Академический Документы
Профессиональный Документы
Культура Документы
The real analysis prelim will be based on the material related to the topics listed below. This list
is not meant to be exhaustive, but is intended to be a guide to subjects to be studied thoroughly.
1
Real Analysis Prelim January 2008
1. Let (X, B, ) be a measure space. Suppose f and g are non-negative integrable functions
such that A B implies Z Z
f d = gd.
A A
Prove that f = g a.e.
3. Let A [0, 1] be a measurable set of positive measure. Prove that there exist x, y A
such that x 6= y and x y is rational.
R1 R1
4. Let f, g be two nonnegative measurable functions on [0, 1] such that 0 f (x)dx = 1, 0 g(x)dx =
R1
2, 0 (g(x))2 dx = 5. Let E = {x|g(x) f (x)}. Show that m(E) 51 where m is Lebesgue
measure.
5. Set Z
h(x) = (x + y)1 f (y)dy
0
for f L1 (0, ). Show that h is C away from the origin. Show as well that h0 L1 [a, )
for all a > 0.
Show that
1
lim x 2 f (x)
x0
2
Real Analysis Prelim August 2007
for almost all x [0, 1]. Here m denotes Lebesgue measure on [0, 1].
2. Let f Lp [0, 1], where 2 < p . Show that
|f (x)|
Z
dx < .
[0,1] x
3. If 1 < p < , let {fn } be a sequence of elements of Lp [0, 1] such that there exists M > 0
with ||fn ||p M for all n > 0. Suppose that the sequence {fn } converges pointwise almost
everywhere to a function f defined on [0, 1].
(a) Prove that f Lp [0, 1].
(b) Let q be the conjugate exponenet to p. Prove that for each function g Lq [0, 1],
Z Z
lim fn (x)g(x)dx = f (x)g(x)dx.
n [0,1] [0,1]
3
(b) Define : [0, ) [0, ] by (t) = m({x R : f (x) t}). Prove that is a
decreasing function of t and that
Z Z
(t)dt = f (x)dx.
[0,) R
4
Real Analysis Prelim January 2006
Ut (f )(x) = f (x t), x R.
3. Prove that
Z 1
X 1 1
n
= x
dx.
n 0 x
n=1
R
You may use the Gamma function evaluation 0 et tk dt = k!.
4. If f is a real-valued function on R, write Sf0 for the set of all x such that f is differentiable
at x. Call a set S a differentiation set if there exists a function f for which S = Sf0 .
(a) Show that Sf0 +g contains Sf0 Sg0 . Show by example that it can happen that Sf0 +g = R
but Sf0 Sg0 = .
(b) Show that every open interval I is a differentiation set.
(c) Let E be a Lebesgue-measurable subset of R with m(E) < . Show that, for each
> 0, there exists a differentiation set S for which m(ES) < , where m denotes
Lebesgue measure, and ES is the symmetric difference of the sets E and S.
5. Let f and g be real-valued Lebesgue measurable functions on [0, 1], not assumed to be
integrable. Let E = {(x, y) [0, 1] [0, 1] : f (x) = g(y)}.
(a) Prove that E is measurable with respect to the Lebesgue product measure m m
defined on [0, 1] [0, 1].
(b) Suppose in addition that m m(E) = 1. Prove that there is a real constant c such
that f g c, m a.e. on [0, 1].
Show that for such a sequence, if fn f almost everywhere for some f , then fn f in
L1 [0, 1].
5
R
Step 1: First show that supn [0,1] |fn |dx < .
Step 2: Next show thatR the (almost-everywhere) limit f is indeed in L1 [0, 1].
Step 3: Last, estimate [0,1] |f fn |dx, employing Egorovs Theorem to finish the job.
6
Real Analysis Prelim August 2006
1. Let {fn } be a sequence of real-valued measurable functions defined on the real line. Let
E = {x : limn fn (x) exists}. Show E is a measurable set.
4. Let f be a real-valued measurable function defined on [a, b]. Then given > 0 show there
exists a continuous function g on [a, b] such that {x : f (x) 6= g(x)} < . Here is
Lebesgue measure.
(a) Show that the product f g is also absolutely continous on [0, 1].
R1 R1
(b) If g(0) = f (1) = 0 then show 0 f 0 (x)g(x)dx = 0 f (x)g 0 (x)dx. Also, explain why
these integrals exist.
R1 R1
(c) Suppose f and h are in L1 (R). Suppose furthur that 0 h(x)g(x)dx = 0 f (x)g 0 (x)dx
for all absolutely continous functions g with g 0 L [0, 1] and g(0) = 0. Show f is
absolutely continous and h = f 0 , -almost everywhere.
Here is Lebesgue measure.
6. Prove the following. Let (X, S, ) and (Y, T ) be -finite measure spaces. Suppose K(x, y)
is a real-valued measurable function with respect to . Suppose there exists a C > 0
such that Z
|K(x, y)|d C, for all x X
Y
Z
|K(x, y)|d C, for all y Y .
X
7
Real Analysis Prelim August 2005
You may use the fact that the space CC (R) of continuous, Rcomplactly supported functions
on R is dense in L1 (R), with respect to the norm ||g||1 = R |g(x)|dx.
2. For this problem, you may take it as given that, if f and g are measurable functions on R,
then F (x, y) = f (x y)g(y) defines a measurableR function on R2 . The convolution f g
of measurable functions f, g on R is the function f g, of a real variable x, defined by
Z
f g(x) f (x y)g(y)dy.
R
The domain of f g is the set of all x such that f g(x) is defined (that is, such that
F (x, y) = f (x y)g(y) defines a Lebesgue integrable function of y).
(a) Show that, if f, g are continuous on R and vanish outside R+ = (0, ) then
Z x
f g(x) = f (x y)g(y)dy,
0
and then use this to show that f g is continuous and vanishes outside R+ . Hint for
continuity: Note that
Z x0 Z x
f g(x) f g(x0 ) = (f (x y) f (x0 y))g(y)dy + f (x y)g(y)dy.
0 x0
(b) Show that, if f, g L1 (R), then f g L1 (R) (in particular, f g(x) is defined for
almost all x R. , and ||f g|| 1 ||f ||1 ||g||1 .
3. For this problem, you may assume that for any function f on R to R the set of all points
of discontinuity is an F set.
For the following three questions, if your answer is yes, explicitly construct such an exam-
ple. If your answer is no, give an argument that no such function exists.
(a) Does there exist a function from R to R that is continuous at exactly one point?
(b) Does there exist a function from R to R that is continuous at every irrational point
and discontinuous at every rational point?
(c) Does there exist a function from R to R that is continuous at every rational point,
and discontinuous at every irrational point?
4. (a) Prove Tchebyschevs inequality, which says the, if f is nonnegative and integrable on
[0, 1] and is Lebesgue measure on [0, 1], then
Z
1
({x [0, 1] : f (x) c}) f (x)dx.
c [0,1]
8
(b) We define an ultraCauchy sequence in a metric space (X, d) to be a sequence a1 , a2 , ...
in X such that, for some c > 0,
c
d(ak , ak+1 )
k3
for all k Z+ .
Let h1 , h2 , ... be an ultraCauchy sequence in L2 (0, 1), the space of square-integrable
functions on [0, 1], with metric
sZ
d(f, g) = ||f g||2 = |f (x) g(x)|2 dx.
[0,1]
Show that the hk s converge pointwise almost everywhere: that is, show that limk hk (x)
exists for almost all x [0, 1].
Hint: For each k 1, define Gk = {x [0, 1] : |hk (x) hk+1 (x)|2 k14 }. Use part
(a) to show that, given > 0, there is an N Z+ such that HN = k=N Gk has a
measure less than . Then show that, if x / N HN , {hk (x)} is a Cauchy sequence.
9
Real Analysis Prelim January 2004
1. Let be Lebesgue measure on [0, 1] and let : [0, 1] [a, b] be measurable. Define a set
function on the Borel sets of [a, b] by (A) = (1 (A)).
d
R R
Prove that dt R cos(tx)f (x)dx = R sin(tx)xf (x)dx.
10
Real Analysis Prelim August 2004
1. Let (X, B, ) be a measure space. Suppose f and g are non-negative integrable functions
such that A B implies Z Z
f d = gd.
A A
Prove that f = g a.e.
3. Let A [0, 1] be a measurable set of positive measure. Prove that there exist x, y A
such that x 6= y and x y is rational.
R1
4. Let f, g be two nonnegative measurable functions on [0, 1] such that 0 f (x)dx = 1,
R1 R1 2 1
0 g(x)dx = 2, 0 (g(x)) dx = 5. Let E = {x|g(x) f (x)}. Show that m(E) 5
where m is Lebesgue measure.
5. Set Z
h(x) = (x + y)1 f (y)dy
0
for f L1 (0, ). Show that h is C away from the origin. Show as well that h0 L1 [a, )
for all a > 0.
Show that
1
lim x 2 f (x)
x0
11
Real Analysis Prelim January 2002
1. Let (X, ) be a separable metric space. Suppose S X. Show that there exists a countable
set F S such that F is dense in S, i.e., S F = closure of F .
2. Let (X, S, ) be a measurable space for which (X) < and E S implies (E) 0. Let
f : X R be S-measurable. Suppose g : R R+ = {x R : x 0} is Borel measurable.
For a Borel set E R, define (E) = (f 1 (E).
4. Let f : [0, ) R be a bounded measurable function such that limx x2 f (x) = 1. Find
an integral expression for R
(1 cos(x))f ( x )dx
lim 0 .
0+ 2
Justify your answer.
5. Let {fn } and f be measurable functions on [0, 1] with Lebesgue measure. Suppose that
fn converges in measure to f , and that for some p > 1, supn ||fn ||p < .
6. Let (X, S, ) be a measure space such that (X) > 0. Recall that L (X, S, ) is a metric
space under the L-infinity norm. An element E of S is called an atom if (E) > 0 and
whenever F S satisfies F E, then either (F ) = 0 or (E \ F ) = 0.
12
Real Analysis Prelim August 2002
1. (16 pts) Let (X, d) be a metric space, and let A X. Recall that a point x X is said to
be an accumulation point of the set A if every neighborhood of x contains a point a A
with a 6= x.
(a) Let x X. Prove that the set {x} is nowhere dense in X if and only if x is an
accumulation point of X.
(b) State, but do not prove, the Baire Category Theorem.
(c) Prove that if (X, d) is complete and each x X is an accumulation point of X, then
X is uncountable.
(d) Using part (iii), deduce that R is uncountable.
2. (17 pts)
(a) Define simple functions, taking their domain to be a bounded interval [a, b] in R.
(b) Show that sums and products of simple functions are simple.
(c) Define step functions, taking their domain to be a bounded interval [a, b] in R.
(d) Let f be a bounded measurable function on a bounded interval [a, b]. Define the
Lebesgue integral of f on [a, b].
(e) Let f be a bounded measurable function on a bounded interval [a, b], which is Riemann
integrable. Show that the Riemann and Lebesgue integrals of f on [a, b] coincide.
3. (16 pts) Let E be a Lebesgue-measurable subset of R, and assume that the measure of E
is 0. Let AE be the subset of R2 consisting of the points (x, y) for which x2 + y E.
4. (17 pts) Let f L1 ([0, 1]), and suppose that f1 , f2 , ... is a sequence of absolutely continuous
functions on [0, 1] such that f (x) = limn fn (x) for every x [0, 1].
(a) Suppose the sequence f10 , f20 , ... converges in L1 to a function g. Prove that f is
absolutely continuous.
(b) Suppose that there exists an L1 function g for which |fn0 (x)| g(x) for every x [0, 1]
and for every n N. Prove that f is absolutely continuous.
(c) Show by giving an example, that the L1 norms of the functions fn0 can be uniformly
bounded, and yet f is not absolutely continuous.
5. (9 pts) Prove that it is impossible to find bounded sequences of numbers, a1 , a2 , ... and
b1 , b2 , ... such that limn [an cos(nx) + bn sin(nx)] = 1a.e. on [0, 2].
6. (16 pts) Prove Egoroffs Theorem, as stated here (m denotes Lebesgue measure): Let
f1 , f2 , ... be a sequence of measurable functions from [0, 2] to R that converges a.e. to a
real-valued measurable function f . Then for every > 0, there is a subset A [0, 2] with
m(A ) < , such that f1 , f2 , ... converges uniformly to f on [0, 2]\A .
7. (9 pts) Prove the following form of the Riemann-Lebesgue Lemma. If f L1 (R), then
Z
lim f (x) cos(nx)dx = 0
n R
13
and Z
lim f (x) sin(nx)dx = 0.
n R
Hint: As a first step, prove the statements for the case in which f is the characteristic function
of a finite interval.
14
Real Analysis Prelim January 2001
1. Let m be Lebesgue measure on [0,1] and let {fn } be a sequence of measurable functions
on [0,1]. Either prove or give a counter-example to each of the following statements:
2. Let R and set f (x) = x sin( x1 ) if x > 0 and f (0) = 0. For which is f of bounded
total variation on [0,1]? Prove that your answer is correct.
3. Let m be Lebesgue measure on R, and let {fn } be a sequence of functions in L2 (m) with
||fn ||2 1 for all n.
a) If fn converges to f in measure (m) prove that f L2 (m) and that ||f ||2 1.
b) If in addition ||fn ||2 ||f ||2 prove that fn f in L2 (m).
5. Let X = Y = [0, 1] and let B be the Borel sets in [0,1]. Let (X, B, ) be Lebesgue measure
on B and letR R(Y, B, ) be counting
R R measure onR B. Let D = {(x, y) X Y : x = y}.
Prove that ( D d) d, ( D d) d, and D d( ) are all unequal. Here D is
the indicator (characteristic) function of the set D.
6. a) Give an example of a measure space (X0 , B0 , ) for which 1 p < q implies that
Lp () Lq (). (No proof required)
b) Give an example of a measure space (X1 , B1 , ) for which 1 p < q implies that
Lq () Lp (). (No proof required)
c) Give an example of a measure space (X2 , B2 , ) for which for any 1 p < q both
Lp () 6 Lq () and Lq () 6 Lp (). (No proof required)
d) Choose one of your examples in a), b), or c) and prove that it is an example having
the claimed property.
15
Real Analysis Prelim August 2001
2. Let (X, d) be a metric space and let x1 , x2 , ... be a Cauchy sequence in X. Prove that if
the sequence has a cluster point, x, then it converges to x.
3. A function f : [0, 1] R is said to be Holder continuous of order 1/2 provided there exists
1
a constant C such that |f (x) f (y)| |x y| 2 for all x and y in [0, 1]. Let f1 , f2 , ... be
a sequence of Holder continuous functions of order 1/2, taking [0,1] to [0,1], such that the
same constant C suffices for all of them. Show that there is a subsequence of f1 , f2 , ... that
converges uniformly on [0,1] to a continuous function f .
(a) Show that a function f satisfying a Lipschitz condition on [0,1] is absolutely contin-
uous.
(b) Show that an absolutely continuous function f satisfies a Lipschitz condition iff |f 0 |
is bounded.
5. Let f : R R be continuous and of period 2, i.e. f (x + 2) = f (x) for all x. Show that
for every > 0, there is a finite Fourier series
N
X
(x) = a0 + (an cos(nx) + bn sin(nx))
n=1
6. Let f L1 (R) and suppose that g : R R has compact support and is infinitely differen-
tiable. Define h : R R by Z
h(x) = f (t)g(x t)dt,
R
for x R. Show that this makes sense and that the resulting function is everywhere
differentiable, with Z
h (x) = f (x)g 0 (x t)dt
0
for x R.
7. Let f, g : R R be Borel measurable functions such that f (x) = g(y) for almost all
pairs (x, y) relative to Lebesgue measure in the plane. Prove that f and g agree almost
everywhere with constant functions.
16
Real Analysis Prelim August 2000
2. Suppose f is a measurable
Rb real-valued function defined on R. Suppose for all pairs of real
numbers a < b, 0 a f (x)dx b a. Show that 0 f (x) 1 for almost all x.
4. Let f be a continuous function on [0, ). Suppose limx f (x) exists (and is finite).
Prove that f is uniformly continuous on [0, ).
5. A function f is continuously differentiable on (0,1). Let fn (x) = n[f (x + n1 ) f (x)] and let
0 < a < b < 1. Show that fn (x) converges uniformly on (a, b). Find out the limit function.
6. Suppose fn ,gn are non-negative finite functions on [0,1] related to each other by the (non-
linear) integral equation
Z 1
fn (x) + k(x, y)fn2 (y) dy = gn (x),0 x 1,
0
where k(x, y) is a non-negative continuous function on [0, 1] [0, 1]. Suppose fn are mono-
tone decreasing in n (and so are gn ). Put f = limn fn , g = limn gn . Prove that
Z 1
f (x) + k(x, y)f 2 (y) dy = g(x).
0
17
Real Analysis Prelim January 6, 1999
2. Let X be a set equipped with a -algebra A. If mi is a measure on (X, A) for each i, such
X 1
that mi (X) 1. Define m : A [0, +) by m(E) = mi (E) for all E A. Prove
2i
i=1
that m is a measure on (X, A).
3. Compute Z 1
1
lim 1 dt.
n 0 (1 + nt )n t n
Justify your answer
4. If f (x) is a Lipschitz function on R, i.e. there exists M such that |f (x) f (y)| M |x y|
for all x, y R, prove that f 0 (x) exists almost everywhere.
5. Let I = [0, 1]. Prove that the set of all polynomials with rational coefficients is dense in
Lp (I), 1 p < .
6. Suppose that I = [0, 1], 0 < p < 1, p1 + 1q = 1. |f (x)|p dx < , g(x) 6= 0 for all x I and
R
Z Z 1 Z 1
p q
q p q
R
|g(x)| dx < . Prove |f (x)g(x)|dx |f (x)| dx |g(x)| dx .
I I I
18
Real Analysis Prelim August 1999
1 1
1. If 1 < p < and 1 < q < such that p + q = 1 and fn f in Lp , gn g in Lq , then
show that fn gn f g in L1 .
Prove that fn (x) converges uniformly on every closed finite interval [a, b]. Find the limiting
function.
3. If A R2 is measurable and every vertical line meets A in a countable set, what is the
Lebesgue measure of A?
R
5. Let f be a positive continuous function such that f (x) dx is finite. Assume that, for
some T > 0, f is decreasing on [T, ) and it is increasing on (, T ]. Show that
X
F (x) = f (x + k)
k=
19
6. Let n , n = 1,2,. . . be a complete orthonormal system in the space L2 (R, ), where
stands for Lebesgue measure.
(a) Show that, for every Borel set B R of strictly positive Lebesgue measure, one has
Z X
1 |n (x)|2 dx.
B 1
Hint: Assuming (B) < , apply the Parceval identity to 1B . Then use Schwartzs
inequality.
(b) Show that 2
P
1 |n (x)| = a.e.
20