0% found this document useful (0 votes)
164 views3 pages

Value Principal Distribution

This document contains solutions to homework problems from a class on distributions and Fourier analysis. 1) It shows that the principal value integral of 1/x exists as a distribution of order 1. 2) It proves that a specific summation defines a distribution on (0,∞) but not on all of R. 3) It demonstrates that a distribution u of order 0 if hu,φi ≥ 0 for all non-negative test functions φ.

Uploaded by

Claudia_orduzsto
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
164 views3 pages

Value Principal Distribution

This document contains solutions to homework problems from a class on distributions and Fourier analysis. 1) It shows that the principal value integral of 1/x exists as a distribution of order 1. 2) It proves that a specific summation defines a distribution on (0,∞) but not on all of R. 3) It demonstrates that a distribution u of order 0 if hu,φi ≥ 0 for all non-negative test functions φ.

Uploaded by

Claudia_orduzsto
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/ 3

MA54200 HOMEWORK

ASSIGNMENT 1: SOLUTIONS

1.3. Show that the principal value integral


Z Z − Z ∞ 
φ(x) φ(x) φ(x)
p.v. dx = lim dx + dx
x →0+ −∞ x  x
exists for all φ ∈ Cc∞ (R), and is a distribution. What is its order?
Solution. Note that we can write
Z ∞
φ(x) − φ(−x)
Z
φ(x)
p.v. dx = dx.
x 0 x
The latter integral exist, since
φ(x) − φ(−x)
≤ 2 sup |φ0 |.
x
Moreover, if supp φ ⊂ [−a, a], then
Z
p.v. φ(x) dx ≤ 2a sup |φ0 |,

x
which implies that the p.v. of 1/x is a distribution of order at most 1. Finally, the
order cannot be 0. Indeed, if 0 ≤ φ ≤ 1 is such that supp φ ⊂ [, 4] and φ = 1
on [2, 3] then Z
φ(x) 1
p.v. ≥ sup |φ |
x 4

1.5. Show that

X
hu, φi = ∂ k φ(1/k)
k=1
is a distribution on (0, ∞), but that there is no v ∈ D0 (R) whose restriction to
(0, ∞) is equal to u.
Solution. Let φ be such that supp φ ⊂ [1/N, N ]. Then
N
X N
X
k
hu, φi = ∂ φ(1/k) ≤ sup |∂ k φ|.
k=1 k=1 [1/N,N ]
Since the compacts [1/N, N ] exhaust (0, ∞), it follows that u is a distribution on
(0, ∞).
Assume now u = v (0,∞) for v ∈ D0 (R). Then there must exist N0 and C0 such
that
N0
X
|hv, φi| ≤ C0 sup |∂ k φ|, supp φ ⊂ [−1, 1].
k=1
So, if we take N > N0 , we will have
1
2 ASSIGNMENT 1: SOLUTIONS

N0
X
|∂ N φ(1/N )| = |hu, φi| = |hu, φi| ≤ C0 sup |∂ k φ|
k=1

if supp φ ⊂ ( N1+1 , N 1−1 ). N


This would imply that ∂ δ1/N is of order at most N0 < N
and consequently that ∂ N δ is of order at most N0 on a small interval (−, ). We
claim that this is impossible. Indeed, let ψ ∈ Cc∞ ((−, )) be such that ∂ N ψ(0) 6= 0.
Consider then the test functions
ψλ (x) = λN ψ(x/λ)
for small λ > 0. We have supp ψλ ⊂ (−λ, λ). Moreover,
∂ N ψλ (0) = ∂ N ψ(0)
and
∂ k ψλ = λN −k ∂ k ψ.
Thus, we would have an estimate
N0
X
|∂ N ψ(0)| ≤ C0 λN −N0 sup |∂ k ψ|
k=1

for any λ > 0. This is clearly a contradiction for small λ. 

1.6. Let u ∈ D0 (Rn ) have the property that hu, φi ≥ 0 for all real valued nonnega-
tive φ ∈ Cc∞ (Rn ). Show that u is of order 0.
Proof. Let K ⊂⊂ Rn and ψK ∈ Cc∞ (Rn ) be a nonnegative cut-off function such
that ψK = 1 on K. Then for real-valued test functions φ with supp φ ⊂ K we have

sup |φ| ψK (x) − φ(x) ≥ 0.
K

Hence

hu, sup |φ| ψK (x) − φ(x)i ≥ 0.
K
This implies
hu, φ(x)i ≤ hu, ψK i(sup |φ|).
K
For complex valued φ we obtain
|hu, φ(x)i| ≤ 2hu, ψK i(sup |φ|)
K

by considering the real and imaginary parts of φ. 

1.9. Let (ck )k∈Z be complex numbers which satisfy


|ck | ≤ C(1 + |k|)m , k ∈ Z,
for some constants C ≥ 0 and m. Show that
X
u= ck eikx
k∈Z
0
converges in D (R).
MA54200 HOMEWORK 3

Proof. Let φ ∈ Cc∞ (R). Then


(−1)m+2
Z Z
ikx
e φ(x)dx = eikx ∂ m+2 φ(x)dx, k 6= 0.
R (ik)m+2 R
Hence Z
C
e φ(x)dx ≤ 2 sup |∂ m+2 φ|, k 6= 0.
ikx

ck

R k
thus, the series hu, φi converges and defines a distribution of order at most m + 2.


You might also like