0% found this document useful (0 votes)
19 views14 pages

Series

Chapter 6 discusses the convergence and divergence of infinite series of real constants, introducing key concepts such as Cauchy sequences and specific series types like geometric and p-series. It outlines conditions for convergence, including Cauchy's Convergence Criterion and provides examples to illustrate these principles. The chapter concludes with applications of these concepts, demonstrating their relevance in analyzing series behavior.
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)
19 views14 pages

Series

Chapter 6 discusses the convergence and divergence of infinite series of real constants, introducing key concepts such as Cauchy sequences and specific series types like geometric and p-series. It outlines conditions for convergence, including Cauchy's Convergence Criterion and provides examples to illustrate these principles. The chapter concludes with applications of these concepts, demonstrating their relevance in analyzing series behavior.
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/ 14

Chapter 6

Infinite Series of Real Constants

6(A) Convergence of Infinite Series, Divergence, Cauchy sequence etc

6.1 Introduction

Let an} be a sequence of real numbers.


Then the symbolic expression

Or, an or, simply an (6.1.1)


n=l

will be called an infinite series of real numbers (also called infinite series of real con
stants).
The numbers a1, 2, ,an, are called its terms (an is the nth term of the series)
sequence
and the sequence {S}, where S, =) a, = a1 t agt t ans is called the
r=l

of partia! sums of the series

L. The infinite series (6.1.1) is said to be convergent if


lim S, exists and = a finite real number S.
n+00

337
338 AN INTRODUTION TO ANALYSIS: DIFFERENTIAL
We then say that the series Sa has the sum 5 and we write
CALCULUS
S=an
M:
e.g., To prove that
n=]

1 1 1
+
r(r + 1) 1.2 2:3 3.4
is convergent and that its sum is 1.
We consider,
+
1 1 1
Sn = +
1.2 2.3 3.4 n(n + 1)
=

(-)*(-)+(-)*-)
1
1 n+1"

Hence, as n ’ oo, Sn ’ 1. This proves that the series is convergent and that its
Sum is 1,

M
2. If lim S, = +oo or -oo, then the seriesan is said to
diverge properly to to
or -oo as the case may be.

e.g., ) r=1+ 2+3+***+n+ diverges to +oo.


r=l

For,
Sn = 1+2+3++n
nn+1) +O0; as ’ O0.
2
3. Again, if {Sn} is o8cillating finitely then
the series San is said to oscillate finitely.
e.g., >(-1)"+ =1-1+1-1+1-...
T=l
Here

1, if n is odd;
0, if n is
even.
So {Sn} oscillates between 0 and 1 and
betweern 0 and 1. consequently also oscllates
CHAPTER 6: INFINITE SERIES OF REAL CONSTANTS 339

4. Lastly if {S,} is oscillating ininitely then an is said to oscillate infinitely.


e... )(-1)"+p =l - 2+3-4+5- 6+ oscillates infinitely.
T=|

Here
(n+ 1), if nisodd;
S,
-sn, if n is even.
Hence S,} oscillates between -oo and oo and consequentlyan does so.
The present chapter willinvestigate the convergence or divergence of infnite serics
of real constants.

6.2 Two Important Series


I. Geometric Series: The infinite Geometric Series
at ar + ar +. + ar + (a > 0) is

(a) convergent if the common ratio r ies between -1 and +1 (i.e., |r| < 1); in this
case the sum of the series is a/(1- r).

(b) properly divergent (to +oo) ifr 1.


ifr=-l and
() oscillates finitely,
oscillates infnitely, ifr< -1
Improperly divergent.

Proof:
Sn = a+ ar + ar'+ . . + arn-]
(r1).

(a) Now if lrl < 1then r ’0as n ’ 0.

S= lim Sn =1

(0) (i) If r > l. r 0 as n ’ 0 and then Sn ’ o0 as n ’ 0; hence the


geometric series diverges to +oo, if r >1.
(ii) If r=1,5, =a+a++ = na 0 as n’ 0.
the series properly diverges to +o0.
ANALYSIS: DIFFERENTIAL
340 AN
INTRODUCTION TO
CALCULUS
(c) (1) If r<-1,
{Sn} {p} oscillates
oscillates infinitely.
infinitely and the series also behaves in the same way.

(ii) Ifr=-1, we have


a - a + a - a + :to n terms.
S, =

a, if n is odd;
0, if n is even.

the series oscillates finitely.

Thus the Geometric series )ar-l converges only when r< 1.


T=l

iM8
II. The p-series: The infinite series
1 1
+ +
1p 2P
n=l

(a) converges ip > 1.

(b) diverges if p< 1.


1
Proof: a) Suppose p > 1. To prove that m?
Converges.
m=1

We examine partial sums of order 2 - 1 where n is a certain positive integer.


1
Syn-1 = 1+(t
2P
++

1 1
+ ++
9P 15P
1 1
(2n-1 + 1)P (27 1)P
s 1+ + 1
4p t + +

1 1 1

1 1 1
(2-1jp (22-1)P
1+2. + t 8 8P + t 27-1 1
1 1 1
(2n-1)2
1+ + + 1
2p-1 (2p-12 (22-1)3 +
(22-)7-1
1-() i.e., <
1
1 1-1 Wnich is a constant K (say) V
W4PTER 6: INFINITE SERIES OF RRAL CONSTANTS 341

Now for any poaitive integer m, 3a poaitive integer n such that


2-1> m.

IS}is clealy monotone increasing (. the terms are all positive).


Sm < Sy-< K, m.

ie. (Sm} is m.i. and bounded above and hence convergent.


the seriesconverges
mP if p> .
1
b) Supposce pS1. To prove 2diergs.
1
First obaerve, when pS 1.nsn (ie.. 1/n> l/n), nbeing a positive integer.
We now examine the sequence of partinl sums of orcder 2,

S - 1+

(*)"(6)·(+}
+ +t

()
1+;+2it4 +

Sy >Gwhenever 1+>Gie., whenever n> XG-2.


ien say G>0, we have
increasing (m.i.).
Ihe fartial euna ae obviously monotone
Gfor m > 27-2
if m> 2, then S > Sya >
(5)is .4. and unbounded above and hence
e, The sequence of partial sunas

Osequently the series 0/n)diverges to too, when ps1.


342 AN INTRODUCTION TO ANALYSIS:
DIFFERENTIAL
In particular, the Harmonic series. CALCULUS
1
Si1++;t+
n=1
+

diverges to +oo. (Write down an independent proof).


1 1
Q. Which of the three series - is convergent?

6.3 Cauchy's Convergence Criterion


applied to infinite series
Theorem 6.3.1. The infinite series an converges if and.only f given any [ >0,
n=l
there erists a positive integer N(E), such that whenever n> N¬)

iM: lan+1 t an+2ttantpl <E (p= 1,2, 3, ).


Alternative Statement: A necessary and sufficient condition that an infinite series
an converges is that, for each [ > 0, there erists a positive integer
n=l
N(E) for uhich

an+1t Qn+2+tam <E


whenever m,n > N. (one may take m>n> N).
For m, if we take n +p, the previous form of
the statement follows.
Proof: Theorem 1. The series an converges if and
sums {Sn} converges. By Cauchy's General Principle ofonly if the sequenceol pal
the sequence {Sn} converges if and only if, given any e Convergence of a sequen
such that whenever n > NE) and p= a >0. apositive integer
positive integer, the inequality
|Sh+p -Sn <e holds.
|n+p
i.e.,
|244-a <e holds
k=1 k=1
i.e., \an+1 t ant2
++ an+l <e holds.
Taking p=lin
Theorem.6.3.1 we obtain the following result:
Theorem 6.3.2. Ijan coverges then lim an = 0.
CHAPTER 6: INFINITE SERIES OF REAL CONSTANTS 343

Proof: In fact, putting p=lin Cauchy's convergence criterion we get:


IfSan converges, then for any e > 0, N(e) such that
lan+ll < e, n N(E)
which implies, lim an+1 =0, or lim a7=0 or we can write lim an = 0.
k+o0

Alternative Proof:

lim an = lim (S, Sn-1)


n-’00

lim S, - lim S,-1 =S-S=0.

(:} an converges » lim S, =S)


Observations:
This theorem is useful in its contrapositive form.

1. According to Theorem 6.3.2,


0.
IF an Converges, then an ’ 0 as n ’

.. the contra-positive statement must be true:

If an does not tend to zero as n ’ 00, then the series > an does not
converge.

convergence of the series


Note: an ’ 0 as n - 00 is the necessary condition of
an
cannot converge, because
Example 6.3.1. n!
n=l

>1whenever n> 1.
! 1.2·3.n
>an cannot converge.
an cannot tend to zero as n ’ o0 and consequently
C.H. 1992]

does not converge because un ’ 1 as n ’ O0.


Example 6.3.2. n+1
n=l
AN INTRODUCTION TO ANALYSIS:
344
DIFFERENTIAL
2. The condition that a, must tend to zero when an converges is not
CALCULUS
a
condition.
1
suf icient
For e.g., the Harmonic Series diverges, but its nth term 1 0 as
n=1
’ 0.

Consider the Harmonic series


n
n=l
For any neN.
1 1 1
Sn =
(1+++it
=

(1+)")+(*) .1
( ++ +
2n-1 4 2
1
2r

> 1+
)· )+
88
+
2n
++

= 1+
(++*) -1+7
It follows that {S} is monotonic increasing and unbounded. For obvious details
1
See art. 6.2 the proof of p-series (b) and hence diverges. Therefore= diverges
yet,
1
lim - =0.
n00 n

Q. What can you say about the series ? ?


n=l
2n + 1 n+ 103
n=l
(Ans. They do not converge since as n oo and l
o)
Theorem 6.3.3. Abel's Theorem: (also knoun as Abel-Pringsheim's Theorem).
is m.d., the conver
If {an} condition lim na, = 0 is a necessary condition for the
gence of the posttive neries a. (C.H. 1983, 93)
such
For, if u, converges, cor reBponding to any e >0, m
that
a positive integer
ln+|t an+? t t a, <[/2 if n > m.
Now each termof the left side > 4n.
CHAPTER 6: INFINITE SERIES OF REAL CONSTANTS 345

Therefore ( - mn)an S an+1 t Qm+2 + + an < [/2, if n> m.


We know that if>an Converges, then an 0as m ’ o.
. we can choose N> m so that m an < [/2 if n> N.
Thus 7 an - m an < e/2 (n > m) gives n an < e/2 + s/2 Vn> N.
n an ’ 0aS n ’ O0.
1
Note 6.3.1. That this condition is not sufficient, follous from Abel's Series ) n log n
1
uwhich is a divergent series (See Erample 6.6.27) althongh lim n: n+oo
n log n
= 0and
1
is m.d.
n logn
1
Example 6.3.3. The series ) an + b
(a, b > 0) diverges.
1
Let an = Here {an} is m.d.
an + b
1
lim n an = lim = lim 0.
n+0 a + b/n
no0 an + b

)an Cannot converge, for if it would, then n an Would tend tozero as n -’ oo.
Being a positive series >an cannot oscillate, so the only conclusion is that an
diverges.
This theorem gives a negative test:
If nan does not tend to zero as n tends to infinity, then an does not
Converge.

6.4 Applications of Cauchy's Principle


Example 6.4.1. Prove by using Cauchy criterion, that the series:
1 1 1
=1+i + L2 +
L3
+
4
++
Ln
+ converges.
n=0

Solutio. 1 1
1
+
(n+ 1)! (n+2)! (n+3)! (n+p)!
lk=n+1
1 1
+n+1 t + 2n+p-1
2n

11-(1/2)P 1 1
which is < 21- p.
2r 1- (1/2)
346 AN INTRODUCTION TO ANALYSIS: DIFFERENTIAL
ntp
CALCULUIS
ieG):k=n+1
(4)" 0 as n 0, therefore for each e > 0 3 N(e) for which 2(4.
if n> N.
<E, whenever n > Nand p >land the series converges by Cauchy's
lkn+1
riterion.
Example 6.4.2. Use Cauchy 's criterion to prove that the series
1 1 1
1 - + ; - - +(-1n + s convergent.

Solution.
n+p
1 1

lkn+| n+1 +(-1)"*42 ++(-1yn+p-l.


ntp
1
n+ 2 n +3
**+(-1)P-!.
n+p

-3)- hich is <;


=

n+]
1

E-eiln+l> ad Vp=1,2, .
i.e., if n > N andp= 1,2,3, where N = integral part of (1/e - 1).
Example 6.4.3. Using Cauchy's principle prove that the
Harmonic series

cKANet converye. E-+i**..


Seluthun

when p=n

1
2r
2n (n terms) =n 2n
ie.,
Vn ud p=n.
CHAPTER 6: INFINITE SERIES OF REAL CONSTANTS 347

if we choose [ <then n and p s.t. |Stp - Sn >¬.


In other words, Cauchy's condition is not satisfied.
ie. the series does not converge.
[Being a positive series it cannot oscillate also; the only conclusion is that the seriesL(1/n)
diverges to too.]

6.5 A Few General Theorems on Convergence or Diver


gence

IM:
Theorem 6.5.1. Let N be a fred positive integer.

(a) The series IM:


ak 3nd
k=l k=N+1
a7 are both convergent or both divergent.

(b) f the series ak Converges to S, then


k=1

a7 converges to S-(aj + a2 + ag + + aN).


k=N+1

Proof: (a) Let n > N and T, = a7 and


k=N+1

let
K= a1 + a 9 + +aN = a fixed number.

Then, t + an
S, = ai + a2 t t a N + a+1
= K+Tn.

can be easily verified. In fact, taking


Letting n ’ 0, the results of the theorem
limit as we obtain a =S- (a1 + a2 +taN).
k=N+1
infinite
stemoval or Addition of a fnite umber of terms at the beginning of an
series does not affect the convergence or divergence of a series.
ANALYSIS: DIFFERENTIAL
AN INTRODUCTION TO
348
CALCULUS
Theorem 6.5.2. Let A be any positive constant. Then the infinite series a7 and
k=l

are both convergent or both divergent.


k=l

Proof: Let S, = aj + a+ +an and T, = Aaj + Aag ++dan.


Then, T, = \Sn. Now make n ’ 0. The results will follow from definitions.

Corollary 6.5.1. If Abe any non-zero constant (i.e., A>0 or A< 0) and ) a,
k=]

diverges, then Aa_ also diverges.


k=l

Thus, multiplication of each term of a series by TLON-zeTO Constant does not affect
the behaviour of the series.

Theorem 6.5.3. (Associativity). If the terms of a convergent series are grouped


in parentheses in any manner to form new terms (the order of the terms remaining
unaltered), then the resulting series will converge and converges to the same sum.
i.e., if the series aj + a2 t a3 +.+ ant converges to S, then the series
(a1 + a2t+an)+ (an1+1 t an+2+tang)t
also converges to S.
Proof: Let S, = aj + a2 t t an. Then,
S. ’S as ’ 00. Hence Sn since >an converges to 5, we naye
S as p ’ 0.
The partial sum of the second series is

u1 t u2t+up = Sn; hence etc.


Caution. It is not true, however, that if a
inserted, it will still converge when they are series converges after parenthesis are
For example, removed:
1. The series (1 - 1) +(1- 1) +
(1 - 1) + (1- 1) +
series Zero but the
converges to

does not 1-1+l-1+1-1+1-1+.


converge (it oscillates between 0 and 1).
HAPTER 6: INFINITE SERIES OF REAL. CONSTANTS 349

()-(-)
&convergent series; because

S, - a where a =(2k2Ak+ 1 )-()


--)--) 2n + 2
2n + 1

-(4-)-)a
1 1
2n 2n + 1

(since the sum is finite we can remove parenthesis).


1 1
S - S, 2rn +2 2rn +3 >0 i.e., the sequence (S,} is m.i.
And S, =
1
V.
G-)-(-)--) 2T + 1

i.e., (S) is bounded.


(S} converges and consequently the series with parenthesis converges.
But the neries (removing the parenthesis) becomes:
3 4 5 6
i-t+(--i+2
I nth term does not tend to zero as n 00 and the series, therefore, cannot
tsesge. Rather it is fiuitely oscillating.
* 1 is t be renenbered, however, that if a series with pareuthesis diverges, then
would still
Se ice without parenthesis cannot converge: for if it couverged, it
RVeg after growping the terus ia parenthesis.
What is interesting to sote is khat:
to
aseries whose separate terno are suma in parenthesis converges
3, then the serles with parenthesls omitted wll eonverge to the same
uEn S, if it eonverges at all.
i*tis kullows from Theojeus 6.6.3 abuve; if the serles without parenthesis would converge
= S.
tbea ibserting pareuttesis will wt change its sn by this Theorem; hence S'
AN INTRODUCTION TO ANALYSIS: DIFFERENTIAL
CALCUUI
The next post, is of gret utility. aS Wiwill be observed in some of the
theorea th.
Thaoren 6.5.4. Let {z,}o be a sequence of postive real numbers. Then the we
B coe f ond only f the seuence (S} of partial sums is bounded nbet

nS, sup{S, : n N}

Since >0. the equence S, of partial sums is monotone increasing:


S < S, < S;< <S, <:.
and hoace by thcretn d gence of monotone increasing sequence, the sequence
(S.}vergs i#i bounded. i conwrges iff (S} is bounded.]

You might also like