Numerical Sequences and Series

Written by Men-Gen Tsai

email: [email protected]

1. Prove that the convergence of {sn } implies convergence of {|sn |}. Is

the converse true?
Solution: Since {sn } is convergent, for any  > 0, there exists N such
that |sn − s| <  whenever n ≥ N . By Exercise 1.13 I know that
||sn | − |s|| ≤ |sn − s|. Thus, ||sn | − |s|| < , that is, {sn } is convergent.
The converse is not true. Consider sn = (−1)n .

√
2. Calculate limn→∞ ( n2 + n − n).
Solution:
√ n
n2 + n − n = √
n2 + n + n
1
= q
1/n + 1 + 1
1
→
2
as n → ∞.

√
3. If sn = 2 and
q √
sn+1 = 2+ sn (n = 1, 2, 3, ...),

prove that {sn } converges, and that sn < 2 for n = 1, 2, 3, ....

r thatq{sn } is strictly increasing. It is trivial that

Proof: First, I show
q √ √ √
s2 = 2 + s1 = 2 + 2 > 2 = s1 . Suppose sk > sk−1 when

1
 k < n. By the induction hypothesis,
q √
sn = 2+ sn−1
q √
> 2+ sn−2
= sn−1

By the induction, {sn } is strictly increasing. Next, I show that {sn } is

bounded by 2. Similarly, I apply the induction again. Hence {sn } is
strictly increasing and bounded, that is, {sn } converges.

4.

5.

6.
P
7. Prove that the convergence of an implies the convergence of
√
X an
n
if an ≥ 0.
Proof: By Cauchy’s inequality,
k k k √
X X 1 X an
an ≥ an
n=1 n=1 n2 n=1 n
√
1 Pk an
for all n ∈ N . Also, both
P P
an and n2
are convergent; thus n=1 an n
√ √
an an
≥ 0 for all n. Hence
P
is bounded. Besides, n n
is convergent.

8.

9. Find the radius of convergence of each of the following power series:

X 2n X 2n X n3
n3 z n , (b) z n , (c) z n , (d) zn.
X
(a) 2
n! n 3n
2
 Solution: (a) αn = (n3 )1/n → 1 as n → ∞. Hence R = 1/α = 1.
(b) αn = (2n /n!)1/n = 2/(n!)1/n → 0 as n → ∞. Hence R = +∞.
(c) αn = (2n /n2 )1/n → 2/1 = 2 as n → ∞. Hence R = 1/α = 1/2.
(d) αn = (n3 /3n )1/n → 1/3 as n → ∞. Hence R = 1/α = 3.

10.
P
11. Suppose an > 0, sn = a1 + ... + an , and an diverges.
P an
(a) Prove that 1+an
diverges.
(b) Prove that
aN +1 aN +k sN
+ ... + ≥1−
sN +1 sN +k sN +k
P an
and deduce that sn diverges.
(c) Prove that
an 1 1
2
≤ −
sn sn−1 sn
P an
and deduce that s2n
converges.
(d) What can be said about
X an X an
and ?
1 + nan 1 + n 2 an

Proof of (a): Note that

an 1
→0 ⇔ 1 →0
1 + an an
+1
1
⇔ →∞
an
⇔ an → 0
an
as n → ∞. If converges, then an → 0 as n → ∞. Thus for
P
1+an
some 0 = 1 there is an N1 such that an < 1 whenever n ≥ N1 . Since
P an
1+an
converges, for any  > 0 there is an N2 such that
am an
+ ... + <
1 + am 1 + an

3
all n > m ≥ N2 . Take N = max(N1 , N2 ). Thus
am an
 > + ... +
1 + am 1 + an
am an
> + ... +
1+1 1+1
am + ... + an
=
2
for all n > m ≥ N . Thus

am + ... + an < 2

an
for all n > m ≥ N . It is a contradiction. Hence
P
1+an
diverges.
Proof of (b):
aN +1 aN +k aN +1 aN +k
+ ... + ≥ + ... +
sN +1 sN +k sN +k sN +k
aN +1 + ... + aN +k
=
sN +k
sN +k − sN
=
sN +k
sN
= 1−
sN +k
P an
If sn
converges, for any  > 0 there exists N such that
am an
+ ... + <
sm sn
for all m, n whenever n > m ≥ N . Fix m = N and let n = N + k.
Thus
am an
 > + ... +
sm sn
aN aN +k
= + ... +
sN sN +k
sN
≥ 1−
sN +k

4
for all k ∈ N . But sN +k → ∞ as k → ∞ since an diverges and
P
P an
an > 0. Take  = 1/2 and we obtain a contradiction. Hence sn
diverges.
Proof of (c):
1 1
sn−1 ≤ sn ⇔ 2
≤
sn sn sn−1
an an sn − sn−1
⇔ 2 ≤ =
sn sn sn−1 sn sn−1
an 1 1
⇔ 2 ≤ −
sn sn−1 sn
for all n.
Hence
k k
X an X 1 1
2
≤ ( − )
n=2 sn sn−1
n=2 sn
1 1
= − .
s1 sn
1 an
→ 0 as n → ∞ since
P P
Note that sn
an diverges. Hence s2n
converges.
P an P an
Proof of (d): 1+nan
may converge or diverge, and 1+n2 an
con-
an 1 P an
verges. To see this, we put an = 1/n. 1+nan
= 2n
, that is, 1+nan
=
P
2 1/n diverges. Besides, if we put
1
an =
n(log n)p
where p > 1 and n ≥ 2, then
an 1
=
1 + nan n(log n) ((log n)p + 1)
2p

1
<
2n(log n)3p
P an
for large enough n. By Theorem 3.25 and Theorem 3.29, 1+nan
converges. Next,
X an X 1
2
=
1 + n an 1/an + n2

5
 X 1
< .
n2
P 1 P an
for all an . Note that n2
converges, and thus 1+n2 an
converges.

P
12. Suppose an > 0 and an converges. Put
∞
X
rn = am .
m=n

(a) Prove that

am an rn
+ ... + >1−
rm rn rm
P an
if m < n, and deduce that rn
diverges.
(b) Prove that
an √ √
√ < 2( rn − rn+1 )
rn
√an
P
and deduce that rn
converges.
Proof of (a):
am an am + ... + an
+ ... + >
rm rn rm
r m − rn
=
rm
rn
= 1−
rm
P an
if m < n. If rn
converges, for any  > 0 there exists N such that
am an
+ ... + <
rm rn
for all m, n whenever n > m ≥ N . Fix m = N . Thus
am an rn
+ ... + > 1−
rm rn rm
rn
= 1−
rN

6
for all n > N . But rn → 0 as n → ∞; thus armm
+ ... + arnn → 1 as n → ∞.
If we take  = 1/2, we will get a contradiction.
Proof of (b): Note that
√ √
rn+1 < rn ⇔ rn+1 < rn
√ √ √
⇔ rn + rn+1 < 2 rn
√ √
rn + rn+1
⇔ √ <2
rn
√ √
√ √ rn + rn+1 √ √
⇔ ( rn − rn+1 ) √ < 2( rn − rn+1 )
rn
rn − rn+1 √ √
⇔ √ < 2( rn − rn+1 )
rn
an √ √
⇔ √ < 2( rn − rn+1 )
rn

since an > 0 for all n.

Hence,
k k
X an X √ √
√ < 2( rn − rn+1 )
n=1 rn n=1
√ √
= 2( r1 − rk+1 )

Note that rn → 0 as n → ∞. Thus √an √an

P P
rn
is bounded. Hence rn
converges.
P P
Note: If we say an converges faster than bn , it means that
an
lim = 0.
n→∞ bn
According the above exercise, we can construct a faster convergent se-
ries from a known convergent one easily. It implies that there is no
perfect tests to test all convergences of the series from a known con-
vergent one.

7
13. Prove that the Cauchy product of two absolutely convergent series con-
verges absolutely.

P P Pn P
Note: Given an and bn , we put cn = k=0 ak bn−k and call cn
the Cauchy product of the two given series.

Pn Pn Pn
Proof: Put An = k=0 |ak |, Bn = k=0 |bk |, Cn = k=0 |ck |. Then

Cn = |a0 b0 | + |a0 b1 + a1 b0 | + ... + |a0 bn + a1 bn−1 + ... + an b0 |

≤ |a0 ||b0 | + (|a0 ||b1 | + |a1 ||b0 |) + ...
+(|a0 ||bn | + |a1 ||bn−1 | + ... + |an ||b0 |)
= |a0 |Bn + |a1 |Bn−1 + ... + |an |B0
≤ |a0 |Bn + |a1 |Bn + ... + |an |Bn
= (|a0 | + |a1 | + ... + |an |)Bn = An Bn ≤ AB

where A = lim An and B = lim Bn . Hence {Cn } is bounded. Note

that {Cn } is increasing, and thus Cn is a convergent sequence, that
is, the Cauchy product of two absolutely convergent series converges
absolutely.

14. If {an } is a complex sequence, define its arithmetic means σn by

s0 + s1 + ... + sn
σn = (n = 0, 1, 2, ...).
n+1
(a) If lim sn = s, prove that lim σn = s.
(b) Construct a sequence {sn } which does not converge, although lim σn =
0.
(c) Can it happen that sn > 0 for all n and that lim sup sn = ∞,
although lim σn = 0?

8
(d) Put an = sn − sn−1 , for n ≥ 1. Show that
n
1 X
sn − σn = kak .
n + 1 k=1

Assume that lim(nan ) = 0 and that {σn } converges. Prove that {sn }
converges. [This gives a converse of (a), but under the additional as-
sumption that nan → 0.]
(e) Derive the last conclusion from a weaker hypothesis: Assume M <
∞, |nan | ≤ M for all n, and lim σn = σ. Prove that lim sn = σ, by
completing the following outline:
If m < n, then
n
m+1 1 X
sn − σn = (σn − σm ) + (sn − si ).
n−m n − m i=m+1

For these i,

(n − i)M (n − m − 1)M
|sn − si | ≤ ≤ .
i+1 m+2
Fix  > 0 and associate with each n the integer m that satisfies
n−
m≤ < m + 1.
1+
Then (m + 1)/(n − m) ≤ 1/ and |sn − si | < M . Hence

lim sup |sn − σ| ≤ M .

n→∞

Since  was arbitrary, lim sn = σ.

Proof of (a): The proof is straightforward. Let tn = sn −s, τn = σn −s.

(Or you may suppose that s = 0.) Then
t0 + t1 + ... + tn
τn = .
n+1

9
Choose M > 0 such that |tn | ≤ M for all n. Given  > 0, choose N
so that n > N implies |tn | < . Taking n > N in τn = (t0 + t1 + ... +
tn )/(n + 1), and then

|t0 | + ... + |tN | |tN +1 + ... + |tn

|τn | ≤ +
n+1 n+1
(N + 1)M
< + .
n+1
Hence, lim supn→∞ |τn | ≤ . Since  is arbitrary, it follows that limn→∞ |τn | =
0, that is, lim σn = s.

Proof of (b): Let sn = (−1)n . Hence |σn | ≤ 1/(n + 1), that is,
lim σn = 0. However, lim sn does not exists.

Proof of (c): Let






1 , n = 0,
1/4 −1
sn =  n + n , n = k 2 for some integer k,
 n−1

, otherwise.


It is obvious that sn > 0 and lim sup sn = ∞. Also,

j√ k j√ k
s0 + ... + sn = 1 + nn−1 + n n1/4 = 2 + n n1/4 .

That is, √
2 b nc n1/4
σn = +
n+1 n+1
The first term 2/(n + 1) → 0 as n → ∞. Note that
√
b nc n1/4
0≤ < n1/2 n1/4 n−1 = n−1/4 .
n+1
It implies that the last term → 0. Hence, lim σn = 0.

10
Proof of (d):
n
X n
X n
X n
X
kak = k(sk − sk−1 ) = ksk − ksk−1
k=1 k=1 k=1 k=1
Xn n−1
X
= ksk − (k + 1)sk
k=1 k=0
n−1
X n−1
X
= nsn + ksk − (k + 1)sk − s0
k=1 k=1
n−1
X n
X
= nsn − sk − s0 = (n + 1)sn − sk
k=1 k=0
= (n + 1)(sn − σn ).

That is,
n
1 X
sn − σn = kak .
n + 1 k=1

Note that {nan } is a complex sequence. By (a),

n
1 X
 
lim kak = n→∞
lim nan = 0.
n→∞ n + 1 k=1

Also, lim σn = σ. Hence by the previous equation, lim s = σ.

Proof of (e): If m < n, then

n
X
(sn − si ) + (m + 1)(σn − σm )
i=m+1
n
X
= (n − m)sn − si + (m + 1)(σn − σm )
i=m+1
X n m
X 
= (n − m)sn − si − si + (m + 1)(σn − σm )
i=0 i=0
= (n − m)sn − (n + 1)σn + (m + 1)σm + (m + 1)(σn − σm )
= (n − m)sn − (n − m)σn .

11
 Hence,
n
m+1 1 X
sn − σn = (σn − σm ) + (sn − si ).
n−m n − m i=m+1

For these i, recall an = sn − sn−1 and |nan | ≤ M for all n,

 n n n
M (n − i)M
 X  X X

|sn − si | =  ak  ≤
 |ak | ≤ =
k=i+1 i + 1 i+1

k=i+1 k=i+1
(n − (m + 1))M (n − m − 1)M
≤ = .
(m + 1) + 1 m+2

Fix  > 0 and associate with each n the integer m that satisfies
n−
m≤ < m + 1.
1+
Thus
n−m n−m−1
≥  and < ,
m+1 m+2
or
m+1 1
≤ and |sn − si | < M .
n−m 
Hence,
1
|sn − σ| ≤ |σn − σ| + (|σn − σ| + |σm − σ|) + M .

Let n → ∞ and thus m → ∞ too, and thus

lim sup |sn − σ| ≤ M .

n→∞

Since  was arbitrary, lim sn = σ.

15.

12
 √
16. Fix a positive number α. Choose x1 > α, and define x2 , x3 , x4 , ...,
by the recursion formula
1 α
xn+1 = (xn + ).
2 xn
√
(a) Prove that {xn } decreases monotonically and that lim xn = α.
√
(b) Put n = xn − α, and show that
2n 2
n+1 = < √n
2xn 2 α
√
so that, setting β = 2 α,
1 n
n+1 < β( )2 (n = 1, 2, 3, ...).
β

(c) This is a good algorithm for computing square roots, since the
recursion formula is simple and the convergence is extremely rapid.
1
For example, if α = 3 and x1 = 2, show that 1 /β < 10
and therefore

5 < 4 · 10−16 , 6 < 4 · 10−32 .

Proof of (a):
1 α
xn − xn+1 = xn − (xn + )
2 xn
1 α
= (xn − )
2 xn
2
1 xn − α
= ( )
2 xn
> 0

since xn > α. Hence {xn } decreases monotonically. Also, {xn } is

bounded by 0; thus {xn } converges. Let lim xn = x. Hence
1 α 1 α
lim xn+1 = lim (xn + ) ⇔ x = (x + )
2 xn 2 x
2
⇔ x = α.

13
 √ √
Note that xn > 0 for all n. Thus x = α. lim xn = α.
Proof of (b):
1 α
xn+1 = (xn + )
2 xn
√ 1 α √
⇒ xn+1 − α = (xn + ) − α
2 xn
√
√ 2
1 xn − 2xn α + α
⇒ xn+1 − α =
2 x
√ n2
√ (xn − α)
⇒ xn+1 − α =
2xn
2 2
 
⇒ n+1 = n < √n .
2xn 2 α

Hence
1 n
n+1 < β( )2
β
√
where β = 2 α by induction.
Proof of (c):
√
1 2− 3 1 1 1
= √ = √ √ = √ < .
β 2 3 2 3(2 + 3) 6+4 3 10

Thus
1 4 √
5 < β( )2 < 2 3 · 10−16 < 4 · 10−16 ,
β
 1 25 √
6 < β( ) < 2 3 · 10−32 < 4 · 10−32 .
β

Note: It is an application of Newton’s method. Let f (x) = x2 − α in

Exercise 5.25.

17.

18.

14
19.

20.

21.

22. Suppose X is a complete metric space, and {Gn } is a sequence of dense

open subsets of X. Prove Baire’s theorem, namely, that ∞
T
1 Gn is not
empty. (In fact, it is dense in X.) Hint: Find a shrinking sequence of
neighborhoods En such that En ⊂ Gn , and apply Exercise 21.
Proof: I’ve proved it in Chapter 2 Exercise 30.

23. Suppose {pn } and {qn } are Cauchy sequences in a metric space X.
Show that the sequence {d(pn , qn )} converges. Hint: For any m, n,

d(pn , qn ) ≤ d(pn , pm ) + d(pm , qm ) + d(qm , qn );

it follows that
|d(pn , qn ) − d(pm , qm )|

is small if m and n are large.

Proof: For any  > 0, there exists N such that d(pn , pm ) <  and
d(qm , qn ) <  whenever m, n ≥ N . Note that

d(pn , qn ) ≤ d(pn , pm ) + d(pm , qm ) + d(qm , qn ).

It follows that

|d(pn , qn ) − d(pm , qm )| ≤ d(pn , pm ) + d(qm , qn ) < 2.

Thus {d(pn , qn )} is a Cauchy sequence in X. Hence {d(pn , qn )} con-

verges.

15
24. Let X be a metric space. (a) Call two Cauchy sequences {pn }, {qn } in
X equivalent if
lim d(pn , qn ) = 0.
n→∞

Prove that this is an equivalence relation.

(b) Let X ∗ be the set of all equivalence classes so obtained. If P ∈ X ∗ ,

Q ∈ X ∗ , {pn } ∈ P , {qn } ∈ Q, define

4(P, Q) = n→∞
lim d(pn , qn );

by Exercise 23, this limit exists. Show that the number 4(P, Q) is
unchanged if {pn } and {qn } are replaced by equivalent sequences, and
hence that 4 is a distance function in X ∗ .

(c) Prove that the resulting metric space X ∗ is complete.

Proof of (a): Suppose there are three Cauchy sequences {pn }, {qn },
and {rn }. First, d(pn , pn ) = 0 for all n. Hence, d(pn , pn ) = 0 as n → ∞.
Thus it is reflexive. Next, d(qn , pn ) = d(pn , qn ) → 0 as n → ∞. Thus it
is symmetric. Finally, if d(pn , qn ) → 0 as n → ∞ and if d(qn , rn ) → 0
as n → ∞, d(pn , rn ) ≤ d(pn , qn ) + d(qn , rn ) → 0 + 0 = 0 as n → ∞.
Thus it is transitive. Hence this is an equivalence relation.
Proof of (b):
Proof of (c): Let {Pn } be a Cauchy sequence in (X ∗ , 4). We wish to
show that there is a point P ∈ X ∗ such that 4(Pn , P ) → 0 as n → ∞.
For each Pn , there is a Cauchy sequence in X, denoted {Qk n}, such
that 4(Pn , Qk n) → 0 as k → ∞. Let n > 0 be a sequence tending
to 0 as n → ∞. From the double sequence {Qk n} we can extract a
subsequence Q0n such that 4(Pn , Q0n ) < n for all n. From the triangle
inequality, it follows that

4(Q0n , Q0m ) ≤ 4(Q0n , Pn ) + 4(Pn , Pm ) + 4(Pm , Q0m ). (1)

16
 Since {Pn } is a Cauchy sequence, given  > 0, there is an N > 0 such
that 4(Pn , Pm ) <  for m, n > N . We choose m and n so large that
m < , n < . Thus (1) shows that {Q0n } is a Cauchy sequence in X.
Let P be the corresponding equivalence class in S. Since

4(P, Pn ) ≤ 4(P, Q0n ) + 4(Q0n , Pn ) < 2

for n > N , we conclude that Pn → P as n → ∞. That is, X ∗ is

complete.

25.

17