Scribd
Upload
164 views6 pages

Distribution: Assignment I

This document is an assignment submitted by five students to their professor. It contains solutions to five problems involving distributions and derivatives. The first problem shows that a given expression defines a distribution by proving linearity and continuity. The second problem proves an identity relating the derivative of the natural logarithm to the Cauchy principal value distribution. The third problem finds the nth derivative of the absolute value function. The fourth problem shows the convergence of a sequence of functions to the Dirac delta distribution. The fifth problem proves identities relating products of continuous functions and the Dirac delta and its derivatives.

Uploaded by

Miliyon Tilahun
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
164 views6 pages

Distribution: Assignment I

This document is an assignment submitted by five students to their professor. It contains solutions to five problems involving distributions and derivatives. The first problem shows that a given expression defines a distribution by proving linearity and continuity. The second problem proves an identity relating the derivative of the natural logarithm to the Cauchy principal value distribution. The third problem finds the nth derivative of the absolute value function. The fourth problem shows the convergence of a sequence of functions to the Dirac delta distribution. The fifth problem proves identities relating products of continuous functions and the Dirac delta and its derivatives.

Uploaded by

Miliyon Tilahun
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/ 6

Assignment I

DISTRIBUTION
November 30, 2016

Prepared By

Getachew Fetene(GSR/1411/08)
Miliyon Tilahun(GSR/1401/08)
Fatuma Zeynu(GSR/1398/08)
Adem Aman(GSR/1397/08)
Eshetu Dadi(GSR/1524/08)

X
Submitted to

Tamirat Temesgen (PhD)

Problems from worksheet


1. (Problem 1) Show that following expressions define a distribution.
hf , i =

m
X

n (0),

n=1

Solution. i) Linearity: Let 1 , 2 D and , R.


hf , 1 + 2 i =

m
X

(1 + 2 )n (0)

n=1

m
X

1n (0) + 2n (0)

n=1

m
X

1n (0) +

n=1

m
X

2n (0)

n=1

= hf , 1 i + hf , 2 i
ii) Continuity: Let k in D (). Then there is exists E compact in such that suppk E for
all k = 1, 2, . . .. Moreover
D k D
Now,


m
m
X
X

kn (0)
n (0)
|hf , k i hf , i| =
n=1

n=1


m 

X
=
kn (0) n (0) 0 as k
n=1

This is because uniform continuity implies pointwise continuity. Thus,
hf , k i = hf , i
Therefore,
hf , i =

m
X

n (0),

n=1

is indeed a distribution.

2. (Problem 3) Prove that

d
dx

ln x = P 1x .

Solution. Let the function f (x) = ln |x|, x R\{0}. This function is locally integrable, so f L1loc (R). To
the considered function we assign the linear functional defined by the formula
Z
hln |x|, i =
(x) ln |x|dx, D (R)
R

We shall show that the functional Tf = ln |x| is a distribution on D 0 (R). We have


"Z

Z
hln |x|, i =

(x) ln |x|dx = lim

+0

Z
(x) ln |x|dx +

#
(x) ln |x|dx = lim

+0

[(x) (x)] ln xdx

Considering supp() [a, a], a > 0, then we can write


Z
|(x) (x)|| ln x|dx
hln |x|, i lim
+0

Because

lim [(x) (x)] ln x = 2 lim [ 0 (x )x ln x] = 2 0 (0) lim x ln x = 0

x+0

x+0

we have

x+0

hln |x|, i 2 sup |x ln x| sup | 0 (x)|


x[0,a]

hence

hln |x|, i c sup | 0 (x)|,

x[0,a]

where c = 2 sup |x ln x|.

x[0,a]

x[0,a]

Consequently, the functional Tf is continuous and the linearity follows from the linearity of integrals.
Hence,Tf = ln |x| is a distribution on D 0 (R).
Now, we have
0

h(ln |x|) , i = h(ln |x|), i = 0 (x) ln |x|dx


R
Z
= lim
[ 0 (x) 0 (x)] ln xdx
+0
Z
(x) (x)
= lim [() ()] ln + lim
dx
0
+0
x
Because lim0 [() ()] ln = 0 and
Z
Z
(x) (x)
(x)
1
lim
dx = hp.v. , i = p.v.
dx
0
x
x
R x
we can write

1
1
h(ln |x|)0 , i = hp.v. , i = hP , i
x
x

Therefore,
(ln |x|)0 = P

1
.
x

3. (Problem 4) Find the nth derivative of f (x) = |x|.


Solution. Let D (R), supp [a, a] for some a > 0.
f (x) = |x|
h|x|0 , i = h|x|, 0 i
Z
=
|x| 0 (x)dx
ZRa
=
|x| 0 (x)dx
Z

a
0

x (x)dx

x 0 (x)dx

a Z a
0 Z 0
(x)dx x(x) 0 +
(x)dx
= x(x) a
0
| {z } a
| {z }
Z

0
0

=
(x)dx +
(x)dx
0
Z aa
=
sgn(x)(x)dx
a

= hsgn(x), i
Therefore,

|x|0 = sgn(x)

Now,
h|x|, i = h(sgn(x))0 , i = hsgn(x), 0 i
Za
=
sgn(x) 0 (x)dx
Z

a
0

0 (x)dx

0 (x)dx

0
a
= (x) a (0) 0
= (0) + (0) = 2(0)
= h2, i
Hence, in a similar fashion

f (n) (x) = |x|(n) = 2n1 ,

for n 2.

4. (Problem 5) Consider the sequence of functions on R defined by


fn (x) =

n
,
(1 + n2 x2 )

n = 1, 2, . . .

Show that {fn } converges to in the sense of distribution.


Solution. We have
Z
lim hfn , i = lim

n(x)
dx = lim
n
(1
+ n2 x 2 )

n(x)
dx,
(1 + n2 x2 )

since is test function (and therefore (x) = 0 for x < [M, M]).
By substituting t = nx, we have
nM

Z
lim hfn , i = lim

nM

( nt )
dt = lim
n
(1 + t 2 )

[nM,nM] (t)( nt )

(1 + t 2 )

dt,

from which we get by Lebesgue dominated theorem,


Z
lim hfn , i =

lim

[nM,nM] (t)( nt )
(1 + t 2 )

(0)
(0)
dt =
arctan t|
= (0)
2

(1 + t )

dt =

lim hfn , i = (0) = h, i

Thus, {fn } converges to in the sense of distribution.


5. (Problem 8) Prove that for every continuous function a we have a = a(0), and, if a C 1 then a0 =
a(0)0 a(0). In particular for a(x) = x, show that x = 0, x(m) = mm1 for m N.
Solution. Let a(x) be a continuous function and (x) D .
i)
ha(x)(x), (x)i = h(x), a(x)(x)i
= a(0)(0)
= a(0)h, i
Therefore,
a(x)(x) = a(0)(x)

(1)

ii)
ha(x)0 (x), (x)i = h0 (x), a(x)(x)i = h0 (x), (a(x)(x))0 i
= h(x), a0 (x)(x) + a(x) 0 (x)i
= h(x), a0 (x)(x)i h(x), a(x) 0 (x)i
= ha0 (x)(x), (x)i ha(x)(x), 0 (x)i
= ha0 (0)(x), (x)i ha(0)(x), 0 (x)i
0

= ha (0)(x), (x)i + ha(0) (x), (x)i


= ha(0)0 (x) a0 (0)(x), (x)i
Therefore,

a(x)0 (x) = a(0)0 (x) a0 (0)(x)

iii) In particular if a(x) = x, a0 (x) = 1.


x(x) = 0 = 0
x0 (x) = 0 0 (x) 1 (x) = (x)
Induction hypothesis(I.H), suppose it holds for m = r 1 i.e.
xr1 (x) = (r 1)r2 (x)

(By 1)

We need to show that it holds for m = r.


hxr (x), (x)i = hr (x), x(x)i
= hr1 (x), (x(x))0 i
= hr1 (x), (x) + x 0 (x)i
= hr1 (x), (x)i hr1 (x), x 0 (x)i
= hr1 (x), (x)i hxr1 (x), 0 (x)i
= hr1 (x), (x)i h(r 1)r2 (x), 0 (x)i
= h

r1

= h

r1

= hr
Hence

(x), (x)i + h(r 1)


(x) (r 1)

r1

r1

r1

(x), (x)i

(x), (x)i

(x), (x)i

xr (x) = rr1 (x)

Therefore, by the principle of mathematical induction, we conclude that


xm (x) = mm1 (x),

The End

m N.

(By I.H)

You might also like