Series Examples - Part II Instructor: Prof.

Cesar Aguilar
MATH 222 - Calculus II Department of Mathematics, SUNY Geneseo
Comparison, Ratio, Absolute Tests South 325A, [email protected]

1. Comparison Tests
The direct Comparison Test has two parts.

Comparison Test. Suppose that {an } and {bn } are sequences that are non-negative, in
other words, an ≥ 0 and bn ≥ 0. Suppose further that an ≤ bn for all n ≥ 1.
P P
(i) If bn converges then an also converges.
P P
(ii) If an diverges then bn also diverges.

P P
The Comparison Test is very intuitive. For example, if bn = B converges and an ≤ bn then an ≤ B.
P P
So the sum an is finite and less than or equal to B. On the other hand, if an diverges then since
P P
an ≥ 0 then an = ∞. Therefore, since an ≤ bn then the sum bn also diverges. These are not
precise mathematical statements but they capture the main idea. The proof of the Comparison Test
would use partial sums of the sequences.

The truth about the Comparison Test: To be frank, the Comparison Test is a good method to
apply when you already have a pretty good idea about whether the series you are given converges or
P
diverges. If you have a pretty good idea that an converges, then to apply the Comparison test, you
P
want to find a series bn that you know converges and also that an ≤ bn . Similarly, if instead you
P P
have a pretty good idea that an diverges then you want to find a series bn that you know diverges
and also an ≤ bn . By now, you should have a list of series that you know converge or diverge.

Example 1: Determine if the series converges or diverges.

∞
X n+2
n=1
3n2 + 1

Solution: As a first test, we compute that

 
n+2
lim = 0,
n→∞ 3n2 + 1

and so we cannot conclude based on this that the series diverges. Now, when n is very large, n + 2 ≈ n
and 3n2 + 1 ≈ 3n2 . Thus, when n is very large
n+2 n 1
2
≈ 2 =
3n + 1 3n 3n
P
We know that ∞ 1
n=1 3n diverges because it is a constant multiple of the Harmonic series. So, we have
a good idea that the series we are given probably diverges. Let’s now be precise and find a sequence

pg. 1
 n+2
P
cn such that cn ≤ 3n2 +1
and such that cn diverges. First of all, n < n + 2 and therefore

n n+2
< 2 Inequality (1)
3n2 +1 3n + 1
1 1
Now, 3n2 + 1 ≤ 3n2 + n2 for all n = 1, 2, 3, . . .. Then, 3n2 +n2
≤ 3n2 +1
for all n, and multiplying both
sides by n we get
n n
2 2
≤ 2 Inequality (2)
3n + n 3n + 1
Combining Inequality (1) and (2), we get
n n n+2
≤ 2 < 2 .
3n2 +n2 3n + 1 3n + 1
n n 1
Now, = = and thus
3n2 + n2 4n2 4n
1 n
≤ 2 .
4n 3n + 1
P 1
P n
The series 4n
diverges and therefore by Part (ii) of the Comparison test, 3n2 +1
also diverges.

Example 2: Determine if the given sequence is convergent of divergent.

∞
X 1
√
n=1
n+ n

Solution: We first compute that limn→∞ n+1√n = 0, and so the sequence might be convergent. Now,
√
when n is very (very) large, n is very small compared to n. So, when n is large
1 1
√ ≈ .
n+ n n
P 1
So, we have reason to believe that √ is divergent. To be precise about this, it is clear that
n+ n
√ √
n ≤ n for all n = 1, 2, 3, . . .. Therefore, n + n ≤ n + n = 2n, and therefore,
1 1
≤ √ .
2n n+ n
P 1 P 1
The series 2n
is divergent, and therefore by Part (ii) of the Comparison Test, the series √ is
n+ n
also divergent.

Example 3: Determine if the given sequence converges or diverges.

∞
X 1
n
.
n=1
n3
P
Solution: The denominator n3n grows faster than 3n and we know that the Geometric series ∞ 1
n=1 3n
converges. So, we have good reason to believe that the given series converges. To see this, we first note
that 3n ≤ n3n for n = 1, 2, 3, . . .. Thus,
1 1
n
≤ n
n3 3

pg. 2
 P
The series ∞ 1
n=1 3n is a convergent Geometric series, and thus by Part (i) of the Comparison Test, the
P
given series ∞ 1
n=1 n3n is also convergent.

From the examples above, we can see that applying the Comparison Test takes some ingenuity and
insight. The next test for convergence/divergence, called the Limit Comparison Test, is a direct appli-
cation of the Comparison Test but is a bit more easy to apply provided you can do limits. The Limit
P
Comparison Test is particularly useful for series an where an is a rational function of n.

The Limit Comparison Test. Suppose that an > 0 and bn > 0 for all n. Compute
an
lim =c
n→∞ bn
P P
(i) If c > 0 then an and
bn both converge or both diverge.
P P
(ii) If c = 0 and bn converges then an converges.
P P
(iii) If c = ∞ and bn diverges then an diverges.

Let’s now use the Limit Comparison Test to redo the three previous examples where we used the
Comparison Test. You can decide which test is easier to apply.

Example 4: Determine if the given series converges or diverges.

∞
X n+2
n=1
3n2 + 1
P
Solution: Here an = 3nn+2 n+2 n 1
2 +1 . When n is very large, 3n2 +1 ≈ 3n2 = 3n , and we know that
1
3n
diverges.
1
So we have good reason to believe the given series diverges. Let bn = 3n . Then
n+2
an 3n2 +1
lim = lim 1
n→∞ bn n→∞
3n

(n + 2)3n
= lim
n→∞ 3n2 + 1

3n2 + 6n
= lim
n→∞ 3n2 + 1

=1
P 1
P∞ n+2
Therefore, since 3n
diverges, then by Part (i) of the Limit Comparison Test, n=1 3n2 +1 diverges
also.

Example 5: Determine if the given sequence is convergent of divergent.

∞
X 1
√
n=1
n+ n

pg. 3
 1√ √ √
Solution: Here an = n+ n
. When n is very large, n is much smaller than n, so n + n ≈ n. Hence,
1√
P1
when n is large, an = n+ n
≈ n1 , and we know that n
diverges. Now,
an n
lim = lim √
n→∞ 1 n→∞ n + n
n

n
= lim
n→∞ n(1 + √1 )
n

1
= lim
n→∞ 1(1 + √1 )
n

=1 could’ve used L.H.R. also

P 1
P∞ 1√
Therefore, since n
diverges, then by Part (i) of the Limit Comparison Test, n=1 n+ n also diverges.

Example 6: Determine if the given sequence converges or diverges.

∞
X 1
.
n=1
n3n
1 1
Solution: Here an = n3n
. Let bn = 3n
. Then
1
an n3n
lim = lim 1
n→∞ bn n→∞
3n

3n
= lim
n→∞ n3n

1
= lim
n→∞ n

=0
P 1
P 1
Therefore, since 3n
converges, then by Part (ii) of the Limit Comparison Test, n3n
also converges.

Example 7: Determine if the given sequence converges or diverges.

∞
X 2n + n
n=1
7n2n
2n +n
Solution: Here an = 7n2n
. Let’s first test for divergence:
2n + n
lim an = lim
n→∞ n→∞ 7n2n

2n (1 + 2nn )
= lim
n→∞ 7n2n
(1 + 2nn )
= lim
n→∞ 7n

=0

pg. 4
Thus, we cannot conclude that the given series diverges and we need to do further analysis. Now, when
n is large 2n + n ≈ 2n because exponentials grow much faster than polynomials. So, when n is large
n 2n
P 1
an = 27n2+n 1
n ≈ 7n2n = 7n . We know that 7n
diverges so we have good reason to believe that the given
1
series diverges. Let bn = 7n . Then
2n +n
an 7n2n
lim = lim 1
n→∞ bn n→∞
7n

(2n + n)7n
= lim
n→∞ 7n2n
2n + n
= lim
n→∞ 2n
2n (1 + 2nn )
= lim
n→∞ 2n
 n
= lim 1 + n
n→∞ 2

=1
P 1
Therefore, since 7n
diverges, then by Part (i) of the Limit Comparison Test, the given series
P∞ 2n +n
n=1 7n2n also diverges.

Example 8: Determine if the given sequence converges or diverges.

∞
X (ln n)2
n=1
n3

(ln n)2
Solution: Here an = n3
. Let’s first test for divergence:

(ln n)2 ∞
lim an = lim so use L.H.R.
n→∞ n→∞ n3 ∞

2(ln n) n1
= lim
n→∞ 3n2
2(ln n) ∞
= lim again so use L.H.R.
n→∞ 3n3 ∞

2 n1
= lim
n→∞ 9n2

2
= lim
n→∞ 9n3

=0

So, we cannot conclude that the series diverges and we need to do further analysis. A useful fact
√
that we can use is that for every n > 0, it holds that ln(n) < n. The graph of f (x) = ln(x)
√
and g(x) = x on the interval [1, 10] are shown in Figure 1.

pg. 5
 3.5

sqrt(x)
3 ln(x)

2.5

1.5

0.5

0
1 2 3 4 5 6 7 8 9 10
x

√
Figure 1: Graph of f (x) = ln(x) (red-solid graph) and g(x) = x (blue-dashed graph)

√
Therefore, since ln(n) < n then squaring both sides yields (ln n)2 < n. Therefore,

(ln n)2 n 1
an = 3
< 3 = 2.
n n n
P 1 P (ln n)2
The series n2
converges, and therefore, by the Comparison Test Part (i), the given series n3
also converges.

Example 9: Determine if the given series converges or diverges.

∞
X 1
n1/2
n=1
P 1
Solution: The given series is a p-series np
with p = 21 . We know that a p-series converges only
P1
if p > 1, so this series diverges. Let’s use the Limit Comparison test to show this. The series n
diverges and
1
n1/2 n
lim = lim
n→∞ 1 n→∞ n1/2
n

= lim n1/2
n→∞

=∞
P 1
Therefore, by Part (iii) of the Limit Comparison test, the series n1/2
diverges. In fact, this same
P 1
procedure shows that if 0 < p < 1 then the p-series np
diverges because
1
np n
lim = lim = lim n1−p = ∞
n→∞ 1 n→∞ np n→∞
n

whenever 1 − p > 0, which is the same as saying that p < 1.

pg. 6
 2. Absolute Convergence Test
P (−1)n (−1)n
Consider the series ∞ n=1 n2 , so that here an = n2 . This is an example of an alternating series;
the terms of the series alternate between positive and negative. This series is not Geometric, we cannot
apply the Integral Test because some of the series terms are negative (go read the Integral Test to
see), and we cannot apply the comparison tests because they require that the series terms also be
non-negative (go read the comparison tests to see). However, we know that the series of the absolute
P P∞ 1
values ∞ n=1 |an | = n=1 n2 converges and so we might use this fact to conclude that the original series
P∞
n=1 an converges. This is in fact true and it is called the Absolute Convergence Test.

P
Absolute
P Convergence P Test. Let an be a given series. If the series of absolute values
|an | converges then an also converges.

P
The Absolute Convergence Test is useful when some of the terms of the series an are negative, which
P
is the case for alternating series. Because the series |an | has non-negative terms, we may use the
P
Integral Test or the Comparison Tests on the series |an | to show that it converges, if possible.

Example 10: Determine whether the given series converges.

∞
X (−1)n
n=1
n2

(−1)n |(−1)n | P
Solution: Here an = n2
and thus |an | = |n2 |
= n12 . The p-series ∞ 1
n=1 n2 converges. Therefore,
P∞ (−1)n
by the Absolute Convergence test, the series n=1 n2 also converges.
P P P
A series an for which |an | converges is somewhat special; not only does |an | converge but by
P P
the Absolute Convergence Test, an also converges. Hence, for such a series, we will say that an is
absolutely convergent or that it converges absolutely.

Example 11: Determine whether the given series converges.

∞
X 8(−1)n
n=0
(3 + (1/n))2n

Solution: The series contains negative terms so we cannot apply the Integral Test or any of the
Comparison Tests. It is clear that limn→∞ an = 0, so we cannot conclude that the series diverges.
Consider the absolute value series
∞ ∞
X X 8
|an | =
n=0 n=1
(3 + (1/n))2n

1
When n is large n
≈ 0, and so the terms of the series behave like

8 8 8
≈ =
(3 + (1/n))2n (3 + 0)2n 32n

pg. 7
 P 8 1
We know that the series 32n
converges (it is Geometric with r = 32
) and so we have good reason to
P
believe that |an | converges. In fact, by direct comparison
8 8
2n
< 2n
(3 + (1/n)) 3
P 8
P
The series 32n
converges and therefore by the Comparison Test Part (i), the series |an | =
P 8
P P 8(−1)n
(3+(1/n))2n
also converges. Thus, by the Absolute Convergence test, the series an = (3+(1/n))2n
converges also.

Example 12: Determine whether the given series converges.

∞
X (−1)n 1 1 1 1
= −1 + − + − +···
n=1
n 2 3 4 5
P∞ P∞ 1
Solution: This is an alternating Harmonic series. The absolute value series n=1 |an | = n=1 n
P∞ (−1)n
diverges but we cannot conclude that n=1 n
diverges because the Absolute Convergence
Test is a test for convergence and not for divergence. Thus, we cannot conclude anything
about the given series with the known tests thus far.

Example 13: Determine whether the given series converges or diverges.

∞
X cos(n)
n=1
n2

Solution: Because cos(n) takes on both negative and positive values for n = 1, 2, 3, . . ., we cannot
apply the Integral Test or the Comparison Test. Using the Squeeze Theorem, we can show that

cos(n)
lim =0
n→∞ n2
and therefore we cannot conclude that the series diverges. Consider the series of absolute values
P P | cos(n)|
|an | = n2
. For all n = 1, 2, 3, . . . , it holds that | cos(n)| ≤ 1 and therefore dividing this
inequality by n2 we obtain
| cos(n)| 1
2
≤ 2
n n
P 1 P | cos(n)|
The series n 2 converges and therefore by Part (i) of the Comparison Test, the series n2
P cos(n)
converges. Therefore, the series n2
is absolutely convergent.

pg. 8
 3. The Ratio Test
The Ratio Test can be applied to any series, it does not matter if it takes on negative values. However,
the test is not full proof; in some cases it is inconclusive. The Ratio Test is also very useful when
factorials n! are present in the series.

P
The Ratio Test. Let an be any given series and suppose that
 
 an+1 
lim  =ρ
n→∞  an 

The following hold:

P
(a) If ρ < 1 then the series an converges absolutely.
P
(b) If ρ > 1 or if ρ = ∞ then the series an diverges.

(c) If ρ = 1 then the test is inconclusive; it may converge or it may diverge and further
analysis is necessary.

Example 14: Determine whether the given series converges or diverges.

∞
X 2n
n=1
n!

2n 2n+1
Solution: Here an = n!
and therefore an+1 = (n+1)!
. Therefore,
   2n+1 
 an+1   
lim   = lim  (n+1)! 
all terms are positive
2n 
n→∞ an  n→∞  n!

2n+1 n!
= lim
n→∞ 2n (n + 1)!

2n · 2 · n!
= lim
n→∞ 2n (n + 1) · n!

2
= lim
n→∞ n + 1

=0

Therefore, since ρ = 0 < 1, by the Ratio Test, the given series is convergent.

Example 15: Determine whether the given series converges or diverges.

∞
X n3
n=1
en

pg. 9
 3
Solution: Here an = n3
en
and therefore an+1 = (n+1)
en+1
. Therefore,
   
 an+1   (n + 1)3 n3 
lim   = lim  n+1 ÷ n 

n→∞  an  n→∞ e e

(n + 1)3 en
= lim
n→∞ en+1 n3
(n + 1)3
= lim
n→∞ e · n3

n3 + 3n2 + 3n + 1
= lim
n→∞ e · n3
1
= .
e
1
Since ρ = e
< 1, by the Ratio Test, the given series converges.

Example 16: Determine whether the given series converges or diverges.

∞
X n2 2n (−1)n
n=1
3n

n2 2n (−1)n 2 n+1 n+1

Solution: Here an = 3n
and therefore an+1 = (n+1) 23n+1(−1) . Therefore,
   
 an+1   (n + 1)2 2n+1 (−1)n+1 n2 2n (−1)n 
lim   = lim  ÷ 
n→∞  an  n→∞  3n+1 3n 
 
 (n + 1)2 2n+1 (−1)n+1 3 n 
= lim  × 
n→∞ 3n+1 n2 2n (−1)n 
 
 (n + 1)2 · 2 · (−1) 
= lim  
n→∞ 3n2 

2(n + 1)2
= lim
n→∞ 3n2
2
=
3
2
Therefore, since ρ = 3
< 1, by the Ratio Test the given series converges.

Example 17: Determine whether the given series converges or diverges.

∞
X (−1)n (2n)!
n=1
n!(n + 1)!

(−1)n (2n)!
Solution: Here an = n!(n+1)!
and therefore

(−1)n+1 (2(n + 1))! (−1)n+1 (2n + 2)!

an+1 = = .
(n + 1)!((n + 1) + 1)! (n + 1)!(n + 2)!

pg. 10
Therefore,
   
 an+1   (−1)n+1 (2n + 2)! (−1)n (2n)! 
lim   = lim 
 ÷ 
n→∞  an  n→∞ (n + 1)!(n + 2)! n!(n + 1)! 
 
 (−1)n+1 (2n + 2)! n!(n + 1)! 
= lim 
 ×
n→∞ (n + 1)!(n + 2)! (−1)n (2n)! 

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

(2n + 2)(2n + 1)(2n)! · n!

= lim
n→∞ (n + 2)(n + 1)n! · (2n)!

(2n + 2)(2n + 1)
= lim
n→∞ (n + 2)(n + 1)

4n2 + 4n + 2
= lim
n→∞ n2 + 3n + 2

=4
Therefore, by the Ratio Test, since ρ = 4 > 1, the given series diverges. Note that here we used that
fact that (2n + 2)! = (2n + 2)(2n + 1)(2n)! and similarly that (n + 2)! = (n + 2)(n + 1)n!.

Example 18: Determine whether the given series converges or diverges.

∞
X n!
n=1
nn

Solution: It is not clear whether limn→∞ nn!n = 0 because n! grows very fast but nn grows faster. In
fact, for n ≥ 2, it holds that n! < nn and therefore nn!n < 1 for n = 2, 3, . . . ,, and so it is possible that
limn→∞ nn!n = 0. To see that n! < nn :
n! = 1 · 2 · 3 · · · n < |n · n ·{zn · · · n} = nn
n times
n! (n+1)!
Anyways, let’s try the Ratio Test. Here an = and therefore an+1 =
nn (n+1)n+1
. Therefore,
   
 an+1   (n + 1)! n! 
lim   = lim 
 ÷ n
n→∞  an  n→∞ (n + 1)n+1 n
 
 (n + 1)! nn 
= lim 
 × 
n→∞ (n + 1)n+1 n!

(n + 1)n!nn
= lim
n→∞ (n + 1)n (n + 1)n!

nn
= lim
n→∞ (n + 1)n

 n
n
= lim get indeterminate power 1∞
n→∞ n+1

pg. 11
The last limit gives the indeterminate power 1∞ , so we consider
 n  
n n
lim ln = lim n ln get ∞ · 0, so rewrite
n→∞ n+1 n→∞ n+1
n


ln n+1 0
= lim 1 now get , so apply L.H.R.
n→∞ 0
n

1 1
n
− n+1
= lim
n→∞ − n12

−n2
= lim
n→∞ n(n + 1)

= −1

Therefore,  
 an+1 
lim   = e−1 = 1 .
n→∞ an  e
1
Hence, since ρ = < 1, then by the Ratio Test, the given series converges. Note that this means that
e
n!
indeed we have that lim an = lim n = 0.
n→∞ n→∞ n

pg. 12