Real Analysis Homework 10, due 2007-11-28 in class

1. (10 points) Given the function f (x, y) = x2 y 2 , (x2 + y 2 )2 (x, y) I = (0, 1) (0, 1)

compute the following iterated integrals (hint: use trigonometric substitution) : Z 1 Z 1 Z 1 Z 1 f (x, y) dx dy and f (x, y) dy dx.
0 0 0 0

Is f (x, y) L (I) or not? Give your reasons. Solution: For xed x we have Z 1 Z f (x, y) dy =
0

= Hence

1 x

tan1

1 x

Z tan1 1 x x2 y 2 x2 x2 tan2 dy = 2 2 d (x tan ) 2 2 (x2 + x2 tan2 ) 0 0 (x + y ) 1 1 1 1 1 1 cos 2d = sin tan . cos tan = x x x 1 + x2
1

1 Z 1 0

x2 y 2 dy dx = . 2 + y 2 )2 4 (x 1 1 + y2

Similarly (by symmetry) we have Z and so Z

1

dx (x2 + y 2 )2

x2 y 2

1 Z 1 0

x2 y 2 dx dy = . 2 + y 2 )2 4 (x

Since the two iterated integrals are dierent, by Fubini theorem, f (x, y) 6 L (I) . 2. (10 points) Do Exercise 1 in p. 96. Solution:

RR R By assumption we know Ex E (x, y) dy = 0 a.e. in x R and so E E (x, y) dxdy = |E| = 0. By Tonellis Theorem again, we have |Ey | = 0 a.e. in y R. (b) . Let E = (x, y) R2 : f (x, y) = . It is a measurable set in R2 . Set Ex = {y R : f (x, y) = } and Ey {x R : f (x, y) = }. By (a) we have |Ex | = 0 in R for a.e. x R 2 |E| = 0 in R |Ey | = 0 in R for a.e. y R. 1

(a) . We set Ex = {y R : (x, y) E}and Ey = {x R : (x, y) E} . By Tonellis Theorem, we have ! Z Z Z Z ZZ E (x, y) dxdy = E (x, y) dy dx = E (x, y) dx dy.
E R Ex R Ey

3. (10 points) Do Exercise 2 in p. 96. Solution: Let h1 (x, y) = f (x) . As a function on R2n , it is measurable since f (x) : Rn R measurable. More precisely, for any a R we have (x, y) R2n : f (x, y) > a = {x Rn : f (x) > a} Rn S {} is

is also a measurable function on R2n . Given E1 Rn , E2 Rn , both are measurable in Rn . By E1 (x) E2 (y) = E1 E2 (x, y) , we know that E1 E2 (x, y) 0 is a measurable function on R2n . Hence the set E1 E2 is measurable in R2n . By Tonellis Theorem ZZ Z Z E1 E2 (x, y) dxdy = E1 E2 (x, y) dy dx |E1 E2 | = E1 E2 Rn Ex Z Z Z Z E1 E2 (x, y) dy dx = [E1 (x) E2 (y)] dy dx = E2 E1 E2 ZE1 Z = E1 (x) dx E2 (y) dy = |E1 | |E2 | .
E1 E2

where by repeated application of Lemma 5.2, we know that the RHS is a measurable set in R2n . Similarly, the function h2 (x, y) = g (y) is also a measurable function on R2n . Then by Theorem 4.10, we know that [ h1 (x, y) h2 (x, y) = f (x) g (y) : R2n R {}

4. (10 points) Do Exercise 3 in p. 96. Solution: We rst know that f (x)f (y) is measurable on (0, 1)(0, 1) . By Fubini Theorem, if F (x, y) = f (x)f (y) is integrable on (0, 1)(0, 1) , then for a.e. y (0, 1) , F (x, y) L1 (0, 1) (as a function of x). Hence f (x) L1 (0, 1) .

Real Analysis Homework 11, due 2007-12-5 in class

1. (10 points) Do Exercise 4 in p. 96. Solution: Choose a = x, b = x and integrate over [0, 1] to get Z 1Z 1 |f (t + x) f (t x)| dtdx c.
0 0

Hence the function F (t, x) = |f (t + x) f (t x)| is integrable on E = [0, 1] [0, 1] . Consider the linear transformation = t + x, = t x. We have ZZ ZZ ZZ + 1 1 , dd = F (t, x) dtdx = F |f () f ()| dd 2 E 2 2 2 E E

0) where E is the diamond-shaped region in (, ) space with vertices (0, 0) , (1, 1) , (1, 1), (2, . In ) . By Fubini Theorem, there exists 1 , 3 such particular we know that |f () f ()| L (E 0 2 2 that as a function of we have |f (0 ) f ()| L E0 , E0 = { : (0 , ) E } , E0 > 1. By |f ()| |f (0 ) f ()| + |f (0 )| , we see that |f ()| L E0 . Since f () is periodic with period 1 and E0 > 1, we know that f L (0, 1) . 2. (10 points) Do Exercise 6 in p. 97. Solution: By denition, we have (assume f L1 (R)) Z Z (x) = f (t) cos xtdt i f (t) sin xtdt, f

x R.

and by Tonellis Theorem we have ZZ ZZ + [f (t y) g (y)] dtdy

R2

We claim that if f, g L1 (R) , then the function f (t y) g (y) L1 R2 (a a function of (t, y)). To see this, note that [f (t y) g (y)]+ |f (t y)| |g (y)| , (t, y) R2 |f (t y)| |g (y)| dtdy = Z |g (y)| dy Z |f ()| d < .

Similarly

[f (t y) g (y)] dtdy < . Hence the claim is true. Now Z Z \ (f g) (t) cos xtdt i (f g) (t) sin xtdt (f g) (x) = Z Z Z Z = f (t y) g (y) dy cos xtdt i f (t y) g (y) dy sin xtdt
R2

RR

R2

= I1 + iI2

where by Fubini Theorem Z Z Z Z I1 = g (y) f (t y) cos xtdt dy = g (y) f () cos x (y + ) d dy Z Z = g (y) f () [cos xy cos x sin xy sin x] d dy Z Z Z Z = f () cos xd g (y) cos xydy f () sin xd g (y) sin xydy h i = Re f (x) g (x) . 1

Similarly we have Z Z I2 = g (y) f (t y) sin xtdt dy Z Z = g (y) f () [sin xy cos x + cos xy sin x] d dy Z Z Z Z = f () cos xd g (y) sin xydy f () sin xd g (y) cos xydy h i = Im f (x) g (x) . We conclude the identity \ (f g) (x) = f (x) g (x) , 3. (10 points) Do Exercise 10 in p. 97. Solution: Let Vn (r) be the volume of the ball in Rn with radius r > 0. For convenience, denote Vn (1) = Vn . We have Lemma 0.1 There holds the formula Vn = Z
1

x R.

Vn1

Proof. Let S = (x1 , , xn ) : x2 + + x2 1 . Then n 1 # Z Z 1 "Z Z dx1 dxn = dx1 dxn1 dxn = Vn =
S 1 Exn

p 1 2 d.

|Exn | dxn

where Exn = {(x1 , , xn1 ) : (x1 , , xn1 , xn ) S} = (x1 , , xn1 ) : x2 + + x2 1 x2 1 n1 n Vn = Z

1

and so

Vn1

p 1 x2 dxn . n

Remark 0.2 Similarly we have Vn (r) = Lemma 0.3 We have

Vn1

p r2 2 d for any

for any

r > 0.

(0.1)

Vn (r) = rn Vn

r > 0.

(0.2)

Proof. We can prove (0.2) using (0.1) and induction. Assume (0.2) holds for all dimensions less than or equal to n 1. Then s 2 Z r Z r p Vn1 r2 2 d = Vn1 r 1 Vn (r) = d r r r Z 1 Z 1 p p 2 rd = Vn1 r 1 rn1 Vn1 1 2 rd (by induction hypothesis) =
1

= rn

p Vn1 1 2 d = rn Vn .

 By the above, we have Vn = Z

1

Vn1 Z

= 2Vn1 The proof is done.

Z p 1 2 d = 1 n1 2 2 d.

n1 1 2 2 Vn1 d

4. (10 points) Do Exercise 11 in p. 97. Solution: We already know that (from calculus) Z 2 ex dx = .

By Tonellis Theorem we can easily get Z

e|x| dx =

Rn

n .

Real Analysis Homework 12, due 2007-12-12 in class

1. (10 points) Do Exercise 1 in p. 123. Solution: f is dened and measurable on Rn . Assume |f | > 0 on a set E with|E| > 0. We |f | > 1 > 0. Denote the set |f | > 1 by EN . We know that there exists N N such that N N 1 have N EN (x) |f (x)| on EN (and on Rn also) so 1 (x) f (x) N EN on Rn .

Hence by estimate (7.7) in p. 104, there exists a large number d such that 1 c1 |EN | 1 (x) f (x) for all |x| d. n N |x| N EN On the compact set S = {x Rn : 1 |x| d} , since f (x) is lower semicontinuous on Rn with f (x) > 0 everywhere, it attains its positive minimum on S (see exercise 7, p. 61). Hence there exists a small c2 positive constant c2 such that f (x) |x|n on S. Let c = min c1 |EN | , c2 . We have N f (x) c |x|n for all |x| 1. 2. (10 points) Do Exercise 2 in p. 123. Solution: Assume || M on Rn and x is in the Lebesgue set of f. We have Z |(f ) (x) f (x)| |f (x y) f (x)| | (y)| dy n Z ZR M |f (x y) f (x)| | (y)| dy n |f (x y) f (x)| dy = B (O) B (O) where B (O) = {|y| } has measure C (n) n . Hence 1 |(f ) (x) f (x)| C (n) M |B (O)| and we know that 1 lim 0 |B (O)| Z Z

B (O)

|f (x y) f (x)| dy

B (O)

|f (x y) f (x)| dy = 0

due to Theorem 7.16.

3. (10 points) Let C0 (Rn ) be the space of all continuous functions on Rn with compact support. We know that it is dense in the space L1 (Rn ) (Lemma 7.3 of the book). It is also clear that each g (x) C0 (Rn ) is uniformly continuous on Rn . Use this dense property to show that if f L1 (Rn ) , then we have the following property called "Continuity of Translation in L1 ": Z lim |f (x + y) f (x)| dx = 0.
y0 Rn

Solution: For any > 0, choose a function g (x) C0 (Rn ) with Then for any y Rn we also have Z |f (x + y) g (x + y)| dx < . 3 Rn

Rn

|f (x) g (x)| dx < 3 .

Let Br (O) be the ball centered at the origin with large radius r > 0 such that it contains the support of g. Then g is uniformly continuous on Br (O) and so there exists 0 < < 1 such that if |y| , then |g (x + y) g (x)|
3

Now if |y| < < 1, we have Z |f (x + y) f (x)| dx Rn Z Z Z |f (x + y) g (x + y)| dx + |g (x + y) g (x)| dx + |g (x) f (x)| dx Rn Z Rn Rn |g (x + y) g (x)| dx + . + 3 3 Br+1 (O) 4. (10 points) There are many applications of the use of convolution in analysis. One easy example is the following. Let 0 if t 0 h (t) = 1 e t if t > 0. It is known that h (t) is a C function on R. Next let g (x) = h 1 |x|2 , x Rn , then g (x) C0 (Rn ) . One can divide it by its integral over Rn so that the new function (x) R x 1 (Rn ) satises C0 Rn (x) dx = 1. For any number > 0, let (x) = n . Then it R satises (x) C0 (Rn ) , (x) 0, (x) > 0 |x| < , Rn (x) dx = 1. Show that: (a) If f C (Rn ) , then (f ) (x) converges uniformly to f (x) on compact subsets of Rn as 0+ . (It is easy to see that (f ) (x) C (Rn ) . You do not have to show this.)

|Br+1 (O)|

for all x Rn .

(b) If f C0 (Rn ) , then for any > 0, (f ) (x) also has compact support.

Solution: (a) . We know that f is uniformly continuous on compact subsets of Rn . Let S be a compact subset of Rn . For any > 0 there exists > 0 such that |f (x y) f (x)| for all x S, |y| . For x S we have |(f ) (x) f (x)| Z Z = [f (x y) f (x)] (y) dy = [f (x y) f (x)] (y) dy n R B (O) Z Z |f (x y) f (x)| (y) dy (y) dy =
B (O) B (O)

where B (O) = {|y| } . Hence (f ) (x) converges uniformly to f (x) on S as 0+ . 2

(b) . Assume S is the compact support of f. Then Z Z (f ) (x) = f (x y) (y) dy =

Rn

B (O)

f (x y) (y) dy.

From above we see that if x S with dist (x, S) > , then we also have xy S for any y B (O) . / / For such x, we have Z (f ) (x) =
B (O)

f (x y) (y) dy = 0.

Hence (f ) (x) also has compact support.

Real Analysis Homework 13, due 2007-12-26 in class

1. (10 points) Do Exercise 21 in P. 144. Solution: Let f Lp (Rn ) , where 0 < p < is a constant. We rst recall the following elementary calculus inequality: Lemma 1 For any 0 < p < , there exist positive constants c (p) and C (p) depending only on p such that (1) c (p) (ap + bp ) (a + b)p C (p) (ap + bp ) for all a > 0, b > 0. Let {rk } be the set of all rational numbers. For any k and any 0 < p < , by (1) the k=1 function |f (y) rk |p is clearly locally integrable on Rn . By Theorem 7.11, for each k we have Z 1 |f (y) rk |p dy = |f (x) rk |p (2) lim Q&x |Q| Q S for a.e. x Rn . Let Zk be the set such that (2) is not valid, |Zk | = 0, and set Z = Zk , k=1 |Z| = 0. If x Z, then by the inequality (in below, Q is centered at x and k is arbitrary) / Z Z 1 1 |f (y) f (x)|p dy {|f (y) rk | + |rk f (x)|}p dy |Q| Q |Q| Q Z Z C (p) C (p) |f (y) rk |p dy + |rk f (x)|p dy |Q| Q |Q| Q Z C (p) |f (y) rk |p dy + C (p) |rk f (x)|p . = |Q| Q Hence lim sup
Q&x

By choosing rk approximating f (x) (note that f (x) is nite almost everywhere in Rn ; without loss of generality, we can assume it is nite everywhere), we obtain Z 1 lim sup |f (y) f (x)|p dy = 0 for all x Z / Q&x |Q| Q for any 0 < p < .

1 |Q|

|f (y) rk |p dy 2C (p) |rk f (x)|p

for all x Z. /

2. (10 points) Prove the following more general version of the Tchebyshev inequality: Assume R f 0 is measurable on E satisfying E f p dx < , where 0 < p < is a constant. Then for any > 0 we have Z 1 f p dx. |{x E : f (x) > }| p E Solution: Over the set S := |{x E : f (x) > }| we have f p p and so Z Z p f dx f p dx p |S| .
E S

The conclusion follows. 1

3. (10 points) Let H (x) be the Heaviside function given by 1 if x > 0 1 if x = 0 H (x) = 2 0 if x < 0. Find the set of those x R such that 1 lim + 2h h0 Z
x+h

f () d = f (x) .

(3)

xh

Determine if the point x = 0 is in the Lebesgue set of f or not. Solution: Clearly for any x R with x 6= 0 we have (3). For x = 0, we have
h0+

lim

1 2h

x+h

f () d = lim

xh

h0+

1 2h

f () d = lim

h0+

1 2h

f () d =

1 = f (0) . 2

Hence x = 0 is not in the Lebesgue set of f.

Hence the set of those x R such that (3) holds is the collection of all real numbers. To determine if x = 0 is in the Lebesgue set of f or not, we compute Z h Z h 1 1 1 f () d = 1 6= 0. lim |f () f (0)| d = lim 2 2 h0+ 2h h h0+ 2h h 4. (10 points) Let g (x) be the function given by g (x) = (
1 sin x

if x 6= 0 Rx
0

0 if x = 0. g (s) ds,

Determine if the point x = 0 is in the Lebesgue set of g or not. Also let G (x) = x R. Do we have G0 (0) = g (0) or not. Solution:

(This solution is provided by TA Yu-Chu Lin) We claim that x = 0 is not in the Lebesgue set of g. Compute Z h Z h Z h 1 1 1 1 sin d = lim 1 sin d |g () g (0)| d = lim lim h0+ 2h h h0+ 2h h h0+ h 0 Z |sin y| = lim q dy . q y2 q We look at q = 2p, p N. Over the intervals 3 3 , I2 = 2p + 2 + , 2p + 2 + I1 = 2p + , 2p + 4 4 4 4 3 I3 = 2p + 4 + , 2p + 4 + , 4 4

we have |sin y| Z

2 2 ,

which implies

and so

To estimate the sum of the series (denote the series by S) we note that (by comparing areas) Z 1 1 S dx = 3 2 2 2p + 3 p 2x + 4 4 2p which implies that Z
2p

Z |sin y| 2 1 dy dy 2 2 I1 I2 I3 y 2 2p y ! ! ) ( 1 2 1 1 1 = + + 2 2p + 2p + 2 + 2p + 3 2p + 2 + 3 4 4 4 4 ( ) 1 2 1 1 2 + 2 + 2 + . 2 2 2p + 3 2 (p + 1) + 3 2 (p + 2) + 3
4 4 4

|sin y| 2 2p dy y2 2 2 2 2p +
q

Z lim q

|sin y| dy y2

2 3 8 4 6= 0.

as p

The claim is proved. Next we estimate 1 lim + h h0 where Z

Z sin y 1 sin d = lim q dy q y2 q 1 cos q d ( cos y) = 2 2 2 y q = lim Z cos q 2q q Z

sin y dy = y2

cos y dy y3 = 0.

and so

Z lim q

sin y dy y2

cos y dy y3

Similarly we have 1 lim h h0 and so G0 (0) = g (0).

1 sin d = 0

Real Analysis Homework 14, due 2008-1-2 in class

1. (10 points) Let E Rn be a measurable set (|E| < or not). If for any 0 < p < we have f Lp (E) and kf kp K, where K is a constant independent of p. Show that f L (E) and kf k K also. Solution: We rst assume |E| < . In this case, we have (see Theorem 8.1) limp kf kp = kf k . Hence f L (E) and kf k K. S If |E| = , decompose E = Ek , where each |Ek | < . On each Ek , we have f L (Ek ) k=1 and kf k,Ek K. Since the constant K is independent of the set Ek , we have f L (E) and kf k,E K. 2. (10 points) Do Exercise 5 in P. 143. Solution: We have 1 p < , 0 < |E| < , and Np [f ] = 1 |E|

|f |

1/p

1 |E|1/p

kf kp .
s s1

If p1 < p2 , 1 p1 , p2 < , then set s = (1, ) . By Hlder inequality we get Z p1 |f (x)|p1 1dx kf kp1 = Z
E E

p2 p1

(1, ) . Its conjugate exponent is t =

p2 p2 p1

|f (x)|

p1 s

1/s Z 1/t Z p1 /p2 p2 t dx 1 dx = |f (x)| dx |E|(p2 p1 )/p2 .

E E

1 Taking p1 power on both sides gives the conclusion. Also by Minkowski inequality we have i 1 1 h kf kp + kgkp = Np [f ] + Np [g] . Np [f + g] = kf + gkp |E|1/p |E|1/p

Furthermore, Hlder inequality implies Z 1 1 1 |fg| kf kp 0 kgkp0 = Np [f ] Np0 [g] , 1/p |E| E |E| |E|1/p Finally since 0 < |E| < , we have
p

1 1 + 0 = 1. p p

lim Np [f ] = lim

1 |E|1/p

kf kp

= lim kf kp = kf k .
p

3. (10 points) Prove the converse of Hlder inequality (Theorem 8.8) for the case p = 1 and p = .

Solution: Assume p = 1. Clearly we have kf k1 sup Z

fg

(1)

for all g L (E) with kgk 1. Conversely, take g = sign f L (E) . Then kgk 1 and Z Z kf k1 = |f | = fg
E E

which implies RHS of (1) LHS of (1). For p = , again we have

kf k sup

for all g L1 (E) with kgk1 1. Conversely, if kf k = 0, then it is clear. For 0 < kf k < , we may assume kf k = 1. Let 1 En = x E : |f (x)| > 1 , n N . n R Then |En | > 0 for all n. On each En one can choose gn (x) satisfying gn 0, En gn = 1, and let gn = 0 outside En . Now Z Z Z Z 1 1 |f | gn = |f | gn 1 gn = 1 , gn = 1 n n E En En E and so kf k = sup Z |f | g

fg

for all g L1 (E) with kgk1 1. Finally it is easy to see that sup For the case kf k = , just repeat the above process with En = {x E : |f (x)| > n,

E |f | g = sup

f g.

n N} .

4. (10 points) Assume 1 p < . Do Exercise 12 in P. 144. Solution: In this problem, we assume 1 p < (in fact, as long as 0 < p < , we have the same conclusion). (=) By Minkowski inequality we have kf kp kfk kp kf fk kp 0 as k . 2p |f |p + 2p |fk |p |f fk |p 0 we have (by Fatous Lemma) Z Z p p p p p lim inf [2 |f | + 2 |fk | |f fk | ] lim inf [2p |f |p + 2p |fk |p |f fk |p ]
E k k E

Hence we have kfk kp kf kp as k .

(=) We assume that fk f a.e. and kfk kp kf kp as k . By the inequality

where

and lim inf

k

Z Z
p

lim inf [2p |f |p + 2p |fk |p |f fk |p ] =

k

[2p |f |p + 2p |f |p ] Z

[2 |f | + 2 |fk | |f fk | ] = R

[2 |f | + 2 |f | ] lim sup
k

|f fk |p .

Since we assume f Lp , the integral E [2p |f |p + 2p |f |p ] is nite (this is essential). Hence we conclude Z |f fk |p 0. lim sup
k E

The proof is done. Remark 1 When p = , the conclusion in (=) fails. Just take f = 1 on R and fk = (k,k) .

Solutions to Homework 1
1. (10 points) Let Q be the set of all rationals in the interval [0, 1] . Let S = {I1 , I2 , ..., Im } be a nite collection of closed intervals covering Q. Show that
m X k=1

v (Ik ) 1.

(0.1)

On the other hand, for any > 0, one can nd S = {I1 , I2 , ..., Im , } , which is a countable collection of closed intervals covering Q, such that
X k=1

v (Ik ) < .

(0.2)

In particular, (0.2) implies that |Q|e = 0. (Now you see the dierence between the use of nite cover and countable cover.) Solution: For (0.1), we rst assume that the intervals I1 , I2 , ... , Im are nonoverlapping. In such case we clearly have (0.1). For arbitrary intervals I1 , I2 , ... , Im with overlapping, one can throw away the overlapping part and the remaining nonoverlapping part, which we denote it as J1 , J2 , ... , Jn , P satises n v (Jk ) 1. Therefore we have (0.1). k=1 For (0.2), it has been done in class. 2. (10 points) Find a set E R with outer measure zero and a function f : E R such that f is continuous on E and f (E) = [0, 1] . This exercise says that a continuous function can map a set with outer measure zero onto a set with outer measure one. Solution: Choose E = C to be the Cantor set contained in [0, 1] and let f (x) be the continuous Cantor Lebesgue function dened on [0, 1] (see book p. 35). We know that when restricted to C, f (x) is still a continuous function on C. Moreover, one can easily see that f (C) = f ([0, 1]) = [0, 1] (for example, we have f 1 , 2 = f 1 , 2 , etc.). 3 3 3 3 3. (10 points) Let E1 and E2 be two subsets of Rn such that E1 E2 and E2 E1 is countable. Show that |E1 |e = |E2 |e . Solution: Clearly we have |E1 |e |E2 |e . Also |E2 |e |E1 |e + |E2 E1 |e = |E1 |e which implies |E1 |e = |E2 |e .

4. (10 points) Find a continuous function f (x) dened on [0, 1] such that f (x) is dierentiable on a subset E [0, 1] with |E|e = 1 and f 0 (x) = 0 for all x E, but f (x) is not a constant function.

Solution: Let f (x) be the Cantor Lebesgue function dened on [0, 1] as given in p. 35 of the book. We know f (x) is dierentiable on the open set O = O1 O2 O3 , where 7 8 1 2 1 2 1 2 , , O2 = , , , O3 = , , O4 = O1 = 3 3 9 9 9 9 27 27 The total length of these open intervals is given by # " 2 3 2 2 1 2 + + = 1. 1+ + 3 3 3 3 The example below shows you how to obtain a set which is measurable, but not Borel measurable. Let f (x) : [0, 1] [0, 1] be the Cantor-Lebesgue function and let g (x) = x + f (x) . It is easy to see that g (x) : [0, 1] [0, 2] is a strictly increasing continuous function. Hence g (x) is a homeomorphism of [0, 1] onto [0, 2] . On each interval I1 , I2 , I3 , ..., removed in the construction 1 2 of the Cantor set, say the interval I1 = 3 , 3 , the function g (x) becomes g (x) = x + 1 . Hence 2 g (x) sends I1 onto an open interval with the same length. Using this observation one can see that X X |g (Ik )| = |Ik | = 1 |g ( Ik )| = | g (Ik )| = k=1 k=1
k=1 k=1

which implies |g (C)| = 2 1 = 1, where C is the Cantor set. Since g (C) has positive measure, by Corollary 3.39 in the book, there exists a non-measurable set B g (C) . Now consider the set A = g 1 (B) C. It has measure zero, hence it is measurable. However it can not be Borel measurable. If A were Borel measurable, then since g (x) is a homeomorphism, it would imply that B = g (A) is also Borel measurable. But this is impossible since B is a non-measurable set.

Real Analysis Homework 2, due 2007-9-25 in class

Show Your Work to Each Problem 1. (10 points) Use Lemma 3.16 to prove Lemma 3.15. Note that in proving Lemma 3.16 we do not have to use Lemma 3.15. Solution: Let {Ik }N be a nite family of nonoverlapping intervals. It suces to show that k=1 N N [ X Ik |Ik | .
k=1 k=1

The proof is done. 2. (10 points)

For each k, consider smaller interval Ik Ik with |Ik | |Ik | /2k . Since {Ik }N are nonoverk=1 and N I has positive distance. We can apply Lemma 3.15 to get lapping, the set I1 k=2 k N N N N [ [ X X Ik Ik = |Ik | |Ik | . k=1 k=1 k=1 k=1

(a) (5 points) Assuming the validity of Theorem 3.30 and the existence of non-measurable sets in Rn at this moment. Show that there exist two nonempty sets E1 and E2 in Rn such that E1 E2 = , but |E1 E2 |e < |E1 |e + |E2 |e . Hence the condition d (E1 , E2 ) > 0 in Lemma 3.16 can not be replaced by just E1 E2 = . (b) (5 points) Construct a sequence of nonempty sets Ek [0, 1] , k = 1, 2, 3..., so that lim sup Ek = [0, 1] , lim inf Ek = . (0.1) Solution: (a). This is easy by Theorem 3.30. Choose E Rn to be a nonmeasurable set. Then there exists some nonempty set A Rn such that |A|e < |A E|e + |A E|e (0.2) where A E and A E are disjoint. One can easily see that if (0.2) holds, then both A E and A E are nonempty. (b). Choose for example the sequence E1 = [0, 1] 1 , E2 = 0, 2 1 E4 = 0, , 3 1 E7 = 0, , 4 . Then we have (0.1). 1 1 E3 = ,1 2 1 2 E5 = , , 3 3 1 2 E8 = , , 4 4

2 E6 = ,1 3 2 3 E9 = , , 4 4

E10

3 ,1 = 4

3. (10 points) Assuming there exists a non-measurable set contained in [0, 1] , do Exercise 17 in p. 48. Solution: Let E = f 1 (A) , where A is a nonmeasurable set contained in [0, 1] . Here f : C (Cantor set) [0, 1] is the Cantor-Lebesgue function in p. 35. It is continuous. Then since |E| = 0, E is measurable. But f (E) = A is nonmeasurable. 4. (10 points) Do Exercise 18 in p. 48. Solution: We rst asume that 0 |E|e < . It is clear that for any interval I, |Ih | = |I| . For any > 0, choose a sequence of intervals Ik , covering E, such that X |Ik | < |E|e + . Now the translation Eh S
k

which gives |Eh |e |E|e . Conversely |E|e = (Eh )h e |Eh |e . We are done.

(Ik )h and so X X |Eh |e |(Ik )h | = |Ik | < |E|e +

k k k

If |E|e = , then we must have |Eh |e = also (note that from the above observation, if |Eh |e < , it will force |E|e < ). Assume E is measurable. For any > 0, there exists open set G E such that |G E|e < . This implies |Gh Eh |e = |G E|e < where Gh is also an open set. Hence Eh is also measurable.

Real Analysis Homework 3, due 2007-10-3 in class

Show Your Work to Each Problem 1. (20 points) Let f : Rn R be a continuous function. Dene the collection of sets X = B R : f 1 (B) is measurable . P as

P Does form a -algebra? If B R is a Borel set, does it follow that f 1 (B) is a Borel set? Give your reasons. P P Solution: Clearly , R . If E , by c f 1 (E c ) = f 1 (E) (0.1) P P we know that E c also. Similarly if {Ek } is a collection of subsets of , then by k=1 ! ! [ [ \ \ f 1 Ek = f 1 (Ek ) and f 1 E = f 1 (E ) (0.2)
k k k k

forms a -algebra. For any open set O R, the set f 1 (O) is also open. Hence f 1 (O) . The same for closed set. Hence all open sets and closed sets are contained in . Since is a -algebra, if B R is a Borel set, then f 1 (B) must also be a Borel set. 2. (20 points) Do Exercise 12 in p. 48. Hint: You can use Theorem 3.29. Solution: We rst show that E1 E2 R2 is measurable. Write E1 = H1 Z1 , H1 is G set and |Z1 | = 0. The same for E2 = H2 Z2 . Then E1 E2 = H1 H2 Z1 H2 H1 Z2 where H1 H2 = R2
\

we know that k Ek and k E . Hence is a -algebra. Similarly we can show that the set = B R : f 1 (B) is Borel measurable

On

n=1

n=1

(On and On are open set in We see that H1 H2 R2 is measurable. is a G set in Assume rst that |H2 | < . One can nd open sets G1 Z1 , G2 H2 such that |G1 | < , |G2 | < |H2 | + . Write R1 ). G1 = G2 =
[

On

\ On On = n=1

Ij Ik

(nonoverlapping closed intervals in R1 ) (nonoverlapping closed intervals in R1 )

j=1 [

k=1

to see that
[ Ij Ik Z1 H2 G1 G2 = j, k=1

(nonoverlapping closed intervals in R2 )

with |G1 G2 | =

which implies |Z1 H2 | = 0. S If |H2 | = , decompose H2 = n= (H2 [n, n + 1)) , disjoint union, and see that |Z1 H2 | = 0. The same for |H1 Z2 | = 0. Hence E1 E2 R2 is measurable. Next we show that |E1 E2 | = |E1 | |E2 | Case 1: |E1 | < , |E2 | < . From (0.3) we see that |G1 G2 | = |G1 | |G2 | for any two open sets in R1 . Choose open sets G1 E1 , G2 E2 with |G1 E1 | < , |G2 E2 | < . Then E1 E2 G1 G2 and so |E1 E2 | |G1 G2 | = |G1 | |G2 | (|E1 | + ) (|E2 | + ) which implies that Write G1 = S |E1 E2 | |E1 | |E2 | . = S N1 , where Ij , N1 =
[ M [

X X X |Ij | Ik = |G1 | |G2 | < (|H2 | + ) Ij Ik =

j,k

(0.3)

(0.4)

j=1 Ij

S= satises

Ij

(M is some large number)

j=1

j=M+1

|E1 | |S| |E1 | + , Then one can express E1 as E1 = (S N1 ) (open set) N2 ,

|N1 | < . |N2 | < .

(0.5)

where N2 = G1 E1 ,

Do the same for G2 and E2 to get E2 = S N1 (open set) N2 , with

N2 = G2 E2 , N1 < .

Now

|E2 | S |E2 | + ,

N2 <

(0.6)

E1 E2 = (S N1 ) S N1 N2 S N1 (S N1 ) N2 := A B 2

where A and B are measurable sets with B A, |B| < . Hence |E1 E2 | = |A| |B| with and by (0.4) |A| = |S N1 | S N1 = (|S| + |N1 |) S + N1 |B| = N2 S N1 + (S N1 ) N2 C |E1 E2 | = |E1 | |E2 | . (0.7) (0.8)

for some nite constant C. By (0.4), (0.7), (0.8), (0.5)and (0.6), we get the equality

Case 2: |E1 | = , |E2 | = 0. Consider E1 := E1 [n, n + 1), n Z, then we have |E1 E2 | = 0 = |E1 | |E2 | (we view 0 = 0). Case 3: |E1 | = , |E2 | = > 0. P (n) (n) Consider E1 := E1 [n, n + 1), then we have |E1 E2 | = E1 |E2 | = = n= |E1 | |E2 | . Case 4: |E1 | = , |E2 | = . Easy to see that |E1 E2 | = = |E1 | |E2 | . 3. (10 points) It has been proved in class that if E Rn is an arbitrary measurable set (|E| = is allowed). We have |E| =
GE, G open in Rn (n)

inf

|G| =

sup
F E, F closed in Rn

|F | .

Show that if E Rn is an arbitrary set satisfying the following ( |E|e < inf GE,
G open in Rn

|G| = supF E,

F closed in Rn

|F |

then E must be measurable. Use an example to explain that the condition |E|e < is necessary. That is, there exists a set E with |E|e = , inf GE, G open in Rn |G| = supF E, F closed in Rn |F | , but it is not measurable. Solution: By the assumption for any xed > 0 one can choose open set G E and closed set F E so that (0.9) |E|e < |F | |G| < |E|e + . |G F | = |G| |F | < 2. In particular we have |G E| |G F | < 2 and so E is measurable. Let E = (, 0) A, where A (1, 2) is some nonmeasurable set. Clearly E satises |E|e = , but it is nonmeasurable. 3
GE, G open in Rn

Since |F | |E|e < and then (0.9) gives

inf

|G| =

sup
F E, F closed in Rn

|F | =

Real Analysis Homework 4, due 2007-10-9 in class

Show Your Work to Each Problem 1. (20 points) (a) (10 points) Use denition (do not use Theorem 3.33) to show that the CantorLebesgue function f (x) : [0, 1] [0, 1] is not a Lipschitz continuous function.

(b) (10 points) Show that the Cantor-Lebesgue function f (x) : [0, 1] [0, 1] satises the following |f (x) f (y)| 2 |x y| , x, y [0, 1] where (0, 1) is a constant given by = log 2/ log 3. (Hint: Use the fact that if x, y [0, 1] with |x y| 3k for some k N, then the dierence |f (x) f (y)| is at most 2k . For arbitrary x, y [0, 1] one can choose an unique k N such that 3k1 < |x y| 3k , which implies |f (x) f (y)| 2k . Rewrite the estimate without involving k.) Solution: (a) . For example, look the function near the point x = 7 at f 2 = 1 . We also have f 27 f 2 = 1 , which gives 9 4 9 8 3 7 f 27 f 2 3 9 = . 7 2 2 27 9 Keep going to get (note that
6 27 2 9

6 27

18 81 .

We have

18 7 81 , 27

= 21 ) 81 4 19 f 81 f 2 3 9 = 19 2 2 81 9

, etc. Thus it is impossible to nd a nite constant M > 0 such that f (y) f (x) M for all x, y [0, 1] . yx

(b) . First note that if x, y [0, 1] with |x y| 3k for some k N, then the dierence |f (x) f (y)| is at most 2k . For arbitrary x, y [0, 1] one can choose largest k N such that |x y| 3k ; therefore 3k1 < |x y| 3k (0.1)

which will gives the best estimate |f (x) f (y)| 2k . (0.1) is equivalent to (k + 1) log 3 < log |x y| k log 3 which gives k < 1 + log |x y| . log 3 1 (0.2)

By (0.2) we get log |f (x) f (y)| k log 2 < log 2 + log |x y| , and so |f (x) f (y)| 2 |x y| , 2. (10 points) Do Exercise 20 in p. 48. Solution: Let E be a non-measurable subset of [0, 1] , as established in Corollary 3.39 of the book. We know |E|e > 0. Let r0 = 0, r1 , r2 , ..., be the set of all rationals in [0, 1] and let Ek = E+rk for k = 0, 1, 2, 3, .... Each Ek [0, 2] for all k with |Ek |e = |E| , and Ei Ej = for dierent i and j. We clearly have [ X |Ek |e = . Ek 2 <
k e k

log 2 log 3

(0.3)

x, y [0, 1] .

3. (10 points) Do Exercise 21 in p. 48. Solution: Following the notation of Ex 20, we let ! ! [ [ [ A1 = Ek E0 , A2 = Ek E0 E1 A3 = [
k k

Ek

[ [ E0 E1 E2

then Ak & , |Ak |e 2 < (due to Ek [0, 2] for all k), and |Ak |e |E|e > 0 for all k. Hence we have limk |Ak |e > 0. 4. (10 points) Do Exercise 23 in p. 49. Solution: For each n N let we have |Zn | = 0 for all n. Also let Tn (x) = \

Zn = Z

[n, n] .

x2 n2

if x [n, n] otherwise. S
n Tn Zn ,

Then Tn : R R is Lipschitz continuous on R. Hence |T Zn | = 0 for all n. As T Z we have |T Z| = 0.

Real Analysis Homework 5, due 2007-10-16 in class

1. (10 points) Do Exercise 2 in p. 61. Solution: Assume f takes distinct values a1 , , aN on disjoint sets E1 , , EN . We can express the function f (x) as f (x) = a1 E1 (x) + + aN EN (x) , xE=
N [

Ek .

k=1

If each Ek is measurable, then the characteristic function Ek (x) : E R is a measurable function on E. Hence if E1 , , EN are all measurable, so is f (x) . Conversely, assume f (x) is measurable. Then by denition we know that (we may assume a1 < a2 < < aN ) the set {f > aN1 } is measurable. Since {f > aN 1 } = EN , the set EN is measurable. Similarly by EN1 = {f > aN 2 } EN 2. (10 points) Do Exercise 3 in p. 61. Solution: Let F (x) = (f (x) , g (x)) , x Rn . F : Rn R2 . (=) Assume F is measurable. For any open set Gx R, let G = Gx R. Then G R2 is open and by the identity F 1 (G) = f 1 (Gx ) we know that f 1 (Gx ) is measurable for any open set Gx R. Hence f : Rn R is measurable. The same for g : Rn R. (=) Assume f : Rn R and g : Rn R are both measurable. Let G R2 be an arbitrary open set. We can express G as a countable union of nonoverlapping closed intervals S G = Ik , where Ik = [ak , bk ] [ck , dk ] . Note that k=1 \ F 1 (Ik ) = f 1 [ak , bk ] g 1 [ck , dk ] (both sets are measurable) F
1

we know that EN1 is measurable. Keep going to conclude that E1 , , EN are all measurable.

and so F 1 (Ik ) is measurable. By

(G) =

k=1

the set F 1 (G) is measurable in Rn . Hence F is a measurable function. 3. (10 points) Do Exercise 4 in p. 61. Solution: For any a R, we have

F 1 (Ik )

As T 1 : Rn Rn is linear and Lipschitz continuous, (0.1) is a measurable set. Hence f (T x) is a measurable function. 1

[ {x Rn : f (T x) > a} = {x Rn : f (T x) (a, )} {x Rn : f (T x) = } [ = x Rn : T 1 f 1 (a, ) x Rn : T 1 f 1 ({}) . (0.1)

4. (20 points) Do Exercise 5 in p. 61. First solution: We rst prove the following Lemma 0.1 The Cantor-Lebesgue function f (x) : [0, 1] [0, 1] satises f (x1 ) = f (x2 ) , where x1 and x2 are in C, if and only if both x1 and x2 are endpoints of some interval removed. Proof. The direction (=) is trivial. (=) Assume at least one of x1 , x2 is not endpoint, say x2 . On the interval (x1 , x2 ) , x1 < x2 , x1 , x2 C, there exists some p 6 C. Let I = (y1 , y2 ) be the maximal open interval containing p such that I C = (note that the complement of C is open). We now have y1 , y2 C and y2 < x2 (otherwise if y2 = x2 , then x2 must be an endpoint, impossible). Similarly one can nd an open interval J = (z1 , z2 ) such that J (y2 , x2 ) with y2 < z1 (otherwise y2 is an isolated point of C, impossible). Now the open interval J is on the right hand side of the open interval I with a positive distance away. These two distinct intervals must be exactly equal to some removed intervals in the process of constructing the Cantor set. Hence f (I) < f (J) , which gives f (x1 ) < f (x2 ) . Corollary 0.2 If x C but x is not endpoint of some interval removed (say x = 1 ), then there 4 is no x C, x 6= x, such that f () = f (x) . x Corollary 0.3 Let C = C {all right endpoints of the removed intervals} , where C is the Can [0, 1] is 1-1 and onto, and is strictly increasing on C. Here f is the tor set. Then f : C Cantor-Lebesgue function. By the above corollary we have g (y) : [0, 1] C [0, 1] , strictly increasing on [0, 1] , [0, 1] . For any a [0, 1] the set which is the inverse of f : C Ea := {y [0, 1] : g (y) a} = {y [0, 1] : y f (a)} is measurable. Hence g is a measurable function on [0, 1] . Let A [0, 1] be a nonmeasurable set. Its image under g has measure zero, hence measurable. Let = g(A) : [0, 1] R then is a measurable function on [0, 1] . But the composite function g : [0, 1] R is not measurable since the set 1 y [0, 1] : (g (y)) > =A 2 is not measurable. Second solution (much easier): Let f (x) : [0, 1] [0, 1] be the Cantor-Lebesgue function and let g (x) = x + f (x) . It is easy to see that g (x) : [0, 1] [0, 2] is a strictly increasing continuous function. Hence g (x) is a homeomorphism of [0, 1] onto [0, 2] . We denote its continuous inverse by h : [0, 2] [0, 1] . On each interval I1 , I2 , I3 , ..., removed in the construction of the Cantor set, say the interval

 I1 = 1 , 2 , the function g (x) becomes g (x) = x + 1 . Hence g (x) sends I1 onto an open interval 3 3 2 with the same length. Using this observation one can see that ! X [ [ X Ik = g (Ik ) = |g (Ik )| = |Ik | = 1 g
k=1 k=1 k=1 k=1

which implies |g (C)| = 2 1 = 1, where C is the Cantor set. Since g (C) has positive measure, by Corollary 3.39 in the book, there exists a non-measurable set B g (C) . Now consider the set A = h (B) C. It has measure zero, hence A is measurable. Let = A : [0, 1] R, |A| = 0

then is a measurable function on [0, 1] . We now have two measurable functions h : [0, 2] [0, 1] (continuous) and : [0, 1] R. But the composite function h : [0, 2] R is not measurable since the set 1 =A [0, 2] : (h ()) > 2 is not measurable.

Real Analysis Homework 6, due 2007-10-24 in class

1. (10 points) Do Exercise 7 in p. 62. Solution: For any x0 E (compact) and Mx0 > f (x0 ) one can choose x0 > 0 such that f (x) < Mx0 for all x E with |x x0 | < x0 .

The collection {x E : B (x0 ; x0 )} forms an open cover of E and so there exist a nite cover B (x1 ; x1 ) , , B (xn ; xn ) of E. Set M = max {Mx1 , , Mxn } . We have f (x) < M for all x E. Hence f is bounded above. Let M = supxE f (x) . For k = 1, 2, 3, ... choose xk E such that 1 M < f (xk ) < M . k (0.1)

As E is compact, by passing to a subsequence if necessary, we may assume that xk E converges to some x0 E. Assume f (x0 ) < M . Choose M with f (x0 ) < M < M . Since f is usc at x0 , there exists some > 0 such that f (x) < M for all x B (x0 ; ) E. (0.2)

For k large enough, we have xk B (x0 ; ) E. (0.1) will contradict to (0.2). Hence we must have f (x0 ) = M , i.e., the maximum is attained. 2. (10 points) Show that the limit of a decreasing sequence of functions (with common domain E) usc at x0 E is also usc at x0 . Give an example of a decreasing sequence of functions continuous at x0 E but its limit is not continuous at x0 (by the rst part of the problem we know that the limit is at least usc at x0 ). Solution: Denote the decreasing sequence of functions by fk (x) and the limit by f (x) . We have f1 (x) f2 (x) fk (x) , x E and each fk (x) is usc at x0 . Since we have f (x) fk (x) on E for each k, we have
xx0 ;xE

lim sup f (x) lim sup fk (x) fk (x0 )

xx0 ;xE

for each k N.

As k is arbitrary. Letting k gives the conclusion. Let E = [0, 1] and let {xk : k = 0, 1, 2, 3, ...} be the set of all rationals in E. Set f0 (x) 1 and set for each k N the function 0 at x = x1 , x2 , , xk fk (x) = 1 otherwise. Then fk (x) is a decreasing sequence of functions on E; all are continuous at 2/2. But the limit f (x) is not continuous at 2/2. However it is usc at 2/2. 3. (10 points) Do Exercise 11 in p. 62. 1

Solution: We only show that h (x) is usc on Rn . The proof of the other case is similar. For any x0 Rn , we rst assume that h (x0 ) < (otherwise we are done). It suces to show that for any M > h (x0 ) = inf {f (y) : y B (x0 )} there exists > 0 such that h (x) < M for all x B (x0 ; ) . For any M > h (x0 ) choose y0 B (x0 ) such that f (y0 ) < h (x0 )+ < M. Set = |y0 x0 | < r and = 1 (r ) > 0. Then for all x with |x x0 | < we have 2 |x y0 | |x x0 | + |y0 x0 | + < r. Hence y0 B (x) and by denition we get h (x) = inf {f (y) : y B (x)} f (y0 ) < M. Thus h (x) is usc at x0 . For the case of using closed balls, consider in R2 the function 0 if p = (1, 0) f (p) = , p R2 1 if p 6= (1, 0)

and take r = 1, B (x) = p R2 : |p x| 1 . Then h (0, 0) = 0 and for any > 0 we have h (, 0) = 1. Hence h is not usc at the point (0, 0) . 4. (10 points) Do Exercise 12 in p. 62. Solution: Assume f (x) : [a, b] R is continuous a.e. on [a, b] . Let k be a sequence of partitions of [a, b] with norms tending to zero. We also assume that each k+1 is a renement (k) (k) of k . For each k, if x1 < x2 < h are the partitioning points of k , let lk (x) and uk (x) be h i
(k) (k) (k) (k)

dened in each semi-open interval xi , xi+1 as the inf and sup of f on xi , xi+1 . Note that lk (x) f (x) uk (x) for all x [a, b) and all k. It is easy to see that for each k, lk (x) and uk (x) are measurable functions on [a, b) (even if it is possible that lk (x) = or uk (x) = + on some intervals). Let x0 (a, b) at which f (x) is continuous. For any > 0 there exists > 0 such that x (x0 , x0 + ) implies |f (x) f (x0 )| <

and when k is large enough, the interval in k containing x0 must lie inside (x0 , x0 + ) . This implies |lk (x0 ) f (x0 )| < and |uk (x0 ) f (x0 )| < for all k large enough. Hence limk lk (x) = f (x) a.e. on [a, b] (we also have limk uk (x) = f (x) a.e. on [a, b]). By Theorem 4.12 of the book, we know that f (x) is measurable on [a, b] . Remark 1 (be careful) If g (x) is a continuous function on [a, b] and f (x) = g (x) a.e. on [a, b] , it does not, in general, imply that f (x) is continuous a.e. on [a, b] . For example, take g (x) = 1 and let 1, x is irrational in [0, 1] f (x) = 0, x is rational in [0, 1] . We see that f (x) = g (x) a.e. on [0, 1] , but f (x) is discontinuous everywhere on [0, 1] .

Real Analysis Homework 7, due 2007-10-31 in class

1. (10 points) (a) (7 points) Do Exercise 15 in p. 62. (b) (3 points) Use Exercise 15 in p. 62 to prove the following: Let f : E R {} be a measurable function where |E| < and |f | < a.e. on E. Show that for any > 0 there exists a constant M > 0 and a closed set F E such that |E F | < and |f (x)| M for all x F.

This says that a nite function is, up to a set of small measure, a bounded function. Solution: For (a) . For each n = 1, 2, ..., let En = {x E : |fk (x)| n for all k} . Then by |fk (x)| Mx < for all k and all x E, we have En % E as n with limn |En | = |E| . Since |E| < , we also have limn |E En | = 0. Choose M such that |E EM | < /2 and choose a closed set F EM such that |EM F | < /2. Then we have |E F | < and the following holds |fk (x)| M for all x F and all k.

For (b) . By (a) , if we choose fk (x) = f (x) for each k N, then we are done. 2. (10 points) Do Exercise 16 in p. 63. Solution: (=). By denition, for any > 0 we have
k

lim |{x E : |f (x) fk (x)| > }| = 0.

Hence for the same > 0 we have |{x E : |f (x) fk (x)| > }| < if k K, for some large K > 0. (=). Fixed an arbitrary > 0 rst. We want to show that for any > 0 there exists K > 0 such that if k K we have |{x E : |f (x) fk (x)| > }| < . (0.1) Now for any > 0, if , then by the assumption we automatically have the existence of a K > 0 such that if k K we have |{x E : |f (x) fk (x)| > }| < . Hence we assume < . Again by the assumption we have the existence of a L > 0 such that if k L we have |{x E : |f (x) fk (x)| > }| < . But the set {x E : |f (x) fk (x)| > } {x E : |f (x) fk (x)| > } , and so we have (0.1). The Cauchy criterion is: For any > 0 there exists K > 0 such that if m, n K we have |{x E : |fm (x) fn (x)| > }| < . (0.2) 1

3. (10 points) Do Exercise 18 in p. 63. Solution: Given f : E R {} measurable and let f (a) = |{f > a}| , where < a < .

As a function of a, f (a) is decreasing on (, ). If fk % f on E, then set Ek = {fk > a} , k = 1, 2, 3, ...

We have E1 E2 E3 and if f (x) > a, we will have fk (x) > a for all k large enough (since fk % f on E). Thus [ Ek {f > a} =
k=1

and so fk (a) % f (a) for all a (, ). If fk f in measure on E, given > 0 let S A1 = {|f fk | > } , k A2 = {|f fk | } . k

We have E = A1 k

T where limk Ek A1 = 0 (due to convergence in measure) and k Ek We have and so \ A2 = {fk > a} k \

A2 (disjoint union). Hence for each xed a R we have k \ [ \ Ek A2 , Ek = {fk > a} Ek = Ek A1 k k {|f fk | } {f > a } .

\ \ \ fk (a) = |Ek | = Ek A1 + Ek A2 Ek A1 + f (a ) k k k lim sup fk (a) f (a )

k

for any > 0.

Similarly we have {f > a + } which gives

k

{|f fk | } {fk > a} for any > 0.

lim inf fk (a) f (a + ) We conclude

f (a + ) lim inf fk (a) lim sup fk (a) f (a )

k k

for any > 0.

Thus if f (x) is continuous at x = a, we have limk fk (a) = f (a) . 4. (20 points) Do Exercise 19 in p. 63.

Solution: Let S = [0, 1] [0, 1] . The idea is to separate the x variable from the y variable. For each n = 1, 2, 3, ...., dene fn (x, y) = where for each k, rk k1
n n1 X k=1

f (rk , y) [ k1 , k ) (x) + f (rn , y) [ n1 ,1] (x) ,

n n n

(x, y) S

k , n and rk n1 , 1 are arbitrary constants. n 2

For each k, the function hk (x, y) := f (rk , y) [ k1 , k ) (x) : S R satises

n n

hk (x, y) = 0 on and on k1
n

k , n [0, 1] we have

k1 k , [0, 1] [0, 1] n n

k1 k , [0, 1] : hk (x, y) > a (x, y) n n k1 k = , {y [0, 1] : f (rk , y) > a} n n

which is a measurable set (due to Exercise 12, p. 48). Hence we can conclude that hk (x, y) is a measurable function on S for each k. As a consequence the function fn (x, y) is also measurable on S. k For each (x0 , y0 ) S we have (assume that x0 k1 , n ) n k1 k , |fn (x0 , y0 ) f (x0 , y0 )| = |f (rk , y0 ) f (x0 , y0 )| , rk n n and so fn (x0 , y0 ) f (x0 , y0 ) as n . Hence f (x, y) is a measurable function on S.

Real Analysis Homework 8, due 2007-10-31 in class

R S 1. (10 points) Let f : E R {} be a nonnegative measurable function such that E f < . Show that for any > 0 there exists > 0 such that for any measurable subset E1 E with |E1 | < we R have E1 f < . Solution: Let fk (x) = ( f (x) , k, if f (x) < k , if f (x) k x E.

Then 0 fk (x) % f (x) on E and by the Monotone Convergence Theorem we have Z Z fk dx = f < lim
k E

and so for any > 0 there exists N such that E f E fN < /2. Note that fN N on E and so if E1 E R with |E1 | < := 2N we would have E1 fN N |E1 | /2. Therefore for any E1 E with |E1 | < we get Z Z Z Z f=
E1 E1

fN +

fN < .

E1

E1

2. (10 points) Do Exercise 3 in p. 85. Solution: Since fk f a.e. on E (both are nonnegative), we have hand, by Fatous lemma we get Z Z Z lim inf fk = f lim inf fk
E k E k E

E fk

f for all k. On the other

which implies

Hence we have limk

3. (10 points) Let fk : E R {} be a sequence of nonnegative measurable function satisfying 0 as k . Show that fk 0 in measure as k . Solution: For any > 0 by Tchebyshevs inequality we have |{x E : |fk 0| > }| = |{x E : fk > }| 1 Z fk .

E fk

f lim inf
k

f.

fk lim sup
k

fk dx

f. R

E fk

Letting k , the conclusion follows. R Remark 1 (be careful) E fk 0 as k does not, in general, imply that fk 0 a.e. on E. 4. (10 points) Compute the limit
n 0

lim

and justify your answer. Solution: Let fn (x) =

x n x/2 e dx n

( x n 1 n ex/2 , 0,

if x [0, n]

if x > n.

One can check that fn (x) % f (x) = ex/2 on E = [0, ). By Monotone Convergence Theorem we have Z Z Z fn = f= ex/2 dx = 2. lim
n E E 0

Real Analysis Homework 9, due 2007-11-21 in class

1. (10 points) Do Exercise 6 in p. 85. Solution: In this problem we assume bounded function on I. Let
x f

(x, y) exists on I = [0, 1] [0, 1] . we also know that it is a , n = 1, 2, 3, ....

Fn (x, y) =

1 f x + n , y f (x, y)
1 n

we see that for each xed x, Fn (x, y) is a sequence of bounded (use mean value theorem to see this) measurable functions of y. By f (x, y) = lim Fn (x, y) n x
we know that for each xed x, x f (x, y) is a measurable function of y. Now by the Bounded Convergence Theorem, we obtain Z 1 Z 1 Z 1 d f (x, y) dy. f (x, y) dy = lim Fn (x, y) dy = n 0 dx 0 0 x

2. (10 points) Do Exercise 9 in p. 85. Solution: For any > 0 we have 1 |{|fk f | > }|
p

|fk f |p 0 as k

due to the Tchebyshev inequality. Hence fk converges to f in measure on E. 3. (10 points) Do Exercise 10 in p. 85. Solution:

By Exercise 9 we know that fk f in measure. In particular, there exists a subsequence fkj such that it converges to f a.e. on E. Fatous lemma implies Z Z Z p p p fk = fk M. lim inf j |f | lim inf j
E j E j E

4. (10 points) Do Exercise 20 in p. 85. Solution: Case1: If f (x) = E1 , E1 E, then LHS of the identity is |E1 | , and the RHS of the identity is given by Z |det T | E1 (T x) dx = |det T | T 1 E1 .
T 1 E

We see that the identity holds by Theorem 3.35. Case2: Assume f 0. Then there exists a sequence of simple functions 0 sn % f on E where sn = a1 E1 + + ak(n) Ek(n) , k (n) depends on n. Now by Case1 sn (y) dy = a1 E1 (y) dy + + ak(n) Ek(n) (y) dy E E Z Z E1 (T x) dx + + ak(n) |det T | Ek(n) (T x) dx = a1 |det T | 1 E T 1 E Z T a1 E1 + + ak(n) Ek(n) (T x) dx = |det T | 1 ZT E = |det T | sn (T x) dx
E T 1 E

and by the Monotone Convergence Theorem we obtain Z Z Z sn (y) dy = f (y) dy, lim |det T | lim
n E E n

T 1 E

sn (T x) dx = |det T |

f (T x) dx.

T 1 E

The conclusion follows. For general f, use Z Z Z Z Z + + f = f f = |det T | f (T x) dx |det T | f (T x) dx E E E T 1 E T 1 E Z f (T x) dx. = |det T |
T 1 E

Real Analysis Midterm Exam, November 13, 2007

Show detailed argument to each problem. 1. (10 points) Suppose E is a Lebesgue measurable subset of R with |E| < . Prove that |E| = sup {|K| : K E and K is compact} . solution: T Let En = E [n, n] , n N. We have En % E and so |E En | 0 as n (note that |E| < ). For any > 0 one can nd m N so that |E En | < /2 for all n m. For Em one can nd a closed subset F Em such that |Em F | < /2. Hence |E F | |E Em | + |Em F | < . In particular F is compact since Em is bounded. We also have |E| |E F | + |F | < + |F | where F E and F is compact. (0.1) is proved. (0.1)

2. (10 points) Assume E (t) is a continuously dierentiable increasing function on [0, ) such that 0 E (t) C for all t [0, ), where C is some positive constant. Show that for any > 0, we have t [0, ) : E 0 (t) > C .

solution: By Tchebyshev inequality we have

since 0 E (t) C for all t [0, ).

t (0, ) : E 0 (t) > 1

E 0 (t) dt

i C 1h lim E (t) E (0) t

3. (10 points) Let fn : E R be a sequence of measurable functions dened on a measurable set E Rn . Let o n A = x E : lim fn (x) exists .
n

Is A a measurable set or not? Give your reasons.

solution: Let F (x) = lim supn fn (x) and F (x) = lim inf n fn (x) . We know that both functions are measurable on E. Hence the sets ( S1 := {x E : F (x) = and F (x) = }

S2 := {x E : F (x) = and F (x) = } S are all measurable. Let S = S1 S2 . Now E S is also measurable, and F F is a measurable function on E S. By the relation A = {x E S : F (x) F (x) = 0} 4. (10 points) Give an example of a sequence of measurable functions {fk } dened on a measurable set E Rn such that the following strict inequalities hold: Z Z Z Z lim inf fk dx < lim inf fk dx < lim sup fk dx < lim sup fk dx.
E k k E k E E k

we know that A is a measurable set.

solution:

On the interval [0, 1] , let fk (x) = and fk (x) = then 0= and Z ( ( 1, 0, 0, 2, x 0, 1 2 , x 1, 1 2 x 0, 1 2 , x 1, 1 2

k

k = 1, 3, 5, 7, ....

k = 2, 4, 6, 8, .... Z

E k

lim inf fk dx < lim inf Z Z

fk dx =

1 2

1 = lim sup
k

fk dx <

3 lim sup fk dx = . 2 E k
a

5. (15 points) Assume f L [a, b] and let h (x) = function on [a, b]? Give your reasons. solution:

Rx

f, x [a, b] . Is the function h (x) a measurable

By looking at f + and f , without loss of generality, we may assume that f 0 on [a, b] . Now there exists a sequence of simple functions 0 fk % f a.e. on E, where fk L [a, b] also for all k. Monotone Convergence Theorem implies Z Z
x x a

fk

f = h (x)

as k

for all x [a, b] . As fk is a bounded function on [a, b] , we know that x [a, b] . Hence h (x) is a measurable function on [a, b] .

Rx
a

fk is a continuous function of

Another solution: Rx Rx Rx Rx We have h (x) = a f + a f and both a f + and a f are increasing functions on [a, b] . We know that an increasing function on [a, b] is continuous a.e. on [a, b] . Hence h (x) is a measurable function on [a, b]. 6. (15 points) Suppose E R is measurable with |E| = > 0, where is a nite number. Show that for any t with 0 < t < , there exists a subset A of E such that A is measurable and |A| = t. That is, the Lebesgue measure || on R satises the Intermediate Value Theorem. solution: Dene f (x) = |(, x) E| , x R. Then for x < y we have f (x) f (y) and \ \ \ f (y) f (x) = (, y) E (, x) E = (x, y) E y x. T

This means that f (x) is a continuous function on R with limx f (x) = 0, limx f (x) = . For any 0 < t < , by the mean value theorem of continuous functions, we have f (s) = t for some s R. Hence T the set A = (, s) E satises |A| = t. 7. (15 points) Let y = T x be a nonsingular linear transformation of Rn . If we have the following change of variables formula: Z Z f (y) dy = |det T | f (T x) dx.
E T 1 E

f (y) dy exists, show that

(0.2)

solution: Case1: If f (x) = E1 , E1 E, then LHS of (0.2) is |E1 | , and the RHS is given by Z |det T | E1 (T x) dx = |det T | T 1 E1 .
T 1 E

We see that (0.2) holds by Theorem 3.35.

Case2: Assume f 0. Then there exists a sequence of simple functions 0 sn % f on E where sn = a1 E1 + + ak(n) Ek(n) , Now by Case1 Z Z k (n) depends on n. Z

and by the Monotone Convergence Theorem we obtain Z Z Z sn (y) dy = f (y) dy, lim |det T | lim
n E E n

sn (y) dy = a1 E1 (y) dy + + ak(n) Ek(n) (y) dy E E Z Z E1 (T x) dx + + ak(n) |det T | Ek(n) (T x) dx = a1 |det T | 1 E T 1 E Z T a1 E1 + + ak(n) Ek(n) (T x) dx = |det T | 1 ZT E = |det T | sn (T x) dx
E T 1 E

T 1 E

sn (T x) dx = |det T |

The conclusion follows. For general f, use Z Z Z Z Z + + f = f f = |det T | f (T x) dx |det T | f (T x) dx E E T 1 E T 1 E Z E f (T x) dx. = |det T |
T 1 E

f (T x) dx.

T 1 E

8. (15 points) Let p > 0 be a constant and let f, fk , k = 1, 2, 3, ..., be measurable functions on R R E. If E |fk f |p 0 as k and E |fk |p M (M > 0 is a constant) for all k, show that R p E |f | M also. solution: For any > 0 we have |{|fk f |p > }| 1 Z

|fk f |p 0 as k

due to the Tchebyshev inequality. Hence fk converges to f in measure on E. In particular, there exists a subsequence fkj such that it converges to f a.e. on E. Fatous lemma implies Z Z Z p p p fk = fk M. lim inf j |f | lim inf j
E j E j E

The proof is done.

Remark 1 (be careful) E |fk f |p 0 as k does not, in general, imply that |fk f |p 0 a.e. on E as k . Also for p > 0 the inequality |f |p |fk f |p + |fk |p is wrong in general. It holds only for 0 < p 1. 3

Real Analysis Sample Exam, November 6, 2007

Show detailed argument to each problem. 1. A totally unlucky number is one that contains no sevens in any decimal expansion. Compute the Lebesgue measure of the totally unlucky numbers in [0, 1] . solution:
1 Among the numbers 0.1...., 0.2...., , 0.9...., the lucky numbers has measure s = 10 . Among 1 the numbers 0.11...., 0.12...., , 0.19...., the lucky numbers has measure 10 s. Based on this observation, all of the lucky numbers has measure

s + (1 s) s + [1 s (1 s) s] s + [1 {s + (1 s) s + [1 s (1 s) s] s}] s + = s + (1 s) s + (1 s)2 s + (1 s)3 s + s = = 1. 1 (1 s) Hence the totally unlucky numbers has measure zero. We are lucky!!!

2. Let f : [a, b] R be a nite increasing function. Show that f is a measurable function on [a, b]. For any p (a, b) , evaluate the following limits: Z 1 lim f (Lebesgue integral) h0+ h [p,p+h] and 1 lim h0+ h Z f (Lebesgue integral).

[ph,p]

solution: If f : [a, b] R is an increasing function, the number of x [a, b] such that f is discontinuous at x is at most countable (see Rudin, p.96). Hence f is continuous a.e. on [a, b] and so measurable. We also know that both f (p+) and f (p) exist for any p (a, b) . For xed p (a, b) , there exists a number A such that for any > 0 there exists > 0 so that if x (p, p + ) then |f (x) A| < . Hence Z Z Z 1 1 1 (A ) A f A (A + ) A h [p,p+h] h [p,p+h] h [p,p+h] R R 1 1 for all 0 < h < . Therefore limh0+ h [p,p+h] f = A = f (p+) . Similarly we have limh0+ h [ph,p] f = f (p) . 3. Let En R be a sequence of measurable sets. Let A = {x R : x En for innitely many n} . Is the set measurable or not? Give your reasons. solution: We actually have A= Hence the set A is measurable
j=1 \

[ 1

Ek

k=j

4. Give an example of a measurable function h : E R R such that for some measurable set B R the inverse image h1 (B) is NOT measurable. solution: Let f (x) : [0, 1] [0, 1] be the Cantor-Lebesgue function and let g (x) = x + f (x) . It is easy to see that g (x) : [0, 1] [0, 2] is a strictly increasing continuous function. Hence g (x) is a homeomorphism of [0, 1] onto [0, 2] . On each interval I1 , I2 , I3 , ..., removed in the construction of the Cantor set, say the interval I1 = 1 , 2 , the function g (x) becomes g (x) = x + 1 . Hence g (x) 3 3 2 sends I1 onto an open interval with the same length. Using this observation one can see that ! [ X [ X g Ik = g (Ik ) = |g (Ik )| = |Ik | = 1
k=1 k=1 k=1 k=1

which implies |g (C)| = 21 = 1, where C is the Cantor set. Since g (C) has positive measure, there exists a non-measurable set A g (C) . Now consider the set B = g 1 (A) C. It has measure zero, hence it is measurable. Let h = g 1 . Then it is a measurable function and h1 (B) = A is not measurable. 5. Let E be a measurable set in Rn with |E| < and f is a measurable function on E. Let En = {x E : |f (x)| n} , P Show that f L (E) if and only if |En | < . n=0 n = 0, 1, 2, 3....

solution:

(=) Assume f L (E) . Then f is nite a.e. in E (without loss of generality, we can assume f is nite everywhere in E). It is not hard to see that Z lim |f | = 0
{|f |}

By the relation E = E0 E1 E2 , |E0 | < , we can decompose E as E = (E0 E1 ) (E1 E2 ) (E2 E3 ) and observe that n |En En+1 | Hence
X n=0

which is like the case of an absolutely convergence sequence. This also implies (by Tchebyshevs inequality) Z |f | = 0. (0.1) lim n |En | lim
n n {|f |n}

(disjoint union)

En En+1

|f | (n + 1) |En En+1 | ,
XZ n=0

n = 0, 1, 2, 3, ...

(0.2)

n |En En+1 |

En En+1

|f | =

|f |

where (in below we need to use the fact that |E| < , and so |En En+1 | = |En | |En+1 |) by (0.1) we have X n |En En+1 | = (|E1 | |E2 |) + 2 (|E2 | |E3 |) + 3 (|E3 | |E4 |) +
n=0

= |E1 | + |E2 | + |E3 | + |E4 | + .

|En | < . P (=) Conversely if |En | < , then the set {x E : |f (x)| = } must have measure zero, n=0 implying f is nite a.e. in E . Again we assume f is nite everywhere in E. Since f is bounded on En En+1 , it is integrable on it. Note that Therefore
n=0 X n=0

(n + 1) |En En+1 | = 2 (|E1 | |E2 |) + 3 (|E2 | |E3 |) + 4 (|E3 | |E4 |) +

= 2 |E1 | + |E2 | + |E3 | + |E4 | + < . By (0.2) we know |f | must be integrable on E. Hence f is integrable on E.

6. Assume h (x) is a dierentiable function on R. Show that h0 (x) is a measurable function on R. solution: Let fn (x) = 1 h x + n h (x)
1 n

x R.

For each n = 1, 2, 3, ..., fn (x) is a nite measurable function on R with fn (x) h0 (x) for all x R. Hence h0 (x) is a measurable function on R. 7. In the Lebesgue Dominated Convergence Theorem (Theorem 5.36) if we replace the condition fk f a.e. in E by fk f in measure on E, is the theorem still correct or not? Give your reasons. solution: The theorem is still correct. First note that by Theorem 4.22 there exists a subsequence fkj f a.e. on E. This implies that |f | a.e. in E and so f L (E) (since L (E)). By the usual Lebesgue Dominated Convergence Theorem we have Z Z fkj f as j . (0.3)
E E

R R We then use contradiction argument to show that E fk E f as k . Assume not. Then there exists a subsequence of fk , still denote it as fkj , j = 1, 2, 3, ..., so that Z Z fkj f > 0
E E

for all j. But then this subsequence has a further subsequence so that (0.3) holds, a contradiction.

Excercise 1: Prove that the convergence of {sn } implies convergence of |sn |. Is the converse true? Proof. (Im assuming the sn are real numbers since Rudin doesnt say, and in an arbitrary metric space |x| is meaningless). Assume sn converges to L. We claim that |sn | converges to |L|. To see this, let > 0 be an arbitrary real number. We know that there is an N N such that for any n N with n N , |sn L| < , and we claim the same N works for |sn |. To see this, let n N . Then ||sn | |L|| |sn L| < (See homework 2 problem 4 for the part). Thus, |sn | converges to L. The converse, however, is FALSE. To see this, consider the sequence sn = (1)n . Then |sn | = 1 for all n, so of course sn converges to 1. However, sn cant converge. To see this, one can just use the fact quoted in Rudin in denition 3.5: It is clear the {pn } converges to p i every subsequence of {pn } converges to p. Thus, since a subsequence of sn is simple 1, 1, 1, ..., its clear this converges to -1. While another subsequence is 1, 1, ..., which clearly converges to 1. It follows that sn has two subsequences which converge to dierent numbers, and hence sn cannot converge. Excercise 2: If s1 = 2 and sn+1 = 2 + sn , prove that {sn } converges and that sn < 2 for all n. Proof. Well show the sequence is monotonely increasing and bounded above by 2. It will follow from theorem 3.14 that {sn } converges. To see its bounded by 2, notice that s1 clearly is. Now, assume induc tively that sn < 2. Then s2 = 2 + sn < 2 + 2 < 4, so that sn+1 < 2. n+1 Then, by the principle of mathematical induction, it follows that sn < 2 for all n. To see that its increasing, well again prove it inductively. Notice that 2 > 2 = s1 , so we start increasing. Now, assume sn1 < sn . s2 = 2 + Then sn1 < sn so 2 + sn1 < 2 + sn . Finally, it follows from this that 2 + sn1 < 2 + sn , but this just says sn < sn+1 . Thus by the principle of mathematical induction, sn < sn+1 for all n. Excercise 3: For any two real sequences an and bn , prove that limsupn an + bn limsupn an + limsupn bn , provided the sum on the right is NOT of the form . Proof. Assume that limsupn an = L < . (The proof with limsupn bn nite works similarly).

Now, we break into three cases depending on whether or not limsupn bn is nite, , or . If its , We claim that both limsupn an + bn and limsupn an + limsupn bn are . Note that its clear that limsupn an +limsupn bn = L = , so well just show limsupn an + bn = . To see this, let r be any real number. We must show an + bn < r for all n large enough, for then it follows that any subsequence of an + bn eventually gets below r, and hence limsupn an + bn r. Doing this for all r R then says limsupn an + bn = . To see that an + bn < r for suciently large n, just notice rst that since limsupn bn = , there must be an N1 such that for all n N1 , bn < r|L|. This is because otherwise, for each N1 N, we could nd an xnN1 with xnN1 r. Then the subsequence xnN1 would show that limsupn bn r, giving a contradiction. There must also be an N2 such that for any n N2 , an < L, since otherwise wed nd L wasnt the limsup of an . If we let N be the max of N1 and N2 , then for any n N , we have that an + bn < L + r |L| r. Thus, the claim is established and it follows that limsupn an + bn = . Now, if limsupn bn = L1 < , then as before, for any > 0, there is an N such that for any n N , an L + /2 and bn L1 + /2. Then we have that for any n N , an + bn L + L1 + . It follows that any subsequence of an + bn has limit as most L + L1 . Thus, we see that limsupn an + bn L + L1 = limsupn an + limsupn bn . Finally, in the case that limsupbn = , notice there is nothing to prove since the right hand side of the inequality is and of course, anything stuck on the left will be less than or equal to that. Excercise 4. If {sn } is a Cauchy sequence, dene its arithmetic mean n 1 +...+s by n = s0 +sn+1 n . a) If limsn = s, prove that limn = s. b) Construct a sequence {sn } which does not converge although limn = 0. c) Can it happen that sn > 0 for all n and that limsupsn = although limn = 0? 1 d) Let an = sn sn1 . Show that sn n = n+1 n kak . Assume that k=1 lim(nan ) = 0 and that n converges. Prove that {sn } converges. e) Derive the last conclusion from a weaker hypothesis: Assume M < |nan | M for all n and that lim n = . Prove that limsn = by completing the following outline....

Proof. a) Let > 0. We wish to nd an N N such that for any n N , well have |n s| < . We know there is an N such that for any n N , N n=1 s |sn s| < /4. Now, choose N > N such that N +1 n < /4 and such that N s/(N + 1) < /4. I claim this N works. For if n N , then |n s| = | sn |+| | N +1
N k=1 N k=1

sn +

N k=N +1

sn N s

N +1
N k=N +1

sn s | + |N s/(N + 1)| N +1
N

/4 + /4 +
k=N +1 N

|sn s|/(N + 1)

/2 +
k=N +1

/4(N + 1) = /2 + (N N 1) /4(N + 1) < /2 + /4 = 3 /4 < .

Thus N prime works. 1 b) Consider sn = (1)n . Then even = 0 while 2n+1 = 2n+2 . Thus we see that lim n = 0. However, cleary sn doesnt converge (see problem 2). k n = k 2 for some k N c) Yes it can, consider sn = Then if you 2n otherwise look at the subsequence of the form sk2 , its clear this has limit . Thus, limsupn sn = . Further, its clear that sn > 0 for each n. What about n ? n n 1/4 3/4 1/4 = 1+1/2+...+1/2 + n +n = Well, s1 +...+sn 1+1/2+...+1/2n +( n+1)n n n n 1+...+1/2n 1/4 3/4 +n +n . Now, for the second two terms, the limit as n n is clearly 0, so we must simply argue the rst term goes to 0 as well. But (n+1) 1 + 1/2 + ... + 1/2n = 12 = 2 2(n+1) . Thus (1 + ... + 1/2n )/n = 11/2 2/n 2(n+1) /n, both of which clearly go to 0. d)Well prove it inductively. If n = 1, then 1 = (s0 + s1 )/2 so we see s1 1 = s1 /2 s0 /2 while 1/2 k ak = 1/2(s1 /2 s0 /2), so they agree. Now, assume its true for n, well show its true for n + 1. Then n+1 n 1 1 n+1 kak = kak + an+1 n + 2 k=1 n + 2 k=1 n+2 n+1 1 = n+2n+1
n

kak +
k=1

n+1 an+1 n+2

n+1 n+1 (sn n ) + an+1 n+2 n+2 n+1 = (sn n + sn+1 sn ) n+2 n+1 = (sn+1 n ) n+2 n+1 1 sn+1 n = sn+1 n+2 n+2 1 = sn+1 (s0 + ... + sn+1 ) = sn+1 n+1 . n+2 Ill skip the next part since well prove a stronger result in part e) e) Assume M < and |nan | M for all n. Let n . Now, if m < n, then = 1 m+1 (n m ) + (sn si ) nm nm 1 1 m + 1 s0 + .... + sn s0 + ... + sm ( )+ (n(m+1)+1)sn = si nm n+1 m+1 nm n m i=m+1 = sn + 1 m+1 ( (s0 + ... + sn ) s0 ... sm si ) nm n+1 i=m+1
n n

1 m+1 ( (s0 + ... + sn ) s0 ... sn ) nm n+1 1 (m + 1)(s0 + ... + sn ) (n + 1)(s0 + ... + sn ) = sn + ( nm n+1 n+1 1 m + 1 (n + 1) ( (s0 + ... + sn )) = sn + nm n+1 = sn n . = sn + Now, for the i in the sum, we have |sn si | = |sn sn+1 + sn+1 sn2 + .... si | = |an + ... + ai+1 | |an | + |an1 | + ... + |ai+1 | M/n + M/(n 1) + ... + M/(i + 1) (n i)M/(i + 1). I claim that (n i)M/(i + 1) < (n m)M/(m + 1) (this is obtained by just plugging in i = m into the expression. This is easiest to see by writing (n i)M/(i + 1) = (n + 1 i 1)M/(i + 1) = (n + 1)M/(i 1) N . But then its clear that one can make this larger, by taking i smaller - that is i = m. Thus, we have |sn si | < (nm)M . m+1 Now, pick > 0 and let m(n) be dened as the integer satisfying m(n) 1 n < m(n). 1+ 4

Then 0 n m = nm (m+1) so that 0 n m (m + 1). Thus, 1+ 1+ nm . Then, we see that |sn si | < (nm)M < M . m+1 m+1 Thus we see that |sn | = |sn n + n | |sn n | + |n |. Now, we can make the second term as small as we want since n , so lets focus on the rst term. 1 m+1 The rst term is |sn n | = | nm (n m ) + nm n i=m+1 (sn si )| m+1 1 | nm (n m )| + nm n |(sn si )|. Now, with the rst term of this, i=m+1 since n it follows that n is bounded, and hence that the rst term 1 goes to 0 for large enough n. Finally, we deal with nm n i=m+1 |(sn si )|. n 1 We just showed |sn si | < M . Then we have nm i=m+1 |sn si | nm+1 |sn si | < M . Thus, we can make this term as small as we like. nm Finally, since we can make all the terms as small as we like, it follows that we can make |sn | as small as we like. It follows that sn . Excercise 5. Suppose {pn } is Cauchy in a metric space X and some subsequence pnk converges to p X. Prove pn converges to p as well. Proof. Let > 0. We seek an N such that for all n N , |pn p| < . To that end, we know that there is an N1 such that for any k N1 , |pnk p| < /2. Likewise, we know there is an N2 such that for all n, m N2 , |pn pm | < /2 (since pn is a Cauchy sequence). Let N = max{N1 , N2 }. Let n N . Then |ppn | = |ppnn +pnn pn | |p pnn | + |pnn pn | < /2 + /2 = . Excercise 6. Prov ethe following analogue of Theorem 3.10(b): If En is a sequence of nonempty bounded sets in a complete metric space X, and if En+1 En and if limdiam(En ) = 0, then E = En has exactly one n=1 point. Proof. For each n, let pn En . I claim the pn form a cauchy sequence. To see this, let > 0. Then there is an N such that for any n N , diam(En ) < . Then for any n, m N , pn , pm EN and diam(EN ) < so that d(pn , pm ) < epsilon. Thus the pn form a cauchy sequence. Since X is complete, pn p X. Next, notice for each n, the sequence pn , pn+1 , ... is a sequence in En and also a subsequence of the original sequence. Hence, each subsequence covnerges to p and since En is closed, p En . Thus, p En . Thus, n=1 E = has at least one point in it. n=1 If it had two distince points p and q, then d(p, q) would be a lower bound for the diameter of E. But since E En for all n, it follows that d(p, q) is 5

a lower bound for the diameter of En for all n. But we know the diameter of all the En approaches 0, and hence gets smaller than d(p, q). This gives a contradiction and thus E consists of exactly one point. (We never used the fact that E was bounded anywhere....?) Excercise 7: Suppose {pn } and {qn } are Cauchy sequences in a metric space X. Show that d(pn , qn ) converges. Proof. Since d(pn , qn ) takes values in R, which is complete, its enough to show this sequence is Cauchy. Thus, let > 0. We know there is an N so that for n, m N1 , |pn pm | < /2 and |qn qm | < /2. I claim this same N works for d(pn , qn ). So, let n, m N and consider |d(pn , qn ) d(pm , qm )|. Assume wlog that d(pn , qn ) d(pm , qm ). Then |d(pn , qn )d(pm , qm )| = d(pn , qn )d(pm , qm ) d(pn , pm )+d(pm , qn ) d(pm , qm ) d(pn , pm ) + d(pm , qm ) + d(qm , qn ) d(pm , qm ) = d(pn , pm ) + d(qm , qn ) < /2 + /2 = .