Real Analysis Final Exam, 2008-1-8 Name:

Show detailed argument to each problem. 1. (10 points) Let E Rn be a measurable set and q (x) 0 be a nonnegative measurable function on E. Show that for any constant > 0 we have the inequality Z Z Z qdx q dx |E| q 1+ dx.
E E E

ID #:

solution: By Hlder inequality Z qdx Z q

1 Z 1+ 1+ dx dx ,
E

q dx

(q )

1+

Z 1 1+ 1+ dx dx
E

the proof is done.

2. (10 points) Assume |E| < and u is a measurable function on E, which is everywhere positive. For 0 < p < , set 1/p 1/p Z Z 1 1 p p p (u) := |u (x)| dx = u (x) dx . (0.1) |E| E |E| E Assume that (1) . up (x) 1 uniformly on E as p 0+ . R d R d (2) . dp E up (x) dx = E dp (up (x)) dx. (3) . |log u (x)| C for all x E, for some constant C > 0.

Evaluate the limit limp0+ p (u) in terms of an integral involving u. (Compare with the limit limp p (u) = kuk .) solution: As p 0+ , p (u) has the form 1 . Hence we can use Lopital rule. Note that R 1 log |E| E up (x) dx 0 form log p (u) = p 0 and so lim log p (u) = lim
1 |E| d dp

log

p0+

p0+

1 |E|

= lim

p0+

1 |E|

up (x) dx E
1 |E|

= lim

p0+

d E dp

Hence we have the formula

(up (x)) dx up (x) dx = lim

lim p (u) = exp

p E u (x) log u (x) dx R 1 p |E| E u (x) dx

1 d p |E| dp E u (x) dx R 1 p |E| E u (x) dx

1 |E|

log u (x) dx.

1 |E|

log u (x) dx .

3. (10 points) If f Lp (Rn ) , where 0 < p < is a constant, show that Z 1 lim |f (y) f (x)|p dy = 0 Q&x |Q| Q for almost everywhere x Rn . (This is a previous homework problem.) solution: Let {rk } be the set of all rational numbers. For any k and any 0 < p < , the function |f (y) rk |p k=1 is clearly locally integrable on Rn . By Theorem 7.11 of the book, for each k we have Z 1 |f (y) rk |p dy = |f (x) rk |p (0.2) lim Q&x |Q| Q S for a.e. x Rn . Let Zk be the set such that (0.2) is not valid, |Zk | = 0, and set Z = Zk , |Z| = 0. If k=1 x Z, then by the inequality (in below, Q is centered at x and k is arbitrary) / Z Z 1 1 p |f (y) f (x)| dy {|f (y) rk | + |rk f (x)|}p dy |Q| Q |Q| Q Z Z C (p) C (p) p |f (y) rk | 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 1 lim sup Q&x |Q| Z |f (y) rk |p dy 2C (p) |rk f (x)|p for all x Z /

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 |f (y) f (x)|p dy = 0 for all x Z / lim sup |Q| Q Q&x for any 0 < p < . Hence (0.2) holds for a.e. x Rn .

4. (10 points) Give an example of a bounded continuous function f on (0, ) such that limx f (x) = 0 but f Lp (0, ) for any p > 0. Give your reasons. (This is exercise 12 in p. 86.) / solution: For example, one can take f (x) =
1 log 2 , 1 log x ,

x (0, 2) x [2, ).

Then f (x) is bounded continuous on (0, ) with limx f (x) = 0, but for any p > 0 we have p Z Z y Z e 1 p f (x) dx = dx = dy = . log x yp 2 2 log 2 Hence f Lp (0, ) for any p > 0. /

5. (10 points) Give an example of a set E R such that the number x = 0 is a point of density of E, but it is not a Lebesgue point of the function E (x) . Give your reasons.

solution: Let E = (1, 1) \ {0} . We have T |E Q| = 1. lim Q&0 |Q|

Hence x = 0 is a point of density of E. Since 0 E, we cannot have / Z 1 lim |E (y) E (0)| dy = 0. Q&0 |Q| Q Hence x = 0 is not a Lebesgue point of the function E (x) . 6. (15 points) (a) (10 points) Let 0 < p < . Assume fk , f Lp (E) , fk f a.e. on E, and kfk kp kf kp . Show that kfk f kp 0 as k . (This is a slight modication of exercise 12 in p. 144.)

(b) (5 points) If we replace 0 < p < in above by p = , do we have the same conclusion or not? Give your reasons. solution: For (a): For any 0 < p < , we have the inequality |f fk |p 2p |f |p + 2p |fk |p 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

where

and lim inf

lim inf (2 |f | + 2 |fk | |f fk | ) =

(2 |f | + 2 |fk | |f fk | ) = the integral R

Since we assume f

Lp ,

2p+1 |f |p lim sup

2p+1 |f |p Z |f fk |p .

p+1

|f | lim sup
k

is nite. Hence we conclude |f fk |p 0.

This implies kfk f kp 0 as k . For (b): For p = , the conclusion fails. Take E = R and fk = (k,k) , f = 1. Then fk , f L (R) , fk f a.e. on R, and kfk k = kf k = 1. But we do not have kfk f k 0 as k . Another example is E = (0, 1) , fk = (1/k,1) , f = 1, k N. Then we have the same conclusion. 7. (15 points) Let E = (0, ) (0, 1) R2 and f (x, y) = yexy sin x, (x, y) E. (b) (10 points) Evaluate the integral (a) (5 points) Show that f (x, y) is integrable on E, i.e., f L (E) . ZZ f (x, y) dxdy.

solution: For (a), note that ZZ ZZ

|f (x, y)| dxdy

yexy dxdy

and by Tonelli Theorem we have ZZ Z yexy dxdy =

Z yexy dx dy =

dy = 1

which implies that f L (E) . For (b), we apply Fubini Theorem and get ZZ f (x, y) dxdy = Z
1

exy sin xdx dy.

Recall the elementary formula Z eax (a sin bx b cos bx) eax sin bxdx = + C, a2 + b2 we see that for each y > 0 Z
0

a, b are constants

xy

and so

x= exy (y sin x cos x) 1 sin xdx = = 2 2+1 y y +1 x=0

xy

Z sin xdx dy =

y2

1 y dy = log 2. +1 2

8. (20 points) Answer Yes or No to each statement. Just provide the answers. (a) If x is a point of density of a measurable set E, then x E. ANS: _________

(b) Let 0 < p < . If kfk f kp,E 0 as k , then there exists a subsequence fkj f almost everywhere in E as j . ANS: _________ (c) Let f and g be two nonnegative measurable functions on measurable set E satisfying |{x E : f (x) > y}| = |{x E : g (x) > y}| R R ANS: _________ Then we must have E f = E g. for all y > 0.

(d) kfk f k,E 0 as k if and only if fk f uniformly on E except on a set Z E of measure zero. ANS: _________ solution: The answers to the ve problems are: NO, YES, YES, YES.