Scribd
Upload
28 views1 page

Exercise Set B: Some Exercises For TMA4230 Functional Analysis

The document contains 7 exercises related to functional analysis. Exercise B.1 asks to show that a metric dp defined on Lp spaces is indeed a metric and that Lp is complete under this metric. Exercise B.2 defines ordered vector spaces and asks to show some properties. Exercise B.3 defines order unit spaces and asks to show that a defined quantity k.k is a norm on such spaces. The remaining exercises ask to show various properties related to order unit spaces.

Uploaded by

Sadia ahmad
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
28 views1 page

Exercise Set B: Some Exercises For TMA4230 Functional Analysis

The document contains 7 exercises related to functional analysis. Exercise B.1 asks to show that a metric dp defined on Lp spaces is indeed a metric and that Lp is complete under this metric. Exercise B.2 defines ordered vector spaces and asks to show some properties. Exercise B.3 defines order unit spaces and asks to show that a defined quantity k.k is a norm on such spaces. The remaining exercises ask to show various properties related to order unit spaces.

Uploaded by

Sadia ahmad
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/ 1

Exercise set B

Some exercises for TMA4230 Functional analysis


2005–03–18

Exercise B.1. Let 0 < p < 1 be given. Consider Lp = Lp [0, 1]: The space of Lp functions on the unit interval
with Lebesgue measure. The Lp “norm” is not really a norm in this case, since it fails to satisfy the triangle
inequality. Instead, we can make a metric dp on Lp by
Z 1
p
dp (u, v) = ku − vkp = |u − v|p dx.
0

Show that dp is a metric. Hint: It is enough to show |u + v|p ≤ |u|p + |v p | and then integrate this inequality.
Since |u+v| ≤ |u|+|v|, it is sufficient to show the inequality when u, v ≥ 0. That is, show that (u+v)p ≤ up +v p
for u, v ≥ 0. This is an equality for v = 0. Differentiate wrt v.
It can be shown that Lp is complete in this metric, and the topology induced by this metric makes Lp into
a topological vector space as well. (You are not expected to show this, but you are welcome to do it anyway.)
The main purpose of this problem is to show that the only continous linear functional on Lp is the zero
functional. (Why does this prove that Lp is not locally convex?)
First, show that a linear functional f on Lp is continuous if and only if it is bounded, in the sense that
supkukp =1 |f (u)| ≤ ∞. Next, note that because of this, if there exists a nonzero continuous linear functional
on Lp , there is one such functional for which
sup |f (u)| = 1.
kukp =1

Pick any u ∈ Lp with kukp = 1. Now, if you can write u = u1 + u2 where u1 u2 = 0, then kukpp = ku1 kpp + ku2 kpp .
(Why?) Explain how to take advantage of the non-discrete nature of the Lebesgue measure to split up u in
this way so that ku1 kpp = ku2 kpp = 21 kukpp = 21 . Conclude that |f (u)| ≤ 21−1/p , and obtain a contradiction.

Exercise B.2. A ordered vector space is a real vector space with a partial ordering ≤ so that: Whenever
x ≤ y then cx ≤ cy for any real number c ≥ 0; and x + v ≤ y + v for any v ∈ X.
If X is an ordered vector space and X + = {x ∈ X : x ≥ 0}, show that 0 ∈ X + , cx ∈ X + whenever c ≥ 0 is
real and x ∈ X + , X + ∩ (−X + ) = {0}, x + y ∈ X + whenever x, y ∈ X + , and X + is convex. (Also show that
the two final conditions are equivalent, given the first three.) We say that X + is a proper convex cone in X.
Show also that if X + is a proper convex cone in X, we can make X into an ordered vector space by saying
x ≤ y ⇔ y − x ∈ X +.
Some examples of ordered vector spaces include: All sequence spaces and all function spaces (with real-valued sequences
and functions), and the space of bounded Hermitian operators on a Hilbert space, where S ≤ T if and only if T − S
is nonnegative definite. For examples of order unit spaces, as will be defined next, let e be constant function on any
function space consisting of bounded functions, or the identity operator in the space of bounded Hermitian operators.

Exercise B.3. An order unit space is an ordered vector space with an order unit e ∈ X + which is supposed
to satisfy two requirements: First, if x ∈ X there exists some real number c with −ce ≤ x ≤ ce, and second,
if x ∈ X and cx ≤ e for all c ≥ 0 then x ≤ 0.
From now on through the remaining problems, let X be an order unit space with order unit e.
Show that if we define
kxk = inf{c : − ce ≤ x ≤ ce} (x ∈ X)
then k·k is in fact a norm, and −kxke ≤ x ≤ kxke.

Exercise B.4. Define the state space S of X to be the set of linear functionals
S = {f ∈ X ∗ : f (x) ≥ 0 for all x ≥ 0, and f (e) = 1}.
Show that kf k = 1 for all f ∈ S, and S is compact in the weak* topology.

Exercise B.5. Show that if x ∈ X then x ≥ 0 if and only if f (x) ≥ 0 for all f ∈ S.
Hint: If x 6≥ 0, use Hahn–Banach to separate x from X + .

Exercise B.6. Show that kxk = sup{|f (x)| : f ∈ S}.


Hint: Assume that (kxk − )e − x 6≥ 0 for every  > 0. (If not, replace x by −x.) Apply the previous exercise.

Exercise B.7. Show that the unit ball of X ∗ is X1∗ = co(S ∪ −S).
Hint: If not, separate some point in X1∗ from co(S ∪ −S) using the Hahn–Banach theorem, and obtain a
contradiction.

You might also like