MA2108S Solutions for Tutorial 5

Group2: Wen Shisi, Rong Cheng, Wang Kunzhen

April 29, 2013

Section 3.1
2n 2
17 Claim: for all n ≥ 3, 0 < ≤ 2( )n−2 .
n! 3
Proof of the claim:

23 4 2 4 23 2
when n = 3, = , and 2( )3−2 = . Hence 0 < ≤ 2( )3−2 .
3! 3 3 3 3! 3
2k 2 k−2
Assume that 0 < ≤ 2( ) for some k ∈ N, k ≥ 3.
k! 3
Then
2k+1 2k 2 2 2 2 2 2
= ≤ 2( )k−2 ≤ 2( )k−2 = ( )k−1 .
(k + 1)! k! k + 1 3 k+1 3 3 3

2k+1
And > 0, as 2k+1 > 0 and (k + 1)! > 0.
(k + 1)!
Hence
2n 2
0< ≤ 2( )n−2 , ∀n ≥ 3.
n! 3
2 2
Since lim(( )n ) = 0, lim(2( )n ) = 2 · 0 = 0. Also, lim(0) = 0. By the
3 3
2n
squeeze theorem, lim = 0.
n!
x
18 Since lim(xn ) = x > 0, consider ² = > 0. Then ∃N ∈ N, such that
2
x
|xn − x| < ∀n ≥ N.
2
Hence,
−x x
xn − x > ⇒ xn >
2 2

1
 and
x 1
xn − x < < x ⇒ x < xn < 2x ∀n ≥ N.
2 2

Section 3.2
n 1 1
1. (a) lim(xn ) = lim( ) = lim(1 − ) = 1 − lim( )
n+1 n+1 n+1
= 1 − 0 = 1.
(b) The sequence is divergent.
Proof: Suppose on the contrary that lim(xn ) = x. Let ² = 12 . Then
(−1)n n 1
there exists k ∈ N, such that | −x| < for all n ≥ k, n ∈ N.
n+1 2
1 m 1 m
For m > k, m is odd, we have − − <x< − .
2 m+1 2 m+1
Since m + 1 is even, we have
m+1 1 m+1 1
− <x< + .
m+2 2 m+2 2
However, note that
m+1 1 1 m
− −( − )
m+2 2 2 m+1
1 1 1 1
= (1 − ) + (1 − )− −
m+2 m+1 2 2
1 1
=1− −
m+2 m+1
1 1
≥ 1 − − = 0.
2 2
1 m m+1 1
Hence x < − < − < x, ⇒⇐ .
2 m+1 m+2 2
n2 n2
(c) Note that ∀n ∈ N, xn = > = n.
n+1 n
∀x ∈ R, ∃N ∈ N such that N > x by the Archimedean Propetry.
Hence xN > N > x ⇒ (xn ) is not bounded.
Hence the sequence is divergent.
(d) Homework

2. (a) X = (n) and Y = (1 − n). Then X and Y diverge, and X + Y =

(1) ⇒ lim(X + Y ) = 1.
(b) X = ((−1)n ) and Y = ((−1)n ). Then X and Y diverge, and
XY = (1) ⇒ lim(XY ) = 1.

2
3. Note that Y = (X+Y) − X, and both X+Y and X are convergent.
Hence we have lim Y = lim[(X + Y ) − X] = lim(X + Y ) − lim(X),
which shows lim Y exists. Therefore Y converges.

4. WLOG, we assume that x > 0.

Then since x2 > 0, ∃N ∈ N, such that
x x x
∀n ≥ N, |xn − x| < . ⇒ xn > x − = > 0.
2 2 2
xn y n lim XY
Hence lim Y = lim(yn : n ≥ N ) = lim( : n ≥ N) = ,
xn lim X
which shows that the limit of Y exists. Therefore Y converges.

5. (a) By the Archimedean Property, ∀M ∈ R+ , ∃n ∈ N, such that

n > M . Since

2n = (1 + 1)n ≥ 1 + n > n > M,

the sequence is not bounded. Hence it is not convergent.

(b) ∀M ∈ R+ , by the Archimedean Property there exists an even
number n ∈ N such that n > M .
Then (−1)n n2 = n2 > n > M ⇒ the sequence is not bounded ⇒
it is not convergent.
−1 (−1)n 1
6. (b) Note that ≤ ≤ , and
n+2 n+2 n+2
−1 1
lim( ) = lim( ) = 0.
n+2 n+2
(−1)n
Hence by the Squeeze Theorem, lim( ) = 0.
n+2
n+1 1 1
(d) lim( √ ) = lim( √ + √ ) = 0 + 0 = 0.
n n n n n
7. Since (bn ) is bounded, we assume that |bn | ≤ M, ∀n ∈ N, with M > 0.
²
Since lim(an ) = 0, ∀² > 0, ∃N ∈ N such that |an − 0| < , ∀n ≥ N .
M
²
⇒ |an bn − 0| = |an − 0||bn | < M = ², ∀n ≥ N.
M
⇒ lim(an bn ) = 0.
Theorem 3.2.3 does not apply here, because lim(bn ) may not exist.
For example, (bn ) = ((−1)n ).

3
 9. Since
√ √ √ √
√ √ ( n + 1 − n)( n + 1 + n) 1 1
0 < n + 1− n = √ √ =√ √ < √ ,
n+1+ n n+1+ n 2 n
√
and lim(0) = 0√= lim(2 √ n),
lim(yn ) = lim( n + 1 − n) = 0 by the squeeze theorem.
r
n+1 1 n+1
Note that lim( ) = lim(1 + ) = 1 + 0 = 1 ⇒ lim( ) = 1.
n n n
And that
√ √ √ √ n 1
nyn = n( n + 1 − n) = √ √ =q .
n+1+ n n+1
+ 1
n

√ 1 1
Hence lim( nyn ) = = .
1+1 2
11. (a) Note that
√ 1 1
(3 n) 2n ≥ 3 2n ≥ 1
and that
√ 1 √ 1 1 √ 1 1
(3 n) 2n = ((3 n) 2 ) n ≤ ((3 n)2 ) n = (9n) n .
1 1
Since lim(9 n ) = 1 and lim(n n ) = 1,
1 1 1 1 1
lim((9n) n ) = lim(9 n n n ) = lim(9 n ) lim(n n ) = 1.

On the other hand, lim(1) = 1. Hence by the squeeze theorem,

√ 1
lim((3 n) 2n ) = 1.

(b) Note that

1 ln(n+1)
(n + 1) ln(n+1) = e ln(n+1) = e
Therefore,
1
lim((n + 1) ln(n+1) ) = lim(e) = e

13.
(a + b)n + ab a + b + ab
n a+b
lim(xn ) = lim( √ ) = lim( q )=
n2 + an + bn + ab + n 1 + a+b + nab2 + 1 2
n

4
14. (a)
1 1 1
1 ≤ n n2 ≤ (n nn )n = n n .
But
1
lim n n = 1 = lim 1.
1
Hence, by the Squeeze Theorem, we have lim(n n2 ) = 1.
(b) Notice that
1 1 1 1
1 ≤ n! n2 = (n(n − 1)(n − 2)... × 2 × 1) n2 ≤ nn n2 = n n .

and
1
lim 1 = lim n n = 1.
By the Squeeze Theorem we have
1
lim n! n2 = 1.

15. Note that 1 1 1

b = bn n ≤ (an + bn ) n ≤ (2bn ) n .
But
1 1 1
lim(2bn ) n = lim 2 n × lim(bn ) n = 1 × b = b = lim b.
By the Squeeze Theorem we have
1
lim(an + bn ) n = b.

16. (a)
an+1 an+1
lim( ) = lim( n ) = lim a = a.
an a
Since a < 1,
lim(an ) = 0.
(b)
bn+1
2n+1 b
lim( bn
)= .
2n
2
If b < 2, then
bn
lim = 0.
2n
bn b b
Otherwise, the limit does not exist since n
= ( )n and > 1.
2 2 2

5
 (c)
n+1 1
bn+1 n+1 1+ n 1
lim n = lim = lim = < 1.
bn
bn b b
(d)
23(n+1)
32(n+1) 8 8
lim 23n
= lim = <1
32n
9 9
Therefore,
23n
lim =0
32n
17. (a) xn = 3
(b) xn = n

18. Since L > 1, the sequence (Ln ) is divergent.

Take L > r > 1, ² = L − r > 0, then ∃K ∈ N such that ∀n ≥ K,
xn+1
| − L| < ²
xn
Therefore,
xn+1
> L − ² = r for n ≥ K.
xn
Hence, for n ≥ K,

xn+1 > rxn > r2 xn−1 > ... > rn−K+1 xK .

Since r > 1, rn−K+1 is not bounded. Therefore, xn+1 is not bounded,

and hence so is xn .
Hence, (xn ) is not convergent since all convergent sequence must be
bounded.

19. (a) Since

(n + 1)2 an+1 (n + 1)2 a n+1 2

lim( 2 n
) = lim( 2
) = lim(( ) )×lim(a) = 12 ×a = a < 1
na n n
the sequence converges to 0.
(b) Since
bn+1
(n+1)2 b × n2 n 2
lim( bn ) = lim( 2
) = lim(b)×lim(( ) ) = b×12 = b > 1
n2
(n + 1) n+1

the sequence diverges.

6
 (c) Since
bn+1
(n+1)! b
lim( bn ) = lim( )=0<1
n!
n+1
the sequence converges to 0.
(d) Since
(n+1)!
(n+1)n+1 n n
n!
=( ) < 1,
nn
n+1
n!
the sequence is decreasing. Also notice that 0 < n , we have
n
n!
lim( n ) converges.
n
20. Let 0 < L < r < 1. By definition, since ² = r − L > 0, ∃N ∈ N such
that
1
|xn n − L| ≤ ² ∀n ≥ N.
Therefore,
1
xn n ≤ L + ² = r ∀n ≥ N.
Hence 0 < xn < rn for n ≥ N . Note that lim(rn ) = 0 since r < 1, by
the squeeze theorem lim(xn ) = 0.

21. (a) (a) Here is a counter example: (xn ) = (1)

(b) Here is a counter example: (xn ) = (n)

22. The sequence (yn ) is convergent.

Proof:
Let lim(xn ) = L. Then ∀² > 0, since ²/2 > 0, ∃N1 ∈ N such that

|xn − L| < ²/2 ∀n ≥ N1 .

Moreover, we have a M > 0 such that

|xn − yn | < ²/2 for all n ≥ M.

Hence

|yn −L| = |yn −xn +xn −L| ≤ |xn −L|+|yn −xn | < ² for n ≥ max{M, N1 }.

Therefore, lim(yn ) = L = lim(xn ).

7
 23. Proof:
Since
|a − b| + a + b a + b − |a − b|
max{a, b} = min{a, b} =
2 2
we conclude that
|xn − yn | + xn + yn
(un ) = (max{xn , yn }) = ( )
2
xn + yn − |xn − yn |
(vn ) = (min{xn , yn } = ( )
2
Since (xn ), (yn ) are convergent, it follows from the limit theorems that
(un ), (vn ) are also convergent.

24. Proof:
Denote lim(xn ) = x lim(yn ) = y lim(zn ) = z.
WLOG, we may assume that x > y > z.
|y − x| |z − y|
If x > y > z, then for ² = min{ , }, ∃N1 ∈ N , such that
2 2
|xn − x| < ² |yn − y| < ² |zn − z| < ² ∀n > N1 .

Therefore
x+y y+z
xn > > yn > > zn ∀n > N1 .
2 2
In this case,

wn = mid{xn , yn , zn } = yn ∀n > N1 .

So (wn ) is convergent since (yn ) is convergent.

Indeed, note that mid{a, b, c} = a + b + c − min{a, b, c} − max{a, b, c}.

Since both min{xn , yn , zn } and max{xn , yn , zn } converge (by repeated
application of Q23), it then follows from limit theorem that (| {xn , yn , zn })
converges.

Section 3.3

8
2. Proof:
Since
1 −(xn − 1)2
xn+1 − xn = 2 − − xn = < 0,
xn xn
the sequence (xn ) is decreasing. Now we prove that the sequence is
bounded below by Principle of Mathematical Induction.
Note that x1 > 1, we suppose that xk > 1 for some k ∈ N, then
1
xk+1 = 2 − > 2 − 1 = 1,
xk
so (xn ) is bounded below by 1.
Since (xn ) is decreasing and bounded below, (xn ) is convergent. Denote
its limit as L. Then by passing by limit,
1
L=2− since L > 1.
L
By solving the equation, we get L = 1.

5. Proof:
First we use Principle of Mathematical Induction to show (yn ) is in-
creasing. Since q
√ √
y2 = p + p > p = y1 ,
the first case is true. Suppose yk > yk−1 for some k ∈ N, then
√ √
yk+1 = p + yk > p + yk−1 = yk ,

and thus the sequence is increasing.

√
We use the same method to conclude (yn ) is bounded above by 1+2 p.
Note that
√ √
y1 = p < 1 + 2 p.
√
Suppose yk < 1 + 2 p for some k ∈ N, then
√ √ √
(yk+1 )2 = p + yk < p + 1 + 2 p < 4p + 1 + 4 p = (1 + 2 p)2 .
√
Thus (yn ) is bounded above by 1p+ 2 p.
r limit as L, thenrL = p + L. By solving the equation, we
Denote its
1 1 1 1
get L = p + + or − p + + .
4 2 4 2 r
1 1
Since yn > 0 ∀n, its limit L > 0, so the limit should be p + + .
4 2

9
7. The sequence (xn ) is divergent.
Proof: Suppose that (xn ) is convergent. Then it is bounded by a
positive natural number M . Since x1 = a > 0, suppose xk > 0 for
some k ∈ N, then
1
xk+1 = xk + > 0.
xk
Thus xn > 0 ∀n ∈ N. So
1 1
xn+1 − xn = > ∀n ∈ N.
xn M
Therefore, ∀N1 ∈ N, ∃m > N1 , such that
1 1 1 1
xM +m+1 − xm = + + ... + >M× = 1,
xm xm+1 xM +m M
which contradicts to the Cauchy Criterion.
Thus the sequence is divergent.

A. Let (xn ), (yn ), (an ), (bn ) be Cauchy sequences of rational numbers. If

(xn ) ∼ (an ) and (yn ) ∼ (bn ), then we claim that (xn yn ) ∼ (an bn ) and
thus multiplication of ”real numbers” is well defined.

Proof: Since (xn ) is a Cauchy sequence, (xn ) is bounded by a positive

rational number M . Similarly, (bn ) is bounded by a positive rational
number N.
∀ positive rational number ², since (xn ) ∼ (an ), ∃ a natural number
N1 , such that
²
|xn − an | < ∀n > N1 .
2N + 1
Similarly, since (yn ) ∼ (bn ), ∃ a natural number N2 , such that
²
|yn − bn | < ∀n > N2 .
2M + 1

Therefore,
M² N²
|xn yn − an bn | < |xn ||yn − bn | + |bn ||xn − an | < +
2M + 1 2N + 1
² ²
< + =² ∀n > K = max{N1 , N2 }.
2 2
So (xn yn ) ∼ (an bn )

10
C. Let z > 0. Show that for any n ∈ N, there exits x > 0 such that xn = c.
Proof:
Case 1: c > 1.
For an arbitrary k ∈ N, k > 1, consider set S = {x ∈ R : x > 0, xk < c}.
First, S is bounded above by c.
Because if x ≤ 1, clearly x ≤ 1 < c. If x > 1 and x ∈ S, then
x < xk < c.
Since S is bounded above, it must have a least upper bound.
Let u = supS. Claim: uk = c.
Proof of the claim: Suppose on the contrary that uk 6= c.
If uk < c, then

1 k k k(k − 1) 1
(u(1 + )) = uk (1 + + 2
+ ... + k )
n n 2n n
k k k(k − 1) 1
≤ u (1 + + + ... + )
n 2n n
k 2k − 1
= u (1 + ).
n

By the Archimedean Property, there exits n ∈ N such that

1 1 c
< k ( k − 1).
n 2 −1 u
Hence
1 k 2k − 1 1 c
(u(1 + )) ≤ uk (1 + ) < uk (1 + (2k − 1) k ( k − 1)) = c.
n n 2 −1 u
1
Note that u(1 + ) > u and u(1 + n1 ) ∈ S.
n
However, this contradicts that u = supS.
1
If uk < c, then consider (u(1 − ))k .
n
1 k k k
(u(1 − )) ≥ u (1 − ) by the Bernuolli’s Inequality.
n n
Note that by the Archimedean Property, there exits n ∈ N,
1 1 c
such that < (1 − k ).
n k u
Hence
1 k k c
(u(1 − )) ≥ uk (1 − ) > uk (1 − (1 − k )) = c.
n n u

11
 1
Note that u(1 − ) > u.
n
However, this contradicts that u = supS.
Hence we conclude uk = c.
Case 2: c < 1.
1
Then > 1. Hence for any n ∈ N, there exist x > 0 such that
c
1 1 1 1
xn = ⇒ for > 0, ( )n = n = c.
c x x x
Case 3: c = 1. Then for any n ∈ N, 1 > 0 and 1n = 1 = c .

12