Вы находитесь на странице: 1из 15

# Convergence of Series

DEFINITION 1 (Series). Let (an ) be a sequence of real numbers. Then an expression of the form

X
a1 + a2 + a3 + · · · + an + · · · denoted by an , is called a series.
n=1

X 1 1 1 1
Example 1. = 1 + + + ··· + + ···
n 2 3 n
n=1

X 1 1 1 1
Example 2. 2
= 1 + + + ··· + 2 + ···
n 4 9 n
n=1

## DEFINITION 2 (Partial Sums). Sn = a1 + a2 + a3 + · · · + an is called the nth partial sum of the

X
series an .
n=1

X
DEFINITION 3 (Convergence). An infinite series an = a1 + a2 + a3 + · · · + an + · · · converges
n=1
to S if the sequence of partial sums

sn = a1 + a2 + · · · an

X
converges to a limit S. If (Sn) does not converge, then we say that the series an diverges.
n=1

X 1 1 1 1
Example 3. = 1 + + + · · · + + · · · does not converges.
n 2 3 n
n=1

X 1 1 1 1
Example 4. = 1 + + + · · · + 2 + · · · converges.
n2 4 9 n
n=1

## Necessary condition for convergence

X
THEOREM 1. If an converges then an → 0.
n=1
n
X n
X
Proof. As Sn = an and Sn+1 = an , Sn+1 − Sn = an+1 .
n=1 n=1

X
Given that an converges to S, so Sn → S and Sn+1 → S
n=1
So, Sn+1 − Sn = S − S = 0. Thus, an+1 → 0 or we can say an → 0.

USE 1. The test says that if the terms ai do not go to zero, then there is no way for the series of
partial sums to converge. So, you can claim series Does NOT converge.

X 1 1
Example 5. converges and also n2
→0
n2
n=1

0-1

X
Example 6. 2n diverges as 2n 9 0
n=1

X
Example 7. sin n diverges as sin n 9 0
n=1

X n
Example 8. Show that diverges.
n+1
n=1
We compute the limit:
n
= 1 6= 0.
lim
n+1 n→∞

Looking at the first few terms perhaps makes it clear that the series has no chance of converging:
1 2 3 4
+ + + + ···
2 3 4 5
will just get larger and larger; indeed, after a bit longer the series starts to look very much like
· · · + 1 + 1 + 1 + 1 + · · · , and of course if we add up enough 1’s we can make the sum as large as we
desire.

X n2
Question 1. 1. Explain why diverges.
2n2 + 1
n=1
2 2
Hint: lim n /(2n + 1) = 1/2
n→∞

X 5
2. Explain why diverges.
n=1
21/n + 14
Hint: lim 5/(21/n + 14) = 1/3
n→∞

X 3
3. Explain why diverges.
n
n=1

X
P∞ 1 1
Hint: n=1 n diverges, so 3 diverges
n
n=1

X
Remark. Converse of this result does not holds, that is, if an → 0 does not mean an converges.
n=1

X 1 1
Example 9. diverges, however, n →0
n
n=1

X n+1 n+1
Example 10. log diverges, however, n →0
n
n=1

## Necessary and sufficient condition for convergence

X
THEOREM 2. Suppose an ≥ 0∀n. Then, an converges if and only if (Sn ) is bounded above.
n=1

Proof. As an ≥ 0∀n, (Sn ) is an increasing sequence. Moreover, every increasing and bounded above
X∞
sequence converges. So, an converges.
n=1

0-2
USE 2. The test says that if the positive term series have sequences of partial sum bounded above,
then the series always converge. So, you can claim series convergence using bounded above condition.

X
Example 1. If an = kxn , an is called a geometric series. A typical partial sum is
n=0

## sn = k + kx + kx2 + kx3 + · · · + kxn = k(1 + x + x2 + x3 + · · · + xn ).

We note that

sn (1 − x) = k(1 + x + x2 + x3 + · · · + xn )(1 − x)
= k(1 + x + + x3 + · · · + xn )1 − k(1 + x + x2 + x3 + · · · + xn−1 + xn )x
x2
= k(1 + x + x2 + x3 + · · · + xn − x − x2 − x3 − · · · − xn − xn+1 )
= k(1 − xn+1 )

so

sn (1 − x) = k(1 − xn+1 )
n+1
sn = k 1−x
1−x .

## If |x| < 1, lim xn = 0 so

n→∞
1 − xn+1 1
lim sn = lim k =k .
n→∞ n→∞ 1−x 1−x
Thus, when |x| < 1 the geometric series converges to k/(1 − x). When, for example, k = 1 and
x = 1/2:

X
1 − (1/2)n+1 2n+1 − 1 1 1 1
sn = = n
= 2 − n and n
= = 2.
1 − 1/2 2 2 2 1 − 1/2
n=0

## We began the chapter with the series

X 1
,
2n
n=1

namely, the geometric series without the first term 1. Each partial sum of this series is 1 less than
the corresponding partial sum for the geometric series, so of course the limit is also one less than the
value of the geometric series, that is,
X∞
1
= 1.
2n
n=1

X
Example 11. For 0 < x < 1, the following series converges xn as an ≥ 0∀n and (Sn ) is bounded
n=1
1
above, which could be easily check as |Sn | ≤ 1−x

X∞
1
Example 12. The Harmonic series diverges because
n
n=1
1 1 1 1 k
S2k ≥ 1 + + 2 + 4 + · · · + 2k−1 k = 1 + , ∀k
2 4 8 2 2

X 1
Example 2. Show that diverges.
n
n=1

0-3
Here the theorem does not apply: lim 1/n = 0, so it looks like perhaps the series converges. Indeed,
n→∞
if you have the fortitude (or the software) to add up the first 1000 terms you will find that
1000
X 1
≈ 7.49,
n
n=1

so it might be reasonable to speculate that the series converges to something in the neighborhood of
10. But in fact the partial sums do go to infinity; they just get big very, very slowly. Consider the
following:
1 1 1 1 1 1 1 1
1+ + + >1+ + + =1+ +
2 3 4 2 4 4 2 2
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
1+ + + + + + + >1+ + + + + + + =1+ + +
2 3 4 5 6 7 8 2 4 4 8 8 8 8 2 2 2
1 1 1 1 1 1 1 1 1 1 1 1 1 1
1 + + + ··· + > 1 + + + + + ··· + + + ··· + =1+ + + +
2 3 16 2 4 4 8 8 16 16 2 2 2 2
1 1 1 1 1 1 1 1
1+ + + >1+ + + =1+ +
2 3 4 2 4 4 2 2
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
1+ + + + + + + >1+ + + + + + + =1+ + +
2 3 4 5 6 7 8 2 4 4 8 8 8 8 2 2 2
1 1 1 1 1 1 1 1 1 1 1 1 1 1
1 + + + ··· + > 1 + + + + + ··· + + + ··· + =1+ + + +
2 3 16 2 4 4 8 8 16 16 2 2 2 2
and so on. By swallowing up more and more terms we can always manage to add at least another
1/2 to the sum, and by adding enough of these we can make the partial sums as big as we like. In
fact, it’s not hard to see from this pattern that
1 1 1 n
1+ + + ··· + n > 1 + ,
2 3 2 2
so to make sure the sum is over 100, for example,
P we’d add up terms until we get to around 1/2198 ,
59
that is, about 4 · 10 terms. This series, (1/n), is called the harmonic series.

The Cauchy convergence test is a method used to test infinite series for convergence. It relies on
bounding sums of terms in the series. This convergence criterion is named after Augustin-Louis
Cauchy who published it in his textbook Cours d’Analyse 1821.

## Definition 1 (Cauchy criterion for series). A series

X
ai is convergent if and only if for every ε > 0ε > 0 there is a natural number n0 such that
i=0
|an+1 + an+2 + · · · + an+p | < ε holds for all n > n0 and p ≥ 1.

X ∞
X
THEOREM 3. If |an | converges, then an converges.
n=1 n=1

X ∞
X
Proof. Since |an | converges the sequence of partial sums of |an | satisfies the Cauchy criterion,
n=1 n=1
i.e. ||an+1 | + |an+2 | + · · · + |an+p || < ε
Moreover, |an+1 + an+2 + · · · + an+p | ≤ ||an+1 | + |an+2 | + · · · + |an+p ||
⇒ |an+1 + an+2 + · · · + an+p | < ε

X ∞
X
Therefore, the sequence of partial sums of an satisfies the Cauchy criterion. Hence, an con-
n=1 n=1
verges.

X ∞
X
1 (−1)n
Example 13. converges, thus .
n2 n2
n=1 n=1

0-4
TEST 1 (Integral Test). Let ai = f (i), where f(x) is a continuous function with f (x) > 0, and is
decreasing. Then

X Z ∞
the series ai converges if the improper integral f (x)dx < ∞.
i=1 1

X Z ∞
the series ai diverges if the improper integral f (x)dx = ∞.
i=1 1

USE 3. Solution of improper integral can help us to tell about convergence of series

X∞
1
converges if p > 1, and diverges if p ≤ 1
np
n=1
P P∞
TEST 2 (Comparison Test). Suppose that ∞ 1 ai and 1 are series with all terms positive - so
ai ≥ 0 and bi ≥ 0.

X ∞
X
bi is convergent and ai ≤ bi for all i =⇒ ai is convergent.
i=1 i=1

X ∞
X
bi is divergent and ai ≥ bi for all i =⇒ ai is divergent.
i=1 i=1

USE 4. This is the “squeeze test” for infinite series. Use it to justify the “cover-up” method of
guessing whether a series converges or diverges.

X ∞
X
1 1 1 1
Example 14. 1. converges because (n+2)(n+3) ≤ (n)(n) and converges.
(n + 2)(n + 3) n2 )
n=1 n=1

X ∞
X
1 1 1
2. p diverges because n ≤ √1 and diverges.
(n + 3) (n+3 n
n=1 n=1

X 1
3. converges as n2 < n!, ∀n > 3
n!
n=1

X ∞
X
1 1 1 1
4. For given p > 1, diverges because n ≤ (log n)p and diverges.
(log n)p n
n=1 n=1

X ∞
X
TEST 3 (Limit Comparison Test). Suppose that ai and bi are series with all terms positive.
1 1

X ∞
X
ai
lim =c>0 =⇒ ai and bi either both converge, or both diverge.
i→∞ bi
i=1 i=1

X ∞
X
ai
lim = 0 and ai converges =⇒ the series bi converges.
i→∞ bi
i=1 i=1

X ∞
X
ai
lim = ∞ and bi diverges =⇒ the series ai diverges.
i→∞ bi
i=1 i=1

USE 5. This is one of the most powerful tests, because it squeezes the two series “in the limit”. Just
be sure to use it right! Part one is clear, but don’t mix up the second and third parts.

0-5

X 1 1
Example 15. 1. (1 − nsin ) converges. Take bn = n2
and use Limit Comparison Test.
n
n=1

X 1 1 1
2. log(1 + ) converges. Take bn = n2
and use Limit Comparison Test.
n n
n=1

THEOREM 4 (Cauchy Test or Cauchy condensation test). If an ≥ 0 and an+1 ≤ an ∀n, then

X ∞
X
an converges if and only if 2k a2k converges.
n=1 n=1

X 1
Example 16. 1. converges if p > 1 and diverges if p ≤ 1.
np
n=1

X ∞ ∞
1 p X 1 (1−p)k X
Hint: Check 2k (
) = ( ) , Which is a geometric series rk , with r = ( 21 )(1−p) .
2k 2
k=0 k=0 k=1
It converges for 0 < r < 1, otherwise diverges.

X 1
2. converges if p > 1 and diverges if p ≤ 1.
n(log n)p
n=2
P P∞ 1
Hint: Check ∞ k 1 p
k=0 2 ( 2k log 2k ) = k=0 kp

X 1
3. diverges.
n(log n)(log(log n))
n=4

Definition 2 (Alternating Series). For all ai > 0, the series of the form

X
(−1)i+1 ai = a1 − a2 + a3 − a4 + a5 − a6 + · · ·
i=1

## is called Alternating Series.

TEST 4 (Leibniz Test). For the alternating series - where all ai > 0

X
(−1)i ai = a1 − a2 + a3 − a4 + a5 − a6 + · · ·
i=1

X
ai ≥ ai+1 for all i and lim ai = 0 =⇒ (−1)i ai converges.
i→∞
i

X (−1)i 1
Example 17. 1. converges as lim = 0 and series is alternative series
i i→∞ i
i=1

X∞
cosiπ 1
2. √ converges as lim √ = 0 and series is alternative series.
i=1
i i→∞ i

X 1 1
3. (−1)i sin( ) converges as lim sin( ) = 0 and series is alternative series
i i→∞ i
i=1

X
TEST 5 (Ratio Test). Consider the series an , an 6= 0 ∀n
n=1

0-6

X ∞
X
ai+1
lim =L<1 =⇒ |ai | converges =⇒ ai converges.
i→∞ ai
i=1 i=1

X
ai+1
lim =L>1 =⇒ ai diverges.
i→∞ ai
i=1

Example 18. The ratio test is particularly useful for series involving the factorial function. Consider

X
5n /n!.
n=0
5n+1 n! 5n+1 n! 1
lim n
= lim n
= lim 5 = 0.
n→∞ (n + 1)! 5 n→∞ 5 (n + 1)! n→∞ (n + 1)
Since 0 < 1, the series converges.

Example 3. If
X X nπ  n
an = 2 + sin r ,
2
an+1 2 + sin (n+1)π
2
= r
an 2 + sin nπ
2

which assumes the values 3r/2, 2r/3, r/2, and 2r, each infinitely many times; hence,
an+1 an+1 r
lim sup = 2r and lim inf = .
n→∞ an n→∞ an 2
P
Therefore, an converges if 0 < r < 1/2 and diverges if r > 2. The ratio test is inconclusive if
1/2 ≤ r ≤ 2.

X
TEST 6 (Root Test). Consider the series an , an ≥ 0 ∀n
n=1

X ∞
X

lim n
an = L < 1 =⇒ |an | converges =⇒ an converges.
n→∞
n=1 n=1

X

lim n
an = L > 1 =⇒ an diverges.
n→∞
n=1

X∞
5n
Example 19. Analyze .
nn
n=0
The ratio test turns out to be a bit difficult on this series (try it). Using the root test:
 1/n
5n (5n )1/n 5
lim = lim = lim = 0.
n→∞ nn n→∞ (nn )1/n n→∞ n

## Since 0 < 1, the series converges.

Example 4. If
X X nπ n n
an = 2 + sin r ,
4
then  nπ 
lim sup a1/n
n = lim sup 2 + sin r = 3r,
n→∞ n→∞ 4
P P
and so an converges, if r < 1/3 and
P an diverges if r > 1/3. The test is inconclusive if r = 1/3,
but then |a8m+2 | = 1 for m ≥ 0, so an = ∞

0-7
Question 2. Determine whether the series converge.

X 3n
1. (−1)n
5n
n=0

X n!
2.
nn
n=1

X n5
3.
nn
n=1

X (n!)2
4.
nn
n=1

0-8
Summary of Convergence and Divergence Tests for Series
Test Series Convergence or Divergence Comments
P
1
Divergence an Diverges if lim an 6= 0 Inconclusive if lim an = 0
n=1 n!1 n!1

## Useful for comparison tests

P
1
n a
Geometric ar (1) Converges to S = if |r| < 1; if the nth -term an of a series
n=0 1 r
series (2) Diverges if |r| > 1 is similar to arn .

## Useful for comparison tests

P1 1
p-series p
(1) Converges if p > 1; if the nth -term an of a series
n=1 n
1
(2) Diverges if p 6 1 is similar to .
np

## The function f obtained

P
1 R1
Integral an (1) Converges if 1 f (x) dx converges; from an = f (n) must be
n=1 R1
an = f (n) (2) Diverges if 1 f (x) dx diverges continuous, positive,
integrable.

P
1
(1) If bn converges and an 6 bn The comparison series
n=1
P
1 P
1
for every n, then an converges; bn is often a geometric
n=1 n=1
P
1 P
1 P
1
Comparison an , bn (2) If bn diverges and an > bn series or a p-series. To find
n=1 n=1 n=1
P1
an > 0, bn > 0 for every n, then an diverges; bn in (3), consider only
n=1
an
(3) If lim > 0 (not 1), then the terms of an that have
n!1 bn
both series converge or both diverge. the greatest e↵ect on the
magnitude.

Inconclusive if L = 1. Useful
an+1
If lim = L (or 1), the series if an involves factorials or
n!1 an
P
1
Ratio an (1) converges (absolutely) if L < 1; nth powers. If an > 0 for
n=1
(2) diverges if L > 1 (or 1) every n, the absolute value
sign may be disregarded.

p Inconclusive if L = 1. Useful
If lim n |an | = L (or 1), the series if an involves nth powers.
n!1
P
1
Root an (1) converges (absolutely) if L < 1; If an > 0 for every n, the
n=1
(2) diverges if L > 1 (or 1) absolute value sign may be
disregarded.

P
1
Alternating ( 1)n an Converges if ak > ak+1 for every k Applicable only to an
n=1
Series an > 0 and lim an = 0. 0-9 alternating series.
n!1
Practice Problems on Convergence of Series

For each of the following, say whether it converges or diverges and explain why.
P∞ n3
1. n=1 n5 +3
n3 n3 1
5
< 5
= 2
n +3 n n
P 1 P n3
for all n. Therefore, since n2 converges (it’s a p-series with p = 2 > 1), the series n5 +3 also
converges by the comparison test.
P∞ 3n
2. n=1 4n +4
3n 3n 3
< =
4n + 4 4n 4
P 3 n 3
P 3n
for all n. Therefore, since 4 converges (it’s a geometric series with r = 4 < 1), the series 4n +4
also converges by the comparison test.
P∞ n
3. n=1 2n
r √ √
n
n n
n n
n 1
lim n = lim √ = lim = .
n→∞ 2 n→∞ n
2n n→∞ 2 2
Since the limit is less than 1, the Root Test says that the series converges absolutely.
P∞ np
4. For what values of p does the series n=1 2+n 3 converge?
P 1
Answer: Doing a limit comparison to n3−p , I see that

np
2+n3 n3 1
lim 1 = lim 3
= .
n→∞ 3−p n→∞ 2 + n 2
n
P 1
Therefore, the series converges if and only if the series n3−p converges. This happens when 3−p > 1,
which is to say when p < 2. So the given series converges when p < 2.
P∞
5. We would like to estimate the sum of the series n=1 n41+3 by using the sum of the first ten terms. Of
P∞
course, the exact error is the sum of all the terms from the 11th on, i.e., n=11 n41+3 . Show that this
error is less than 1/3000 by comparing this with the sum of 1/n4 and then by estimating this latter
sum using an appropriate integral.
1 1
< 4
n4 +3 n
for all n, so

X ∞
X
1 1
< .
n=11
n + 3 n=11 n4
4

Z ∞ ∞
1 −1 1
4
dx = = ,
10 x 3x3 10 3000

## so we see that the error is less than 1/3000.

6. Does the series
X∞
n!(n + 1)!
n=1
(3n)!
converge or diverge?

(n+1)!(n+2)!
(3n+3)! (n + 1)!(n + 2)! (3n)! (n + 1)(n + 2)
lim n!(n+1)! = lim · = lim
n→∞ n→∞ (3n + 3)! n!(n + 1)! n→∞ (3n + 3)(3n + 2)(3n + 1)
(3n)!

n2 + 3n + 2
= lim .
n→∞ 27n3 + 54n2 + 33n + 6

## Dividing numerator and denominator by n3 yields

1 3 2
n + n2 + n3
lim 54 33 6 = 0.
n→∞ 27 + n + n2 + n3

Since 0 < 1, the Ratio Test says that the series converges absolutely.
7. Does the series  

X
n1
(−1) cos
n=1
n
converge absolutely, converge conditionally, or diverge?
     
1 1 1
lim cos = lim cos = cos lim = cos(0) = 1
n→∞ n x→∞ x x→∞ x

## since cosine is a continuous function. Therefore, the terms

 
n 1
(−1) cos
n
are not going to zero, so the Divergence Test says that the series diverges.
8. Check the convergence of the series
X∞
n3
n=0
n4 + 1

Answer: Using the comparision Test, one can see it is similar to the following

X 1

n=0
n
(−1)n +n
9. Consider the sequence defined by an = (−1)n −n . Does this sequence converge and, if it does, to what
limit?
Answer: Dividing numerator and denominator by n, we have that
1 (−1)n
(−1)n + n n ((−1)n + n) n +1 1
lim = lim 1 = lim (−1)n
= = −1,
n→∞ (−1)n − n n→∞ ((−1)n − n) n→∞ −1 −1
n n

## so the sequence converges to −1.

10. Find the value of the series
X∞
1 + 2n
n−1
.
n=1
3

## Answer: I can re-write the terms as:

 n−1  n−1
1 + 2n 1 2n 1 2
= + = +2 .
3n−1 3n−1 3n−1 3 3

Therefore,
X∞ X∞  n−1 X∞  n−1
1 + 2n 1 2
n−1
= + 2 .
n=1
3 n=1
3 n=1
3
Shifting the indices of the sums down by one yields
∞  n
X X∞  n
1 2
+ 2 .
n=0
3 n=0
3

These are both geometric series, so I can sum them using the formula for geometric series:
∞  n
X X∞  n
1 2 1 2 3 15
+ 2 = 1 + 2 = +6= .
n=0
3 n=0
3 1− 3 1− 3
2 2

## 11. Does the series

X n+5

n=1
n n+3
converge or diverge?
P
Answer: Do a limit comparison to √1 :
n

n+5
√ √
n n+3 (n + 5) n n3/2 + 5n1/2
lim = lim √ = lim √ .
n→∞ √1 n→∞ n n + 3 n→∞ n3 + 3n2
n

## Dividing numerator and denominator by n3/2 yields

1

n3/2
n3/2 + 5n1/2 1 + n5 1 + n5
lim 1
√ = lim q = lim q = 1.
n→∞ n 3 + 3n2 n→∞ 1 3 + 3n2 ) n→∞ 3
n 3/2
3 (n n 1 + n

P 1 P 1
Therefore, since √ =
n n1/2
diverges (it’s a p-series with p = 1/2 < 1), the Limit Comparison
Test says that the given series also diverges.

3
12. Does the series
X∞
3 + cos n
n=1
en
converge or diverge?
|3 + cos n| ≤ 4
for all n, so  n
3 + cos n |3 + cos n| 4 1
= ≤ n =4
en en e e
P 1 n P 3+cos n
for all n. Since 1
< 1, the series 4 e converges and so, by the comparison test, n also
e e
converges.
P 3+cos n
Hence, the series en converges absolutely.
13. Does the series

X 1
(−1)n √
n=0
n2 + 1
converge absolutely, converge conditionally, or diverge?
Answer: The terms √n12 +1 are decreasing and go to zero (you should check this), so the Alternating
Series Test says that the series converges.
To see that the series does not converge absolutely, it suffices to show that the series
X∞ X ∞
1 1
(−1) √
n = √
2
n +1 2
n +1
n=0 n=0
P1
diverges. To see this, do a limit comparison with the divergent series n:

√ 1 1
2 n nn
lim n1 +1 = lim √ √
= lim
n→∞
n
n→∞ n2 + 1 n→∞ n1 n2 + 1
1
= lim q
n→∞ 1 2
n2 (n + 1)
1
= lim q
n→∞ 1
1+ n2

= 1.
P
Since the limit is finite and non-zero, the limit comparison test says that the series √ 1 diverges.
n2 +1

## 14. Does the series

X n!
(−1)n
n=1
πn
converge absolutely, converge conditionally, or diverge?