A EIR is bounded if 7M ntAifM M


A xd is bounded if 7M so An yfA d aint M

Fatx V30 st A I Bylair

an we define cauchy sequente in metri spare Yes replace Innnm with d un nm


need not in metalspate
Cauchy sequinie converge

x d is complete if every Cauchy sequence converges

Ex X R deniy In yl R d complete metricspace LVBofR1

d
R d
finit
n na y yn

R fun Mn Mint

42 42 Uni Man

ni Mi Mi Min

nini in
un IR is Cauchy so INEN rm N d un un CE

Finnit CE

Inni Rmit E ti 1,2 in


uni touchy since UnifR unit converges to R

we need to prove un converges to n

EE 5
dicnn.nl

sinie nail converges to id for 7N s.tn Niluni hilc


let N may Ni Nk

n N dfnn.nl LEE E
Henie Cant R all is complete
ags to n in Rn

let dpfn.nl
then Rn dp also converges

Rr Hpl 14110 is complete


cheek for IR do First of all it is not a metricspace

l Sequeries

8 0 bourdsequence cent dlnn.tl 0

12
h un dz un o co

FI Innio we have to chalk if this series converges

T.is
a

fi
Exist let s ktz Ken
S is non empty let us suppose 5 is empty
4 21 Kon
n arts as lowerbound for 2
but no 1 bfor Z
contradiction

n is nib for S
i lab exists in S
let α be lab s
ESO α E cannot be ub ofs
let E I Kts s t α 1 11

This k satisfies Ken


since n is an ubI is a lab

α n we'redone

α U
if
teh then kt ktifs
fffadiation
i
HIM Kitsu
key
Unis let us suppose m n satisfiesthe condition
n t emen nichen

let us suppose min


let men mile n e n
Mtl En m n 1 contradiction

i Man

of nthroot of a real no given at Rinfix


Existenie α sit α n

S
ut
y R yren
satisfiesthis condition
S is non empty for her
H.IE
n a n117 an 4111
n is an ab of S
i lab exists ins ocacb
let α be lab s b an b all
b arc b a bnt n
Suppose n let k 9ft a
Cfca
Kex
α Kk k n

non _an
s
n K K1 αk is an ab ofS contradiction

Suppose anan let


hayf.fm ochel

hi
gift
ch.am n a 6th

am
I
c n an
1 41 ca
i th t s contradiction

i n since lab is unique this nthroot is unique

mplete metrics IR d Rn Ap X daisi


Not complete X 0,11 id
induced meth

In R B a n d is not complete
I.fi In un i

A is finite if either A is empty or nfN sit Ff In A sit f is bijeitive

A is countable if Jf N A that is bijeltive


1081 f In A that is bijeltive
cover of a set Cover is a colleition ofsets F Ax th
Fis cover of x.tl if
Lta
if YEX then F is cover of 4
Opencover All A's are open

Ex for R F An ntz For x d I Blair rfR


An ninti
Bfair afx
Bn ntz
Bn 1min21

QQ
R

d nitric spate Completeness Compactness Connectedness

Every caully sequente should converge to point in X


Completeness some

Given any open ball B air in X say A then the open ball Bla r in Y
is AMY
Openset x fatx 8 0 St Blair EX UBlairl aeX
For x d I Blair 8fR

Blair afx arts as covers of the set


Prove 0 1 is not compact cheek compartness def

011 B i 1 2 0

B E E t n 2

0,1 is not open in R but 0,1 is open in 0,1

image of compactset is compart True False

conneitedness in metricspate

Any similar results for continuous funi

Given x dl.ly d be a metri space f Y is a continuous font if KEX is


compart in then f k is compact in Y

Prof let us at a be an open cover of ffk in Y

flk 4
WK T if f A EB A ÉB
i FINE Ux I KE f
YU
WK T f AUB I A OÉ B Provethis alsofor intersection

i K Flat
sinie Us are open I Val are also open

i Étual in acts as open cover for


Read series offunctions

we say for f if the graph offn approines similar to the graph of f

i.e

Mairi
Hutton VE 0 FNtNS.tVn N.Ifnfnl flullcE
domain offun
1
works
thisLef
Inta flat
let 4 be a metri with d at distance funi
then nt oil 30 NtN St n N d faint fful CE

only limited no of points


Defining companies helps us to cheek for contlude about

the whole set

We can generalise above def s for fn x all 4d

Ex fn 10,1 IR
utfoil fatal k we cantakefful k Antfoil as the fun to which
this sequente converges

fn 0,1 R
fretfo I fatal a flu n

In 0,1 R

fnful nn diverges

fn 0,1 IR
fatal converges to fful 0
fatal n converges since nF foil a sequence converges to

Infal nth for n 1 a sequence converges

to 1 as n α
not converges
1
49 n
fatal n mod n
MEN µ
Inful n mod n

for each ntN friful eventually


ends a
up as
in constant
seq
so it converges

so it
for each nfN we end
up with an oscillating seq
doesnot converge

iiii
I let fn ab IR be a sequente offunctions
in if
we
say fat converges a b
in X

it I
Ant b He 0 not sit mn I
he
Hausdorff
T2 property in topology

Limiting funt need not be continuous for given sequence ofcontinuousfunctions


incase of pointwise convergente ex.fr ul nn vntCoil7
Just change
teso.fnotN.tns.no Autlain
fatal fful CE

For frfr an Ant 0,1 this sequente donot converge uniformly to fful l
i 7 70 nofN n no 7nF 0.11sit frful flu E

Ex fn 0,1 IR fatal fin O


let us fix n no

let Ital be a sequence of funitions


We
say fnl convergee pointwise in ab if 7 f ab R st

at a b SO Froth sit n no fatal flu E

let Ifn be a sequente offunitions


we say Ifn converge uniformly in ah if f ab R st

0 nofN S t n no utfa b fnfn flat CE

Q.fr al Sinf 0.21

511 Guess the limit finl O Rt 0,21T

01 pint
Estimate faint final sing

for given SO we can find no St ICE


frontari
i This fat converges uniformly to fail O affo
21T principle
ch ris.it AIX
IndicatorFuns Xa R
1 nta
Aful
0 H A

Q.fr nl Xanco on oil

finess the limit flu O nf 0,1

Estimate fatal flat Halal fall XQnco.tn 0

and.tn
but when n O it converges to 1

guess is wrong
i Our

Another
guess 1 fin
o go.is
for n O Xanco An Xango 1 0 E

given kt 0,1 not NS.t not


then n no Xangoing D forgiven n
i fatal flu OLE for givenn

o n f o
t it I 7

uniformly convergesto fln 0


i e for f 0
I f t
49
In

n't
In flat 0

In
O In

I 1

In flat n

i
O In

n 3

Eh In flu

0
I
1h

BAR d

Given f are bounded sup gin fful nts dfg


g
191m flat glut hla h1n1 flu
d gift dgh dfk f
the d fig defined above is metri on B x.IR f X IR f is bounded

ie BCX.IR d is a metri space fis bounds


7M ntx flu M
Sup metric do
o n 1 it converges to fin 0 011
I Consider faint n then
1 M I

i Continuousfuni's can converge to a discontinuous fun

2 Consider
frin converges to flat O

i Sequence of discontinuous fun s canconverge to a continuous fun

If X M then B x IR bounded sequences 1

1,2 n BCX.IR Rn

If we consider is finite then pointwise convergence uniformconvergence for


Sequente of funi's from IR

4 9nF B IR do sit gn is a sequente then


gn g in Y
T 30 notN n no do 9 19 CE
suph an 91 ntx E

ntx 19min girl CE

f R IR is continuous at a if ESO 7830 s t n aks fin flat CE

f Xian 4dy is cont at a if so 78 0 sit at Blais wrt dn


flat B fla E
wrtdy
If fn are continuous fn If then is f continuous

since fn I f fa f flat fn n

i flu f a flat Infn fn n flat


Ifn n flu fnful f a

Ifn n fatal fnla flat


Ifn n flu fn n fnla fnla flat
from ptwise conv continuous ptwiseconv

f initimconvergence
In for 7mi man Ifalus flat leg

Fnla flat for Elz nz Unanz Ifn al f a


since fn is cont at a for 6 3 78 s t nFBla8
flu fatal

v aex.mltahtofntnl n.tkItafatal if fatal If on

Ifthe sequence offunctions are continuous at a thenthe limitingfuni is also


continuous at a
f o1 IR
then f is diff at atx if
a flathh.FI flat
SO 7870 1h0168 f
fa flat

flain fla hf a E 1h

flain flail interms of h


Estimate
Efl into eanff
let for If
then f nth flu flush forfath fornth fn n forful fln

firth fornihl fulmint faint fmln find

f ab R
f is differentiable at atfaibl it
all fkthl.tl
exists

it 7k SO 7850 octales
tell k E

Fk Veso 3870 octn l


k cE
cks.fr

film int is at C

f
cont

K R I

f is diff at s if fifa b IR sit film f a s mn


figs
If Ff ab 112 sit f a f 1 the film In C f is cont at
then we can say f is diff at c

Diff.at I cont at a

E is diff at s
Fg's fig
fig which satisfies the 19

fig

f fff.it fig is diff.at


f7 a c film
fg g c in c g.nl
f al.gr In 1 film glitt gint fil
n 1 finl.g.lul
8

suppose for
1M
Q1 fn I f then is f diff
Q2 fn's are diff In If then is f diff
Qs Qap is f diff
Qa Q2P
it in 10,1
fn's are diff f f
fn f are diff fn f sinnin
Q1 Q6iP

Suppose Ffn if utD Ifnful Mn Mn 0 then

fnt
iiit .is
mmat sinfn
W
sinn sinn sin
sirth
n J R is uniformly sauchy in J
6 0 nofN In m no rtJ fn nl fmln E

fEminil E

do fn.fm
We know that Cauchy sequence converges uniformly in lace ofreal no's

In lase offunt s we need to chelk

Define net flu n


fnlnl

we lan follow the sameproless of real no s in this case offunctions

If fn's are diff fn g f I

T3 Intermediate ValueTheorem
continuous

T4 Mean ValueTheorem
differentiability

fill
flbbiff
f's b a f bl flat

ffn.fm d n d fn fm t where tt d
fn Fm n a

fnlul.fm nl in d filth fm t focal fmidt

nd in ti fm'ftl fold fmidt


lb.at fn'lts fm'lts fnldl fmldl
For we can use uniform convergente of film
For we can assume pointwise convergence of fn at thepoint d
then fnin fmln b a
Ita E

i Statement
If fn's are diff in ab In I g fold are cauchy converging

for some de ab then fn is uniformly cauchy in ab


question

what about a 453


i e is for uniformly caushy
in a b

for given flnl we ian choose faint as flute 1 f net


or or combined
as sequence of fun s which uniformly converge to flnl

Right now we know that fn If on T.fr are cont f is cont on J

fn I g ont f are diff fn is candy at some


point it aib
fn I
f is cont
To know if f is diff we need i film f n c

i film is
chinious
ati
flat
fatal
fn are diff
at c
i fniful.fm f n C
if at 1 for is cont c

fn.cn at 1
f frics a C

fn.fm's are cont at all points

Now we know film is cont at all points nfc


Toshow about cont at c we need to prove fn ul's are uniformly convergingto
fiful

let Onn1 fn Inl


101141 0m14 at a c it converges since fn g
at 1 10min 0min

m
t.fm
if
eF fn
h
ii
for
Ifn fm Z
i f is diff flg converge
to
g
25111

fn's are cont fn If Hafnal nita data

f
get n'to
fat Ian
Iota Ifn
fn's are integrable
Guess fn Tg
n Ifn Into
fnlm.tn flat.de faint flu an IR
g F a b IR
is linear
fnlns flnil.tn ff g If g
K
If IF monotone
K b a

if we assume fn If then the kn's s t lfn.tl Cnverges kn's are


bounded we can find such k we need provedbefore kn Mnl

The above holds if we define for on some intervals sinie we can bound du
so if fn's are cont on compact sets the above holds

Integration
Dividingthedomain triven a h finitesetwithorder
partition

No ki hn ntN Ro a unz

nif1Itont
Refinement of apartition p a partitionp is called refinement
of p if p p

UpperSum LowerSum
Mi f ki where kit ni ni ni

uppersum qq.uflntxfri ni.it


Max min exists only if f is cont
lowersum or f is bounded
ri ni i
ftp.uigffulx
Mi f ki where kit nii.mil
i U fp Mi ni ni e

fipl mi ni ni i

let us consider iniluding another point n in no n

then Uff Pr Mo n no M n n

M n no Uffip u fire Uffip


siteone of Mo M is max Mo in Uffpl other one isless

Similarly lff.pl fin


w.k.TL fipl uff.pl
then L Fip fp Uffin u fip

Partition of ab is an ordered finite set uo.me an with urea un b


Refinement of partition P is partition Q if PEQ
L fP milui ni il where
mini mi
p
uff.pl Mi ni ni i EMiani where Mi
fykf.tl
If PEQ L f p L FQ
fp U FQ
Forany P
Lffipl Ulfip
Foranyp Q fip UffQ fp Lf puQ off PUQ u fiQ

A fip p is apartition of ab

A IR non empty

A is bounded above one v B Ufffaib or FIP


i IR sit luh A α
Similarly for set B off p p is partition of ab
βtR glb B β
exists

llaim β prove bycontradiction or Uffip sinie u fip is ubofA


acts as lb for B B
Now if α β we saythat f is integrable Riemann Darboux

fluida β Frida
a

Ex flu k on a b α kba
β kb a f is integrable
findn kb a

flat n on a b 0,1

tip urini ni il E 0t tt
Itf
U Fip milki air In It

1
Fip C

UIF.pl
71 In
flat 1 neanco.lt

f 11 β O not integrable
9
0 atQincong

Jfinidn flatdn α
p
a

f aib IR is integrable if Eso partition sit Uffip Fip CE


def
By this we can prove is integrable

Pnexists for given


Ulfip fip I
0511
Ro Theorem let f ab IR be a continuous differentiable fun ffa flbl
Then 7 if a b sit f 11 0

Prof Sinie f is continuous over compact set a b So f attains its max


I min in a b
let m net a b sit flail is min fluil is max i.e flail fluteflu
Int ab

if n or net aib then we lan find interval n r n r for 8 0 n n or ne

say Faib since flails flu nF a b


flail flat Vat n 8 n.to ab

flui is local minimum at me in n 8 npr


f nil D
c ry f a b f s 0
Similar argument can be made for Nz

if nz b or U b 12 a
bothElaib m a
flats flnl ffb or flat fful f a nt aib

by law of trichotomy flnl flat F b M say data


fln M uffab
f al D Hatfaib
aib f 1 D
Hence proved

NElutheorem Let f ab R be a continuous diff funi then f ab


f b flat f s b a
Proof Let Flat flat flat k al
ftp.ff
sinie flul is cont Flu is continuous on ab
fin is diff F ul is differentiable on ab

Flat 0 Fb 0 Flat F b
Applying Rolie's Theorem we get 7 ab sit F i D
f'm ftp.aflal O

f i ba f b flat

intermediatevluthorem let f IR be a continuous fans


ab beany real no
between flat flat then t a b sit f c X

emma_ let f ab IR be a continuous funi flat f b has opposite

set flit 0
sign i e flat f b 0 then it a b

Proofoftemme.tt let us assume flat co fth 0 W 04


consider set E ut a b F 740 ain
sinie EEIR E is bounded flako f is cont at a
a s at8
by local maintenance of sign 7850 s.tt nlcovnt
f n LO nF a S at81
In co at a at

at 1 E

since at E E is non empty


E a b CTR E is bounded above
i By LVB Axiom α Aub E

Now At b f ab

claims f 1 to
if ffil 40 then flako or ffc 30

850 s.tt t7cO Vtf 18,48


El bylocalmaintenanie of sign
cc is its sit flil co te contradiction

f 850 s t f 750 f 18,48


bylocalmaintenanie of sign
7 scice s t f i so
i
contradiation
but I'Ll F flat o

i fc 0

Now given flatcx cffbl flat XCO flbl so


fini x 7 a bl s t
forgin g a co
g b so 961 O
ffi X D
ffc X

Ees
A cut α is subset ofQ satisfying α Q
if I α qtQ 9 1 then qtα
If ptx 7qtQ qtx sit pcq
i e Acut contains nomax element

Ex α ntQ n 3 Realno s cuts


α ntQ n223 not a cut

B Cp in cuts
Additioninct Xp atb atd bep

Gm β is a realno xp is a cut
atp sine 40 1340
βI Q

I
If atb for some at x
pf β plop P bfβ
plop pl bcp b pl bca
plb f
p Igbtby p'tap
let
ptarp p atb for some aft btB
sinie α is a cut a'fQ sit a'tα a's a
alb atb
β is a cut b'to sit blap b's
p a'tb Extp plop
Given α ptQ FREQ rso st
p r fx
Claims β is a cut

β 0
1 0 7ptx Ptp βtQ
Q p α rt Q since α is a lutt
p.atα for some
ptp β
2 Ptβ gap treat sit p r α

q842
r p v
q get
3 PFP rtQ s t p 8 α
let q p rfα
11 q p 2 q p qtr

ptαtp p atb for some atx btp afp btp


a b
athco pt0
vfO D SO Then
by Archimedian property on Q
Fnt2 s.tn WE α nellWfd
Take Int2 w a2 w
p p
p w Inti w pW pep Int2 we β
nuts nt2lwtp nw net we
xp
2W f x p Vt atp

ab afx.be B a20,620 UGatQ


xp uco
Similarly Xp is a cut
Setofcuts is a fieldunderaddition multiplication

Theorem Thereexists an ordered fieldin whith every non empty set that is boundedabove
has least upper bound Moreover thisfield contains
Qas subfield
Consider A to be collection of cutsthat is boundedabove A α α is acut
we call
β as ub of A if α
β α FA
consider I 8 is an ub of A
Yadi
if 8 8 SC
Yadi pf8 butp S EX forsome α
Sexi
8 is not an ub of A
i I is tub A

081

2 can t an R
fissiemaffi
URS Jfuldn yo iM ni ni e ipfpulf.pl
LRS fluida n
Efmifnini.nl Llfipl

we say f is riemann integrable if URS RS


ESO 7pfP a b u fip LF f LE

Prove n is R I with
fp Ifip ni ni ni i Eni i ni ni i

E ni ni e
E179
montoninity of
Chein linearity of ntD finlegial If g
assuming gave
If f is cont then f is RI RI

Ietheorem let f x IR be a continuous monotoni fit I is compact

set If f x IR be cont fun sit fn contverges to f pointwise


thenfor converges uniformly to f
Proof gn fn f
In a hutx anfulze
is llosed set
Em In e
Unie nex glut E

To prove the is open we need to find B nr for some 830 NfRhe


s't Blair the y Blairl then In yler yeah e
sinie
gn is cont Eso 78 0 n y 18 gnlnl gn.ly CE
i In a is closed set
nested
Sinie closed subset of compact set is compact In e is compactly
I Intine Rn E
Continuous funt on a compart set attains it's maximum

let f K IR A fin ntk img f


if M fin MEA
y
A compait set is bounded
Let K be a compact 818
Assume K is unbounded

no be fixed point in K.dfuo.nl R RID


B no R ytk d no g CR
B no Ri
KEY
sinie K is compact K 1Bno Ri
R R Rn

A no in max R 1 Contradiction

K is bounded
Given a compactset K if we take infinite subset EEK then E has a limit
point in K
let us assume there is no limitpoint of E in K
ie qtk
Profthestatemnt f K IR
wherek is compait
Continuous imageof compact set is compair
A is compact

A is bounded

By lub Axiom Pflat M

Define offal'ss t M flu M


seq
E flu M Incflu m

sinie E is an infinite subset limit point in A

flan hae limit M 4 ME A


Iheem A continuous image of a connetted set is always conneited

hiventwo metric spares x dx y dy a cont func f X 1Y a EEX sit


E is conneited f E is conneited
Assume f E is not connected
P Q sit PUQ F E PhQ 0 a P Q are open

Pna f pin flal if nF F plnF a finlf Pna


E
sinie f is cont we can pave FPI Ifa are open
bythe continuity

def
F union of disjoint open sets E is not connected contradiction

Inverse image of an open set under a cont funi is always open


let x dy Y dy be two metric spaces fix y f is contfunt
let AEY be open
consider not sit fluoltA
sinie A is open Irso sit B fful r 1A
Nowtake E r
Then 8 0 sit at B no8 flu f B f no E fluit A
nt FYA
i B no s FA I a is
open
Them A continuous funt ofa compact set is uniformly continuous

let x dx 4 dy bethe metric spares f X Y is a cont funt

You might also like