0% found this document useful (0 votes)
204 views9 pages

MA2108S Tutorial 4 Answer Sheet: Qian Yujie, Shi Yangyang, Xu Jingwei 21 February 2011

This document contains solutions to exercises from a real analysis tutorial. It includes 3 proofs: 1) A proof that if S is a nonempty bounded set, the closed interval [inf S, sup S] contains S. 2) A proof of the principle of iterated suprema, establishing that supx,y h(x,y) = supx supy h(x,y) = supy supx h(x,y) for functions h with bounded range. 3) A proof that the intersection of nested closed intervals In = [0, 1/n] for n ∈ N equals {0}. The document provides detailed step-by-step workings for each
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
204 views9 pages

MA2108S Tutorial 4 Answer Sheet: Qian Yujie, Shi Yangyang, Xu Jingwei 21 February 2011

This document contains solutions to exercises from a real analysis tutorial. It includes 3 proofs: 1) A proof that if S is a nonempty bounded set, the closed interval [inf S, sup S] contains S. 2) A proof of the principle of iterated suprema, establishing that supx,y h(x,y) = supx supy h(x,y) = supy supx h(x,y) for functions h with bounded range. 3) A proof that the intersection of nested closed intervals In = [0, 1/n] for n ∈ N equals {0}. The document provides detailed step-by-step workings for each
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 9

MA2108S Tutorial 4 Answer Sheet

Qian Yujie, Shi Yangyang, Xu Jingwei


21 February 2011

Section2.4
9. Perform the computations in (a) and (b) of the preceding exercise for the function
h : X × Y → R defined by (
0 if x < y,
h(x, y) :=
1 if x ≥ y.

Proof. (a) Note that h(y, y) = 1, hence f (x) := sup{h(x, y) : y ∈ Y } = 1 for all x ∈ X.
Thus inf{f (x) : x ∈ X} = 1.
1+x 1+x
(b) For each x ∈ X, we take y = , then we have h(x, ) = 0,then g(y) =
2 2
inf{h(x, y) : x ∈ X} = 0 for all y ∈ Y . Thus sup{g(y) : y ∈ Y } = 0.

10. Let X and Y be nonempty sets and let h : X × Y → R have bounded range in R. Let
f : X → R and g : Y → R be defined by

f (x) := sup{h(x, y) : y ∈ Y }, g(x) := inf{h(x, y) : x ∈ X}.

Prove that

sup{g(y) : y ∈ Y } ≤ inf{f (x) : x ∈ X}.

Proof. Since f (x0 ) ≥ h(x0 , y)∀ y ∈ Y , and g(y0 ) ≤ h(x, y0 ) ∀ x ∈ X. Take any x0 ∈ X,
then f (x0 ) ≥ h(x0 , y) ≥ g(y)∀ y ∈ Y , namely f (x) ≥ g(y) ∀ x ∈ X ∀ y ∈ Y .
Hence f (x) is an upper bound of {g(y) : y ∈ Y }, sup{g(y) : y ∈ Y } ≤ f (x) ∀ x ∈ X,
Therefore sup{g(y) : y ∈ Y } is a lower bound of {f (x) : x ∈ X}.
Hence sup{g(y) : y ∈ Y } ≤ inf{f (x) : x ∈ X}.

11. Let X and Y be nonempty sets and let h : X × Y → R have bounded range in R.
Let F : X → R and G : Y → R defined by

F (x) := sup{h(x, y) : y ∈ Y }, G(y) := sup{h(x, y) : x ∈ X}.

1
Establish the Principle of the Iterated Suprema

sup h(x, y) = sup sup h(x, y) = sup sup h(x, y).


x,y x y y x

Proof. ∀ x0 ∈ X, y ∈ Y, h(x0 , y) ≤ F (x0 ). Hence sup h(x, y) ≤ sup{F (x) : ∀ x ∈ X}.


x,y
On the other hand, sup h(x, y) ≥ h(x0 , y) ∀x0 ∈ X, y ∈ Y , hence sup h(x, y) ≥ sup{F (x) :
x,y x,y
x ∈ X}. Hence sup h(x, y) = sup{F (x) : ∀ x ∈ X} = sup sup h(x, y).
x,y x y
By symmetry, sup h(x, y) = sup sup h(x, y).
x,y y x
Hence sup h(x, y) = sup sup h(x, y) = sup sup h(x, y).
x,y x y y x

12. Given any x ∈ R, show that there exists unique n ∈ Z such that n − 1 ≤ x < n.

Proof. (1) If x = 0, then let n := 1, n − 1 = 0 ≤ x < 1 = n.

(2) If x > 0, Let S := {N : n > x}. By the Archimedean Property, ∃ n ∈ N, such


that n > x, S is nonempty. By the well ordering property, a smallest element n0 can
be picked in S. Since n0 − 1 < n0 , therefore n0 − 1 ∈
/ S. Namely n0 − 1 ≤ x. Hence
n0 − 1 ≤ x < n0 .

(3) If x < 0, then −x > 0. By the Archimedean Property, ∃ m ∈ N with m > −x.
Therefore m + x > 0. By the result of part (2), ∃ n0 ∈ N n0 − 1 ≤ m + x < n0 . Hence
(n0 − m) − 1 ≤ x < (n0 − m) with (n − m) ∈ Z. Take n := n0 − m, then n − 1 ≤ x < n.

Uniqueness: suppose to the contrary, for any x ∈ R, ∃ m, n ∈ Z and m 6= n, such


that n − 1 ≤ x < n and m − 1 ≤ x < m. Without loss of generality, assume m > n,
because m, n ∈ Z, therefore m ≥ n + 1. Then, m − 1 ≥ n > x, which contradicts with
the fact that m − 1 ≤ x, therefore m = n.

Hence given any x ∈ R, there is a unique n ∈ Z such that n − 1 ≤ x < n.

1
13. If y > 0, show that there exists n ∈ N such that < y.
2n
1
Proof. Since y > 0, by The Archimedean Property, ∃n ∈ N, such that < y. Since
µ ¶ n
n 1 1
2n = (1 + 1)n ≥ 1 + > n for n ∈ N , n < < y.
1 2 n
1
Hence if y > 0, there exists n ∈ N such that n < y.
2

2
18. If u > 0 is any real number and x < y, show that there exists a rational number r
such that x < ru < y. (Hence the set {ru : r ∈ Q is dense in R.)
x y y x
Proof. Since u > 0, and x < y, therefore < . Since − > 0, by the Archimedean
u u u u
1 y x x·n y·n
Property, ∃ n ∈ N, such that < − . Hence 1 + < . Apply Corollary
n u u u u
x·n
2.4.6 (or result of Exercise 2.4.12), ∃ m ∈ N, such that m − 1 ≤ < m. Therefore,
u
x·n y·n x·n y·n m
m≤ +1 < , whence <m< . Thus the rational number r :=
u u u u n
satisfies x < ru < y.

ExtraA Let A and B be nonempty subsets of R. Let

A + B = {a + b : a ∈ A, b ∈ B}. and AB = {ab : a ∈ A, b ∈ B}.

Show that inf(A+B) = inf A+inf B. If A and B are both subsets of positive real number,
show also that inf(AB) = inf A inf B.
Proof. (1) ∀a ∈ A, ∀b ∈ B, inf A + inf B ≤ a + b. Thus inf A + inf B is a lower bound of
A + B.
For any ² > 0, ∃ a 2² ∈ A, such that a 2² < inf A+ 2² ; and ∃ b 2² ∈ B, such that b 2² < inf B + 2² .
Therefore, for any ² > 0, ∃ a 2² ∈ A, ∃ b 2² ∈ B, such that a 2² + b 2² < inf A + inf B + ². By
Lemma 2.3.4 inf A + inf B = inf(A + B).

(2)∀ a ∈ A a ≥ inf A ≥ 0. ∀ b ∈ B, b ≥ inf B ≥ 0. Hence ∀ a ∈ A, ∀b ∈ B, inf A·inf B ≤


a · b. Hence inf A · inf B is a lower bound of AB. Let u be a lower bound of AB.
u u
Then ∀ a ∈ A, ∀b ∈ B, ab ≥ u. a, b > 0, thereforea ≥ , is a lowerbound of A.
b b
u
Henceinf A ≥ which implies u ≤ inf A · b. Since inf A ≥ 0,
b
u u
Case 1: if inf A > 0 then ∀ b ∈ B b ≥ . Therefore inf B ≥ then inf A·inf B ≥ u;
inf A inf A
Case 2: if inf A = 0, y > 0, therefore u = 0, thus ∴ inf B · inf A = 0 = u. u is any
lower bound of AB, hence inf A · inf B = inf(AB)

Hence inf A · inf B = inf(AB).

3
Section 2.5
Question 1
If I := [a, b] and I 0 := [a0 , b0 ] are closed intervals in R, show that I ⊆ I 0 if and only if
a0 ≤ a and b ≤ b0 , which leads to contradiction.

Proof. If ∀a, b ∈ I ⇒ a, b ∈ I 0 , then a0 ≤ a, b ≤ b0


Conversely, if a ≤ a, b ≤ b0 , then a0 ≤ x ≤ b0 whenever a ≤≤ x ≤ b. Thus I ⊆ I 0 .

Question 2
If S ⊆ R is nonempty, show that S is bounded if and only if there exists a closed bounded
interval I such that S ⊆ I.
Proof. (1) Claim: The set S is bounded if there exists a closed bounded interval I such
that S ⊆ I.
Take I = [a, b], then for any x ∈ S, a ≤ x ≤ b.
Hence for any x ∈ S, a ≤ x ⇒ S is bounded below.
For any x ∈ S, x ≤ b ⇒ S is bounded above.
Hence S is bounded.
The set S is bounded if there exists a closed bounded interval I such that S ⊆ I.
(2) Claim: The set S is bounded only if there exists a closed bounded interval I such that
S ⊆ I.
As S is bounded above, there exists a number u ∈ R such that s ≤ u for all s ∈ S.
Also, S is bounded below, there exists a number w ∈ R such that w ≤ s for all s ∈ S.
Hence w ≤ s ≤ u.
Hence there exists closed bounded interval I = [w, u] such that S ⊆ I.
Hence S is bounded only if there exists a closed bounded interval I such that S ⊆ I.
Hence S is bounded if and only if there exists a closed bounded interval I such that
S ⊆ I.

Question 3
If S ⊆ R is a nonempty bounded set, and Is := [inf S, supS], show that S ⊆ Is . Moreover,
if J is any closed bounded interval containing S, show that Is ⊆ J.
Proof. (1) Claim: The set S ⊆ Is .
Take any s ∈ S, then inf S ≤ s ≤ sup S ⇒ s ∈ Is . Hence S ⊆ Is .

(2) Claim: The set Is ⊆ J if S ⊂ J.


Suppose J = [a, b]. Then for any s ∈ S, since S ⊂ J, we have a ≤ s ≤ b. Hence a is a
lower bound of S and b is an upper bound of S. Therefore a ≤ inf S.
For any s ∈ S,s ∈ J, then b ≥ s. Therefore b is a upper bound of S. Hence sup S ≤ b ≤
a ≤ b.
Hence Is ⊆ J by Question 1. Hence if J is any closed bounded interval containing S,
Is ⊆ J.

4
Question 7

\
Let In := [0, 1/n] for n ∈ N. Prove that In = {0}.
n=1

\
Proof. Clearly 0 ∈ In for any n ∈ N . Hence 0 ∈ In .
n=1

\
Claim: The set {0} = In .
n=1

\ ∞
\ ∞
\
Suppose {0} 6= In . Since 0 ∈ In , there exists a x 6= 0, such that x ∈ In .Hence
n=1 n=1 n=1
x ∈ In for all n ∈ N . As In = [0, 1/n], x > 0. Hence by the Archimedean Property,

\
there exists a K ∈ N such that 1/K < x. Hence x ∈ / IK . Hence x ∈
/ In . This is a
n=1
contradiction!

\
Hence In = {0}.
n=1

Question 8

\
Let Jn := (0, 1/n) for n ∈ N. Prove that Jn = ∅.
n=1

\
Proof. Suppose there exists a x, such that x ∈ Jn . Hence x ∈ Jn for all n ∈ N . As
n=1
Jn = (0, 1/n), we have x > 0. Hence by the Archimedean Property, there exists a K ∈ N

\
such that 1/K < x. Hence x ∈ / JK . Hence x ∈
/ Jn . This is a contradiction!
n=1

\
Hence Jn = ∅.
n=1

Question 9

\
Let Kn := (n, ∞) for n ∈ N. Prove that Kn = ∅.
n=1

5

\
Proof. Suppose there exists a x, such that x ∈ Kn . Hence x ∈ Kn for all n ∈ N . Also,
n=1
Kn = (n, ∞), we have x > 0. Hence by the Archimedean Property, there exists a K ∈ N
\∞
such that x < K. Hence x ∈
/ KK . Hence x ∈/ Kn . This is a contradiction!
T n=1
Hence ∞ n=1 Kn = ∅.

Question 10
With the notation in the proofs of Theorems 2.5.2 and 2.5.3, show that we have η ∈

\
T∞
n=1 In . Also show that [ξ, η] = In .
n=1

Proof. Recall that η = inf{bn : n ∈ N} and the fact that for each n ∈ N, an is a lower
bound of the set {bn : n ∈ N}, it is clear that η ≥ an for each n. Thus η ≥ ξ = sup{an :
n ∈ N}.
Next, it is clear that if x ∈ [ξ, η], then an ≤ ξ ≤ Tx ≤ η ≤ bn for any n. Thus
x ∈ [an , bn ] = In for each
T n. It is then clear that [ξ, η] ⊂ n∈N In .
Conversely, if x ∈ n∈N In , then an ≤ x ≤ bn for all n and hence x is an upper
bound of T {an : n ∈ N} and a lower bound of {bn : n ∈ N}. Thus ξ ≤ x ≤ η and hence
[ξ, η] ⊂ n∈N In . T
Thus [ξ, η] = n∈N In .

Question 14
a1 a2 an b1 b2 bm
14.Show that if ak , bk ∈ {0, 1, . . . , 9} and if + 2 +· · ·+ n = + 2 +· · ·+ m 6= 0,
10 10 10 10 10 10
then ak = bk for k = 1, · · · , n.
ak bk
Proof. Suppose not. Then there is a least number i such that k = k when k < i and
10 10
ai bi
6= i .
10i 10
ai bi
Without loss of generality, we may assume i > i . Since ai , bi ∈ {0, 1, . . . , 9}, we
10 10
ai bi 1
have ai ≥ bi + 1. Hence i ≥ i + i . Note that
10 10 10
bi+1 bi+2 bm 9 9 9
i+1
+ i+2 + · · · + m ≤ i+1 + i+2 + · · · + m
10 10 10 10 10 10
9 1 m−i
i+1 [1 − ( 10 ) ] 1 1 1
= 10 1 = i [1 − m−i ] < i .
1 − 10 10 10 10

6
Then we have
ai bi 1 bi bi+1 bi+2 bm
i
≥ i + i > i + i+1 + i+2 + · · · + m .
10 10 10 10 10 10 10
Hence
a1 a2 an a1 a2 ai b1 b2 bn
+ 2 + ··· + n ≥ + 2 + ··· + i > + 2 + ··· + n.
10 10 10 10 10 10 10 10 10
This is a contradiction.
Hence ak = bk for all k.

Question 16
1 2
Express and as periodic decimals.
7 19
1
Solution: By computation, we have = 0.142857 · · · 142857 · · ·
7
2
and = 0.105 263 157 894 736 842 · · · 105 263 157 894 736 842 · · · .
19

Question 17
What rationals are represented by the periodic decimals 1.25137 · · · 137 · · · and 35.14653 · · · 653 · · · ?

Solution: Take x = 1. 25137 · · · 137 · · · . Then we have (100000 − 100)x = 125012. Hence
x = 31253/24975. Hence 1. 25137 · · · 137 · · · = 31253/24975.
Take y = 35. 14653 · · · 653 · · · . Then we have (100000 − 100)x = 3511139. Hence
x = 3511139/99900. Hence 35. 14653 · · · 653 · · · = 3511139/99900.

Extra Question (C):


Let z > 0. Show that for any n ∈ N, there exists x > 0 such that xn = z.
Proof. Case 1: z > 1
Let S = {s ∈ R : s ≥ 0, sn < z}.
Since 1 ∈ S, then S is not an empty set. Also, S is bounded above by z, as z n > z and
sn < z n ⇔ s < z (recall that s, z ≥ 0). Therefore, the Completeness Property implies
that the set S has a supremum in R. Let x = supS. Note that x > 1.
First assume that xn < z.
(x+δ)n −xn = (x+δ−x)((x+δ)n−1 +(x+δ)n−2 x+· · ·+xn−1 ) < δ(x+δ)n−1 n < δ(x+1)n−1 n,
for n ≥ 2, n ∈ N and 0 < δ < 1.
Since z − xn > 0, by the Archimedean Property, there exists N ∈ N such that z − xn > N1 .
1 1
Let δ = n−1
. Then (x + δ)n − xn < δ(x + 1)n−1 n = < z − xn .
N (x + 1) n N

7
Hence (x + δ)n < z. It is a contradiction to the fact x := supS. Hence xn ≮ z.
Next assume that xn > z.
Then xn − (x − δ)n = (x − x + δ)(xn−1 + xn−2 (x − δ) + · · · + (x − δ)n−1 ) < δxn−1 n, for
n ≥ 2, n ∈ N and 0 < δ < 1.
1
Since xn −z > 0, by the Archimedean Property, there exists N ∈ N such that xn −z > .
N
1 n n n−1 1 n
Let δ = . Then x − (x − δ) < δx n = < x − z.
N xn−1 n N
Hence (x − δ)n > z. It is a contradiction to the fact x := supS. Hence xn ≯ z.
Combining these two cases, we conclude that xn = z.

Case 2: z = 1
Clearly, if x = 1, then xn = 1.
Case 3: 0 < z < 1
Then 1/z > 1. By Case (1), there exists x > 1 such that xn = 1/z. Observe that
(1/x)n = z.

Section3.1

Question: 5(a)
n
Use the definition of the limit of a sequence to establish the limit of lim( 2 ) = 0.
n +1
¯ ¯
¯ n ¯
¯
Proof. Observe that ¯ 2 ¯ < n = 1.
n + 1 ¯ n2 n
1
Given any ² > 0, by the Archimedean Property, there exists k ∈ N, such that < ².
k
1 1
Therefore if n ≥ k, then ≤ < ².
n k
n n
Hence, | 2 | < ², for all n ≥ k. Hence lim( 2 ) = 0.
n +1 n +1

Question: 5(d)
n2 − 1 1
Use the definition of limit of a sequence to establish the limit of lim( 2
)= .
2n + 3 2
n2 − 1 1 2n2 − 2 − 2n2 − 3 5 5 5 5
Proof. Observe that | 2
− | = | 2
|=| 2 |< 2 < 2 ≤ .
2n + 3 2 4n + 6 4n + 6 4n n n
1 ²
Given ² > 0, by the Archimedean Property, there exists k ∈ N, such that < .
n 5
5 5
Therefore take n ≥ k, then ≤ < ².
n k
n2 − 1 1 5
Hence, | 2 − | < < ², for all n ≥ k.
2n µ+ 3 2 ¶ n
n2 − 1 1
Hence, lim 2
= .
2n + 3 2

8
Question: 6(c)
1
Show that lim( √ ) = 0.
n+7
√ √
n n 1
Proof. Observe that | |< =√ .
n+7 n n
1
Given any ² > 0, by the Archimedean Property, there exists k ∈ N, such that k
< ²2 .
1 1 1
Therefore if n ≥ k, the ≤ ≤ ²2 , thus √ < ².
√ n k n
n
Hence, | | < ², for all n ≥ k.
n+1

Question: 6(d)
(−1)n
Show that lim( ) = 0.
n2 + 1
(−1)n n n n 1
Proof. Observe that | 2
|≤| 2 | < | 2| = .
n +1 n +1 n n
1 ²
Given any ² > 0, by the Archimedean Property, there exists k ∈ N, such that < .
k 5
1 1
Therefore if n ≥ k, then ≤ < ².
n k
(−1)n n
Hence, | 2 | < ², for all n ≥ k.
n +1

You might also like