0% found this document useful (0 votes)
57 views15 pages

ExercisesEnglish 1

This document contains 10 exercises summarizing key concepts in distributions: 1) Exercise 1 covers properties of continuous functions including differentiability using the mean value theorem. 2) Exercise 2 examines properties that ensure a function is smooth, including finding functions with specified support properties. 3) Exercise 3 investigates properties of sequences of smooth functions, including conditions for a sequence to converge uniformly. 4) The exercises cover topics like characterizing closed, bounded sets in function spaces; conditions for uniform convergence; and characterizing convergence in spaces of smooth functions with seminorms.
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)
57 views15 pages

ExercisesEnglish 1

This document contains 10 exercises summarizing key concepts in distributions: 1) Exercise 1 covers properties of continuous functions including differentiability using the mean value theorem. 2) Exercise 2 examines properties that ensure a function is smooth, including finding functions with specified support properties. 3) Exercise 3 investigates properties of sequences of smooth functions, including conditions for a sequence to converge uniformly. 4) The exercises cover topics like characterizing closed, bounded sets in function spaces; conditions for uniform convergence; and characterizing convergence in spaces of smooth functions with seminorms.
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/ 15

Exercises Master 1: Distributions

Exercise 1. 1. Let g : [0, +∞[→ R be a continuous function, dierentiable on ]0, +∞[ .


Assume the existence of limy→0+ g ′ (y) = ℓ . With the help of the mean value theorem
on [z, y] with 0 < z < y , prove that g is C 1 sur [0, +∞[ and g ′ (0) = ℓ .

2. Let h be an odd C ∞ function on R. Prove the existence of a C∞ function h1 such


that h(y) = yh1 (y) avec h1 .

3. (a) Let f be an even real valued C ∞ function on R . Set for y ∈ [0, +∞[ , g(y) =

f ( y) . Prove that g is C 1 on [0, +∞[ with the help of the rst question.
(b) Prove that g is C∞ .

4. Let f be a C∞ function on R . Prove the existence of two C∞ functions on R , g0 , g1 ,


such that for all x ∈ R , f (x) = g0 (x2 ) + xg1 (x2 ) .

Exercise 2. 1. Let f :R→R be the map dened by f (t) = 0 if t≤0 and f (t) = e−1/t
else. Prove that f is C ∞ .

2. Find a C ∞ ϕ : Rn → R+ function such that ϕ(0) > 0 of which the support is the
closed ball B(0, 1) with center 0 and radius 1 , and another ψ of which the support is
n
included in B(x0 , δ), (x0 ∈ R , δ > 0) and ψ(x0 ) > 0.

3. Let f n
be a continuous map from the open set Ω ⊂ R to R . Assume that for all

R
function ϕ ∈ C0 (Ω; R), f (x)ϕ(x)dx = 0 . Prove that f is the null function.

4. Same question if f ∈ L1loc (Ω, dx; R) .

Exercise 3. Let ϕ : R → R be C ∞ function with support in [−2, 2], and equal to 1 on


n (k)
[−1, 1] . Set for n in N, ϕn (y) = yn! ϕ(y) and An = supk≤n supy∈R |ϕn (y)|.

1. Let (cn )n
be any real sequence. Prove that there exists a real sequence (tn )n in [1, +∞[
P+∞ dk −n
such that for all xed integer k the series n=0 cn dxk ϕn (tn x)tn converges uniformly.
P+∞ xn
2. Prove that the function f (x) = n=0 cn ϕ(tn x) n! is C∞ with n, f (n) (0) = cn (Borel
Theorem).

Exercise 4. For A = {f ∈ C 0 (R) |f (x)| ≤ 1 for all x ∈ Ω} :
0
1. Prove that A is a closed bounded set in C (R) .
2. For fn (x) = sin(nx) , n ∈ N , prove that fn ∈ A for all n ∈ N , but that the sequence
{fn } has no convergent subsequence in C 0 (Ω) . Remember that if (gn )n converges unifomly
on the compact set K to the limit g , then for all sequence (xn )n in K converging to a limit
ℓ , gn (xn ) converge to g(ℓ) .

Exercise 5. In this exercise the convergence of sequences in C0∞ (Ω) with seminorms de-
ned in the course, is characterized. A exhaustive sequence of compacts subsets of
S Ω will
be denoted by (Kn )n with Kn ⊂ K̊n+1 and Kn = Ω .

1. Let (uk )k be a sequence in C0∞ (Ω) . Assume that there exists a compact subset K
of Ω (independent of k ) such that for all k , supp uk is contained in K and such
that (uk )k converges uniformly to zero as well as all its derivatives. Prove that for
any xed continuous function ϵ from Ω to R , positive on Ω , the seminorm pϵ (uk ) =
supx∈Ω,|α|≤ϵ(x)−1 |∂ α uk (x)|ϵ(x)−1 converges to 0 as k goes to innity.

1
2. (a) For all n∈N let ϕn be a C 0 function, equal to 1 on Kn ,
K̊n+1 ,
with support in
with values in sequence of integers with nj+1 ≥ nj + 1 ,
[0, 1] . Let (nj )j∈N be a
and let (rj )j be a real positive decreasing sequence converging to 0 . Set a0 = r0
P+∞
and for j ≥ 1 , aj = rj−1 − rj . Prove that ϵ(x) = j=0 aj ϕnj (x) is a continuous
function on Ω , positive on Ω and such that ϵ(x) ≤ rj−1 on Knj+1 − Knj .

(b) Let (uk )k be a sequence of C0∞ (Ω) such that for all ϵ (pϵ (uk ))k con-
like above,
verges to 0. Assume that there is no compact set K
Ω containing supp uk for
in
all k. Prove the existence of a subsequence (ukj )j of(uk )k , a sequence (nj )j like
above and a sequence of points xj ∈ Knj+1 − Knj such that |ukj (xj )| = rj−1 > 0.
(c) Prove that (pϵ (uk ))k cannot converge to 0 and conclude.

Exercise 6. Let ϕ ∈ C0∞ (Rn ) , and h ∈ Rn − {0} . For t ∈ R − {0} set

ϕ(x + th) − ϕ(x)


ϕt (x) =
t
1. Prove that ϕt ∈ C0∞ (Rn ) .

2. Prove that as t → 0 , ϕt converges in C0∞ to a limit to be specied.

Exercise 7. Let f :R→R be a locally bounded function which satises:

(H)(∀N ∈ N, ∃CN > 0 , ∀z ∈ [0, 1] , |f (z)| ≤ CN z N )



and f R− ≡ 0 .
 
1. Take φ ∈ C0∞ (R) and set F (z, x) = f (z)φ −1 + xz . Prove that for all R > 0 , and

all k ∈ N, there exists an integrable function z → Hk,R (z) such that for all real z and
k
all x ∈ [−R, R] , |∂x F (z, x)| ≤ Hk,R (z) .
Z
2. Set G(x) = f (x(1 + y))φ(y) dy . Prove that G is a C ∞ function on R .

3. Assume that f ≥0 is increasing. Check that φ can be chosen such that G satises
f (x) ≤ G(x) for all x and G(x) = 0 when x < 0.

4. Prove that for all f which fullls the hypothesis (H) , there exists a C∞ function,
vanishing on R− and such that f (x) ≤ G(x) for all x ∈ R .

Exercise 8. Let θ ∈ C0∞ (Rn ) be such thatθ ≡ 1 in a neighborhood of 0 . Prove that for
β
any φ ∈ C0∞ (Rn ) and any m ∈ N there ∞ n
exists ψm ∈ C0 (R ) with ∂x ψm (0) = 0 for |β| ≤ m
such that
X ∂ α φ(0)
x
φ(x) = θ(x) + ψm (x) .
α!
|α|≤m

n
(resp. an open) subset of R with K ⊂ Ω . For
Exercise 9. Let
K (resp.
N
Ω) be a compact

ε > 0 set Kε = x ∈ R , d(x, K) ≤ ε and consider ε < εK,Ω , εK,ω > 0 small enough so
that Kε is a compact subset of Ω .

1. Construct a function θε = ϱε ∗ 1K2ε in C0∞ (Rn ; [0, 1]) such that supp θε ⊂ K3ε , θε ≡ 1
on Kε and supx∈Ω |∂xα θε (x)| ≤ Cα ε−α .

2. Assume that φ ∈ C0∞ (Ω) vanishes at order k − 1 on the compact set K (namely
∂xα φ(x) = 0 for x ∈ K and |α| ≤ k − 1).

(a) Prove that for any β , |β| ≤ k , supx∈K2ε |∂xβ φ(x)| ≤ Ck εk−|β| sup|α|≤k,x∈Ω |∂xα φ(x)|

2
(b) Prove that for any ℓ ∈ {0, . . . , k} ,

sup |∂xα (θε φ)(x)| ≤ Ck εk−ℓ sup |∂xα φ(x)|


|α|≤ℓ,x∈Ω |α|≤k,x∈Ω

Exercise 10. We recall that the set S(Rn ) is the Schwartz space of rapidly decaying func-
tion  
n ∞ n n
S(R ) = f ∈ C (R ), ∀α, β ∈ N , sup |xα ∂xβ f (x)| < +∞
x

and that the space of tempered distributions S ′ (Rn ) is its dual


!
T ∈ S ′ (Rn ) ⇔ ∃N ∈ N, ∃CN > 0, |⟨T , ∀φ ∈ S(Rn ) , φ⟩| ≤ CN |xα ∂xβ φ(x)|

sup .
|α+β|≤N,x∈Rn

1. Check that S ′ (Rn ) ⊂ D′ (Rn ) , that an element of S ′ (Rn ) must have a nite order.
Give an example of a C
∞ function which belongs to D ′ (Rn ) and ′ n
not in S (R ) .

2. In all the following exercises on will discuss when the identities or convergences in
D′ (Rn ) hold in S ′ (Rn ) .

Exercise 11.
Z 1. Let T be the linear form on C0∞ (R) given for all φ in C0∞ (R) by

⟨T, φ⟩ = φ(x + x2 ) dx . Prove that T is a distribution on R and specify its order.


R
Z
2
2. Same question for ⟨T, φ⟩ = φ′ (x)ex dx .
R
Z +∞
3. (a) Prove that T dened by ⟨T, φ⟩ = φ′ (x) log x dx is a distribution on R.
0
(b) Determine the order of T . (Hint : Take χ in C0∞ (R)with values in [0, 1], equal to
1 on [−1, 1] and supported in [−2, 2] , and consider for n ∈ N∗ , the test functions
φn (x) = χ(x)(1 − χ)(nx)).
Z +∞  dx
4. Let 1 < γ < 2. Prove that T dened by ⟨T, φ⟩ = φ(x) − φ(0) γ is a
0 x
distribution or order 1 on R. (Hint : use the same test functions φn (x) as in the
previous question.)

Exercise 12. 1. Let T be the linear form on C0∞ (R2 ) given by


Z
φ(t−2 , sin t) − φ(0, sin t) dt.

⟨T, φ⟩ =
R

(a) Prove that T is a distribution on R2 .


(b) Prove that T is of order 1. Use the last question of the previous exercise.

2. Same questions for


Z  ∂φ ∂φ 
⟨T, φ⟩ = (x2 + y 2 )−α x +y dxdy
x2 +y 2 ≤1 ∂x ∂y
1 3
when 2 <α< 2 . (Hint : use polar coordinates).

Exercise 13. We recall that when f : ]0, a] → R is continuous for a > 0 , the in-
Ra
tegral 0 f (x) dx is said semi-convergent if limx→0+ F (x) exists and is nite, F (x) =
Ra
− x f (y) dy .

3
1. For φ in C0∞ (R). Prove that
Z a
déf
⟨T, φ⟩ = lim f (x)φ(x) dx
ϵ→0+ ϵ
Ra
exists if 0 f (x) dx is semi-convergent. (Hint: Use an integration by parts).

2. Prove that T denes a distribution of order ≤ 1.


3. (a) Let χ be a step function , 0 ≤ χ ≤ 1 , supported in [c, a] with 0 < c < a . Check
that there exists a sequence (χn )n of continuous functions supported in [c, a] ,
n→∞
such that 0 ≤ χn ≤ 1 and χn (x) → χ(x) for a.e. x .
(b) Assume that T is of order 0. Prove that there exists C > 0 such that for all
c ∈]0, a], and all χ like before, ⟨T, χn ⟩ ≤ C .
Ra
(c) With the same assumptions, prove that 0 |f (x)| dx < +∞ .
Z +∞ √
2
Exercise 14. We recall e−x dx = π.
−∞

1. For φ in C0∞ (R) , prove that


Z
1 x2
⟨T, φ⟩ = lim √ e− ϵ φ(x) dx
ϵ→0+ πϵ R

denes a distribution that will be computed.

2. Let f :R→R be a C1 function such thatf (0) = 0, f ′ (0) ̸= 0.


(a) Set for δ ≥ 0 , ϵ > 0 , φ ∈ C0∞ (R) ,
Z
1 (f (x))2
Iδ,ϵ = √ e− ϵ φ(x) dx.
πϵ |f (x)|>δ
Prove that for δ > 0 , limϵ→0+ Iδ,ϵ = 0 .
(b) Assume f (x) ̸= 0 if x ̸= 0 . Prove that ⟨T, φ⟩ = limϵ→0+ I0,ϵ denes a distribution
which will be computed (Hint: x → y = f (x) is a change of variable around 0 ).
3. Assume that f is C1 function with simple zeroes (i.e. f (x) = 0 ⇒ f ′ (x) ̸= 0).
(a) Prove that f has a nite number of zeroes on any compact subset of R.
(f (x))2
− ϵ
R
(b) Prove that ⟨T, φ⟩ = limϵ→0+
Re φ(x) dx denes a distribution which will
be computed in terms of the zeroes of f.
Exercise 15. 1. For φ ∈ C0∞ (R) , prove that

n
X (−1)p−1 1 
lim √ φ
n→+∞ p p
p=1

1 1
 
exists and denes a distribution on R. (Hint: Write φ p = φ(0) + [φ p − φ(0)]).
2. Determine the order and the support of this distribution.

Exercise 16. Set Ω = R − {0}.


1. Let f be a locally integrable function on Ω,
such that there exists m in N∗ such that
for all x in Ω with |x| < 1 , |f (x)| < |x| −m ∞
. If θ ∈ C0 (R) , we set for φ ∈ C0∞ (R),
k ∈ N∗ ,
1
xk
Z Z
⟨Tθ , φ⟩ = θ(x)f (x) φ(k) (tx)(1 − t)k−1 dtdx.
R (k − 1)! 0
Prove that for a large enough integer k , Tθ is a distribution in R.

4
2. Prove that T is a distribution on R such that T |ω = f .

3. Let S
be the distribution on Ω dened by the locally integrable function exp(1/x2 ). Is

there T ∈ D (R) such that T |Ω = S ? Explain.

Exercise 17. 1. Let φ belong to C0∞ (R) . Compute

1−x x2 − ϵ 2 i
Z h
1+x
lim + − 2 φ(x) dx.
ϵ→0+ (x + iϵ)2 (x − iϵ)2 (x2 + ϵ2 )2

2. Prove that for all φC0∞ (R) , the limit


in

ex e−x x2 − ϵ2 i
Z h
L(φ) = lim + − 2 φ(x) dx
ϵ→0 (x + iϵ)2 (x − iϵ)2 (x2 + ϵ2 )2

exists and thatφ → L(φ) is a distribution which will be computed from the function
ch x−1
and another explicit distribution.
x2

Exercise 18. 1. Let λ ∈ C , ℜλ > 0 . Prove that the expression

Z +∞
(1) ⟨Tλ , ϕ⟩ = xλ−1 ϕ(x) dx
0

denes a distribution Tλ on R of which you will specify the order and the support.

2. For all k ∈ N, check

k−1 Z 1 Z ∞
ϕ(j) (0)
(2)
X
λ+k−1
⟨Tλ , ϕ⟩ = + x ψk (x) dx + xλ−1 ϕ(x) dx ,
j!(j + λ) 0 1
j=0

X ϕ(j) (0)  k−1


1
where ψk (x) = k ϕ(x) − xj .
x j!
j=0

3. Prove that (2) denes a distribution Tλ for all λ ∈ C such that ℜλ > −k and
λ ̸= 0, −1, . . . , −k + 1. Determine the order of Tλ with respect to λ.

4. For all ϕ ∈ C0∞ (R) , prove that the map λ 7→ ⟨Tλ , ϕ⟩ dened by (1) has an analytic
extension into a meromorphic function on whole C . Study the poles and residues of
this analytic extension.

Exercise 19. Let T ∈ D′ (Ω) , Ω open subset of Rn , have a compact support.

1. Explain why T has a nite order N.

2. Prove that for any φ ∈ C0∞ (Ω) which vanishes at order N +1 on supp T , ⟨T , φ⟩ = 0 .
(Hint: use ⟨T, θε φ⟩ where θε is the cut-o introduced in Exercise 9.)
(α)
Exercise 20. Remember ⟨δ0 , φ⟩ = (−1)α ∂xα φ(0) . We prove here that any distribution
(α)
T ∈ D′ (Rn ) such that supp T = {0} is a linear combination of the δ0 's.

1. Apply Exercise 19 in order to say that ⟨T , φ⟩ = 0 for any φ ∈ C0∞ (Rn ) vanishing at
order N + 1, N ∈ N.

2. Apply Exercise 8 with m=N +1 and conclude.

Exercise 21. Consider the function f :R→R given by f (x) = 1Q∩R− (x) + 1(R\Q)∩R+ (x) .
What is its distributional derivative ?

5
Exercise 22. 1. Prove that the integral

x2 − y 2
ZZ
A= dxdy
|x|>1,|y|<1 (x2 + y 2 )2
converges.

2. Let (x, y) → φ(x, y) be a C0∞ (R2 ) function and let θ be a C0∞ (R) function such that
ZZ 2 2
x −y
φ(x, y)θ(x)θ(y) = φ(x, y) . Set Iϵ1 = φ(x, y) dxdy . Prove that
|x|>ϵ (x2 + y 2 )2
x2 − y 2 
ZZ
1
Iϵ1

= 2 2 2
1|x|>ϵ φ(x, y) − 1|y|>ϵ φ(y, x) dxdy.
2 (x + y )
Deduce

x2 − y 2 h
ZZ
φ(0, 0) i
Iϵ1 = 1 − 1 |x|<ϵ θ(x)θ(y) dxdy + Rϵ
2 (x2 + y 2 )2 |x|>ϵ
|y|<ϵ |y|>ϵ

with
x2 − y 2 
ZZ
1 
lim Rϵ = 2 2 2
φ(x, y) − φ(y, x) dxdy.
ϵ→0 2 (x + y )
3. Prove that limϵ→0+ Iϵ1 exists and denes a distribution on R2 .
x2 − y 2
ZZ
4.
2
Set Iϵ = φ(x, y) dxdy . Determine explicitely the distribution limϵ→0+ (Iϵ1 −
(x 2 + y 2 )2
|y|>ϵ
I2ϵ ) in terms of the constant A of the rst question.
Exercise 23. Let S be a distribution on R. Let θ in C0∞ (R) be equal to 1 around 0 and
∗ ′ ∞
let k in N . For S ∈ D (R) , dene for all φ ∈ C0 (R)

φ(j) (0) j
φ(x) − k−1
P
j=0 j! x θ(x)
⟨T, φ⟩ = ⟨S, k
⟩.
x
1. Prove that TS,k is a distribution on R . (Use Exercise 8)

2. Prove that TS,k satises xk TS,k = S .


3. Determine all the distributions solutions to xk T = S . (Use Exercise 20.)

Exercise 24. 1. Let T T = (H(x) log x)′ , where H


be the distribution is the Heaviside
function H(x) = 1[0,+∞[ (x) . Prove that T is given by
Z 1 Z +∞
φ(x) − φ(0) φ(x)
⟨T, φ⟩ = dx + dx.
0 x 1 x

2. Compute the distributional derivative of log|x| .


Exercise 25. For all λ in C − {−k; k ∈ N} , call xλ−1
+ the distribution Tλ dened in the
Exercise 18.

1. Compute the derivative of xλ−1


+ .

2. Deduce in particular that for −k < ℜλ < −k + 1 , k ∈ N ,


+∞
xλ+k−1
Z
λ−1
⟨x+ , φ⟩ = lim (−1)k φ(k) (x) dx
ε→0+ ε (λ) · · · (λ + k − 1)
+∞ k−1
φ(j) (ε)
Z X
λ+k−1
= lim φ(x)x dx + (−1)j ελ+j
ε→0+ ε (λ) . . . (λ + j)
j=0

from which comes the name nite part and the other used notation Pf (xa+ ) .

6
3. Compute xxλ−1
+ .

4. When −k < ℜλ < −k + 1 does the relation xk xλ−1


+ = xk+λ−1
+ ∈ L1loc (R) suce for
λ−1
the determination of x+ .

Exercise 26. 1. Prove that the nite part of 1[0,+∞[ x1 , 1


denoted by x (or
+
Pf( x1+ )) and
dened by
Z +∞
1 φ(x)
⟨ , φ⟩ = lim + φ(0) ln(ε)
x+
ε ε→0+ x

belongs to D (R) . Check that it is the distribution derivative of 1]0,+∞[ log(x) .

2. Prove that x−1


+ coincides with the distribution T of Exercise 23.

1
3. Check that the principal value of x , vp( x1 ) dened by
Z
1 φ(x)
⟨vp( ) , φ⟩ = lim dx
x ε→0 R\[−ε,ε] x

belongs to D (R) . Check that it is the derivative of log |x| .
1
4. In view of Exercise 20 does the notation x make sense in D′ (R) ?
1
5. Check that x±iε , ε > 0 has a limit as ε → 0+ in D′ (R) which will be denoted by
1
x±i0 . Check  
1 1
= vp ∓ iπδ0 .
x ± i0 x
Exercise 27. Let θ : R → R be an increasing C ∞ function, equal to 0 on ] − ∞, 0[ and to
d
1 on [1, +∞[ . Besides let f : R → R be a C
∞ function such that Ω = {x ∈ Rd ; f (x) > 0}
d
is not empty. Assume moreover that for all point in Σ = {x ∈ R ; f (x) = 0}, ∇f (x) ̸= 0 .
∇f (x)
We then set N (x) the outer normal vector to Ω at the point x in Σ i.e. the vector −
|∇f (x)| .
Finally let χ be the characteristic function of Ω .
Pd ∂f ∂χ
1. Prove that S = i=1 ∂xi ∂xi is a distribution supported in Σ .

2. For n in N∗ , set θn (t) = θ(nt) . Prove that θn ◦ f ⇀ χ in D′ (Rd ) . Deduce that S is


a non negative measure.

3. (a) Prove for all 1 ≤ j ≤ d,


d
X ∂j f ∂i f ∂ ∂
θn (f (x)) = θn (f (x)).
|∇f |2 ∂xi ∂xj
i=1

(b) Set σ = |∇f |−1 S . Prove for all j , −Nj σ = ∂


∂xj χ.

4. Let F = (f1 , . . . , fd ) ∈ C0∞ (Rd , Rd ) . Check the Gauss-Green formula

⟨χ, div F ⟩ = ⟨σ, F · N ⟩.

Exercise 28. Consider the locally integrable function on R2 , E(t, x) = 1|x|<t . Compute
 
∂2 ∂2
in D′ (R2 ) , ∂t2
− ∂x2
E(t, x) .

Exercise 29. 1. Let x → A(x) be a C∞


Mn (C) (the space of n × n
map from R to
complex matrices). Remember that the Wronskian matrix x → W (x) is the unique
C ∞ function from R to Mn (C) solution to W ′ (x) = A(x)W (x) , W ′ (0) = In .
Prove that if S is a Cn -valued distribution solution to S ′ = AS , then S = W (x)Y0
for some vector Y0 in Cn . (In particular, S is a C ∞ function).

7
2. Let x → a(x), b(x) be two C ∞ C-valued functions. Let T be a C-valued distribution
on R solution to T ′′ + a(x)T ′ + b(x)T = 0 . Prove that T is a C ∞ function.

3. Let f± be two C∞ functions on R solutions to the equation f±′′ +a(x)f±′ +b(x)f± = 0 ,


with the boundary
′ ′
conditions f+ (0) − f− (0) = 0, f+ (0) − f− (0) = 1 . Set F =
f+ 1x>0 + f− 1x<0 . ′′ ′
Compute the distribution F + a(x)F + b(x)F .

4. Determine all the distributional solutions to the equation T ′′ − T ′ − 2T = δ .

Exercise 30.
Pn−1 1. Compute the limit of the sequence of distributions given by Tn =
1
n p=0 δp .
n

2. Compute the limit of the sequence of distributions given by Tn = n δ 1 − δ− 1 .
n n

Exercise 31. Let (an )n be a complex sequence.


P+∞ ′
1. Prove that n=1 an δ 1 converges in D (]0, +∞[).
n
P+∞
2. Assume that the above series converges in D′ (R) and prove that n=1 an converges.
P+∞
⟨ +∞
P
3. Reciprocally assume now that n=1 an converges and prove that n=1 an δ 1 , φ⟩ con-
n
verges for all φ in C0∞ (R) vanishing at second order at 0.

4. We recall Abel's criterion : if (an )n is a sequence with uniformaly bounded partial


PN P
sums AN =
P n=1 an and if (bn )n is a sequence such that n |bn+1 − bn | < +∞ then
a b
n n n converges.
P+∞
Prove that the assumption of the previous question implies that n=1 an δ n
1 converges

in D (R).

Exercise 32. 1. Compute the limit

einx
Z
lim lim dx
n→+∞ r→0+ r<|x|<1 x
sin x
R
in terms of I= R x dx (Recall why this integral semi-convergent).

inx ψ(x) dx
R
2. Prove that for all R > 0, limn→+∞ |x|<R e = 0 for every function ψ in
C ∞ (R) .
 
3. Compute the distributional limit of einx vp x1 in terms of I.

Exercise 33. 1. Let P (x, y) be a degree d polynomial of two variables. Prove that if
Z 2π
k∈Z satises |k| > d , P (r cos θ, r sin θ)e−ikθ dθ = 0 for all r > 0 .
0

2. Let m belong to N∗ . Consider the sequence of distributions (Tnm )n∈N∗ given by Tnm =
1x2 +y2 >n−2
. Prove that as n goes to innity, the sequence converges in D′ (R2 ) to
(x + iy)m
a distribution of which the order will be specied. (Hint : If φ ∈ C0∞ (R2 ) , write
φ(x, y) = φ(x, y)ψ(x, y) for a suitable function ψ and use the Taylor expansion of φ
at the origin).

Exercise 34. Let (an )n∈Z be a complex sequence. Assume that there exists C > 0 and
α≥0 such that
|an | ≤ C(1 + |n|)α , for all n ∈ N..

8
1. Consider for N ∈N the C∞ function

N
X
TN (x) = an einx , x ∈ R.
n=−N
X
Prove that (TN )N ∈N converges in D′ (R) as N → +∞ . Let an einx denote its
n∈Z
limit.

2. Consider the 2π -periodic function f :R→R given on [0, 2π[ by

π−x
f (x) = , x ∈ [0, 2π[ .
2
With its Fourier series expansion prove

N
X sin(nx) π−x
lim = lin L2 ([0, 2π]) .
N →+∞ n 2
n=1


X sin(nx)
Deduce = f (x) in D′ (R).
n
n=1

3. By computing in two dierent ways the distributional derivative of f prove the equality

in D (R)
X X
einx = 2π δ2πn
n∈Z n∈Z

where δz is the Dirac mass at z given by δz : ϕ 7→ ϕ(z) .

ϕ ∈ C0∞ (R) . ϕ̂(n) = R e−inx ϕ(x)dx


R
4. Let For n ∈ Z, write and prove

X X
ϕ̂(n) = 2π ϕ(2πn) Poisson formula.
n∈Z n∈Z

Exercise 35. 1. Let Tn = n2 xeinx . Prove that(Tn )n converges in D′ (R) to a limit to


be specied.

2. Same question with Sn = n2 |x|einx . (Compute ⟨Sn , φ⟩ for a test function φ and use
einx = −in−1 ∂x einx ).

Exercise 36. Let x → A(x) be a C∞ function on R, such that for all integer k there exists
Ck such that

(1) |∂ k A(x)| ≤ Ck (1 + |x|)−k .


Z R
sin x
Set IR (A) = A(x) dx.
−R x
1. Prove that limR→+∞ IR (A) exists and call this limit I(A) .

2. Let (An )n be a sequence of functions such that for all n, all k , An fulls (1) with
constants Ck independent of n and such that there exists a constant B with

∀x ∈ R, lim An (x) = B, ∀x ∈ R, ∀k ∈ N∗ , lim ∂xk An (x) = 0.


n→+∞ n→+∞

Prove lim I(An ) = BI(1) .


n→+∞

9
sin(nx)
3. Determine the limit in D′ (R) of the sequence Tn = x . (Consider ⟨Tn , φ⟩ for a
test function φ and apply the previous results.)

Exercise 37. Use the notation xλ−1


+ introduced in Exercises 18 and 25 for λ ∈ C\
{−k, k ∈ N} .
Pk−1(−1)j (j)
1. Prove that Sλ = xλ−1
+ −

j=0 j!(j+λ) δ0 denes a D (R)-valued holomorphic function
(which means that for any test function λ 7→ ⟨Sλ , φ⟩ is holomorphic) on {λ ∈ C , ℜλ > −k} .
−k′ −1
We call x+ = S−k′ for k ′ {0, . . . , k − 1} .

2. Check that x−1


+ coincide with its denitions in Exercises 23 and 26.

λ−1
3. For k ∈ N compute ∂x x+ + (k + 1)xλ−2
+ rstly for λ ∈ C \ (−N) and then take the

limit in D (R) as λ → −k .

4. Deduce
(−1)k+1 (k+1)
∂x x−k−1
+
−k−2
= −(k + 1)x+ + δ .
(k + 1)! 0
−k−1
5. Give a formula formulas x+ for k ∈ N in the form of Exercise 23 or 26.

k (k)
P∞ ′
Exercise 38. Prove that the series in D (R) but does not con-
k=1 (−1) δk converges

P∞ −k
verge in S (R) . (For the second statement use the test function f (x) = k=1 3 (x −
k)k χ(x − k) with χ ∈ C0∞ (R) and χ ≡ 1 around 0 .)
Exercise 39. Remember that the space Cc0 (R) of compactly supported continuous functions
on R
p
is dense in L (R) when 1 ≤R p < +∞ . Let ρ be a C ∞ function on R, with values

in [0, +∞[ , supp ρ ⊂ [−1, 1] and R ρ(x) dx = 1 . Set for n ∈ N , ρn (x) = nρ(nx) . We
consider here 1 ≤ p < +∞ .

1. Check that when f ∈ Cc0 (R) , the sequence ρn ∗ f converges uniformly to f, and that
supp ρn ∗ f is contained in a xed compact domain.

2. Prove that ∥ρn ∗ f − f ∥Lp goes to 0 as n → ∞.

3. Recover the density of C0∞ (R) in Lp (R) .

4. Prove that C0∞ (R) is dense in E = {f ∈ Lp (R); (1 + |x|)f ′ ∈ Lp (R)} endowed with
the norm
∥f ∥E = ∥f ∥Lp + ∥(1 + |x|)f ′ ∥Lp .

Exercise 40. Let DR′ + (R) = {T ∈ D′ (R); supp T ⊂ [0, +∞[} . The delta function will be
denoted here by δ = δ0 .
1. Prove that DR′ + (R) endowed with the convolution product is a commutative unital
algebra (a convolution algebra).

2. An element T ∈ DR′ + (R) ′


is said invertible if there exist S ∈ DR (R) such that
+
S ∗ T = δ ; and we then write S = T −1 . Prove that T ∈ DR′ + (R) admits at most
′ ′
one inverse in DR (R) . When T1 , T2 ∈ DR (R) are invertible, prove that T1 ∗ T2
+ +

is invertible and specify its inverse. When T ∈ C0 (]0, +∞[) , check that T is not

invertible DR (R) .
+

3. Let T ∈ DR′ + (R) . Prove that the convolution equation S ∗ T = R admits a unique
′ ′
solution S ∈ DR (R) for all R ∈ DR (R) if and only if T is invertible.
+ +

4. For all λ ∈ C , prove that δ ′ − λδ is invertible and specify its inverse Eλ . Write the
′ ′ ′
solution S ∈ DR (R) to the dierential equation S − λS = R for all R ∈ DR (R).

+ +

10
5. For m ∈ N∗ , compute Eλ∗m = Eλ ∗ Eλ ∗ · · · ∗ Eλ (m factors).
P+∞ ′
Exercise 41. 1. Prove that the series k=0 δk ∗(δ1 −δ0 ) converges in D (R) and express
its sum.
P+∞ ′
2. Compute k=0 δk ∗ 1[0,1] .

3. Deduce from the previous question that there exist a distribution E such that supp E ⊂
[0, +∞] and E ∗ 1[0,1] = δ0 .

Exercise 42. For a complex number ℜλ > −1 , call xλ+ = 1[0,+∞[ xλ the locally
λ such that
λ µ λ+µ+1
integrable function. When ℜλ > −1 and ℜµ > −1 prove x+ ∗ x+ = B(λ, µ)x+ , for a
function B(λ, µ) which will be expressed as an integral in terms of λ, µ .

Exercise 43. 1. Prove that if φ belongs to C0∞ (R) and equals 1 around 0, and if F =
xφH (H(x) = 1R+ (x)), ′′ ∞
then F = δ0 + g with g ∈ C0 (R).

2. Let u ∈ D′ (R) . Assume that for all compactly supported continuous function F , u∗F
is a C
∞ function. Prove that u is C ∞ .

3. Same question if F is a compactly supported continuous function such that u∗F ∗F



is C .

x3 d4
Exercise 44. 1. Let E(x) = 12 sgn x . Check that E is an elementary solution to
dx4
.

2. Prove that there exists a unique a ∈]0, 1[ to be specied, such that for any integer k,
0 ≤ k ≤ 3, Z
y k 1[−1,1] (y) − δ−a (y) − δa (y) dy = 0.
 
R

3. Prove that there exists a compactly supported distribution u such that

∂ 4 u = 1[−1,1] − δ−a − δa .

4. Check that u is a C2 function.

Exercise 45. What is the problem in the equalities

0 = (1 ∗ δ0′ ) ∗ H = 1 ∗ (δ0′ ∗ H) = 1 ∗ δ0 = 1?

Exercise 46. Let C be a closed convex cone of Rn with vertex 0 . We call −C = {−x, x ∈ C} .

1. Prove that C +C ⊂C.

2. Prove that the map +:C ×C →C is proper i C ∩ (−C) = {0}

3. When C ∩ (−C) = {0} , explain why


′ (Rn ) = {u ∈ D ′ (Rn ) , supp u ⊂ C}
DC is a
convolution algebra.

In the following exercises with the Fourier transform, the duality products are left anti-
linear and right C-linear. The Fourier transform is dened as Fh u(ξ) = limε→0+ ⟨eiξ.x e−ε|x| , u⟩
2

and Fh−1 v(x) = limε→0+ ⟨e−ix.ξ e−ε|ξ| ⟩ dξ n which generalize


2

(2πh)

Z Z
iξ.x ix.ξ dξ
[Fh u](ξ) = e− h u(x) dx and [Fh−1 v](x) = e h v(ξ)
Rn Rn (2πh)n

when u, v ∈ L1 (Rn )

11
Exercise 47 (Vanishing moments)R. Consider a function f ∈ L1 (R, dx; C) such that for
all k∈ N , xk f ∈ L1 (R, dx; C) and xk f (x) dx = 0 .

1. Prove that if f is compactly supported then f = 0.

2. Without the assumption that f is compactly supported, does the conclusion still hold
true ? Hint: translate the assumption in terms of the Fourier transform.

Exercise 48 (Vanishing convolution product). Consider two functions f, g ∈ L1 (Rd , dx; C) .

1. Does f ∗g =0 imply f =0 and g=0 ?

2. Same question with f = g.

Exercise 49 (Fourier transform of a gaussian function.). The Siegel half-space is dened


by
Hd = τ ∈ Md (C) , t τ = τ , τ = τR + iτI , τI

positive denite .
t xτ x
We want to compute the Fourier transform of the gaussian function Gτ (x) = ei 2 :
Z t xτ x
Ĝτ (ξ) = e−iξ.x ei 2 dx
Rd

1/2 −1/2 −1/2


1. By writing −iτ = τI (Id − iτI p τR τI )τI p
check that τ is inversible and that a

determination of the square root det(−iτ ) or det(iτ −1 ) is specied by Re λ > 0
when Re λ > 0 .

R −λx2 √
2. With a contour deformation method in dimension 1 , prove R e 2 dx = √2π λ
for
Reλ > 0 .

3. Deduce
t ξτ −1 ξ
Ĝτ (ξ) = (2π)d/2 det(iτ −1 )1/2 e−i 2 .

Exercise 50 (Uncertainty principle). Use the Fourier transform


Z
ξ.x
Fh u(ξ) = e−i h f (x) dx .
Rd

1. Prove in S(Rd ) ,

h h h
[ ∂xj , xk ]f = [∂xj (xk f ) − xk ∂xj f ] = δjk f .
i i i

2. By applying the Cauchy-Schwarz inequality on ⟨f , [ hi ∂xj , xk ]f ⟩ deduce

Z 1/2 Z 1/2
d dξ h
∀f ∈ S(R ) , ∥f ∥L2 = 1 ξj2 |Fh f |2 (ξ) x2k |f |2 (x) dx ≥ δjk
Rd (2πh)d Rd 2

Exercise 51 (Paley-Wiener Theorem).f ∈ L1 (R, dx) .


Let
1
For f ∈ L (R, dx) such that suppf ⊂ R+ , prove that F1 f is holomorphic in C− =
{ξ ∈ C , Imz < 0} and continuous and bounded on C− .
Prove the converse statement.

Exercise 52. Compute the Fourier transform F1 u of the following tempered distributions
u:

1. eiax (a ∈ R) , cos(x) , cosk (x) (k ∈ N) .

12
sin(x)
2. x sin(x) , x .
(α) (α)
3. δ0 , δx0 in S ′ (Rd ) ;

4. H(x) = 1R+ (x) in S ′ (R) (start with H(x)e−εx and then take the limit as ε → 0+ ) ;

5. Pf( x1+ ) , 1
x±i0 and vp( x1 ) in S ′ (R) .
t xAx
Exercise 53. For A a symmetric real invertible matrix, verify that GA (x) = ei 2h belongs
′ d
to S (R ) and compute its Fourier transform Fh GA . Hint: use the result for A + iεId ,
ε > 0 , and then take the limit as ε → 0+ .

Exercise 54 (Heat equation). Consider the heat equation ∂t u = ∆u , u(t = 0) = u0 , in


S ′ (Rd , dx) .

1. With the help of the Fourier transform, prove that the heat equation admits a unique
1 ′ d
solution u ∈ C ([0, +∞[; S (R )) .

2. Prove that for t>0 and all s > 0 , u(t) ∈ H s (Rd ) .

3. Deduce that for all t > 0 , u(t) ∈ Cb∞ (Rd ) .

4. Prove that for t > 0 , u(t) = Gt ∗ u0 where Gt is a gaussian function to be specied.

5. Check that 1R+ (t)Gt is an elementary solution of ∂t + ∆ x on Rn+1 = Rt × Rnx .

Exercise 55 (Schrödinger equation). Consider the Schrödinger equation i∂t u = ∆u , u(t =


0) = u0 , in S ′ (Rd , dx) .

1. With the help of the Fourier transform, prove that the Schrödinger equation admits
1 ′
a unique solution u ∈ C (R; S (R)) .

2. Prove that for t∈R and all s ∈ R , u(t) ∈ H s (Rd ) as soon as u0 ∈ H s (Rd ) with
∥u(t)∥H s = ∥u0 ∥H s .

3. Give an elementary solution in D′ (R1+d ) for the Schrödinger operator i∂t − ∆x .

Exercise 56 (Elementary solution of ∆). Let |x| denote the euclidean norm on Rd , d ≥ 3,
d
Pd 2
1 X ∂2
|x| = j=1 xj
2
. Write ∆= .
j=1
∂x2j

1. Let F be a C∞ on ]0, +∞[ ,. Prove that

d−1 ′
∆ F (|x|) = F ′′ (|x|) +

F (|x|).
|x|

1
2. Set f (x) = |x|d−2
for x in Rd − {0} . Check that f denes a distribution on Rd and

that its restriction to R


d − {0} satises ∆f = 0.

3. Deduce that ∆f is a linear combination of derivatives of δ0 (see Exercise 20).

4. For φ ∈ C0∞ (Rd ) , prove


Z
1
⟨∆f, φ⟩ = lim Iϵ where Iϵ = ∆φ(x) dx .
ϵ→0+ |x|>ϵ |x|d−2

13
5. Use Green's formula in order to prove Iϵ = Jϵ − Kϵ with
Z Z
2−d ∂φ ∂
Jϵ = |x| (x) dσϵ , Kϵ = φ(x) |x|2−d dσϵ
|x|=ϵ ∂n |x|=ϵ ∂n

d−1 dω is the surface measure on the sphere with radius ∂


where dσϵ = ϵ ϵ and ∂n =
Pd xj ∂
− j=1 |x| ∂xj is the unit norma derivative.

6. Study the limits of Jϵ and Kϵ as ϵ → 0, according to µ(Sd−1 ) the total measure of


the d−1 unit sphere.

7. Compute ∆f in D′ (Rd ) and deduce an elementary solution of ∆ in dimension d ≥ 3.


1
Exercise 57. In dimension 2 consider the function f (x) = 2πlog |x| for x ̸= 0 and by
following the same method as in dimension d ≥ 3, prove that ∆f = δ0 in D′ (R2 ) .

Exercise 58 (Elementary solution of ∂z̄ ). In Cz ∼ R2(x,y) write ∂z̄ = 21 (∂x + i∂y ) . Check
1
that πz is an elementary solutionn of ∂z .

Exercise 59 (Elementary solution for the wave equation in R3 ). Call dω the surface
2
measure on S . For t∈R and all φ C0∞ (R3 ) set
Z
t
⟨Tt , φ⟩ = φ(tω) dω.
4π S2

1. Check that for all xed t ∈ R , Tt is an order 0 compactly supported distribution

2. Prove that (Tt )t belongs C ∞ (R, E ′ (R3 )) i.e. t → ⟨Tt , φ⟩ is a C∞ function for all
φ ∈ C ∞ (R3 ) .
(ℓ) (ℓ) dℓ
3. For any integer ℓ let Tt denote the distribution given by ⟨Tt , φ⟩ = dtℓ
⟨Tt , φ⟩ .
(1)
Prove T0 = 0, T0 = δ0 .

4. Let dσr = r2 dω
be the surface measure on the sphere with center 0 and radius r > 0
∂ ∞ 3
and ∂n the exterior normal unit vector to this sphere. Prove for all φ ∈ C0 (R ) the
equality Z Z
∂φ
∆φ(x) dx = (x)dσr .
|x|<r |x|=r ∂n

5. Deduce
Z r Z 3 Z
2 2
X ∂φ
∆φ(ρω)ρ dωdρ = r ωj (rω) dω.
0 S2 S2 ∂xj
j=1
Z
6. Check that if for t>0 we set F (t) = ∆φ(tω)t2 dω , then
S2

3 Z
X ∂φ X Z ∂2φ
F (t) = 2t ωj (tω) dω + t2 ωi ωj (tω) dω.
S2 ∂xi S2 ∂xi ∂xj
j=1 1≤i,j≤3

(2)
7. Prove Tt = ∆Tt .
Z +∞
8. If (t, x) → ψ(t, x) ∞ 4
belongs to C0 (R ) , we set ⟨E, ψ⟩ = ⟨Tt , ψ(t, ·)⟩ dt . Prove
0
that E is a distribution on R4 and specify its support.

14
9. Prove Z +∞
⟨E, ∂t2 ψ⟩ = ψ(0, 0) + ⟨Tt , ∆ψ⟩ dt.
0

10. Deduce that E is an elementary solution for ∂t2 − ∆.

Exercise 60 (Parametrix of an elliptic operator). Consider a dierential operator with


constant coecients

X 1
P (Dx ) = aα Dxα with D xk = ∂ xk ,
i
|α|=m

α which is an homogeneous polynomial with


P
and associate its symbol P (ξ) = |α|=m aα ξ
degree m.
The operator P (D) is said elliptic if (P (ξ) = 0) ⇔ (ξ = 0) .

1. Prove that there exists C≥1 such that for all ξ ∈ Rd , C −1 |ξ|m ≤ |P (ξ)| ≤ C|ξ|m .
∞ d
2. Take χ ∈ C0 (R ) such that χ ≡ 1 in a neighborhood of 0 and consider Q =
(1−χ)
F1−1 [ P (ξ)]F1 . Prove P ◦ Q = Q ◦ P = 1 + R where R is a continuous linear
′ d ∞ d
map from S (R ) into Cb (R ) .

3. InC = R2 check that ∂z̄ = 12 (∂x − i∂y ) is elliptic. Deduce that a tempered distribution
u ∈ S ′ (C ∼ R2 ; C) such that ∂z̄ u = 0 is an entire (holomorphic on the whole C)
function.

15

You might also like