Convergence of Sequence and Series

MA1001D Mathematics I Monsoon 2022-23
DEPARTMENT OF MATHEMATICS, NIT CALICUT
CONVERGENCE OF REAL SEQUENCES AND INFINITE SERIES
Handout for MA1001D - Module II (Batch D)
Dr. Krishnan Paramasivam, Course Teacher
JEAN LE ROND D’ALEMBERT Born in Paris, France, on 17th Nov. 1717, died on 29th Oct.
1783; known for fluid dynamics, refraction Memoire sur la refraction des corps solides and D’Alembert’s
paradox: that the drag on a body immersed in an inviscid, incompressible fluid is zero, D’Alembert’s
principle (that the sum of the differences between the forces acting on a system and the time derivatives
of the momenta of the system itself along a virtual displacement consistent with the constraints of the
system, is zero): known as fundamental classical laws of motion.
GOTTFRIED LEIBNIZ Born in Leipzig, Germany, (1646-1716); invented infinitesimal calculus

independently of Newton [Leibniz and Newton calculus controversy: who is the founder of calculus, it
is still unclear.], invented binary system, published his book ‘ On the Art of Combinations’ (at the age
of 20); one of the founders of library science; wrote ‘ Discours de metaphysique’ (1686) - concerning
physical substance, motion and resistance of bodies, and God’s role within the universe, La Monadologie
(1714) - best work in philosophy, Leibniz harmonic triangle; given the ‘ Proofs of Fermat’s little theorem’;
dy
Rb
anticipates modern logic and analysis; founder of the notations: dx , integral symbol a , two dimensional
arrangements of elements (later called as matrix by Joseph J. Sylvester).
AUGUSTIN-LOUIS CAUCHY Born in Paris, France, (1789-1857), single-handed development of

complex function theory; his famous work ‘Cauchy’s integral formula’ (integration of analytic (complex)
function around any simple closed curve is zero.) (you may study in details in MA2002D) - initiator of
permutation groups - the Cauchy-Riemann equations - his works with convergent series - Most number
of concepts and theorems have been named for Cauchy than for any other mathematician. He graduated
in civil engineering.
SEQUENCE, CAUCHY SEQUENCE AND CONVERGENCE
• N = {1, 2, 3, · · · } is set of all positive integers, R is set of all real numbers.
• Open interval (0, 1) = {x ∈ R : 0 < x < 1}, closed interval [0, 1] = {x ∈ R : 0 ≤ x ≤ 1} =

{0} ∪ (0, 1) ∪ {1}.
Note: It is in draft form; Possibly there may be few typos. Please inform if you find any typos. Class email: 1
epsilon.ganitham@gmail.com, Class Whatsapp group: MA1001D D Batch (ECE)
• The interval (0, ∞) is set of all positive real numbers.The interval (−∞, 0) is set of all negative real
numbers, the interval (−∞, ∞) = R, the entire set of real numbers.
• Both −∞ and ∞ are not real numbers.
• Let A and B are two non-empty sets. A mapping f from A to B (denoted as, f : A −→ B) is a
function if every element of A is associated (or mapped) to some element in B.
• A real infinite sequence (or shortly, sequence) is a function f : N −→ R. that is, f (1) is a real
number, say a1 , f (1) is a real number, say a1 , ... f (1000) is a real number, say a1000 , ... f (50000)
is a real number, say a50000 and so on. Since there are infinite number of positive integers, we have
f (1), f (2), f (3), ·f (1000000), · (never end).
• Since f (1) = a1 , f (2) = a2 , f (3) = a3 , ·, f (1000) = a1000 , ·, we denote a sequence a1 , a2 , a3 · · · a100000 , · · ·

by {an }∞
n=1 or shortly (an ). (note that an or a10 are just real numbers, in fact if we write an , then
it just represents n-th term of the sequence (an ).
• Convergence: A real sequence (an ) converges to real number L if for given ϵ > 0 (however small),
we can find a positive integer k such that
ak+1 − L < ϵ,
ak+2 − L < ϵ,
ak+3 − L < ϵ,
ak+4 − L < ϵ, · · · · · ·
or, equivalently, at − L < ϵ, · · · for all t > k (that is t = k + 1 or t = k + 2, or t = k + 3, · · · ) (that

is, AFTER THE STAGE k, the difference between terms of the sequence 4(an ) with L is closer and
closer as epsilon ϵ is so small.
• Illustration of above Definition. Recall that (i) if we say x is any real number such that |x| < 2,
then −2 < x < +2, that is, the values of x lies between −2 to +2, or {x : x ∈ (−2, +2)} or x
assumes/takes any real number from −2 to +2. Recall that (ii) For a fixed if we say x is any real
number such that |x| < 2, then −2 < x < +2, that is, the values of x lies between −2 to +2, or
{x : x ∈ (−2, +2)} or x assumes/takes any real number from −2 to +2.
Thus above condition, at − L < ϵ, can be rewritten that −ϵ < at − L < +ϵ or L − ϵ < at < L + ϵ
• Question: BEFORE THE STAGE k, are the terms of the sequence (an ) ”CLOSER” to L ?
Answer: They NEED NOT be closer to L.!!, that is, a1 − L < ϵ or > ϵ, a1 − L < ϵ or > ϵ
a2 − L < ϵ or > ϵ ... ak−1 − L < ϵ or > ϵ, ak − L < ϵ or > ϵ.
• The real number L is called the limit of the sequence (an ).
• Let ϵ = 0.1, a1 = 100, a2 = 200, a3 = 300, a4 = 400, · · · a9 = 900, a10 = 1

10 , a11 = 1
11 , a12 = 1
12 ,
1 1
a13 = 13 , a14 = 14 · · · . Clearly, L = 0, what is the stage after that the terms of the sequence (an )
with L, are closer? For example (or stage 1 or k = 1), we get a1 − L = 100 − 0 = 100 > 0.1 = ϵ.
NOT CLOSER, stage 9, or k = 9, we get a9 − L = 900 − 0 = 900 > 0.1 = ϵ. But stage 11 or
1 1 1
k = 11, we get a11 − L = 11 −0 = 11 < 10 = 0.1 = ϵ
• Note that all members of sequence a1 , a2 , · · · a1000 , · · · , L and ϵ are all real numbers; but ”the stage”
k is a positive integer. (Now, CLOSER). Again stage 12 or k = 12, CLOSER. Again stage 13 or
k = 13, CLOSER and SO ON ALWAYS CLOSER ! right? Thus our positive integer in this case is
k = 10.
• Sub-sequence Let (an ) be a given sequence. Let m1 < m2 < m3 < m4 < · · · be increasing
sequence of positive integers. The the sequence am1 , am2 , am3 , am4 , · · · is called a sub-sequence of
the sequence (an ). A sub-sequence is denoted by {amn }∞
n=1 or shortly, (amn ).
• Example. Let the real sequence (an ): a1 = 1, a2 = 5, a3 = 1, a4 = 5, a5 = 1, a6 = 5, a7 = 1, a8 =

5, · · · . that is, 1, 5, 1, 5, 1, 5, 1, 5, · · · . Choose increasing sequence m1 = 2 < m2 = 4 < m3 = 6 <
m4 = 8 < · · · . Then the sub-sequence (amk ) : a2 , a4 , a6 , a8 , · · · that is, (amk ) : 5, 5, 5, 5, · · · is one
sub-sequence. Similarly, 1, 1, 1, 1, · · · is another sub-sequence a1 , a3 , a5 , a7 , · · · .
• Cauchy Sequence. A real sequence (an ) is called Cauchy sequence if for given ϵ > 0 (however
small), we can find a positive integer k such that
am − an < ϵ,
for all m, n > k.
• Example. (an ) : 1, 12 , 13 , 14 , 15 , · · · be a given real sequence. (i) Let ϵ = 1. Find a stage k. (ii) Let
ϵ = 0.1. Find a stage k. (iii) Let ϵ = 0.0021. Find a stage k.
• Monotonically Increasing Sequence. A sequence (an ) is monotonically increasing (or shortly,

MI) if we can find a positive integer k such that
ak+1 ≤ ak+2 ,
ak+2 ≤ ak+3 ,
ak+3 ≤ ak+4 , · · · · · ·
or equivalently, am ≤ am+1 for all m = k + 1, k + 2, k + 3, k + 4, · · ·

or equivalently, am ≤ am+1 for all m > k.
• Example: (an ) : a1 = 1, a2 = −1, a3 = −2, · · · , a20 = −19, a21 = 20, a22 = 21, a23 = 22, · · · , a55 =
54, a56 = −55, a57 = −56, a58 = −57, · · · , a102 = −101, a103 = 102, a104 = 103, a105 = 104, a106 =
105, a107 = 106, · · · .
(an ) : 1, −1, −2, · · · , −19, 20, 21, 22, · · · , 54, −55, −56, −57, · · · , −101, 102, 103, 104, 105, 106, · · · .
Prove or disprove: (an ) is monotonically increasing sequence.
(an ) : 1, −1, −2, · · · , −19, 20, 21, 22, · · · , 54, −55, −56, −57, · · · , −101, 102, 103, 104, 105, 106, · · · .
• Monotonically Decreasing Sequence. A real sequence (an ) is monotonically decreasing (or

shortly, MD) if we can find a positive integer k such that
ak+1 ≥ ak+2 ,
ak+2 ≥ ak+3 ,
ak+3 ≥ ak+4 , · · · · · ·
or equivalently, am ≥ am+1 for all m = k + 1, k + 2, k + 3, k + 4, · · ·

or equivalently, am ≥ am+1 for all m > k.
• Example. (an ) : a1 = −1, a2 = +1, a3 = +2, · · · , a20 = +19, a21 = −20, a22 = −21, a23 =
−22, · · · , a55 = −54, a56 = +55, a57 = +56, a58 = +57, · · · , a102 = +101, a103 = +102, a104 =
+103, a105 = −104, a106 = −105, a107 = −106, a108 = −107, a109 = −108, a109 = −108 · · · . Prove
of disprove: (an ) is MI or MD by finding the stage k (or finding the positive number).
• Monotone Sequence A real sequence (an ) is monotone if (an ) is either monotonically increasing
sequence or monotonically decreasing sequence.
• Notation (an ), (bn ), (cn ) denote real sequences. at denote t-th term of a sequence (an ). All
sequence (an )’s are having real numbers. There is no complex number in any of the term of a real
sequence (an ).
• Theorem A sequence (an ) converges if and only if (an ) is Cauchy sequence.

Proof
Let (an ) be convergent and lim an = ℓ.
n−→∞
ϵ
Then a pre-assigned positive ϵ there exists a natural number m such that |an − ℓ| < 2 for all n ≥ m.
ϵ
Therefore |an+p − ℓ| < 2 for all n ≥ m and p = 1, 2, 3, ..........
ϵ ϵ
Now, |an+p − an | ≤ |an+p − ℓ| + |an − ℓ| < 2 + 2 for all n ≥ m and p = 1, 2, 3, ..........
That is, |an+p − an | < ϵ for all n ≥ m and p = 1, 2, 3, ..........
This proves that (an ) is a Cauchy sequence.
We now prove that the sequence (an ) is convergent if it is a Cauchy sequence. First we prove
that the sequence (an ) is bounded. Let ϵ = 1. Then there exists a natural number k such that
|an+p − an | < 1 for all n ≥ k and p = 1, 2, 3, ........
Therefore, |an+p − an | < 1 for all p = 1, 2, 3, ........
or, ak − 1 < ak+p < ak + 1 for p = 1, 2, 3, ......
Let B = max{a1 , a2 , ...., ak , ak + 1}, b = min{a1 , a2 , ...., ak − 1}. Then b ≤ an ≤ B for all n ∈ N .
This proves that (an ) is a bounded sequence. By Bolzano-Weierstrass theorem, the sequence (an )
has a convergent subsequence. Let ℓ be the limit of the sequence (an ). Then ℓ is a subsequential
limit of (an ). Let ϵ > 0. Then by the given condition there exists a natural number m such that
ϵ
|an+p − ℓ| < 2 < ϵ for all p = 1, 2, 3, ......... Hence (an ) is a convergent sequence.
• Theorem If (an ) is monotone sequence, then (an ) is either convergent or divergent. (never oscil-
lates).
• Theorem If the sequence (an ) is monotonically increasing and is bounded above, then (an ) con-
verges.
Proof
Let (an ) be a monotone increasing sequence bounded above and let M be its least upper bound.
Then (i) an ≤ M for all n ∈ N and (ii) for a given ϵ > 0, there exists a natural number k such that
ak > M − ϵ.
Since (an ) is a monotone increasing sequence, M − ϵ < ak ≤ ak+1 ≤ ak+2 ≤ ..... ≤ M .
That is, M − ϵ < an < M + ϵ for all n ≥ k.
This shows that (an ) is convergent and lim an = M .
n−→∞
• Theorem (an ) diverges if and only (an ) monotonically increasing and not bounded above.
Proof
Let (an ) be a monotone increasing sequence, NOT bounded above. Since the sequence is unbounded
above, for a pre-assigned positive number G, however large, there exists a natural number k such
that ak > G.
Since the sequence (an ) is a monotone increasing, G < ak ≤ ak+1 ≤ ak+2 ≤ ......
That is, an > G for all n ≥ k.
This shows that the sequence (an ) is divergent and lim an = ∞.
n−→∞
• Theorem (an ) converges if and only (an ) monotonically decreasing and bounded below.
• Theorem (an ) diverges if and only (an ) monotonically decreasing and not bounded below.
• Infinite Series Let a1 , a2 , a3 , · · · an , · · · be positive real numbers, that is, ai ∈ (0, ∞) for all i’s.
Then a formal sums of those infinitely many real numbers ai ’s (say) a1 + a2 + a3 + ... + an + ... =
∞
X X
an = an is called an infinite series or shortly, a series.
n=1
• Sequence of partial sums of a Series

∞
X
Let an be a given series of positive real numbers. Let
n=1
s1 = a1 ,
s2 = a1 + a2 ,
s3 = a1 + a2 + a3 ,
·········
n
X
sn = a1 + a2 + ... + an = ai ,
i=1
·········
∞
X
Then the sequence (sn ) is called the sequence of partial sums of an .
n=1
∞
X
• Convergence of an infinite series an converges if and only if the sequence (sn ) converges.
n=1
∞
X
• Divergence of an infinite series an diverges if and only if the sequence (sn ) diverges.
n=1
• Note. The series of positive terms

P
an is either convergent or divergent (but it never oscillates).
Note that si are all positive real numbers as all ai are all positive real numbers. Further, since
s1 ≤ s2 ≤ s3 ... ≤ sm ≤ sm+1 ≤ ..., we see that (sn ) is always monotonically increasing sequence.
NECESSARY & SUFFICIENT CONDITION FOR A SERIES TO CONVERGE

∞
X
an is convergent if and only if the sequence of partial sums (sn ) is bounded above.
n=1
Note From the above statement, we can find a single positive real number k dominates all sn ’s i.e., sn ≤ k
for all n. One can conclude that convergence property of the series completely lie on the tail of the series
(but not lie on first few hundreds of terms or first few thousands of terms). Therefore, omitting finite
number of terms from the series will never alter the behavior of convergence or divergence of the series.
GEOMETRIC SERIES
∞
X ∞
X
n
The series r diverges when r ≥ 1 and rn converges to the sum 1
1−r , when 0 < r < 1.
n=0 n=0
Proof
Case 1 r = 1. In this case, sn = 1 + 1 + ...... + 1 = n and lim sn = ∞. Therefore the series is divergent.
n−→∞
r n −1
Case 2 r > 1. In this case sn = r−1 and lim sn = ∞ since lim rn = ∞ in this case. Therefore the
n−→∞ n−→∞
series is divergent.
n n
Case 3 0 < r < 1. In this case sn = 1 + r + r2 + ....... + rn−1 = 1−r 1 r
1−r = 1−r − 1−r
1
lim sn = since lim rn = 0. Therefore the series is convergent and the sum of the series is 1
1−r
n−→∞ 1−r n−→∞
∞ ∞
X 1 X 1
THE p-SERIES p
is convergent if p > 1 and p
is divergent if p ≤ 1.
n=1
n n=1
n
∞
X k
Note In general, is convergent, where k is any finite real number and p > 1.
n=1
np
P 1 P 2 P 5 P 5 2 P q 3 P 1 P 1 P −3 P 5 3
Example n, 3n ,
√ ,
n
√
n
, √ diverge.
n n2 , n3/2
, n , √
n
converge.
P
TEST FOR DIVERGENCE Let an be a series of positive terms. If lim an ̸= 0 (or limit does
n−→∞
P
not exist), then an is divergent.
P
Note. Above test can also be written (logically equivalent) as: if an converges, then lim an = 0.
n−→∞
Note. In the above test, if lim an = 0, then do not use this test; the series may converge or diverge.
n−→∞
P1 P 1 1 1
For example, n is divergent and n2 is convergent though n−→∞
lim = 0 and lim 2 = 0.
n n−→∞ n
INTEGRAL TEST Let a ∈ N. Let f be a continuous, positive, decreasing function on defined on

∞
X
[a, ∞) and let an = f (n) for each n ∈ N. Then an converges if and only if the improper integral
n=a
Z ∞ ∞
X Z ∞
f (x)dx is finite (or convergent!). an diverges if and only if the improper integral f (x)dx is
a n=a a
infinite (or divergent!).
Note. We know that for a definite integral, both lower bound & upper bound are finite real number!
In calculus, an improper integral is the limit of a definite integral i.e., the lower or the upper or both
Z n Z ∞
bounds of integration approaches ∞ or −∞ such as lim f (x)dx = f (x)dx.
n−→∞ 1 1
Z ∞ ∞
X
Caution! If integral converges, then series converges ONLY! it does NOT mean f (x)dx = an .
1 n=1
∞
X 1
Example. Using Integral Test, show that is divergent.
n=1
n
1
Solution Let f (x) = defined on [1, ∞). Clearly, f (x) is positive, continuous, decreasing on [1, ∞).
Z n x Z n n
1
Now, lim f (x)dx = lim dx = lim ln(x) = lim ln(n) − ln(1) = ∞. So, The integral is
n−→∞ 1 n−→∞ 1 x n−→∞ n−→∞
1
X1
divergent. Thus, by Integral Test, given series is divergent.
n
∞
X 1
Example Test the convergence of .
n=2
n(ln2 n)
1
Solution Let f (x) = defined on [2, ∞). Clearly, f (x) is positive, continuous, decreasing on [2, ∞).
Z n x ln2 x n
1 1 −1 1 1 1
Now, lim dx = lim − = lim + = 0+ = . So, The
n−→∞ 2 x(ln2 x) n−→∞ ln(x) 2 n−→∞ ln(n) ln(2) ln(2) ln(2)
∞
X 1
integral is finite (convergent). Thus, by Integral Test, given series is convergent.
n=2
n(ln2 n)
P P
COMPARISON TEST I Let an be a given series of positive terms to be tested. If dn is a (known)
P
divergent series of positive terms such that di ≤ ai for infinitely many i’s, then an is divergent.
Note Roughly speaking, if (smaller series) ≤ (larger series) and (smaller series) diverges, then (larger
series) diverges!
P P
COMPARISON TEST II Let an be a given series of positive terms to be tested. If cn is
P
a (known) convergent series of positive terms such that ci ≥ ai for infinitely many i’s, then an is
convergent.
Note Roughly speaking, if (larger series) ≥ (smaller series) and (larger series) converges, then (smaller
series) converges!
1
P
Example Test the convergence of the series n2 +n .
Solution Clearly, n2 + n ≥ n2 for all n ≥ 1. Then, n21+n ≤ n12 for all n ≥ 1. Now, 1
P
n2 is conver-
P 1
gent(known p-series). So, by Comparison Test II, n2 +n is convergent.
Caution! In the above example, if we compare like n ≤ n2 + n for all n ≥ 1. Then n1 ≥ n21+n . Now,
P1
n
P 1
is divergent (larger series), then we can NOT say (smaller series) n2 +n is divergent! (Wrong usage of
Comparison Test!).
LIMIT COMPARISON TEST

an P P
If lim = c, (c ̸= 0 and c < ∞), then the series an and bn both converge or both diverge.
n−→∞ bn
∞
1 1 1 1 X 1
Example Test the convergence of the series + + + + ... = n−1
.
1 3 7 15 n=1
2
Solution Let an = 2n1−1 . For larger n, the behavior of an is like 21n so we take bn = 21n . Note that
P P 1 an 2n 1 1
bn = 2 n , which converges. Now, lim = lim = lim = = 1 > 0.
n−→∞ bn n−→∞ 2n − 1 n−→∞ 1 − (1/2n ) 1−0
P P
Since bn converges and limit is 1 > 0 and by limit comparison test, an converges.
P
Note Always try bn is known (simple) series (i.e., series whose convergence or divergence is known)!

P an
D’ALEMBERT’S RATIO TEST Let an be a series of positive terms. Let L = lim .
n−→∞ an+1
Suppose that the limit exists (a finite number). Then the given series converges if L > 1 and the series
diverges if L < 1.
Note If L = 1 or the limit fails to exist, then the test is inconclusive (i.e., there exist both convergent
and divergent series satisfying this case. Try yourself such series! ).
Xn+1
Example The series is convergent.
en
(n+1)+1
Solution Let an = n+1
en . Then, an+1 = en+1 .
an n + 1 en+1 n + 1 en · e1
= n
· = ·
an+1 e n+2 en n+2
n+1 n(1 + 1/n) (1 + 1/n)
= e· = e· = e·
n+2 n(1 + 2/n) (1 + 2/n)

an

1 + 1/n
lim (1 + 1/n) 1+0
lim = lim e · = e · n−→∞ =e· =e
n−→∞ an+1 n−→∞ 1 + 2/n lim (1 + 2/n) 1+0
n−→∞

an
Thus, L = lim = e > 1. From D’ Alembert’s Ratio test, given series converges.
n−→∞ an+1
2 2 2
1 1·2 1·2·3
Example Test convergence of the series + + + ...
3 3·5 3·5·7
2 2
1·2·3···n 1 · 2 · 3 · · · n · (n + 1)
Solution Let an = and an+1 = .
3 · 5 · 7 · · · (2n + 1) 3 · 5 · 7 · · · (2n + 1) · (2n + 3)
2
2 + n3 2

an 2n + 3
= =
an+1 n+1 1 + n1
2 + n3 2 2 + n3 1
3 1
an 2+ n
lim = lim = lim · lim
n−→∞ an+1 n−→∞ 1 + 1 n−→∞ 1 + 1 n−→∞ 1 + 1
n n n
2+0 2+0
= · = 2.2 = 4 > 1
1+0 1+0

an
Thus, L = lim > 1.
n−→∞ an+1
Hence by D’Alembert’s ratio test, given series is convergent.

P an
JOSEPH LUDWIG RAABE’S TEST Let an be a series. Let ℓ = lim−1 . n·
n−→∞ an+1
Suppose that the limit exists (a finite number). Then the given series is convergent if ℓ > 1. The given
series is divergent if ℓ < 1.
Proof Case1 ℓ = 1

an
Choose a positive real number ϵ such that ℓ − ϵ > 1. Since lim n · −1 = ℓ, there exists a
n−→∞ a
n+1
an an
natural number m such that | n · an+1 − 1 − ℓl| < ϵ and hence ℓ − ϵ < n · an+1 − 1 < ℓ + ϵ for all
n ≥ m. Let ℓ − ϵ = r. Then r > 1.
We have nan − nan+1 > ran+1 for all n ≥ m,
or, nan − (n + 1)an+1 > (r − 1)an+1 for all n ≥ m.
We have mam − (m + 1)am+1 > (r − 1)am+1
(m + 1)am+1 − (m + 2)am+2 > (r − 1)am+2
..... ....... ......... .............

(n − 1)an−1 − nan > (r − 1)an . Where n > m.
Consequently, mam − nan > (r − 1)(am+1 + am+2 + .......... + an ) for all n > m.
1
or, am+1 + am+2 + .......... + an < r+1 (mam − nan )
1
< r+1 mam
1
or, sn − sm < r+1 mam wheresn = a1 + a2 + ....... + an
1
or, sn < r+1 mam + sm for all n > m.
P
This shows that the sequence (sn ) is bounded above and therefore the series an is convergent.
Case2 ℓ < 1
Let us choose a positiveϵ such that ℓ + ϵ < 1. There exists a natural number k such that ℓ − ϵ <

n · aan+1
n
−1 < ℓ + ϵ for all n ≥ k.

Let ℓ − ϵ = p. Then p < 1. We have n · aan+1
n
−1 < p < 1 for all n ≥ k.
Therefore n(an − an+1 ) < pan+1 < an+1 for all n ≥ k
Or, nan < (n + 1)an+1 for all n ≥ k.
We have kak < (k + 1)ak+1
(k + 1)ak+1 < (k + 2)ak+2
....... ......... .........
(n − 1)an−1 < nan where n ≥ k
Consequently, nan > kak for all n > k (or) equivalently, an > kak n1
P1 P∞ 1
P∞
kak is a positive number and n diverges =⇒ n=1 (kak ) n diverges =⇒ n=1 an diverges (using
comparison test).
P
Therefore, an is divergent by Comparison test.
Note If ℓ = 1 or the limit fails to exist, then the test is inconclusive (there exist both convergent and
divergent series satisfying this case. Try yourself such series!)
Note Most of the cases, Rabee’s test is useful when the ratio test fails. See the following example.
1 1·5 1·5·9 1 · 5 · 9 · 13
Example Test the convergence of the series + + + + ...
3 3 · 7 3 · 7 · 11 3 · 7 · 11 · 15
Solution
1 · 5 · 9 · 13 · · · (4n − 3)
an =
3 · 7 · 11 · 15 · · · (4n − 1)
1 · 5 · 9 · 13 · · · (4n − 3) · (4n + 1)
an+1 =
3 · 7 · 11 · 15 · · · (4n − 1) · (4n + 3)
an 4n + 3 n · (4 + n3 ) (4 + n3 )
= = =
an+1 4n + 1 n · (4 + n1 ) (4 + n1 )

3
lim 4 +
an n−→∞ n 4+0
L = lim = = = 1.
n−→∞ an+1 1 4+0
lim 4 +
n−→∞ n
Thus, D’Alembert’s ratio test fails. Now, let us try, Rabee’s Test.
an (4 + n3 ) (4 + n3 ) − (4 + n1 ) 2
n
−1 = − 1 = =
an+1 (4 + n1 ) (4 + n1 ) (4 + n1 )
2
an n 2
n· −1 = n· 1 =
an+1 (4 + n) (4 + n1 )

an 2 2 2 1
ℓ = lim n · −1 = lim = = = < 1.
n−→∞ (4 + 1 )

n−→∞ an+1 n
1 4+0 2
lim 4 +
n−→∞ n
Therefore, by Raabe’s test, given series diverges.
an O(n) an
Note Rabee’s test is applicable only if = . i.e., both numerator and denominator of
an+1 O(n) an+1
have linear expression in n such as 3n − 2, 5n + 2, n − 1. Otherwise, Raabe’s test also fails.
Note In case Raabe’s test fails. Then, Gauss’ Test may be applicable. (Optional!)

P an ρ 1 P
GAUSS’ TEST Let an be given series. Suppose that = 1 + + O 2 . Then an converges
an+1 n n
if ρ > 1 and diverges if ρ ≤ 1. This test never fails.
P 1
CAUCHY’S ROOT TEST Let an be a given series. Let l = lim ann . Suppose that the limit
n−→∞
exists (a finite number). Then the given series is convergent if l < 1 and is divergent if l > 1.
Note If l = 1 or the limit fails to exist, then the test is inconclusive (there exist both convergent and
divergent series when l = 1).
2 3
1 2 3 2 4
Example Test the series + x+ x + x3 + ... where x > 0.
2 3 4 5
Solution By omitting or adding few terms of the given series, the convergence or divergence of given
series does not affect. Here, just omit 12 .
n
n+1
an = · xn
n+2
n · (1 + n1 ) 1

1 n+1 1+ n
(an ) n = ·x= 2 · x = 2 ·x
n+2 n · (1 + n ) 1+ n
1
1 1+ n
l = lim (an ) n = lim · x = x.
n−→∞ n−→∞ 1 + 2
n
Case (i) When l = x > 1, the given series diverges.

Case (ii) When l = x < 1, the given series converges.
Case (iii) When l = x = 1, Cauchy’s root test fails.
2 3
1 2 3 4
So, put x = 1, then given series becomes + + + + ...
2 3 4 5
n 1 n
(1 + n1 )n

n+1 1+ n
an = = 2 =
n+2 1+ n (1 + n2 )n
n
1
lim 1 +
n−→∞ n e1 1
lim an = n = 2 = ̸= 0.
n−→∞ 2 e e
lim 1 +
n−→∞ n
Therefore, lim an ̸= 0, by test for divergence, given series diverges.
n−→∞
Note So far, the above series we discussed are all series of positive terms. In next section, all the series
we are going to study are series which contains both positive and negative terms.
ALTERNATING SERIES A series whose terms are alternatively positive and negative numbers. Let
∞
X
(−1)n+1 an , where an > 0 for all n ≥ 1, is called an alternating series.
n=1
∞
X
n+1 1 1 1 1 1 1
Example (−1) = (−1)2 . + (−1)3 . + (−1)4 . + ... = 1 − + − ...
n=1
n 1 2 3 2 3
∞
X
LEIBNIZ’S TEST Let (−1)n+1 an be an alternating series to be tested. The alternating series is
n=1
convergent if the following 2 conditions are satisfied.
I. (an ) is a monotonically decreasing sequence. i.e., a1 ≥ a2 ≥ a3 ... ≥ an ≥ an+1 ≥ ...
II. lim an = 0.
n−→∞
Note In the above test, if lim an ̸= 0, then given series oscillates.

n−→∞
Example Test the convergence of the (alternating) series 1 − 12 + 13 − 14 + 51 − ....

∞
1 1 1 1 1 1 1 1 1 X 1
Solution 1− + − + −... = (−1)2 +(−1)3 +(−1)4 +(−1)5 +(−1)6 +... = (−1)n+1 .
2 3 4 5 1 2 3 4 5 n=1
n
Now, (an ) = ( n1 ) is monotonically decreasing i.e., 1 > 21 > 13 > 14 > ... Condition I of Leibniz’s Test is
1
satisfied. Now, lim an = lim = 0. Condition II of Leibniz’s Test is satisfied. Thus it converges.
n−→∞ n−→∞ n
Example Test the convergence of the (alternating) series 1 − 1 + 1 − 1 + 1 − 1 + ....

∞
X
Solution 1−1+1−1+1−1+... = (−1)n+1 ·1. Now, (an ) is 1, 1, 1, ... and lim an = lim 1 = 1 ̸= 0.
n−→∞ n−→∞
n=1
Condition II of Leibniz’s Test is not satisfied. Thus, given series is divergent.
ABSOLUTE CONVERGENCE Let an > 0 for all n ≥ 1. An alternating series a1 − a2 + a3 − a4 + ....

is said to be absolute convergent if a1 + a2 + a3 + a4 + ... is convergent. (or)
∞
X ∞
X ∞
X
n+1 n+1
The alternating series (−1) an is absolute convergent if |(−1) an | = |an | is convergent.
n=1 n=1 n=1
P P
ABSOLUTE CONVERGENCE TEST If |an | is convergent, then an is also convergent.
Note (i) absolute convergence ⇒ convergent.

(ii) convergence ̸⇒ absolute convergence.
1 1 1 1
For example, the series 1 − 2 + 3 − 4 + 5 − ... is convergent but not absolute convergence.
CONDITIONALLY CONVERGENCE Let an > 0 for all n ≥ 1. An alternating series a1 − a2 +

a3 − a4 + .... is said to be conditionally convergent if given series is convergent but a1 + a2 + a3 + a4 + ...
is not convergent. (or)
X∞ ∞
X
n+1
The alternating series (−1) an is conditionally convergent if (−1)n+1 an is convergent but
n=1 n=1
∞
X ∞
X
n+1
|(−1) an | = |an | is divergent.
n=1 n=1
∞
X 1 1 1 1 1 1
Example (−1)n+1 = (−1)2 . + (−1)3 . + (−1)4 . + ... = 1 − + − ... is convergent by
n=1
n 1 2 3 2 3
∞
X 1 1 1 1
Leibniz’s test. But (−1)n+1 = + + + ... is divergent series.
n=1
n 1 2 3
CERTAIN STRATEGY TO PREDICT THE CONVERGENCE OF A SERIES (optional)
We now have lot of tests to verify the convergence of a series; the problem is to decide which test to use
on which series. In this respect testing series is similar to integrating functions. Again there are no hard
and fast rules about which test to apply to a given series, but you may find the following suggestions of
some use. It is not advisable/wise to apply a list of the tests in a specific order until one finally works.
That would be a waste of time and effort. Instead, the main strategy is to classify the series according
to its form.
1. If you see that lim an ̸= 0, then by Test for Divergence, given series diverges.
n−→∞
1
P
2. If the series looks like the famous p-series form np , then it is convergent when p > 1 and divergent
when p ≤ 1.
arn−1 , then it converges when

P
3. If the series looks like another famous geometric series the form
0 < r < 1 and it diverges when r ≥ 1. Some preliminary algebraic manipulation may be required
to bring the series into this form.
4. If the series has a form that is similar to a p-series or a geometric series, then one of the comparison
tests should be considered. In particular, if an is a rational function or algebraic function of n
(involving roots of polynomials), then the series should be compared with a p-series. (The value of
p should be chosen by keeping only the highest powers of n in the numerator and denominator.)
P
The comparison tests apply only to series with positive terms, but if an has some negative terms,
P
then we can apply the Comparison Test to |an | and test for absolute convergence.
(−1)n+1 an , then Leibniz’s Test is useful (one and only only

P
5. If the series is alternating series
choice!).
6. Series that involve factorials or other products (including a constant raised to the nth power) are
often conveniently tested using the D’Alembert’s Ratio Test. Keep in mind that | aan+1
n
| −→ 1 as
n −→ ∞ as for all p-series and therefore all rational or algebraic functions of n. Thus, the Ratio
Test should not be used for such series.
Z ∞
7. If an = f (n) , where f (x)dx is easy to calculate, then the Cauchy’s Integral Test is effective
a
(assuming the initial conditions of this this test are satisfied).

Convergence of Sequence and Series

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Convergence of Sequence and Series

Uploaded by

Copyright:

Available Formats

MA1001D Mathematics I Monsoon 2022-23

DEPARTMENT OF MATHEMATICS, NIT CALICUT

CONVERGENCE OF REAL SEQUENCES AND INFINITE SERIES

Handout for MA1001D - Module II (Batch D)

Dr. Krishnan Paramasivam, Course Teacher

GOTTFRIED LEIBNIZ Born in Leipzig, Germany, (1646-1716); invented infinitesimal calculus

AUGUSTIN-LOUIS CAUCHY Born in Paris, France, (1789-1857), single-handed development of

SEQUENCE, CAUCHY SEQUENCE AND CONVERGENCE

• N = {1, 2, 3, · · · } is set of all positive integers, R is set of all real numbers.

• Open interval (0, 1) = {x ∈ R : 0 < x < 1}, closed interval [0, 1] = {x ∈ R : 0 ≤ x ≤ 1} =

• Both −∞ and ∞ are not real numbers.

• Since f (1) = a1 , f (2) = a2 , f (3) = a3 , ·, f (1000) = a1000 , ·, we denote a sequence a1 , a2 , a3 · · · a100000 , · · ·

it just represents n-th term of the sequence (an ).

or, equivalently, at − L < ϵ, · · · for all t > k (that is t = k + 1 or t = k + 2, or t = k + 3, · · · ) (that

• The real number L is called the limit of the sequence (an ).

• Let ϵ = 0.1, a1 = 100, a2 = 200, a3 = 300, a4 = 400, · · · a9 = 900, a10 = 1

• Example. Let the real sequence (an ): a1 = 1, a2 = 5, a3 = 1, a4 = 5, a5 = 1, a6 = 5, a7 = 1, a8 =

for all m, n > k.

• Monotonically Increasing Sequence. A sequence (an ) is monotonically increasing (or shortly,

or equivalently, am ≤ am+1 for all m = k + 1, k + 2, k + 3, k + 4, · · ·

Prove or disprove: (an ) is monotonically increasing sequence.

• Monotonically Decreasing Sequence. A real sequence (an ) is monotonically decreasing (or

or equivalently, am ≥ am+1 for all m = k + 1, k + 2, k + 3, k + 4, · · ·

• Theorem A sequence (an ) converges if and only if (an ) is Cauchy sequence.

• Sequence of partial sums of a Series

• Note. The series of positive terms

NECESSARY & SUFFICIENT CONDITION FOR A SERIES TO CONVERGE

INTEGRAL TEST Let a ∈ N. Let f be a continuous, positive, decreasing function on defined on

LIMIT COMPARISON TEST

Hence by D’Alembert’s ratio test, given series is convergent.

..... ....... ......... .............

Case (i) When l = x > 1, the given series diverges.

I. (an ) is a monotonically decreasing sequence. i.e., a1 ≥ a2 ≥ a3 ... ≥ an ≥ an+1 ≥ ...

Note In the above test, if lim an ̸= 0, then given series oscillates.

Example Test the convergence of the (alternating) series 1 − 12 + 13 − 14 + 51 − ....

Example Test the convergence of the (alternating) series 1 − 1 + 1 − 1 + 1 − 1 + ....

ABSOLUTE CONVERGENCE Let an > 0 for all n ≥ 1. An alternating series a1 − a2 + a3 − a4 + ....

Note (i) absolute convergence ⇒ convergent.

CONDITIONALLY CONVERGENCE Let an > 0 for all n ≥ 1. An alternating series a1 − a2 +

CERTAIN STRATEGY TO PREDICT THE CONVERGENCE OF A SERIES (optional)

arn−1 , then it converges when

(−1)n+1 an , then Leibniz’s Test is useful (one and only only

You might also like