You are on page 1of 164

DOCTORAL THES I S

Luleå University of Technology
Department of Mathematics
:oo;:¡¡|issx: +¡o:-+¡¡¡|isrx: i+u-i+ -- o;⁄ ¡¡ -- sr
:oo;:¡¡
Carleman type inequalities
and Hardy type inequalities
for monotone functions
Maria Johansson
Carleman type inequalities and
Hardy type inequalities for
monotone functions
by
Maria Johansson
Department of Mathematics
Luleå University of Technology
971 87 Luleå, Sweden
November 2007
Supervisor
Lars-Erik Persson,
Luleå University of Technology, Sweden
Published 2007
Printed in Sweden by University Printing Office, Luleå
To Hans, Emma and Nils
iv
Abstract
This Ph.D. thesis deals with various generalizations of the inequalities by
Carleman, Hardy and Pólya-Knopp. In Chapter 1 we give an introduction
and overview of the area that serves as a frame for the rest of the thesis. In
Chapter 2 we consider Carleman’s inequality, which may be regarded as a
discrete version of Pólya-Knopp’s inequality and also as a natural limiting
inequality of the discrete Hardy inequality. We present several simple proofs
of and remarks (e.g. historical) about this inequality. In Chapter 3 we give
some sharpenings and generalizations of Carleman’s inequality. We discuss
and comment on these results and put them into the frame presented in the
previous chapter. In particular, we present some new proofs and results. In
Chapter 4 we prove a multidimensional Sawyer duality formula for radially
decreasing functions and with general weights. We also state the correspond-
ing result for radially increasing functions. In particular, these results imply
that we can describe mapping properties of operators defined on cones of
such monotone functions. Moreover, we point out that these results can also
be used to describe mapping properties of operators between some corre-
sponding general weighted multidimensional Lebesgue spaces. In Chapter 5
we give a new weight characterization of the weighted Hardy inequality for
decreasing functions and use this result to give a new weight characteriza-
tion of the weighted Pólya-Knopp inequality for decreasing functions and we
also give a new scale of weightconditions for characterizing the embedding
Λ
p
(v) → Γ
q
(u) for the case 1 < p ≤ q < ∞. In Chapter 6 we make a unified
approach to Hardy type inequalitits for decreasing functions and prove a re-
sult which covers both the Sinnamon result with one condition and Sawyer’s
result with two independent conditions for the case when one weight is non-
decreasing. In all cases we point out that this condition is not unique and
can even be chosen among some (infinite) scales of conditions. In Chapter 7
v
vi Abstract
we prove a weight characterization of L
p
ν
[0, ∞)−L
q
µ
[0, ∞) boundedness of the
general Hardy operator (H
s
f)(x) =
_
_
[0,x]
f
s
udλ
_1
s
restricted to the cone of
monotone functions f ≥ 0 for 0 < p, q, s < ∞ with positive Borel σ-finite
measures λ, µ and ν. In Chapter 8 we present some new integral conditions
characterizing the embedding Λ
p
(v) → Γ
q
(w) , 0 < p, q ≤ ∞, including
proofs also for the cases (i) p = ∞, 0 < q < ∞, (ii) q = ∞, 1 < p < ∞ and
(iii) p = q = ∞. Only one condition is necessary for each case, which means
that our conditions are different and simpler than other corresponding con-
ditions in the literature. We even prove our results in a more general frame
namely when the space Γ
q
(w) is replaced by the more general space Γ
q
u
(w) .
In our proof we use a technique of discretization and anti-discretization de-
veloped by A. Gogatishvili and L. Pick, where they considered the opposite
embedding.
Preface
This Ph.D. thesis is written as a monograph. In particular, the contributions
of the author in the papers below are included and put into this general
frame.
• M. Johansson, L.-E. Persson and A. Wedestig, Carlemans olikhet-
historik, skärpningar och generaliseringar, Normat 51:3 (2003), 89-108
(in Swedish).
• M. Johansson, L.-E. Persson and A. Wedestig, Carleman’s inequality:
history, proofs and some new generalizations, J. Inequal. Pure Appl.
Math. 3(4) (2003)(19 pages).
• S. Barza, M. Johansson and L.-E. Persson, A Sawyer duality principle
for radially monotone functions in R
n
, J. Inequal. Pure Appl. Math.
6(2) (2005)(13 pages).
• A. Gogatishvili, M. Johansson, C.A. Okpoti and L.-E. Persson, Char-
acterizations of embeddings in Lorentz spaces, Bull. Austral. Math.
Soc. 76 (2007), 69-92.
• M. Johansson, V.D. Stepanov and E.P. Ushakova, Hardy inequality
with three measures on monotone functions, to appear in Math. In-
equal. Appl. (20 pages).
vii
viii Preface
• M. Johansson, A unified approach to the Sawyer and Sinnamon char-
acterizations of Hardy’s inequality for decreasing functions, Research
report 6, Department of Mathematics, Luleå University of Technology,
(2007) (11 pages).
• M. Johansson, A new characterization of the Hardy and its limit Polya-
Knopp inequality for decreasing functions, Research report 7, De-
partment of Mathematics, Luleå University of Technology, (2007) (14
pages), submitted.
See also [9], [32], [48], [49], [50], [51] and [52].
Acknowledgements
First of all I want to express my deepest gratitude to Professor Lars-Erik
Persson, my main supervisor, for his guidance, support and encouragement.
I consider myself fortunate to be one of his students.
Secondly, I want to express my gratitude to Professor Lech Maligranda
and Vladimir Stepanov, my co-supervisors, for their support and many fruit-
ful discussions. Moreover , I want to Thank Dr. Sorina Barza and Dr. Anna
Wedestig, my co-supervisors up to licentiate degree, for constant support
and generous advices.
I would also like to thank my co-authors Professor Lars-Erik Persson,
Professor Vladimir Stepanov (Peoples Friendship University, Moscow, Rus-
sia), Professor Amiran Gogatishvili (Mathematical Institute, Academy of
Sciences, Czech Republic), Dr. Elena Ushakova (Computing Center FEB
RAS, Khabarovsk, Russia), Dr. Sorina Barza (Karlstad University), Dr.
Anna Wedestig and Dr. Christopher Okpoti (University of education, Win-
neba, Ghana) , for their cooperation and also for their help with this Ph.D.
thesis. Especially I want to thank Dr. Anna Wedestig for all her help and
for just being a good friend.
Furthermore, I thank everyone at the Department of Mathematics at
Luleå University of Technology for the inspiring and challenging atmosphere.
Finally, I want to thank my husband Hans for his constant love and
support, and I thank our children Emma and Nils for bringing so much joy
into our lives.
Luleå, November 2007
Maria Johansson
ix
x Acknowledgements
Conventions and notations
The notations R, Z, N stand for the field of real numbers, the group of integers
and the semigroup of natural numbers, respectively. If n ∈ N¸¦0¦ we denote
R
n
+
:= ¦x = (x
1
, ..., x
n
) : x
i
≥ 0, i = 1, 2, ...n¦ and R
+
:= R
1
+
.
Constants are always positive and denoted by C and may be different
at different places. Throughout this thesis all functions are assumed to be
measurable. Moreover, n ∈ Z
+
, 1 ≤ p < ∞, p

=
p
p−1
(p

= ∞ if p = 1),
v (x) and u(x) are weights (positive and measurable functions on R
n
) and
L
p
v
= L
p
v
(R
n
)
=
_
f : R
n
→R, measurable s.t.
__
R
n
[f (x)[
p
v (x) dx
_1
p
< ∞
_
.
A function f : R
n
→ R is decreasing (increasing) if f (x) ≤ f (y) whenever
x ≤ y. If f : R
n
→ R is decreasing (increasing) separately in each variable
we write 0 ≤ f ↓ (0 ≤ f ↑). A set D ⊂ R
n
is said to be decreasing if its
characteristic function χ
D
is decreasing, and clearly if 0 ≤ h ↓ and t > 0,
then the set D
h,t
= ¦x ∈ R
n
: h(x) > t¦ is decreasing.
The symbol ≈ (c.f. (4.1)) means that the quotient of the right and left
hand sides is bounded from above and below by positive constants, while
expressions as 0 ∞ = ∞ 0 are taken as zero. Other notations will be
introduced and explained when required.
xi
xii Conventions and notations
Contents
Abstract v
Preface vii
Acknowledgements ix
Conventions and notations xi
1 Introduction 1
1.1 On Hardy’s original inequalities and its pre-history . . . . . . 1
1.2 Some further developments of the Hardy inequality . . . . . . 4
1.3 On the limiting inequalities of the discrete Hardy inequality . 5
1.4 On the limiting inequalities of the continuous Hardy inequality 6
1.5 On Hardy and Pólya-Knopp type inequalities on the cone of
decreasing functions . . . . . . . . . . . . . . . . . . . . . . . 8
1.6 On the connection between the one condition case and the
two conditions case . . . . . . . . . . . . . . . . . . . . . . . . 10
1.7 On multidimensional Hardy-type inequalities . . . . . . . . . 12
1.8 On the Hardy inequality with measures . . . . . . . . . . . . 13
2 Carleman’s inequality - history and proofs 15
2.1 Some proofs of Carleman’s inequality . . . 16
2.2 Pólya-Knopp’s inequality . . . . . . . . . . . . . . . . . . . . . 24
2.3 Final remarks about Torsten Carleman and his work . . . . . 28
3 Carleman’s inequality - some new sharpenings and general-
izations 31
xiii
xiv Contents
4 A Sawyer duality principle for radially monotone functions
in R
n
41
4.1 Preliminary results . . . . . . . . . . . . . . . . . . . . . . . . 42
4.2 The Duality Principle for radially monotone functions . . . . 44
4.3 Further results and applications . . . . . . . . . . . . . . . . . 50
5 A new characterization of the Hardy inequality and its limit
Pólya-Knopp inequality for decreasing functions 59
5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
5.2 A new weight characterization of the weighted Hardy inequal-
ity for decreasing functions . . . . . . . . . . . . . . . . . . . 60
5.3 A characterization of the Pólya-Knopp inequality for decreas-
ing functions . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
5.4 Another approach via an equivalence theorem . . . . . . . . . 68
6 A Unified approach to Sawyer and Sinnamon characteriza-
tions of the Hardy inequality for decreasing functions 73
6.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
6.2 The main result . . . . . . . . . . . . . . . . . . . . . . . . . . 75
6.3 Concluding remarks . . . . . . . . . . . . . . . . . . . . . . . 81
7 Hardy inequality with three measures on monotone func-
tions 83
7.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
7.2 Preliminary remarks . . . . . . . . . . . . . . . . . . . . . . . 86
7.3 The case 0 < p ≤ 1 . . . . . . . . . . . . . . . . . . . . . . . . 89
7.4 The case 1 < p, q < ∞ . . . . . . . . . . . . . . . . . . . . . . 104
8 Characterizations of embeddings in Lorentz spaces 107
8.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
8.2 Preliminaries . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
8.3 The main result . . . . . . . . . . . . . . . . . . . . . . . . . . 114
8.4 Proofs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
Chapter 1
Introduction
1.1 On Hardy’s original inequalities and its pre-
history
The dramatic pre-history until G.H. Hardy finally proved his famous inequal-
ity (see (1.2) below) in his 1925 paper was recently described and analyzed in
the paper [61] by A. Kufner, L. Maligranda and L.-E. Persson. In particular
we have included some facts from this description in this section.
Let us first consider the following standard forms of Hardy’s inequality.
The discrete Hardy inequality: if p > 1 and ¦a
k
¦

1
is a sequence of
positive real numbers, then

n=1
_
1
n
n

k=1
a
k
_
p

_
p
p −1
_
p

n=1
a
p
n
, (1.1)
and the continuous Hardy inequality: suppose that p > 1 and f is a non-
negative p-integrable function on (0, ∞). Then f is integrable over the in-
terval (0, x) for all x > 0 and

_
0
_
_
1
x
x
_
0
f(t)dt
_
_
p
dx ≤
_
p
p −1
_
p

_
0
f
p
(x)dx. (1.2)
We make the following remarks:
1
2 Carleman type inequalities and Hardy type inequalities
for monotone functions
1. The inequalities (1.1) and (1.2) are the standard forms of Hardy’s
inequalities, which can be found in many textbooks in analysis and have been
highlighted first in the famous book [44] by G.H. Hardy, J.E. Littlewood and
G. Pólya.
2. By restricting (1.2) to the class of step functions it can easily be proved
that (1.2) in fact implies (1.1). This important fact seems to have been first
mentioned by E. Landau (cf. [43, p. 154] and [61, Section 8], see also section
1.3).
3. The constant
_
p
p−1
_
p
in both inequalities (1.1) and (1.2) is sharp in
the sense that it can not be replaced by any smaller number.
4. The inequalities (1.1) and (1.2) imply the following information:
If

n=1
a
p
n
< ∞, then

n=1
h
p
n
(a) < ∞, (1.3)
where a = ¦a
n
¦ with a
n
≥ 0 and h(a) = ¦h
n
(a)¦ ,
h
n
(a) :=
1
n

n
k=1
a
k
< ∞ is the discrete Hardy operator.
If

_
0
f
p
(x)dx < ∞, then

_
0
(Hf (x))
p
dx < ∞, (1.4)
where f(x) ≥ 0 and Hf (x) :=
1
x
x
_
0
f (x) dx is the continuous Hardy operator.
Note that (1.3) and (1.4) are sometimes (see e.g. comments at the end
of Hardy’s paper [41]) called the weak forms of (1.1) and (1.2), respectively.
5. The inequalities (1.1) and (1.2) together with remark 3 above im-
ply the important information that the Hardy operators H and h map the
spaces L
p
into L
p
and l
p
into l
p
, respectively (p > 1), and their norms are
equal to p

:=
p
p−1
. Here, as usual, the spaces L
p
and l
p
are the Lebesgue
spaces consisting of all classes of measurable functions f = f (x) on (0, ∞) ,
respective sequences a = ¦a
n
¦ of real numbers, such that
|f|
L
p
:=
_
_

_
0
[f(x)[
p
dx
_
_
1/p
< ∞ and |a|
l
p
:=
_

n=1
[a
n
[
p
_
1/p
< ∞.
6. The inequalities (1.1) and (1.2) have been generalized and applied in
analysis and in the theory of differential equations. Some of these develop-
ments, generalizations and applications have been described and discussed
Introduction 3
in several papers and books e.g. [44], [62], [63] and [80], devoted only to this
subject.
Let us now present shortly the inequality that in fact was Hardy’s main
source and motivation when he initiated the research that finally led him
to the inequality (1.1). This was another famous inequality, which was dis-
covered by D. Hilbert in the early 1900s. Here we present it in the most
basic form: If


m=1
a
2
m
< ∞ and


n=1
b
2
n
< ∞ (a
m
≥ 0, b
n
≥ 0) , then the
double series


n=1


m=1
a
m
b
n
m+n
converges (the weak form). More precisely,
the inequality

n=1

m=1
a
m
b
n
m+n
≤ π
_

m=1
a
2
m
_
1/2
_

n=1
b
2
n
_
1/2
(1.5)
holds with π as the sharp constant. In the literature also the following more
general form of (1.5) is referred to as Hilbert’s inequality:

n=1

m=1
a
m
b
n
m+n

π
sin
_
π
p
_
_

m=1
a
p
m
_
1/p
_

n=1
b
p

n
_
1/p

, (1.6)
where p > 1 and p

=
p
p−1
, even if Hilbert himself was not even close to
consider this case (the space L
p
appeared only in 1910). We remark that it
was M. Riesz and G.H. Hardy himself who took the first step towards a proof
of (1.6). Other important mathematicians who take part in this development
was E. Landau and G. Pólya.
Now consider the inequality (1.1). In fact, for p = 2 G. H. Hardy dealt
with this inequality already in his paper [41] from 1919 but his ideas and
motivation (to find an elementary proof of Hilbert’s inequality) traces back
even to his paper [40] from 1915. Moreover, in the paper [42] from 1920 Hardy
described a proof of (1.1) which had been communicated to him by M. Riesz
but without the best constant (in his estimate the constant
_
p
2
p−1
_
p
appears).
Finally, in his famous 1925 paper [43] Hardy wrote: ”In a letter dated 21
June 1921, Landau communicated to me a direct proof of (1.1) which gives
the correct value
_
p
p−1
_
p
”. In the same note he stated and proved his
celebrated integral version of (1.1) namely (1.2) and informed that Landau
already in his 1921 letter also pointed out that in fact (1.2) implies (1.1).
The inequality (1.1) is usually referred to as the discrete Hardy inequality
in the literature but for historical reasons (described above) we could also
call it the Hardy-Riesz-Landau inequality. As you can see in [61] other
mathematicians made very important contributions in the development of
4 Carleman type inequalities and Hardy type inequalities
for monotone functions
the inequalities (1.2) and (1.1), e.g. E. Landau, G. Pólya, M. Riesz and I.
Schur. For example if the results of these persons had been published in
another way it could have changed the situation and it should not be unfair
to call the discrete inequality (1.1) as the Riesz, or the Landau-Riesz or the
Hardy-Landau-Riesz inequality.
1.2 Some further developments of the Hardy in-
equality
Since 1925 the investigations and generalizations of Hardy’s inequality have
developed in several directions. Here we have to mention some books which
are important. The first one is the classical Hardy-Littlewood-Pólya book
[44] from 1934. Moreover, in 1990 B. Opic and A. Kufner published the
book [80], which is completely concerned with the Hardy inequalities. The
newest developments of Hardy-type inequalities can be found in [63] by A.
Kufner and L.-E. Persson and in [62] by A. Kufner, L. Maligranda and L.-E.
Persson. Here we also mention the following books on inequalities (related
to this subject): [28], [29], [31] and [72].
There are also many papers written in this subject. For more information
see the references in the books. Here we also have to mention some recent
Ph.D. theses devoted to this subject, namely those by S. Barza [7], A.A.
Kalybay [55], M. Nassyrova [75], C.A. Okpoti [79], D.V. Prokhorov [89], E.
Ushakova [101] and A. Wedestig [103].
Important developments have been made in, for example, the following
directions:
• Weighted forms of Hardy-type inequalities.
• The mapping properties of the Hardy operator between weighted L
p
or L
q
spaces.
• Compactness properties of the Hardy operator.
• Weighted Hardy-type inequalities of weak type.
• Weighted Hardy-type inequalities on the cone of monotone functions
or sequences.
• Hardy-type inequalities in more general function spaces e.g. Orlicz,
Lorentz or even Banach function spaces.
Introduction 5
• Limiting (Carleman or Pólya-Knopp type) inequalities of the Hardy-
type inequalities (when p → ∞).
• Hardy-type inequalities with more general integraloperators (described
e.g. by an integralkernel satisfying an Oinarov condition).
• Multidimensional Hardy-type inequalities (e.g. for cones or even more
general subsets of R
n
).
• Fractional order Hardy-type inequalities.
• Refined Hardy-type inequalities.
• Higher order Hardy-type inequalities.
• General three-weights Hardy-type inequalities.
• Hardy-type inequalities for negative indices.
• Mapping properties of the Hardy operator with variable limits and its
limiting (geometric mean) operator.
• Various combinations of the developments above.
• Various applications e.g. mapping properties of the maximal function,
imbedding of Lorentz spaces, the theory of partial differential equa-
tions, signal processing, etc.
In the remaining part of this introduction we just comment more on some
points in connection to which contributions have been made in this Ph.D.
thesis.
1.3 On the limiting inequalities of the discrete Hardy
inequality
First we observe that by replacing a
k
in (1.1) by a
1
p
k
and letting k −→ ∞ we
obtain the inequality

n=1
_
n

k=1
a
k
_
≤ e

n=1
a
n
. (1.7)
This inequality was first presented in 1922 in [14] by the Swedish mathemati-
cian Torsten Carleman (1892-1942) and it is called Carleman’s inequality.
6 Carleman type inequalities and Hardy type inequalities
for monotone functions
Carleman discovered this inequality during his important work on quasian-
alytical functions and he could hardly have imagined at that time that this
discovery would be an object for such great independent interest.
More information about Carleman’s inequality is found in Chapter 2,
where we present several proofs and remarks on (1.7). At the end of the
chapter we also present some facts about T. Carleman and his work.
In Chapter 3 we give some examples of some generalizations of (1.7)
and (1.9). We discuss and comment on these results and put them into
the frame presented in Chapter 2. We also include some new proofs and
results. In particular, we prove a new weight characterization of a general
weighted Carleman type inequality for the case 0 < p ≤ q < ∞, i.e. we
prove a necessary and sufficient condition on the weight sequences ¦b
k
¦

k=1
and ¦d
k
¦

k=1
so that the inequality

k=1
(
k

a
1
a
2
...a
k
)
q
b
k
≤ C
_

k=1
a
p
k
d
k
_
1/p
(1.8)
holds for some finite and positive constant C and for all sequences ¦a
k
¦

k=1
of positive numbers.
1.4 On the limiting inequalities of the continuous
Hardy inequality
By making similar calculations as mentioned in the previous section, in the
inequality (1.2) we obtain the corresponding limit inequality

_
0
exp
_
_
1
x
x
_
0
ln f(t)dt
_
_
dx < e

_
0
f(x)dx, (1.9)
where f(t) > 0 and it is sometimes called Knopp’s inequality with reference
to [54] (cf. Remark 2.12). However, it seems that it was G. Pólya who first
discovered this inequality (see Remark 2.3). Therefore we prefer to call it
Pólya-Knopp’s inequality.
In Chapter 2 we prove that (1.9) implies (1.7). We also present some
proofs of (1.9) and thereby some more proofs of (1.7).
Next we mention that in [84] L.-E. Persson and V.D. Stepanov proved a
weight characterization for the inequality (1.14) by first characterizing the
Introduction 7
Hardy inequality
_
_

_
0
_
_
1
x
x
_
0
f(t)dt
_
_
q
u(x)dx
_
_
1
q
≤ C
_
_

_
0
f
p
(x)v(x)dx
_
_
1
p
(1.10)
with a new weight criteria. Their result reads:
Theorem 1.1. Let 1 < p ≤ q < ∞. Then the Hardy inequality (1.10) holds
for f ≥ 0 if and only if
A
PS
= A
PS
(p, q, u, v) := (1.11)
sup
t>0
_
_
t
_
0
u(x)x
−q
_
_
x
_
0
v(y)
1−p

dy
_
_
q
dx
_
_
1
q
_
_
t
_
0
v(x)
1−p

dx
_
_

1
p
< ∞.
Moreover, the best constant C in (1.10) can be estimated as follows:
A
PS
≤ C ≤ p

A
PS.
.
Then by performing a limiting procedure they received the following limit
result:
Theorem 1.2. Let 0 < p ≤ q < ∞. Then the inequality (1.14) holds for all
f ≥ 0 if and only if
D
PS
= D
PS
(p, q, w) := sup
t>0
t

1
p
_
_
t
_
0
w(x)dx
_
_
1
q
< ∞, (1.12)
where
w(x) = exp
_
_
1
x
x
_
0
ln v(y)dy
_
_

q
p
u(x). (1.13)
Moreover, if C is the best possible constant in (1.14), then
D
PS
≤ C ≤ e
1
p
D
PS
.
In Chapter 5 partly guided by these investigations of Persson and Stepanov
we give a new weight characterization to the Pólya-Knopp inequality for de-
creasing functions. More exactly, we proceed as follows: first we give a
8 Carleman type inequalities and Hardy type inequalities
for monotone functions
characterization of the Hardy inequality for decreasing functions, and then
we use this to give a weight characterization of the weighted Pólya-Knopp
inequality
_
_

_
0
_
_
exp
_
_
1
x
x
_
0
ln f(t)dt
_
_
_
_
q
u(x)dx
_
_
1
q
≤ C
_
_

_
0
f
p
(x)v(x)dx
_
_
1
p
(1.14)
for decreasing functions f ≥ 0, for the case 0 < p ≤ q < ∞.
1.5 On Hardy and Pólya-Knopp type inequalities
on the cone of decreasing functions
Recall that, for 0 < p < ∞, the Lorentz space Λ
p
(w) is defined by
Λ
p
(w) :=
_
¸
_
¸
_
f : |f

|
p,w
=
_
_

_
0
(f

(x))
p
w(t)dt
_
_
1
p
< ∞
_
¸
_
¸
_
,
where w is a weight, f is a measurable function on a measure space X (for
example R
n
) and f

denotes the decreasing rearrangement of [f[ defined by
f

(t) = inf ¦λ > 0 : [¦x > 0 : [f (x)[ > λ¦[ ≤ t¦ .
It is well-known that the functional ||
p,w
is a norm if and only if w is
decreasing and p ≥ 1.
In order to be able to study the structure of the Λ
p
(w) −space it is impor-
tant to study the boundedness of the Hardy operator f
∗∗
(t) =
1
t
_
t
0
f

(s) ds
between weighted L
p
-spaces, which means that the Hardy inequality
_
_

_
0
_
_
1
x
x
_
0
f

(t)dt
_
_
p
u(x)dx
_
_
1
p
≤ C
_
_

_
0
f

(x)
p
u(x)dx
_
_
1
p
holds for decreasing functions f

(t). Recall that the rearrangement of the
Hardy-Littlewood maximal function Mf is equivalent to the Hardy operator
of the rearrangement of [f[. To be more precise, if
(Mf) (x) = sup
x∈Q
1
[Q[
_
Q
[f (z)[ dz, x ∈ R
n
,
Introduction 9
where Q is a cube in R
n
with sides parallel to the coordinate axes and [Q[
is its Lebesgue measure, then
(Mf)

(x) ≈
1
t
_
t
0
f

(s) ds, t > 0.
Some historical remarks and further developments of this equivalence can be
found in paper [6] (see also [62, p. 92]).
Hence, to prove that
M : Λ
p
(v) → Λ
q
(u) , 1 < p, q < ∞
is a bounded mapping, or, in other words, to characterize the weight func-
tions u and v for which M is bounded between weighted Lorentz spaces, it
is equivalent to prove that the Hardy operator H, defined by
(Hf) (t) :=
1
t
_
t
0
f (s) ds, t ≥ 0,
on the cone of positive decreasing functions, is bounded fromL
p
(v) to L
q
(u) ,
1 < p, q < ∞. This means that one desires to characterize the weight
functions u and v for which the Hardy inequality
_
_

_
0
_
1
x
_
x
0
f (t) dt
_
q
u(x) dx
_
_
1
q
≤ C
_
_

_
0
f
p
(x) v (x) dx
_
_
1
p
holds for all f ≥ 0 which are decreasing.
In Chapter 8 we present some new integral conditions which characterize
the embedding Λ
p
u
(v) → Γ
p
u
(v), where
Γ
p
u
(v) =
_
f : |f|
Γ
p
u
(v)
=
__

0
f
∗∗p
u
(t) v (t) dt
_1
p
< ∞
_
.
for the cases 0 < p, q ≤ ∞. We include proofs also for the cases (i) p = ∞,
0 < q < ∞, (ii) q = ∞, 1 < p < ∞ and (iii) p = q = ∞. Hence the results
are new also for the case u ≡ 1. Only one condition is necessary for each
case which means that our conditions are different and simpler than other
corresponding conditions in the literature. In our proof we use a technique
of discretization and anti-discretization developed by A Gogatishvili and L.
Pick in [35] and [36], where the opposite embedding was considered.
Other aspects of generalized Lorentz spaces and Hardy type inequalities
for monotone functions have been studied by several authors among them
10 Carleman type inequalities and Hardy type inequalities
for monotone functions
we mention L. Maligranda (see e.g. [26], [45], [57], [56], [68], [69], [70] and
[71]) and for more references see also the Ph.D. Thesis [7] by S. Barza and
the new book [62] by A. Kufner, L. Maligranda and L.-E. Persson.
We also mention that in 1990 E. Sawyer [93] established the following
remarkable duality principle for weighted L
p
spaces on the cone of decreasing
functions:
Suppose that 1 < p < ∞. Let g, v be positive measurable functions on
(0, ∞) with v locally integrable. Then
sup
0≤f↓
_

0
f (x) g (x) dx
__

0
f
p
(x) v (x) dx
_1
p

_
_

0
__
x
0
g (t) dt
_
p
__
x
0
v (t) dt
_
−p

v (x) dx
_
1/p

+
_

0
g (x) dx
__

0
v (x) dx
_
1/p
.
A very nice proof of this principle was presented by V.D. Stepanov in
[99].
1.6 On the connection between the one condition
case and the two conditions case
In 1990 E. Sawyer also made a very important contribution to the study of
the Hardy inequality for decreasing functions, namely he proved the follow-
ing result by using a general approach (see [93, Theorem 2] and also [100,
Theorem 2] for a simpler proof):
Theorem 1.3. Let 1 < p ≤ q < ∞. Then the Hardy inequality
_
_

_
0
_
1
x
_
x
0
f (t) dt
_
q
u(x) dx
_
_
1
q
≤ C
_
_

_
0
f
p
(x) v (x) dx
_
_
1
p
(1.15)
holds for all f ↓≥ 0 if and only if
A
0
= A
0
(p, q, u, v) := sup
t>0
V

1
p
(t)
_
_
t
_
0
u(x)dx
_
_
1
q
< ∞, (1.16)
and
A
1
= A
1
(p, q, u, v) := (1.17)
Introduction 11
sup
t>0
_
_

_
t
u(x)x
−q
dx
_
_
1
q
_
_
t
_
0
x
p

V
−p

(x)v(x)dx
_
_
1
p

< ∞,
where p

=
p
p−1
and
V (t) =
t
_
0
v(x)dx. (1.18)
Moreover, if C is the best possible constant in (1.15), then
C ≈ max(A
0
, A
1
).
The inequality (1.15) is a special case of the general three-weights in-
equality studied in [77] (here the measures are dx, u(x)dx and v(x)dx, re-
spectively). We also note that G. Sinnamon in [97] studied the following
two-weights Hardy inequality
_
_

_
0
__
x
0
fdλ
_
q

_
_
1
q
≤ C
_
_

_
0
f
p

_
_
1
p
,
where λ and µ are measures on R. For the special case when dµ = u(x) dx
and dλ = v (x) dx this result reads (see [97, Corollary 3.4]):
Theorem 1.4. Let 1 < p ≤ q < ∞and let u(x) and v (x) be weight functions
on (0, ∞). Then the inequality
_
_

_
0
__
x
0
f (t) v (t) dt
_
q
u(x) dx
_
_
1
q
≤ C
_
_

_
0
f
p
(x) v (x) dx
_
_
1
p
(1.19)
holds for all non-increasing f ≥ 0 if and only if
A
2
:= sup
t>0
_
_
t
_
0
u(x)
_
_
x
_
0
v (y) dy
_
_
q
dx
_
_
1
q
_
_
t
_
0
v (x) dx
_
_

1
p
< ∞. (1.20)
In Chapter 6 we investigate the connection between the result due to
Sawyer (Theorem 1.3) and the result due to Sinnamon (Theorem 1.4). We
also present a new Hardy type inequality from which we can get both the
Sinnamon one condition result and the Sawyer two conditions result for a
special case.
12 Carleman type inequalities and Hardy type inequalities
for monotone functions
1.7 On multidimensional Hardy-type inequalities
There are few results concerning characterizations of the weights so that
multidimensional Hardy-type inequalities hold in weighted L
p
-spaces. Here
we just mention the following results, which have guided the investigations
in this thesis:
1. Hardy-type inequalities when the standard one-dimensional Hardy-
operator is replaced by spherical means. One early result of this type is due
to P. Drábek, H.P. Heinig and A. Kufner [27] (see our Theorem 4.1). This
idea is further developed and explained in the paper [24] by A. Čižmešija,
L.-E. Persson and A. Wedestig, where general cones in R
n
are considered.
2. In the paper [93] E. Sawyer considered mapping properties of the
two-dimensional Hardy operator H
2
defined by
H
2
(f (x
1
, x
2
)) =
_
x
1
0
_
x
2
0
f (t
1
, t
2
) dt
1
dt
2
.
He proved the remarkable result that it was necessary and sufficient to have
three different integral conditions to characterize the weights u and v so that
a Hardy-type inequality holds in the normal L
p
(u) → L
q
(v) situation when
1 < p ≤ q < ∞.
Moreover, recently in her Ph.D. thesis [103] A. Wedestig proved that if
the weight v is of product type, then in fact we can get a weight charac-
terization in the Sawyer situation described above by only using one of his
conditions. This Wedestig result in fact also holds in dimension n ∈ Z
+
and the corresponding limit (Pólya-Knopp type) inequalities can be derived.
These ideas were also further developed and complemented in the Ph.D.
thesis [101] by E. Ushakova.
3. In the paper [10], S. Barza, L.-E. Persson and J. Soria studied multi-
dimensional Hardy-type inequalities on the cone of functions decreasing (or
increasing) in each variable. In particular, in their characterization these
authors were guided to introduce the new concept of decreasing sets. More
results of this type can be found in the paper [11] by S. Barza, L.-E. Persson
and V.D. Stepanov. These ideas were fully developed in the Ph.D. thesis [7]
by S. Barza.
In Chapter 4 of this Ph.D. thesis we prove a multidimensional version
of the Sawyer duality principle for radially decreasing functions and with
general weights (see Theorem 4.3). We also state the corresponding result
for radially increasing functions (see Theorem 4.5). A number of applications
and related results are also included.
Introduction 13
1.8 On the Hardy inequality with measures
Hardy was originally most interested in the discrete case but his final result in
the 1925 paper was in the continuous case (but he was aware of the fact that
the continuous case implies the discrete case in this unweighted situation).
After that almost all development has been concerning the continuous case
(see e.g. the book [62] and the references given there). However, fairly
late also the discrete case has been considered, by e.g. Bennett [13] and the
Ph.D. thesis by C. Okpoti [79] (and in the references given there). Thus, it
is natural to consider the case with general measures. The first result in this
direction seems to be the following result by B. Muckenhoupt [74]: In 1972
he proved that, in the case 1 ≤ p = q < ∞, the inequality
_
_

_
0
¸
¸
¸
¸
¸
¸
x
_
0
f (t) dt
¸
¸
¸
¸
¸
¸
q
dµ(x)
_
_
1
q
≤ C
_
_

_
0
[f (x)[
p
dν (x)
_
_
1
p
, (1.21)
where µ and ν are Borel measures, holds if and only if
M = sup
r>0
(µ[r, ∞))
1
p
_
_
r
_
0
_
d

v
dx
_
1−p

dx
_
_
1
p

< ∞,
where

v denotes the absolute continuous part of ν. Moreover, if C is the
least constant for which (1.21) holds, then M ≤ C ≤ p
1/p
(p

)
1/p

M for
1 < p < ∞ and C = M for p = 1. Moreover, V. Kokilashvili [60] (see also
[62]) in 1979 announced the general result (without a proof there) that for
1 ≤ p ≤ q < ∞ the inequality (1.21) holds if and only if
MK = sup
r>0
(µ[r, ∞))
1
q
_
_
r
_
0
_
d

v
dx
_
1−p

dx
_
_
1
p

< ∞.
It was proved in [33] that there are infinite many conditions (even four scales
of conditions) for characterizing (1.21) for 1 < p ≤ q < ∞. We remark that
in (1.21) three measures are involved namely dt, dµ(x) and dν (x) .
We also consider the following general two weights Hardy type inequality
_
_
_
_
R
_
_
_
_
(−∞,x]
fdλ
_
_
_
q
dµ(x)
_
_
_
1
q
≤ C
_
_

_
0
f
p

_
_
1
p
, (1.22)
14 Carleman type inequalities and Hardy type inequalities
for monotone functions
and the final general three weights inequality
_
_
_
_
R
_
_
_
_
(−∞,x]
fdσ
_
_
_
q
dµ(x)
_
_
_
1
q
≤ C
_
_

_
0
f
p

_
_
1
p
. (1.23)
By using Sinnamon’s idea about level functions it is possible to characterize
(1.22) and, furthermore, by [78, Theorem 2.1] we can in fact reduce (1.23)
to (1.22).
In Chapter 7 we will study the inequality
_
_
_
_
[0,∞)
_
_
_
_
[0,x]
fudλ
_
_
_
q
vdµ(x)
_
_
_
1
q
≤ C
_
_

_
0
f
p
wdν
_
_
1
p
where u, v and w are weights, for monotone functions and for all cases 0 <
p, q < ∞.
Chapter 2
Carleman’s inequality - history and
proofs
In this chapter we discuss the following remarkable inequality:
a
1
+

a
1
a
2
+... +
k

a
1
a
2
...a
k
< e (a
1
+a
2
+...) , (2.1)
where a
1
, a
2
, ... are positive numbers and

i=1
a
i
is convergent. This inequal-
ity was presented in 1922 in [14] by the Swedish mathematician Torsten
Carleman (1892-1942) and it is called Carleman’s inequality. Carleman dis-
covered this inequality during his important work on quasianalytical func-
tions and he could hardly have imagined at that time that this discovery
would be an object for such great interest. The continuous version of (2.1)
reads

_
0
exp
_
_
1
x
x
_
0
ln f(t)dt
_
_
dx < e

_
0
f(x)dx, (2.2)
where f(t) > 0 and it is sometimes called Knopp’s inequality with reference
to [54] (cf. Remark 2.12). However it seems that it was G. Pólya who first
discovered this inequality (see Remark 2.3). Therefore we prefer to call it
Pólya-Knopp’s inequality.
In Section 2.1 of this chapter we present several proofs of and remarks
on (2.1). In Section 2.2 we prove that (2.2) implies (2.1) and present some
proofs of (2.2) (and thus some more proofs of (2.1)).
15
16 Carleman type inequalities and Hardy type inequalities
for monotone functions
Finally, we include some facts about Torsten Carleman and his work,
which we have found, by studying [1], [58] and the recent article [67] by L.
Maligranda. This description partly complements the information given by
L. Gårding in his book [30].
2.1 Some proofs of Carleman’s inequality
Proof 1. (Rough sketch of Carleman’s original proof)
Carleman first noted that the problem can be solved by finding maximum
of the expression
k

i=1
(a
1
a
2
...a
i
)
1
i
under the constraint
k

i=1
a
i
= 1.
He then substituted a
i
= e
−x
i
and obtained the simpler problem:
Find maximum M
k
for k = 1, 2, ... of
G =
k

i=1
e

x
1
+x
2
+...+x
i
i
under the constraint
H =
k

i=1
e
−x
i
= 1.
This problem can be solved by using the Lagrange multiplier method. Unfor-
tunately this leads to some technical calculations, which, of course Carleman
carried out in an elegant way. We leave out these calculations here, and only
refer to Carleman’s paper [14], where all the details are presented. The result
is that M
k
< e for all k ∈ Z
+
. Carleman then showed separately that the
inequality (2.1) is strict when the sum on the left hand side converges.
Remark 2.1. In the same paper [14], Carleman proved that the inequality
(2.1) does not hold in general for any constant c < e, i.e., that the constant
e is sharp.
Proof 2. (via Hardy’s inequality)
Carleman’s inequality - history and proofs 17
The discrete version of Hardy’s inequality reads (see [43], [44])

k=1
_
1
k
k

i=1
a
i
_
p
<
_
p
p −1
_
p

k=1
a
p
k
, p > 1. (2.3)
Replace a
i
with a
1/p
i
and note that by using that x = e
lnx
and the definition
of the derivative we find that
_
1
k
k

i=1
a
1/p
i
_
p
= exp
1
p
_
ln
k

i=1
a
1/p
i
−ln
k

i=1
a
0
i
_
→ exp
__
D(ln
k

i=1
a
x
i
)
_
x=0
_
( when p → ∞)
= exp
__
k

i=1
a
x
i
ln a
i
/
k

i=1
a
x
i
_
x=0
_
= exp
1
k
k

i=1
ln a
i
=
_
k

i=1
a
i
_
1/k
and we see that (2.3) leads to the non-strict inequality (2.1) since
_
p
p−1
_
p
→ e when p → ∞. Observe that this method does not automati-
cally prove that we have strict inequality in (2.2) and this has to be proved
separately (see for example our later proofs).
Remark 2.2. G.H. Hardy formulated his inequality (2.3) in 1920 in [42]
and proved it 1925 in [43] but it seems that Carleman did not know about the
inequality (2.3) at this time, since he does not refer to the simple connection
that holds according to the proof above. This is somewhat remarkable since
Carleman worked together with Hardy at that time, see for example their
joint paper [15].
Remark 2.3. The above means that (2.1) may be considered as a limit
inequality for the scale (2.3) of Hardy inequalities. This was pointed out by
G.H. Hardy in 1925 in the paper [43, p. 156], but he pronounced that it was
G. Pólya who made him aware of this interesting fact.
We now present two proofs which are based on variations of the arithmetic-
geometric mean inequality (the AG-inequality).
18 Carleman type inequalities and Hardy type inequalities
for monotone functions
Proof 3 Because of the AG-inequality the following holds for every i =
1, 2, ..., every k and all c
i
> 0:
_
k

i=1
a
i
_
1/k
=
_
k

i=1
c
i
_
−1/k
_
k

i=1
c
i
a
i
_
1/k

_
k

i=1
c
i
_
−1/k
1
k
k

i=1
c
i
a
i.
(2.4)
We now choose c
i
=
(1+i)
i
i
i−1
, i = 1, 2, ..., k. Then
_
k

1
c
i
_
1/k
= k + 1 (2.5)
and (2.4) and (2.5) gives that

k=1
k

a
1
a
2
...a
k

k=1
1
k (k + 1)
k

i=1
c
i
a
i
=

i=1
c
i
a
i

k=i
1
k (k + 1)
=

i=1
c
i
a
i
i
=

i=1
a
i
_
1 +
1
i
_
i
≤ e

i=1
a
i
.
The strict inequality holds, since we cannot have equality at the same time in
all terms of the inequality. This can only occur if c
i
a
i
= c for some constant
c > 0 i.e. a
i
= c
_
i
1+i
_
i
1
i
, i = 1, 2, ... but this can not hold since

i=1
a
i
is
convergent (note that
_
i
1+i
_
i
→ e when i → ∞).
Remark 2.4. This idea of proof was presented by G. Pólya (see [86, p. 249])
but here we have more closely followed the presentation which can be found
in L. Hörmander’s book [47], p. 24.
Remark 2.5. We note that Proof 3 shows that for finite sums the inequality
(2.1) holds even for some constant strictly less than e. More precise, for
N = 1, 2, ...we have
N

k=1
k

a
1
a
2
...a
k

N

k=1
_
1 +
1
k
_
k
a
k
.
For historical reasons we also present another variant of Proof 3
Carleman’s inequality - history and proofs 19
Proof 4. We choose c
i
= i, i = 1, 2, ...k, in (2.4) and get that
_
k

i=1
a
i
_
1/k
≤ (k!)
−1/k
1
k
k

i=1
ia
i
. (2.6)
Moreover, it holds that
(k + 1)
k
k!
=
_
1 +
1
1
__
1 +
1
2
_
2
...
_
1 +
1
k
_
k
< e
k
(2.7)
and, by using this inequality and (2.6), we get that

k=1
k

a
1
a
2
...a
k

k=1
(k!)
−1/k
1
k
k

i=1
ia
i

k=1
e
k(k + 1)
k

i=1
ia
i
= e

i=1
ia
i

k=1
1
k(k + 1)
= e

i=1
a
i
.
The strict inequality can be proved in a similar manner as in Proof 3 (equality
demands that a
k
=
c
k
but this contradicts the fact that

k=1
a
k
is convergent).
Remark 2.6. In the paper [42, p. 77], G.H. Hardy presented essentially
this proof but he also pronounced that it was G. Knopp who pointed out this
proof to him.
Proof 5 (Carleson’s proof)
We first note that we can assume that a
1
≥ a
2
≥ ... (because the sum at
the left hand side of (2.1) obviously becomes the greatest if the sequence ¦a
i
¦
is rearranged in decreasing order while the sum at the right hand side will
be the same for every rearrangement). Let m(x) be a polygon through the
points (0, 0) and (k,
k

1
log(1/a
i
)), k = 1, 2, ... The function m(x) is obviously
convex and because of that it holds that for every r > 1
m(rx) ≥ m(x) + (r −1)xm´(x). (2.8)
20 Carleman type inequalities and Hardy type inequalities
for monotone functions
Furthermore
m´(x) = log(1/a
k
), x ∈ (k −1, k) , (2.9)
and since m(0) = 0 and m is convex.
m(x)
x
=
m(x) −m(0)
x

m(k) −m(0)
k
(2.10)
=
m(k)
k
=
1
k
k

i=1
log(1/a
i
) for all x ≤ k.
We now make a substitution and use Hölder’s inequality and (2.8). Then we
find, for every A > 0 and r > 1,
1
r
A
_
0
e
−m(x)/x
dx ≤
1
r
rA
_
0
e
−m(x)/x
dx
=
A
_
0
e
−m(rx)/rx
dx

A
_
0
e
−m(x)/rx−((r−1)/r)m

(x)
dx

_
_
A
_
0
e
−m(x)/x
dx
_
_
1/r
_
_
A
_
0
e
−m

(x)
dx
_
_
(r−1)
r
so that
A
_
0
e
−m(x)/x
dx ≤ r
r
r−1
A
_
0
e
−m

(x)
dx.
We let A → ∞, r → 1+ and note that then r
r
r−1
→ e so that

_
0
e
−m(x)/x
dx ≤ e

_
0
e
−m

(x)
dx. (2.11)
We now use (2.9) and (2.10) and get that

_
0
e
−m

(x)
dx =

k=1
k
_
k−1
e
−m

(x)
dx =

k=1
e
−log(1/a
k
)
=

k=1
a
k
Carleman’s inequality - history and proofs 21
respectively

_
0
e
−m(x)/x
dx =

k=1
k
_
k−1
e
−m(x)/x
dx

k=1
exp
_

1
k
k

i=1
log (1/a
i
)
_
=

k=1
_
k

i=1
a
1/k
i
_
1/k
.
The non-strict inequality (2.1) follows by using these estimates and (2.11).
The strict inequality follows from the fact that we can not have equality in
(2.10) at the same time for all x and k.
Remark 2.7. This is L. Carleson’s proof (see [16]) and in fact he proved
that the even more general inequality (2.11) holds for every piecewise dif-
ferentiable convex function m(x) on [0, ∞] such that m(0) = 0. In fact,
Carleson formulated and proved his inequality in the following slightly more
general form:

_
0
x
p
e
−m(x)/x
dx ≤ e
p+1

_
0
x
p
e
−m

(x)
dx, p > −1. (2.12)
Proof 6 (via Redheffer’s inequality)
R.M. Redheffer proved in 1967 the following interesting inequality (see
[90] and also [91]):
nG
n
+
n

k=1
k (b
k
−1) G
k

n

k=1
a
k
b
k
k
(2.13)
which holds for all n = 1, 2, ... and all positive sequences ¦b
k
¦ and where
G
k
=
_
k

i=1
a
i
_
1/k
.
In particular, we see that if
a) b
k
= 1, k = 1, 2, ... , then
G
n

1
n
n

k=1
a
k
= A
n
,
i.e. (2.13) coincides with the AG-inequality
22 Carleman type inequalities and Hardy type inequalities
for monotone functions
b) b
k
= 1 +
1
k
, k = 1, 2, ... ,then
nG
n
+
n

k=1
G
k

n

k=1
_
1 +
1
k
_
k
a
k
,
which implies that the non-strict inequality (2.1) follows when n → ∞. The
strict inequality can also be proved by using the arguments in the following
proof of the inequality (2.13): We use the elementary inequality
1 +a (x −1) ≤ x
a
, x > 0, a > 1, (2.14)
(a simple proof of (2.14) can be obtained by putting α =
1
a
and replacing x
with x
a
in the following form of the AG-inequality: x
α
1
1−α
≤ αx+(1 −α) 1).
We now use (2.14) with a = k and x =
G
k
G
k−1
b
k
(k ≥ 2) to obtain that
1 +k
_
G
k
G
k−1
b
k
−1
_

_
G
k
G
k−1
b
k
_
k
=
a
k
G
k−1
b
k
k
,
which can be written as
G
k−1
+k (G
k
b
k
−G
k−1
) ≤ a
k
b
k
k
. (2.15)
We now prove (2.13) by simple induction. First we observe that G
1
= a
1
so
that
G
1
+ (b
1
−1) G
1
= a
1
b
1
i.e. (2.13) holds for n = 1. Assume that the inequality (2.13) holds for
n = N ∈ Z
+
. Now us this induction assumption and (2.15) with k = N + 1
and we get that
(N + 1) G
N+1
+
N+1

k=1
k (b
k
−1) G
k
= (N + 1) b
N+1
G
N+1
+
N

k=1
k (b
k
−1) G
k
= NG
N
+
N

k=1
k (b
k
−1) G
k
+ (N + 1) b
N+1
G
N+1
−NG
N

N

k=1
a
k
b
k
k
+a
N+1
b
N+1
N+1
=
N+1

k=1
a
k
b
k
k
i.e. (2.13) holds even for n = N+1 and, by the induction axiom, we conclude
that (2.13) holds for all n ∈ Z
+
.
Carleman’s inequality - history and proofs 23
Remark 2.8. The proof above is somewhat more complicated than the other
proofs but it leads to a better result. In fact, this method of proving inequal-
ities uses a well-known principle, which is sometimes referred to as Redhef-
fer’s recursion principle (see [81]). This principle can also be used to improve
several other classical inequalities.
Let a
(n)
= ¦a
1
, a
2
, ..., a
n
¦ be a positive sequence (n = 1, 2, ...). We define
the powermeans M
r,n
of a
(n)
in the following way
M
r,n
= M
r,n
_
a
(n)
_
=
_
¸
¸
¸
_
¸
¸
¸
_
_
1
n
n

k=1
a
r
k
_
1/r
, r ,= 0,
_
n

k=1
a
k
_
1/n
, r = 0.
Note that A
n
= M
1,n
, G
n
= M
0,n
and H
n
= M
−1,n
are the usual arithmetic,
geometric and harmonic means, respectively. We also look at the following
sequence of powermeans:
M
r,n
= (M
r,1
, M
r,2
, ..., M
r,n
) .
In 1996 B. Mond and J. Pečarić proved the following interesting inequality
between iterative powermeans, see [73]:
M
s,n
(M
r,n
) ≤ M
r,n
(M
s,n
) , (2.16)
for every r ≤ s. We have equality if and only if a
1
= ... = a
n
. The proof in
[73] is fairly complicated so we do not show it here. But we are convinced
that there is a more elementary proof and we leave that as an open question
to the reader. Our next proof is based on the Mond-Pečaric result.
Proof 7 We use (2.16) with s = 1 and r = 0 and get that
1
n
n

k=1
G
k

_
n

k=1
_
1
k
k

i=1
a
i
__
1/n
. (2.17)
By using this inequality and the fact that
k

i=1
a
i

n

i=1
a
i
, k ≤ n,
we find that
n

k=1
G
k

n
n

n!
n

k=1
a
k
. (2.18)
24 Carleman type inequalities and Hardy type inequalities
for monotone functions
We use our previous estimate (2.7) with k = n −1 and get that
n
n
n!
=
n
n−1
(n −1)!
< e
n−1
, i.e.,
n
n

n!
< e
1−1/n
.
By combining this inequality with (2.17) we find that
n

k=1
_
k

i=1
a
i
_
1/k
< e
1−1/n
n

k=1
a
k
. (2.19)
This non-strict inequality (2.1) follows when we let n → ∞. The fact that
the inequality actually is strict follows from the fact that equality in (2.18)
only can occur when all a
i
are equal, but this can not be true under our
assumption that

k=1
a
k
is convergent.
Remark 2.9. More information about how (2.16) can be used to prove and
improve inequalities can be found in the papers [25] and [22].
Remark 2.10. We note that if we, in the proof above, combine (2.17) with
the following variant of the AG-inequality
_
n

k=1
1
k
k

i=1
a
i
_
1/n
=
_
1
n!
_
1/n
(a
1
(a
1
+a
2
) ... (a
1
+a
2
+... +a
n
))
1/n

_
1
n!
_
1/n
(na
1
+ (n −1) a
2
+... +a
n
)
n
we get the following strict improvement of (2.19):
n

k=1
_
k

i=1
a
i
_
1/k
< e
1−1/n
n

k=1
_
1 −
k −1
n
_
a
k
. (2.20)
2.2 Pólya-Knopp’s inequality
We begin by proving that (2.2) implies (2.1). As before we note that it is
enough to prove (2.1) when ¦a
k
¦

1
is a decreasing sequence. Use (2.2) with
the function f(x) = a
k
, x ∈ [k −1, k), k = 1, 2, .... Then

_
0
f(x)dx =

k=1
a
k
(2.21)
Carleman’s inequality - history and proofs 25
and

_
0
exp
_
_
1
x
x
_
0
ln f(t)dt
_
_
dx =

k=0
k+1
_
k
exp
_
_
1
x
x
_
0
ln f(t)dt
_
_
dx. (2.22)
Furthermore, it yields that
1
_
0
exp
_
_
1
x
x
_
0
ln f(t)dt
_
_
dx = a
1
, (2.23)
and, for k = 1, 2, ... ,
_
k+1

i=1
a
i
_
1
k+1
=
k+1
_
k
exp
_
1
k + 1
k+1

i=1
ln a
i
_
dx (2.24)

k+1
_
k
exp
_
1
x
k

i=1
ln a
i
+
x −k
x
ln a
k+1
)
_
dx
=
k+1
_
k
exp
_
_
1
x
x
_
0
ln f(t)dt
_
_
dx.
The crucial inequality in (2.24) depends on the fact that the integrand
is a weighted arithmetic mean of the numbers ln a
i
, i = 1, 2, ..., k + 1, with
weights
1
x
, ...,
1
x
(k − 1 weights) and
x−(k−1)
k
. Here k − 1 ≤ x ≤ k and since
the sequence is decreasing the mean value is smallest for x = k, i.e., when
all weights =
1
k
. Now (2.1) follows by combining (2.21) - (2.24).
We now present some simple proofs of (2.2) (and thereby some more
proofs of (2.1)).
Proof 8 We note that the function m(x) = −
x
_
0
lnf

(t) dt fulfils the condi-
tions to use Carleson’s inequality (2.12) (here f

(t) is the decreasing rear-
rangement of the function f). Therefore, according to (2.12), it holds that

_
0
x
p
exp
_
_
1
x
x
_
0
ln f

(t) dt
_
_
dx ≤ e
p+1

_
0
x
p
f

(x) dx, (2.25)
for every p > −1. Carleson’s argument shows that we in fact have strict
inequality in (2.25) and especially by using (2.25) for p = 0 we get Pólya-
Knopp’s inequality (2.2).
26 Carleman type inequalities and Hardy type inequalities
for monotone functions
Remark 2.11. Carleson did not note this fact explicitly in his paper [16]
since he obviously was mainly interested in giving a simple proof of the in-
equality (2.1).
We now present two other proofs of (2.2) and thereby of (2.1) which, like
Carleson’s proof, only are based on convexity arguments.
Proof 9 First we note that
exp
_
_
1
x
x
_
0
ln f(t)dt
_
_
(2.26)
= exp
_
_
1
x
x
_
0
ln tf(t)dt −
1
x
x
_
0
ln tdt
_
_
= exp
_
_
1
x
x
_
0
ln tf(t)dt
_
_
exp
_
_

1
x
x
_
0
ln tdt
_
_
.
Furthermore, it yields that

1
x
x
_
0
ln tdt = −ln x + 1 (2.27)
and, in view of Jensen’s inequality (or the AG-inequality),
exp
_
_
1
x
x
_
0
ln tf(t)dt
_
_

1
x
x
_
0
tf(t)dt. (2.28)
We integrate, use (2.26) - (2.28) and change the order of integration and find
that

_
0
exp
_
_
1
x
x
_
0
ln f(t)dt
_
_
dx ≤

_
0
e
−ln x+1
1
x
_
_
x
_
0
tf(t)dt
_
_
dx
= e

_
0
1
x
2
_
_
x
_
0
tf(t)dt
_
_
dx = e

_
0
tf(t)

_
t
1
x
2
dx = e

_
0
f(t)dt.
The strict inequality follows since equality in Jensen’s inequality requires that
tf(t) is constant a.e. but this can not occur since

_
0
f(x)dx is convergent.
Carleman’s inequality - history and proofs 27
Remark 2.12. The proof above is partly related to Knopp’s original idea
(see [59] p. 211). However Knopp worked with the interval [1, x] instead of
[0, x] and hence Jensen’s inequality can not be used. Moreover, Knopp never
wrote out the inequality (2.2) explicitly even if it is sometimes referred in the
literature as this is the case, see for example [44, p. 250] and [72, p. 143].
Remark 2.13. By modifying the proof above we can easily prove even some
weighted versions of (2.2) for instance the following

_
0
exp
_
_
1
x
x
_
0
ln f(t)dt
_
_
x
p
dx <
e
1 −p

_
0
f(x)x
p
dx
for every p < 1 which is more general than (2.2), cf. also (2.25).
Proof 10 We first note that if we replace f(t) by f(t)/t in (2.2), then by
(2.27) the left hand side in (2.2) equals

_
0
exp
_
_
1
x
_
_
x
_
0
ln f(t)dt −
x
_
0
ln tdt
_
_
_
_
dx
= e

_
0
exp
_
_
1
x
x
_
0
ln f(t)dt
_
_
dx
x
.
Thus, (2.2) can equivalently be written in the maybe more natural form

_
0
exp
_
_
1
x
x
_
0
ln f(t)dt
_
_
dx
x
< 1

_
0
f(x)
dx
x
. (2.29)
In order to prove (2.29) we use the fact that the function f(u) = e
u
is convex
and Jensen’s inequality :

_
0
exp
_
_
1
x
x
_
0
ln f(t)dt
_
_
dx
x


_
0
1
x
2
_
_
x
_
0
f(t)dt
_
_
dx
=

_
0
f(t)
_
_

_
t
1
x
2
dx
_
_
dt
=

_
0
f(t)
dt
t
.
The strict inequality follows in the same way as in Proof 9.
28 Carleman type inequalities and Hardy type inequalities
for monotone functions
Remark 2.14. There are some great similarities between Proof 3 of Carle-
man’s inequality and Proof 10 of Pólya -Knopp’s inequality. After changing
order of summation and integration, respectively, in the concluding calcula-
tions we get

k=i
1
k (k + 1)
=
1
i
and

_
t
1
x
2
dx =
1
t
,
respectively, and this is crucial for the proofs. We also note that (2.29) does
not follow from Carleson’s inequality (2.12) since p = −1.
2.3 Final remarks about Torsten Carleman and his
work
Remark 2.15. A main reference concerning Torsten Carleman and his
mathematics is of course the book [30] of L. Gårding (see p. 233-276). In
this book Carleman is described in the following way: ”With Torsten Carle-
man (1892-1949) Sweden got their so far most outstanding mathematician.”
It is therefore not curious that Gårding spent the next 30 pages to describe
Carleman and his mathematical work and no other mathematician was given
even close to so much space in the book. It is remarkable that (2.1) was not
explicitly mentioned in Gårding’s book, which can depend on the fact that
he (as well as Carleman himself ) obviously regarded the inequality only as a
necessary tool to prove his important main results concerning quasianalytical
functions. However, as we have seen in this chapter, Carleman’s inequality
(2.1) and its continuous variant (Pólya-Knopp’s inequality (2.2)) has at-
tracted a lot of attention and it is even mentioned in the title of a number
of papers. See our list of references, Chapter 4 in the book [72] (with 174
references), Chapter 1 in the book [63] and the recently published paper [83]
(with 53 references). And the interest seems only to have increased during
the very last years.
Remark 2.16. (About the person Torsten Carleman). There is a lot of
interesting information in Gårding’s book [30] and some complementary in-
formation can be found in [1]. However, the main reference in this con-
nection is the recent biographic paper [67] by L. Maligranda. Tage Gillis
Torsten Carleman was born 8 July 1892. He defended his Ph.D. thesis 1917
at Uppsala University. In 1923 he was appointed as full professor at the
Lund University. Shortly after this, and on an initiative of Professor Gösta
Mittag-Leffler, he was called as professor at the University of Stockholm, in
Carleman’s inequality - history and proofs 29
1924. He died 1949. Carleman was a remarkable person and there are many
remarks concerning him (see e.g. Professor B. Kjellberg’s interesting and
very personal description in [58, p. 93]).
30 Carleman type inequalities and Hardy type inequalities
for monotone functions
Chapter 3
Carleman’s inequality - some new
sharpenings and generalizations
In this chapter we give some sharpenings and generalizations of (2.1) and
(2.2). We discuss and comment on these results and put them into the frame
presented in the previous chapter. We also include some new proofs and
results. In particular, we prove a new weight characterization of a general
weighted Carleman type inequality for the case 0 < p ≤ q < ∞, i.e. we prove
a necessary and sufficient condition on the sequence ¦b
k
¦

k=1
and ¦d
k
¦

k=1
so
that the inequality

k=1
(
k

a
1
a
2
...a
k
)
q
b
k
≤ C
_

k=1
a
p
k
d
k
_
1/p
holds for some finite and positive constant C and for all sequences ¦a
k
¦

k=1
of positive numbers. Moreover, we give upper and lower estimates of the
best constant C in the inequality (the corresponding operator norm).
We have already given some examples of sharpenings of Carleman’s in-
equality for finite sums (see for example remark 2.5 and (2.13), (2.19) and
(2.20)). In this chapter we give some additional sharpenings and generaliza-
tions of the Carleman and Pólya-Knopp inequalities. We start by noting the
following:
Remark 3.1. Proof 9 is of course similar to Proof 10 but it contains the
important information that (2.1) in fact can be equivalently rewritten in the
31
32 Carleman type inequalities and Hardy type inequalities
for monotone functions
form (2.29) with the best constant 1. By using this observation, and by
modifying the proof, we find that, in fact, the following more general theorem
holds (cf. [53], Theorem 4.1):
Theorem 3.1. Let 0 < b ≤ ∞. Let φ be a positive and strictly monotone
convex function on (a, c), −∞ ≤ a < c ≤ ∞. Then
b
_
0
φ
_
_
1
x
x
_
0
f(t)dt
_
_
dx
x
<
b
_
0
φ(f(x))
_
1 −
x
b
_
dx
x
(3.1)
for every realvalued function on (0, b) such that a < f (x) < c.
Proof. We use Jensen’s inequality, change the order of integration and get
b
_
0
φ
_
_
1
x
x
_
0
φ
−1
f(t)dt
_
_
dx
x

b
_
0
_
_
x
_
0
f(t)dt
_
_
1
x
2
dx
=
b
_
0
f (t)
_
_
b
_
t
1
x
2
dx
_
_
dt
=
b
_
0
f (t)
_
1
t

1
b
_
dt
=
b
_
0
f (t)
_
1 −
t
b
_
dt
t
.
We replace f (t) with φ(f(t)) and the inequality (3.1) is proved.
Remark 3.2. For b = ∞ the inequality (3.1) is strict under the condition
that the integral on the right hand side converges and Theorem 3.1 and the
proof essentially coincide with Theorem 4.1 in [53]. The case described in
Theorem 3.1 has been treated even more generally in [23].
Remark 3.3. By choosing φ(u) = e
u
and replacing f (x) with ln f (x) we see
that (3.1) becomes (2.29) and thereby the equivalent inequality (2.1) and by
choosing φ(u) = u
p
we find that (3.1) with b = ∞ implies Hardy’s inequality
in the form

_
0
_
_
1
x
x
_
0
f(t)dt
_
_
p
dx
x
<

_
0
f
p
(x)
dx
x
, p ≥ 1, (3.2)
Carleman’s inequality - some new sharpenings and gener-
alizations
33
which for the case p > 1 (after some substitutions) can be written in the
usual form (see (1.2))

_
0
_
_
1
x
x
_
0
g(t)dt
_
_
p
dx <
_
p
p −1
_
p

_
0
g
p
(x)dx, p > 1, (3.3)
where g (x) = f
_
x
1−1/p
_
x
−1/p
. Note especially that Hardy’s inequality writ-
ten in the form (3.2) holds even when p = 1 but that the inequality (3.3) then
has no meaning.
Remark 3.4. In particular, Theorem 3.1 implies the following improvements
of Pólya-Knopp’s and Hardy’s inequalities for finite intervals (0, b) , b < ∞:
b
_
0
exp
_
_
1
x
x
_
0
f(t)dt
_
_
dx ≤ e
b
_
0
f (x)
_
1 −
x
b
_
dx,
and
b
0
_
0
_
_
1
x
x
_
0
f(t)dt
_
_
p
dx ≤
_
p
p −1
_
p
b
0
_
0
_
1 −
_
x
b
0
_
(p−1)/p
_
dx,
respectively, where b
0
= b
p/(p−1)
and g (x) = f
_
x
(p−1)/p
_
x
−1/p
as before.
These inequalities has by the way been proved in the paper [25] (see also
[22]) with a different method which is built on the inequalities between mixed
means (cf. our Proof 7).The idea in this remark is further developed and
applied in [23].
Another interesting question which has been studied is to find general
weighted versions of the inequality (2.2). Partly guided by the development
concerning Hardy type inequalities (see for example the books [63] and [80])
one has asked:
Let 0 < p, q < ∞. Find necessary and sufficient conditions for the weights
(i.e. the positive and measurable functions) u(x) and v (x) so that
_
_

_
0
_
_
exp
_
_
1
x
x
_
0
ln f(t)dt
_
_
_
_
q
u(x) dx
_
_
1/q
≤ C
_
_

_
0
f
p
(x)u(x) dx
_
_
1/p
(3.4)
holds with a stable estimate of the operator norm (= the smallest constant
so that (3.4) holds).
The following results were recently proved by L.E. Persson and V.D.
Stepanov [84]:
34 Carleman type inequalities and Hardy type inequalities
for monotone functions
Theorem 3.2. Let 0 < p ≤ q < ∞. Then the inequality (3.4) holds if and
only if
D := sup
x>0
x
−1/p
_
_
x
_
0
w(s)ds
_
_
1/q
< ∞,
where
w(s) =
_
_
exp
_
_
1
s
s
_
0
ln
1
v (t)
dt
_
_
_
_
q/p
u(s)
and
D ≤ C ≤ e
1/p
D.
Theorem 3.3. Let 0 < q < p < ∞. Then the inequality (3.4) holds if and
only if
B :=
_
_
_

_
0
_
_
1
x
x
_
0
w(s)ds
_
_
r/p
w(x)dx
_
_
_
1/r
< ∞,
where w(x) is defined as in Theorem 3.2, and C ≈ B.
Remark 3.5. These results were proved in [84]. We also refer the reader to
some earlier results of this type which can be found in the papers [46], [73]
and [82]. A further development of Theorem 3.2 and 3.3 can be found in [76]
and the Ph.D thesis by M. Nassyrova [75].
We will now present an example of a new sharpening of the Carleman
inequality (2.1) (see [53], Theorem 2.1).
Theorem 3.4. Let ¦a
k
¦

1
, be a sequence of positive numbers and put x
i
=
ia
i
_
1 +
1
i
_
i
, i = 1, 2, .... Then the following holds for N ∈ Z
+
:
N

k=1
G
k
+
N

k=1
l
k
k (k + 1)

N

k=1
_
1 −
k
N + 1
__
1 +
1
k
_
k
a
k
, (3.5)
where
G
k
:=
k

a
1
a
2
...a
k
and l
k
:=
[x]

i=1
__
x

k−i+1

_
x

i
_
2
.
Here [x] is the usual integer part of x and ¦x

k
¦ is the sequence ¦x
k
¦
rearranged in decreasing order.
Carleman’s inequality - some new sharpenings and gener-
alizations
35
Remark 3.6. For previous results of this type we also refer to the papers
[2], [3], [4], [83], [104], [105] and the references found there. We note that
by using the estimate l
k
≥ 0,
_
1 +
1
k
_
k
< e and letting N → ∞ we get (2.1)
as a special case of (3.5).
Remark 3.7. Refinements of Carleman’s inequality (2.1) with e replaced by
_
1 +
1
k
_
k
has been known since at least 1967 (see [90] and [91] and compare
with our Proof 6). We also note that the factor 1 −
k
N+1
in (3.5) means that
the ”usual” sum on the right hand side of the inequality has been replaced
by the equivalent Cesarosum, i.e. we have calculated the arithmetic mean of
partial sums. This mean value is of course strictly less than the ”usual” sum
since the terms are positive.
Remark 3.8. In the paper [104] P. Yan and G. Sun proved that Carleman’s
inequality (2.1) can be improved in the following way:

k=1
_
k

i=1
a
i
_
1/k
< e

k=1
_
1 +
1
k + 1/5
_
−1/2
a
k
. (3.6)
This result easily follows from (3.5) by estimating the important factor
_
1 +
1
k
_
k
in the following way:
_
1 +
1
k
_
k
≤ e
_
1 +
1
k +c

_
−1/2
(3.7)
where c

=
8−e
2
e
2
−4
≈ 0, 1802696 < 1/5. The inequality (3.7) does not hold for
numbers smaller than c

. (See [53], Remark 12). This means that by using
(3.5) we see that the Yan-Sun inequality (3.6) actually can be replaced with
the sharper inequality

k=1
_
k

i=1
a
i
_
1/k
+

k=1
b
k
k (k + 1)
< e

k=1
_
1 +
1
k +c

_
−1/2
a
k
.
Remark 3.9. The factor
_
1 +
1
k
_
k
has also been of interest in some other
new papers. For example M. Gyllenberg and P. Yan recently proved in the
paper [39] that
_
1 +
1
k
_
k
= e
_
1 −

n=1
a
n
(1 +k)
n
_
where all a
n
are positive and can be calculated recursively. For example
a
1
=
1
2
, a
2
=
1
24
, a
3
=
1
48
etc. This answers an earlier question raised by
Yang (see [105]).
36 Carleman type inequalities and Hardy type inequalities
for monotone functions
Remark 3.10. We have noted before that Carleson’s inequality (2.12) gives
both (2.1) and (2.2) as special cases. Another inequality with that property
has recently been proved, namely the following see ([53], theorem 3.1):
B
_
0
exp
_
_
1
M(x)
x
_
0
ln f(t)dM(t)
_
_
dM(x)
+e
B
_
0
_
1 −
M

(x)
M(x)
_
f(x)dM(x)
≤ e
B
_
0
_
1 −
M

(x)
M(B)
_
f(x)dM(x).
Here B ∈ R
+,
M(x) is a right continuous and increasing function on (0, ∞)
and M

(x) is a special defined function with the property that M

(x) ≤ M(x).
By using this theorem with M(x) = x and B = ∞ we get (2.2) and by using
it with
M(x) =
_
1
2
, 0 ≤ x ≤ 1
k , k ≤ x ≤ k + 1, k = 1, 2, ...
we get a refinement of (2.1).
In view of questions raised in connection to (3.4) it is natural to ask the
following connected to (2.1): Let 0 < p, q < ∞. Find necessary and sufficient
conditions on the positive sequence ¦b
k
¦

1
and ¦d
k
¦

1
such that
_

k=1
(
k

a
1
a
2
...a
k
)
q
b
k
_1
q
≤ C
_

k=1
a
p
k
d
k
_1
p
(3.8)
holds. We will prove the following general weighted Carleman inequality:
Theorem 3.5. For k = 1, 2..., let a
k
≥ 0, b
k
≥ 0 and d
k
≥ 0. If 0 < p ≤
q < ∞, then the inequality (3.8) holds for some finite constant C > 0, if and
only if
B
1
= sup
N>0
N
−1/p
_
_
N+1

k=1
_
k

i=1
d
i
_
−q/kp
b
k
_
_
1/q
< ∞. (3.9)
Remark 3.11. This result was also stated and proved in the recent Ph.D.
thesis [103] of A. Wedestig. However, our proof here is different.
Carleman’s inequality - some new sharpenings and gener-
alizations
37
Proof. Assume that (3.9) holds. Let first w
1
= 0, and replace a
k
with ˜ a
k
d
−1/p
k
in (3.8). Then (3.8) is equivalent to
_
_

k=1
_
k

i=1
˜ a
i
_
q/k
_
k

i=1
d
i
_
−q/kp
b
k
_
_
1
q
≤ C
_

k=1
˜ a
p
k
_1
p
or, if w
k
=
_

k
i=1
d
i
_
−q/kp
b
k
,
I
1/q
:=
_
_

k=1
_
k

i=1
˜ a
i
_
q/k
w
k
_
_
1
q
≤ C
_

k=1
˜ a
p
k
_1
p
. (3.10)
Now if ¦a

k
¦

k=1
is decreasing rearrangement of ¦˜ a
k
¦

k=1
, then obviously
_
_

k=1
_
k

i=1
˜ a
i
_
q/k
w
k
_
_
1
q

_
_

k=1
_
k

i=1
a

i
_
q/k
w
k
_
_
1
q
. (3.11)
Let f

(x) = a

k
and w(x) = w
k
for x ∈ [k −1, k) , k = 1, 2, .... Then
_
_

k=1
_
k

i=1
a

i
_
q/k
w
k
_
_
1
q
(3.12)
=

k=1
k
_
k−1
_
exp
_
k

i=1
log a

1
k
i
__
q
w
k
dx
=

k=1
k
_
k−1
_
exp
_
1
k
k−1

i=1
log a

i
+
1
k
log a

k
__
q
w
k
dx

k=1
k
_
k−1
_
exp
_
1
x
k−1

i=1
log a

i
+
x −(k −1)
x
log a

k
__
q
w
k
dx
=

k=1
k
_
k−1
_
_
exp
_
_
1
x
x
_
0
ln f

(t)dt
_
_
_
_
q
w(x)dx
=
_
_

_
0
_
_
exp
_
_
1
x
x
_
0
ln f

(t)dt
_
_
_
_
q
w(x)dx
_
_
1/q
.
38 Carleman type inequalities and Hardy type inequalities
for monotone functions
Moreover, it follows from Theorem 3.2 that if
D := sup
x>0
x
−1/p
_
_
x
_
0
w(t)dt
_
_
1/q
< ∞, (3.13)
then the inequality
_
_

_
0
_
_
exp
_
_
1
x
x
_
0
ln f

(t)dt
_
_
_
_
q
w(x)dx
_
_
1/q
≤ C
_
_

_
0
f
∗p
(x) dx
_
_
1/q
(3.14)
holds, and if C is the best possible constant in (3.14), then
C ≤ e
1/p
D.
Hence, by combining (3.11), (3.12) and (3.14) we have
_
_

k=1
_
k

i=1
˜ a
i
_
q/k
w
k
_
_
1
q
≤ C
_
_

_
0
f
∗p
(x) dx
_
_
1/q
= C
_

k=1
a
∗p
k
_1
p
= C
_

k=1
˜ a
p
k
_1
p
i.e. (3.10) and thus (3.8) holds whenever (3.13) holds. In fact, for N ∈ Z
+
,
we have
sup
N<x<N+1
x
−1/p
_
_
x
_
0
w(t)dt
_
_
1/q
≤ N
−1/p
_
N+1

k=1
w
k
_
1/q
.
Hence
sup
x>0
x
−1/p
_
_
x
_
0
w(t)dt
_
_
1/q
(3.15)
≤ sup
N>0
N
−1/p
_
_
N+1

k=1
_
k

i=1
d
i
_
−q/kp
b
k
_
_
1/q
= B
1
,
so that, according to (3.9), (3.13) holds.
Carleman’s inequality - some new sharpenings and gener-
alizations
39
If w
1
,= 0, then, by using what we just have proved and an elementary
inequality, we have
I
1/q
=
_
_
˜ a
q
1
w
1
+

k=2
_
k

i=1
˜ a
i
_
−q/k
w
k
_
_
1/q
(3.16)
≤ max
_
1, 2
1/q−1
__
w
1/q
1
+C
_
_

k=2
˜ a
p
k
_
1/p
.
Therefore, by using (3.9), (3.13), (3.15) and (3.16), we conclude that (3.8)
holds.
On the contrary, assume that (3.8) (and thus (3.10)) holds for all positive
sequences. In particular, let ˜ a
k
= 1, k = 1, ..., N +1 and ˜ a
k
= 0, k > N +1.
Then the left hand side in (3.10) can be estimated as follows:
_
_

k=1
_
k

i=1
˜ a
i
_
q/k
w
k
_
_
1/q

_
_
N+1

k=1
_
k

i=1
˜ a
i
_
q/k
w
k
_
_
1/q
=
_
N+1

k=1
w
k
_
1/q
.
For the right hand side we have
_

k=1
˜ a
p
k
_
1/p
=
_
N+1

k=1
1
_
1/p
= (N + 1)
1/p
≤ (2N)
1/p
and in view of (3.8), it follows that
(2N)
−1/p
_
N+1

k=1
w
k
_
1/q
≤ C,
so that (3.9) holds. The proof is complete.
Remark 3.12. We note that our proof gives that if b
1
= 0, then C ≈ B
1
, or
more exactly,
2
−1/p
B
1
≤ C ≤ e
1/p
B
1
.
Remark 3.13. In [46] the weighted Carleman’s inequality (3.8) was char-
acterized with another condition than B
1
.
40 Carleman type inequalities and Hardy type inequalities
for monotone functions
Chapter 4
A Sawyer duality principle for
radially monotone functions in R
n
As mentioned in the introduction a remarkable duality principle for positive
decreasing functions of one variable was proved by E. Sawyer in [93], and also
some applications are well-known. Here we also refer to the useful proof and
ideas concerning this principle presented by V. Stepanov in [99]. Moreover,
it is natural to look for extensions to functions of several variables. Such
generalizations were recently obtained in [8], [10] and [11]. For 0 < q < p <
∞,
1
r
=
1
q

1
p
, it was shown in [11] that
sup
0≤f↓
__
R
n
f
q
u
_1
q
__
R
n
f
p
v
_1
p

__
R
n
u
_1
q
__
R
n
v
_1
p
+ sup
0≤h↓
_
_
_

0
_
_
D
h,t
u
_r
q
d
_
_
D
h,t
v
_

r
p
_
_
1
p
,
(4.1)
where u and v are weights, i.e. positive and locally integrable functions on
R
n
. For the case 0 < p ≤ q < ∞ cf. [10] .
If q = 1 < p < ∞, then (4.1) with n = 1 is a variant of the duality
theorem given in [93] (cf. also [99]).
An explicit form of (4.1) for n ∈ Z
+
was given in [8], but only when v
is of product type, that is weights of the form v (x) = v (x
1
, x
2
, ..., x
n
) =
v
1
(x
1
) v
2
(x
2
) ...v
n
(x
n
), v
i
≥ 0, i = 1, 2, ..., n.
We say that f : R
n
→ R is a radially decreasing (increasing) function
if f (x) = f (y) when [x[ = [y[ and f (x) ≥ f (y) (f (x) ≤ f (y)) when
41
42 Carleman type inequalities and Hardy type inequalities
for monotone functions
[x[ < [y[ and we write f ↓ r and f ↑ r , respectively (see also [65] for further
explanations and applications of these notions).
In this Chapter we will prove a duality formula of the type (4.1) for
radially decreasing functions and with general weights (see Theorem 4.3).
We also state the corresponding result for radially increasing functions (see
Theorem 4.5). In particular, these results imply that we can describe map-
ping properties of operators defined on cones of such monotone functions.
Moreover, we point out that these results can also be used to describe map-
ping properties of operators between some corresponding general weighted
multidimensional Lebesgue spaces (see Theorem 4.6 and cf. also [99] for the
case n = 1). We illustrate this useful technique for the identity operator (see
Corollary 4.7) and for the Hardy integral operator over the n-dimensional
sphere (see Corollary 4.9).
4.1 Preliminary results
Consider the Hardy integral operator H of the form
(Hf) (x) =
_
B(x)
f(y)dy,
where B(x) is the ball in R
n
centered at the origin and with radius [x[. We
need the following well-known n-dimensional form of Hardy’s inequality (see
[27]):
Theorem 4.1. Let W and U be weights on R
n
and 1 < p ≤ q < ∞.
(i) The inequality
_
_
R
n
W(x)
_
_
B(x)
f(y)dy
_
q
dx
_
1
q
≤ c
__
R
n
U(x)f
p
(x)dx
_1
p
(4.2)
holds for f ≥ 0 if and only if
a := sup
α>0
_
_
|x|≥α
W(x)dx
_1
q
_
_
|x|≤α
U
1−p

(x)dx
_ 1
p

< ∞ .
Moreover, if c is the smallest constant for which (4.2) holds, then
a ≤ c ≤ ap

1
p

p
1
q
.
A Sawyer duality principle for radially monotone func-
tions in R
n
43
(ii) The inequality
_
_
R
n
W(x)
_
_
R
n
\B(x)
f(y)dy
_
q
dx
_
1
q
≤ c
__
R
n
U(x)f
p
(x)dx
_1
p
(4.3)
holds if and only if
b := sup
α>0
_
_
|x|≤α
W(x)dx
_1
q
_
_
|x|≥α
U
1−p

(x)dx
_ 1
p

< ∞ .
Moreover, if c is the smallest constant for which (4.3) holds, then
b ≤ c ≤ bp

1
p

p
1
q
.
In particular we need the following special case of Theorem 4.1 (ii):
Lemma 4.2. Let v be a weight function, V (x) =
_
B(x)
v(y)dy and 1 < p <
∞. Then for f ≥ 0
_
R
n
v(x)
_
_
R
n
\B(x)
f(y)dy
_
p
dx ≤ p
_
R
n
f
p
(x)V
p
(x) v
1−p
(x)dx (4.4)
is satisfied.
Proof. Apply Theorem 4.1 (ii) with W(x) = v(x),
U(x) = v
1−p
(x)
_
_
B(x)
v(y)dy
_
p
and q = p. Denote
v
n
(s) =
_
Σ
n−1
s
n−1
v(sσ)dσ , (4.5)
where Σ
n−1
as usual denotes the unit sphere in R
n
, s ∈ R. We note that
b = sup
α>0
_
_
|x|≤α
v(x)dx
_1
p
_
_
_
|x|≥α
_
_
B(x)
v(y)dy
_
−p

v(x)dx
_
_
1
p

= sup
α>0
V
1
p
(α)
_
_
_
|x|≥α
_
_
|x|
0
v
n
(s) ds
_
−p

v (x) dx
_
_
1
p

= sup
α>0
V
1
p
(α)
_
_

α
_
Σ
n−1
t
n−1
__
t
0
v
n
(s) ds
_
−p

v (tδ) dtdδ
_
1
p

44 Carleman type inequalities and Hardy type inequalities
for monotone functions
= sup
α>0
V
1
p
(α)
_
_

α
__
t
0
v
n
(s) ds
_
−p

_
_
Σ
n−1
t
n−1
v (tσ) dσ
_
dt
_
1
p

= sup
α>0
V
1
p
(α)
_
_

α
__
t
0
v
n
(s) ds
_
−p

v
n
(t) dt
_
1
p

= sup
α>0
V
1
p
(α)
_
_

α
_
d
dt
__
t
0
v
n
(s) ds
_
−p

+1
_
1
1 −p

dt
_
1
p

≤ sup
α>0
1
(p

−1)
1
p

V
1
p
(α)
__
α
0
v
n
(t) dt
_
−p

+1
p

= sup
α>0
1
(p

−1)
1
p

V
1
p
(α) V
1
p

−1
(α) =
1
(p

−1)
1
p

< ∞ .
Therefore, by Theorem 4.1 (ii), (4.4) holds with a constant
c ≤
1
(p

−1)
1
p

(p

)
1
p

p
1
p
= p and the proof is complete.
4.2 The Duality Principle for radially monotone func-
tions
In the sequel we sometimes delete the integration variable and write e.g.
_
R
n
fg instead of
_
R
n
f (x) g (x) dx when it can not be misunderstood. More-
over, as usual, |g|
1
= |g|
L
1
=
_
R
n
[g (x)[ dx. Our main result in this section
reads:
Theorem 4.3. Suppose that v is a weight on R
n
and 1 < p < ∞. If f
is a positive radially decreasing function on R
n
and g a positive measurable
function on R
n
, then
C (g) := sup
f↓r
_
R
n
fg
__
R
n
f
p
v
_1
p
≈ I
1
+I
2
, (4.6)
where
I
1
= |v|
−1
p
1
|g|
1
and
I
2
=
__
R
n
G(t)
p

V (t)
−p

v (t) dt
_ 1
p

with V (t) =
_
B(t)
v (x) dx and G(t) =
_
B(t)
g (x) dx.
A Sawyer duality principle for radially monotone func-
tions in R
n
45
Remark 4.1. Theorem 4.1 for the case n = 1 is just Theorem 1 in [93].
However, our proof below is based on the technique in [99] and our investi-
gation in Section 4.1.
Proof. If f ≡ c > 0, then
C (g) ≥ sup
f=c
_
R
n
+
fg
_
_
R
n
+
f
p
v
_1
p
=
|g|
1
|v|
1
p
1
= I
1
.
Moreover, we use the test function f (x) =
_
|t|>|x|
h(t)dt, where h(t) ≥ 0
(note that f is radially decreasing), Lemma 4.2 and usual duality in L
p
v
-
space to find that
C (g) = sup
0≤f↓
_
R
n
f (x) g (x) dx
__
R
n
f
p
(x) v (x) dx
_1
p
≥ sup
h≥0
_
R
n
_
_
|x|<|t|
h(t) dt
_
g (x) dx
_
_
R
n
_
_
|x|<|t|
h(t) dt
_
p
v (x) dx
_1
p
= sup
h≥0
_
R
n
h(t)
_
|x|<|t|
g (x) dxdt
_
_
R
n
_
_
|x|<|t|
h(t) dt
_
p
v (x) dx
_
1
p

1
p
sup
h≥0
_
R
n
h
_
|x|<|t|
g
__
R
n
h
p
V
p
v
1−p
_1
p
=
1
p
sup
h≥0
_
R
n
_
hV
v
_
G
V
v
_
_
R
n
_
hV
v
_
p
v
_1
p
=
1
p
_
_
R
n
_
G
V
_
p

v
_ 1
p

= I
2
.
To prove the upper bound of C (g) we use the monotonicity of f. Since f is
46 Carleman type inequalities and Hardy type inequalities
for monotone functions
radially decreasing we have
_
R
n
gf =
_
R
n
gfV
1
V
=
_

0
_
Σ
n−1
f (x) g (x)
1
V (x)
_
_
B(x)
v(t)dt
_
dx
=
_
R
n
v (t)
_
_
|x|>|t|
f (x) g (x)
1
V (x)
dx
_
dt

_
R
n
v (t) f (t)
_
_
|x|>|t|
g (x)
1
V (x)
dx
_
dt. (4.7)
To estimate the inner integral we define g
n
(s) and v
n
(s) analogously to
(4.5), note that V (x) = V ([x[) and find that
_
|x|>|t|
g (x)
1
V (x)
dx =
_

|t|
s
n−1
_
Σ
n−1
g (sσ)
1
V (sσ)
dsdσ
=
_

|t|
_
_
Σ
n−1
s
n−1
g (sσ) dσ
_
1
V (s)
ds =
_

|t|
g
n
(s)
1
V (s)
ds
=
_
1
V (s)
_
s
0
g
n
(z) dz
_

|t|
+
+
_

|t|
1
V
2
(s)
_
_
Σ
n−1
s
n−1
v (sσ) dσ
_
__
s
0
g
n
(z) dz
_
ds

1
V (∞)
_

0
g
n
(z) dz
. ¸¸ .
K
1
+
_

|t|
1
V
2
(s)
v
n
(s)
__
s
0
g
n
(z) dz
_
ds
. ¸¸ .
K
2
.
Hence the inner integral can be estimated by K
1
+ K
2
and by substituting
this into (4.7) and applying Hölder’s and Minkowski’s inequalities we get
that
_
R
n
fg ≤
_
R
n
fv (K
1
+K
2
) ≤
__
R
n
f
p
v
_1
p
__
R
n
(K
1
+K
2
)
p

v
_ 1
p


__
R
n
f
p
v
_1
p
_
__
R
n
K
p

1
v
_ 1
p

+
__
R
n
K
p

2
v
_ 1
p

_
.
A Sawyer duality principle for radially monotone func-
tions in R
n
47
Moreover,
__
R
n
K
p

1
v
_ 1
p

=
_
_
R
n
_
1
V (∞)
_

0
g
n
(z) dz
_
p

v (x) dx
_ 1
p

=
1
V (∞)
_
_

0
_
Σ
n−1
s
n−1
g (sσ) dσds
_
__
R
n
v (x) dx
_ 1
p

= |v|
−1+
1
p

1
|g|
1
= |v|

1
p
1
|g|
1
= I
1
,
and, according to Lemma 4.2,
__
R
n
K
p

2
v
_ 1
p

=
_
_
_
R
n
_
_

|t|
1
V
2
(s)
v
n
(s)
__
s
0
g
n
(z) dz
_
ds
_
p

v (t) dt
_
_
1
p

=
_
_
_
R
n
_
_

|t|
_
Σ
n−1
s
n−1
v (sσ)
V
2
(s)
_
_
s
0
_
Σ
n−1
z
n−1
g (zδ) dzdδ
_
dσds
_
p

v
_
_
1
p

=
_
_
_
R
n
_
_
R
n
\B(t)
v (x)
V
2
(x)
_
B(x)
g (y) dy
_
p

v (t) dt
_
_
1
p

≤ p

_
_
R
n
V (t)
−p

_
_
B(|x|)
g
_
v (t) dt
_ 1
p

= p

I
2
.
The upper bound of C (g) follows by combining the last estimates and the
proof is complete.
Remark 4.2. According to our proof we see that the duality constant C (g)
in (4.6) can in fact be estimated in the following more precise way:
max (I
1
, I
2
) ≤ C (g) ≤ I
1
+p

I
2
.
In particular we have the following useful information:
48 Carleman type inequalities and Hardy type inequalities
for monotone functions
Corollary 4.4. Let the assumptions in Theorem 4.3 be satisfied and
_
R
n
v =
∞. Then
I
2
≤ sup
f↓r
_
R
n
fg
__
R
n
f
p
v
_1
p
≤ p

I
2
. (4.8)
The proof above is self-contained and not depending directly on the one-
dimensional result (only on similar arguments as V.D. Stepanov used when
he proved this case). Here we give another shorter proof where we directly
use the (Sawyer) onedimensional result.
Proof. Make the following changes of variables
t = sσ and x = yτ, (4.9)
where s, y ∈ (0, ∞) and σ, τ ∈

n−1
. By using that f (sσ) = f (s) since f
is radial, we get:
_
R
n
fg
__
R
n
f
p
v
_1
p
=
_

0
_

n−1
f (sσ) g (sσ) s
n−1
dσds
_
_

0
_

n−1
f
p
(sσ) v (sσ) s
n−1
dσds
_
1/p
=
_

0
f (s)
¯
G(s) ds
_
_

0
f
p
(s)
¯
V (s) ds
_
1/p
,
and hence, by using the (Sawyer) one-dimensional result we find that
_
R
n
fg
__
R
n
f
p
v
_1
p
≈ (4.10)
__

0
¯
V (s) ds
_

1
p
__

0
¯
G(s) ds
_
+
_
_
_
_

0
_
_
s
0
¯
G(y) dy
_
p

_
_
s
0
¯
V (y) dy
_
p

¯
V (s) ds
_
_
_
1
p

:= I.
Moreover,
A Sawyer duality principle for radially monotone func-
tions in R
n
49
I =
_
_

0
_

n−1
v (sσ) s
n−1
dσ ds
_

1
p
_
_

0
_

n−1
g (sσ) s
n−1
dσds
_
+
_
_
_
_

0
_
_
s
0
_

n−1
g (yτ) y
n−1
dτdy
_
p

_
_
s
0
_

n−1
v (yτ) y
n−1
dτdy
_
p

_

n−1
v (sσ) s
n−1
dσds
_
_
_
1
p

=
__
R
n
v (t) dt
_

1
p
__
R
n
g (t) dt
_
+
_
_
_
_
R
n
_
_
B(t)
g (x) dx
_
p

_
_
B(t)
v (x) dx
_
p

v (t) dt
_
_
_
1
p

(4.11)
The proof follows by combining (4.10) and (4.11).
For completeness and our applications we also state the corresponding
result for radially increasing functions in R
n
:
Theorem 4.5. Suppose that v is a weight on R
n
and 1 < p < ∞. If f
is a positive radially increasing function on R
n
and g a positive measurable
function on R
n
, then
D(g) := sup
f↑r
_
R
n
fg
__
R
n
f
p
v
_1
p
≈ I
1
+I
3
,
where
I
1
= |v|
−1
p
1
|g|
1
,
and
I
3
=
__
R
n
G
1
(t)
p

V
1
(t)
−p

v (t) dt
_ 1
p

,
with G
1
(t) =
_
R
n
\B(t)
g (x) dx and V
1
(t) =
_
R
n
\B(t)
v (x) dx.
Proof. We now use Theorem 4.1 (i) (instead of (ii) as in the proof of Lemma
4.2) and obtain as in the proof of (4.4):
_
R
n
v(x)
_
_
B(x)
f(y)dy
_
p
dx ≤ p
_
R
n
f
p
(x)V
p
1
(x) v
1−p
(x)dx.
By using this estimate the proof follows similarly as the proof of Theorem
4.3 so we leave out the details.
50 Carleman type inequalities and Hardy type inequalities
for monotone functions
Remark 4.3. In fact, similar to Remark 4.2 and Corollary 4.4, we find that
max (I
1
, I
3
) ≤ D(g) ≤ I
1
+p

I
3
and if, in addition to the assumptions in Theorem 4.5,
_
R
n
v = ∞, then
I
3
≤ sup
f↑r
_
R
n
fg
__
R
n
f
p
v
_1
p
≤ p

I
3
.
4.3 Further results and applications
Let T be an integral operator defined on the cone of functions f : R
n
→ R,
which are radially decreasing (0 < f ↓ r) and let T

be the adjoint operator.
Then our results imply the following useful duality result:
Theorem 4.6. Let 1 < p, q < ∞, u, v be weights on R
n
with
_
R
n
v (x) dx = ∞. Then the inequality
__
R
n
(Tf(x))
q
u(x) dx
_1
q
≤ c
__
R
n
f
p
(x)v (x) dx
_1
p
(4.12)
holds for all f ↓ r if and only if
_
_
_
R
n
_
_
B(x)
T

g (y) dy
_
p

V
−p

(x) v (x) dx
_
_
1
p

(4.13)
≤ c
__
R
n
g
q

(x)u
1−q

(x) dx
_ 1
q

holds for every positive measurable function g.
Proof. Assume first that (4.12) holds for all 0 < f ↓ r. Then, by using
Corollary 4.4, duality and Hölder’s inequality, we find that
_
R
n
_
_
B(x)
T

g (y) dy
_
p

V
−p

(x) v (x) dx ≤ sup
f↓r
_
R
n
f (x) T

g (x) dx
__
R
n
f
p
(x) v (x) dx
_1
p
= sup
f↓r
_
R
n
Tf (x) g (x) dx
__
R
n
f
p
(x) v (x) dx
_1
p
A Sawyer duality principle for radially monotone func-
tions in R
n
51
≤ sup
f↓r
__
R
n
(Tf (x))
q
u(x) dx
_1
q
_
_
R
n
g
q

(x) u
−q

q
(x) dx
_ 1
q

__
R
n
f
p
(x) v (x) dx
_1
p
= c
__
R
n
g
q

(x) u
1−q

(x) dx
_ 1
q

.
On the contrary assume that (4.13) holds for all g ≥ 0. Then, by using
Corollary 4.4 again, we have that
p

c
__
R
n
(g (x))
q

(u(x))
1−q

dx
_ 1
p

≥ p

_
_
_
R
n
_
_
B(x)
T

g (t) dt
_
p

V
−p

(x) v (x) dx
_
_
1
p


_
R
n
f (x) T

g (x) dx
__
R
n
f
p
(x) v (x) dx
_1
p
=
_
R
n
Tf (x) g (x) dx
__
R
n
f
p
(x) v (x) dx
_1
p
for each fixed 0 < f ↓ r. Therefore we find that
_
R
n
h(x) Tf (x) u
1
q
(x) dx ≤ p

c
__
R
n
f
p
(x) v (x) dx
_1
p
, (4.14)
where
h(x) =
g (x) u

1
q
(x)
_
_
R
n
_
g (x) u

1
q
(x)
_
q

dx
_ 1
q

.
Since |h|
L
q
= 1 we obtain (4.12) by taking supremum in (4.14) and using
usual duality in L
p
-spaces.
Remark 4.4. By modifying the proof above we see that a similar dual-
ity result also holds for positive radially increasing functions. In fact, in
this case we just need to replace
_
B(x)
by
_
R
n
\B(x)
and V (x) by V
1
(x) =
_
R
n
\B(x)
v (x) dx in (4.13).
For example when T is the identity operator we obtain the following
result:
52 Carleman type inequalities and Hardy type inequalities
for monotone functions
Corollary 4.7. Let 1 < p ≤ q < ∞ and suppose that u, v are weights on R
n
with
_
R
n
v = ∞ and V (x) =
_
B(x)
v (y) dy.
a) The following conditions are equivalent:
i) The inequality
__
R
n
f
q
u
_1
q
≤ c
__
R
n
f
p
v
_1
p
(4.15)
is satisfied for all 0 ≤ f ↓ r.
ii) The inequality
_
_
_
R
n
_
_
B(x)
g (y) dy
_
p

V
−p

(x) v (x) dx
_
_
1
p

(4.16)
≤ c
__
R
n
g
q

(x)u
1−q

(x) dx
_ 1
q

holds for all g ≥ 0.
iii)
sup
α>0
_
_
B(α)
v(x)dx
_

1
p
_
_
B(α)
u(x)dx
_1
q
< ∞ .
b) If V
1
(x) =
_
R
n
\B(x)
v (t) dt, then for 0 ≤ f ↑ r (4.15) is equivalent to that
_
_
_
R
n
_
_
R
n
\B(x)
g (y) dy
_
p

V
−p

1
(x) v (x) dx
_
_
1
p

≤ c
__
R
n
g
q

(x)u
1−q

(x) dx
_ 1
q

which in turn is equivalent to
sup
α>0
_
_
R
n
\B(α)
v(x)dx
_

1
p
_
_
R
n
\B(α)
u(x)dx
_1
q
< ∞ .
Proof. a) The equivalence of i) and ii) is just a special case of Theorem 4.6.
Moreover, the fact that ii) and iii) are equivalent follows from Theorem 4.1
A Sawyer duality principle for radially monotone func-
tions in R
n
53
(i) with f replaced by g, q by p

, p by q

, W by vV
−p

and U by u
1−q

. In
fact, then (4.16) is equivalent to (note that (1 −q) (1 −q

) = 1 )
sup
α>0
_
_
|x|>α
vV
−p

_ 1
p

_
_
|x|<α
u
_1
q
= sup
α>0
_
_
|x|<α
v
_

1
p
(p

−1)
1
p

_
_
|x|<α
u
_1
q
< ∞.
The proof of b) follows similarly by just using Remark 4.4 and Theorem 4.1
(ii).
Remark 4.5. The equivalence of i) and iii) can also be proved using the
technique from [10]. For the case q < p cf. also [11].
The next result concerns the multidimensional Hardy operator, defined
on the cone of radially decreasing functions in R
n
.
Proposition 4.8. Let 1 < p ≤ q < ∞ and suppose that u and v are weights
on R
n
with
_
R
n
v = ∞ and V (x) =
_
B(x)
v (x) dx. If 0 ≤ f is a radially
decreasing function in R
n
, then
_
_
R
n
_
_
B(x)
f (y) dy
_
q
u(x) dx
_1
q
≤ c
__
R
n
f
p
(x)v (x) dx
_1
p
(4.17)
is satisfied if and only if the following conditions hold:
sup
α>0
_
_
B(α)
v(x)dx
_

1
p
_
_
B(α)
u(x) [B(x)[
q
dx
_1
q
< ∞ (4.18)
and
sup
α>0
_
_
B(α)
[B(x)[
p

V
−p

(x) v (x) dx
_ 1
p

_
_
R
n
\B(α)
u(x)dx
_1
q
< ∞.
(4.19)
Here, as usual, [B(x)[ denotes the Lebesgue measure of the ball with
center at 0 and with radius [x[.
Remark 4.6. For the case n = 1 this result is due to V. Stepanov (see [99],
Theorem 2).
Proposition 4.8 can be proved by using the method of reduction to the
one-dimensional case but here we present an independent proof:
54 Carleman type inequalities and Hardy type inequalities
for monotone functions
Proof. Since Tf (x) =
_
B(x)
f (t) dt its conjugate T

is defined by T

g (x) =
_
R
n
\B(x)
g (t) dt. Assume first that (4.18) and (4.19) hold. We note that,
according to Theorem 4.6, (4.17) for 0 < f ↓ r is equivalent to (4.13) for
arbitrary g ≥ 0. Moreover, to be able to characterize weights for which
(4.13) is satisfied, we first compute:
_
B(x)
T

g =
_
B(x)
_
_
|y|>|z|
g (y) dy
_
dz
=
_
B(x)
_
_

|z|
_
Σ
n−1
t
n−1
g (tδ) dδdt
_
dz =
_
B(x)
_
_

|z|
g
n
(t) dt
_
dz
=
_
|x|
0
_
Σ
n−1
s
n−1
__

s
g
n
(t) dt
_
dσds
= [Σ
n−1
[
_
_
|x|
0
s
n−1
_
_
|x|
s
g
n
(t) dt
_
ds +
_
|x|
0
_

|x|
g
n
(t) dtds
_
= [Σ
n−1
[
_
|x|
0
__
t
0
s
n−1
ds
_
g
n
(t) dt +[Σ
n−1
[
_
|x|
0
ds
_

|x|
g
n
(t) dt
=
_
|x|
0
_
_
Σ
n−1

_
t
0
s
n−1
ds
_
Σ
n−1
t
n−1
g (tδ) dδ
_
dt
+
_
_
|x|
0
_
Σ
n−1
dσds
_
_

|x|
_
Σ
n−1
t
n−1
g (tδ) dδdt
=
_
B(x)
[B(y)[ g (y) dy
. ¸¸ .
I
1
+[B(x)[
_
R
n
\|B(x)|
g (y) dy
. ¸¸ .
I
2
= I
1
(x) +I
2
(x) .
This means that (4.17) holds if and only if
__
R
n
(I
1
(x) +I
2
(x))
p

V
−p

(x) v (x) dx
_ 1
p

(4.20)
≤ c
__
R
n
g
q

(x) U
1−q

(x) dx
_ 1
q

.
Moreover, by Theorem 4.1 (i) with q replaced by p

and p replaced by q

, we
A Sawyer duality principle for radially monotone func-
tions in R
n
55
obtain that
__
R
n
(I
1
(x))
p

V
−p

(x) v (x) dx
_ 1
p

(4.21)
=
_
_
_
R
n
_
_
B(x)
[B(y)[ g (y) dy
_
p

V
−p

(x) v (x) dx
_
_
1
p

≤ c
_
_
R
n
([B(x)[ g (x))
q
u(x)
1−q

[B(x)[
q

dx
_1
q
= c
__
R
n
g (x)
q

u(x)
1−q

dx
_ 1
q

,
which holds because, according to (4.18),
sup
α>0
_
_
_
R
n
\B(α)
v (x)
_
_
B(x)
v (y) dy
_
−p

dx
_
_
1
p

_
_
_
B(α)
_
u(y)
1−q

[B(y)[
q

_
1−q
dy
_
_
1
q
≤ sup
α>0
_
_
B(α)
v (y) dy
_

1
p
(p

−1)
1
p

_
_
B(α)
u(y) [B(y)[
q
dy
_1
q
< ∞.
Similarly, according to Theorem 4.1 (ii),
__
R
n
(I
2
(x))
p

V
−p

(x) v (x) dx
_ 1
p

(4.22)
=
_
_
_
R
n
\B(α)
_
[B(x)[
_
R
n
\B(x)
g (y) dy
_
−p

V
−p

(x) v (x) dx
_
_
1
p

≤ c
__
R
n
g
q

(x) u
1−q

(x) dx
_ 1
q

because
sup
α>0
_
_
B(α)
[B(x)[
p

V
−p

(x) v (x) dx
_ 1
p

_
_
R
n
\B(α)
u(x) dx
_1
q
< ∞
which holds by (4.19). Thus, by using (4.20), Minkowski’s inequality and
(4.21)-(4.22), we see that (4.17) holds.
56 Carleman type inequalities and Hardy type inequalities
for monotone functions
Now assume that (4.17) holds i.e. that (4.20) holds. Then, in particular,
__
R
n
(I
1
(x))
p

V
−p

(x) v (x) dx
_ 1
p

≤ c
__
R
n
g
q

(x) u
1−q

(x) dx
_ 1
q

(4.23)
and, by using Theorem 4.1 (i) and arguing as above, we find that (4.18)
holds. Moreover, (8.13) holds also with I
1
replaced by I
2
so that, by using
Theorem 4.1 (ii) and again arguing as in the sufficiently part, we see that
(4.19) holds. The proof is complete.
Remark 4.7. For the case when also the weights are radially decreasing or
increasing some of our results can be written in a more suitable form. Here
we only state the following consequence of Proposition 4.8:
Corollary 4.9. Let 1 < p ≤ q < ∞ and let f (x) and B(x) be positive and
radially decreasing in R
n
. Then the Hardy inequality
_
_
R
n
_
1
[B(x)[
_
B(x)
f (y) dy
_
q
[B(x)[
b
dx
_
1
q
≤ c
__
R
n
f
p
(x) [B(x)[
a
dx
_1
p
holds if and only if −1 < a < p −1, −1 < b < q −1 and
a + 1
p
=
b + 1
q
. (4.24)
Proof. Apply Proposition 4.8 with v (x) = [B(x)[
a
and u(x) = [B(x)[
b
. We
note that some straightforward calculations give that
_
_
B(α)
v (x) dx
_

1
p
_
_
B(α)
u(x) dx
_1
q
≈ α
−(a+1)n
p
+
(b+1)n
q
(4.25)
whenever a > −1, b > −1 and
_
_
B(α)
[B(x)[
p

V
−p

(x) v (x) dx
_ 1
p

_
_
R
n
\B(α)
u(x)dx
_1
q
(4.26)
≈ α
an−anp

+n
p

+
bn−nq+n
q
= α
n
_

a+1
p
+
b+1
q
_
whenever a < p −1 and b < q −1.
Moreover, according to the estimates (4.25) and (4.26), the condition
(4.24) (and only this) gives finite supremum. The proof is complete.
A Sawyer duality principle for radially monotone func-
tions in R
n
57
Remark 4.8. It is easy to see that Theorem 4.1 is true also if R
n
is replaced
by R
n
+
or even some more general cone in R
n
since the proof essentially
consists of using polar coordinates and carrying over the problem to the (well-
known) one dimensional question. Therefore, by modifying our proofs, we see
that all our results in this chapter indeed holds also when R
n
is replaced by
R
n
+
or even general cones in R
n
as defined in [24].
By a general cone C in R
n
we mean a subset of R
n
defined as follows:
Let Ω be a measurable subset of the unit sphere S
n−1
. Then
C = ¦x ∈ R
n
: x = sσ, 0 < s < ∞ and σ ∈ Ω¦ .
58 Carleman type inequalities and Hardy type inequalities
for monotone functions
Chapter 5
A new characterization of the
Hardy inequality and its limit
Pólya-Knopp inequality for
decreasing functions
5.1 Introduction
As mentioned in the introduction E. Sawyer [93] (see Theorem 1.3) proved
the following theorem by performing a general approach for general operators
(In [100] V.D. Stepanov gave a direct proof, which also gave good estimates
of the equivalence constants).
Theorem 5.1. Let 1 < p ≤ q < ∞. Then the Hardy inequality
_
_

_
0
_
_
1
x
x
_
0
f (t) dt
_
_
q
u(x) dx
_
_
1
q
≤ C
_
_

_
0
f
p
(x) v (x) dx
_
_
1
p
(5.1)
holds for all f ↓≥ 0 if and only if
A
0
= A
0
(p, q, u, v) := sup
t>0
V

1
p
(t)
_
_
t
_
0
u(x)dx
_
_
1
q
< ∞, (5.2)
59
60 Carleman type inequalities and Hardy type inequalities
for monotone functions
and
A
1
= A
1
(p, q, u, v) := (5.3)
sup
t>0
_
_

_
t
u(x)x
−q
dx
_
_
1
q
_
_
t
_
0
x
p

V
−p

(x)v(x)dx
_
_
1
p

< ∞,
where p

=
p
p−1
and
V (t) =
t
_
0
v(x)dx. (5.4)
Moreover, if C is the best possible constant in (5.1), then
C ≈ max(A
0
, A
1
).
Moreover, as also mentioned in the introduction L.-E. Persson and V.D.
Stepanov (see Theorem 1.1) proved a new weight characterization for the
Hardy inequality. Their motivation for doing that was the fact that it
was not possible to perform the mentioned limit procedure by using the
usual Muckenhoupt condition so they needed an equivalent characterization.
Partly guided by this fact we will prove a new weight characterization for
the Hardy inequality for decreasing functions, because also in this case it
is not possible to use Sawyer’s characterization to perform a corresponding
limit procedure.
More exactly, in Section 5.2 we present and prove our new alternative
weight characterization of the Hardy inequality for decreasing functions (see
Theorem 5.2). In Section 5.3 we will perform the mentioned limit procedure
in this result and, thus, derive the corresponding Pólya-Knopp inequality
for decreasing functions (see Theorem 5.4). Finally, in Section 5.4 we use
a fairly new equivalence theorem to give an alternative proof of Theorem
5.2 and Sawyer’s result (see Remark 5.3). In fact, this result shows that
there are infinitely many possibilities to characterize the considered Hardy
inequality for decreasing functions.
5.2 A new weight characterization of the weighted
Hardy inequality for decreasing functions
The main result of this section reads:
A new characterization of the Hardy inequality and its
limit Pólya-Knopp inequality for decreasing functions
61
Theorem 5.2. Let 1 < p ≤ q < ∞. Then the Hardy inequality (5.1) holds
for all f ↓≥ 0 if and only if (5.2) and
A
2
= A
2
(p, q, u, v) := (5.5)
sup
y>0
_
_
y
_
0
_
_
x
_
0
t
p

V
−p

(t)v(t)dt
_
_
q
u(x)x
−q
dx
_
_
1
q
_
_
y
_
0
t
p

V
−p

(t)v(t)dt
_
_

1
p
< ∞
hold, where V is defined by (5.4). Moreover, if C is the best possible constant
in (5.1), then
C ≈ max(A
0
, A
2
). (5.6)
For the proof we need the following Lemma:
Lemma 5.3. Let 0 < p ≤ q < ∞. Then the inequality
_
_

_
0
f
q
(x)u(x)dx
_
_
1
q
≤ C
_
_

_
0
f
p
(x)v(x)dx
_
_
1
p
(5.7)
holds for all f ↓≥ 0 if and only if (5.2) holds. Moreover, the constant C = A
0
is the best possible, where A
0
is defined by (5.2).
A proof of this Lemma can be found e.g. in [100] (for more references
see e.g. the Ph.D. thesis [7] by S. Barza).
Proof. (Theorem 5.2). Since f is a decreasing function we can write f(x) =

_
x
h(y)dy, for some h(y) ≥ 0. First note that, by changing order of integra-
tion,
1
x
x
_
0
f(t)dt =
1
x
x
_
0

_
t
h(y)dydt =

_
x
h(y)dy +
1
x
x
_
0
yh(y)dy. (5.8)
Consequently, by using (5.8) and Minkowski’s inequality, we find that
F :=
_
_

_
0
_
_
1
x
x
_
0
f(t)dt
_
_
q
u(x)dx
_
_
1
q
62 Carleman type inequalities and Hardy type inequalities
for monotone functions
=
_
_

_
0
_
_

_
x
h(y)dy +
1
x
x
_
0
yh(y)dy
_
_
q
u(x)dx
_
_
1
q

_
_

_
0
f
q
(x) u(x)dx
_
_
1
q
+
_
_

_
0
_
_
1
x
x
_
0
yh(y)dy
_
_
q
u(x)dx
_
_
1
q
:=
_
_

_
0
f
q
(x)u(x)dx
_
_
1
q
+F
1
.
The first term on the right hand side can be estimated by using Lemma 5.3
and, thus,
F ≤ A
0
_
_

_
0
f
p
(x)v(x)dx
_
_
1
p
+F
1
. (5.9)
Let us estimate the second term F
1
. We have, setting H = hV, where V
is defined by (5.4),
F
q
1
=

_
0
_
_
1
x
x
_
0
yh(y)dy
_
_
q
u(x)dx =

_
0
_
_
1
x
x
_
0
yH(y)
V (y)
dy
_
_
q
u(x)dx.
Moreover, by integrating by parts and setting
y
_
0
H(t)dt = G(y), we get that
x
_
0
yH(y)
V (y)
dy =
xG(x)
V (x)

x
_
0
G(y)
V (y)
dy +
x
_
0
yv(y)
V
2
(y)
G(y)dy

xG(x)
V (x)
+
x
_
0
yv(y)
V
2
(y)
G(y)dy.
Hence, according to Minkowski’s inequality,
F
1

_
_

_
0
_
G(x)
V (x)
_
q
u(x)dx
_
_
1/q
+
_
_

_
0
_
_
1
x
x
_
0
yv(y)
V
2
(y)
G(y)dy
_
_
q
u(x)dx
_
_
1/q
(5.10)
:= F
2
+F
3
.
A new characterization of the Hardy inequality and its
limit Pólya-Knopp inequality for decreasing functions
63
Moreover, by integrating by parts and using an approximation argument, we
find that
F
q
2
=

_
0
_
G(x)
V (x)
_
q
u(x)dx ≤ q

_
0
G(x)
q
v(x)
V (x)
q+1
_
_
x
_
0
u(t)dt
_
_
dx.
Hence, by applying condition (5.2), we have that
F
q
2
≤ qA
q
0

_
0
G(x)
q
v(x)
V (x)
q+1−q/p
dx. (5.11)
Moreover,
G(y) =
y
_
0
h(x)
x
_
0
v(t)dtdx =
y
_
0
v(t)
_
_
y
_
t
h(x)dx
_
_
dt (5.12)

y
_
0
v(t)
_
_

_
t
h(x)dx
_
_
dt ≤
y
_
0
f(t)v(t)dt.
Consequently, by inserting (5.12) into (5.11), we get that
F
q
2
≤ qA
q
0

_
0
_
_
x
_
0
f(t)v(t)dt
_
_
q
v(x)
V (x)
q+1−q/p
dx.
If we apply Theorem 1.1 to the function f (t) v (t) instead of f (t) and the
weights V (x)
q
p
−q−1
v (x) x
q
instead of u(x) and v (x)
1−p
instead of v (x) we
find that the condition (1.11) becomes
A
q
PS
= sup
t>0
_
_
t
_
0
V (x)
q/p−1−q
v(x)
_
_
x
_
0
v
(1−p)(1−p

)
dy
_
_
q
dx
_
_

_
_
t
_
0
v(x)
(1−p)(1−p

)
dx
_
_

q
p
= sup
t>0
_
_
t
_
0
V (x)
q/p−1
v(x))dx
_
_
_
_
t
_
0
v(x)dx
_
_

q
p
=
p
q
,
64 Carleman type inequalities and Hardy type inequalities
for monotone functions
and, thus,
F
q
2
≤ p
_
p

_
q
A
q
0
_
_

_
0
f
p
(x)v(x)dx
_
_
q
p
. (5.13)
Denoting
yv(y)
V
2
(y)
G(y) = Φ(y)
and
x
p

V
−p

(x)v(x) = Ψ(x)
1−p

(5.14)
and, again applying Theorem 1.1, (5.5) and (5.12), we find that
F
q
3
=

_
0
_
_
1
x
x
_
0
Φ(y)dy
_
_
q
u(x)dx ≤
_
p

_
q
A
q
2
_
_

_
0
Φ
p
(x)Ψ(x)dx
_
_
q
p
=
_
p

_
q
A
q
2
_
_

_
0
G
p
(x)v(x)V (x)
−p
dx
_
_
q
p

_
p

_
q
A
q
2
_
_

_
0
_
_
x
_
0
f(y)v(y)dy
_
_
p
v(x)V (x)
−p
dx
_
_
q
p

_
p

_
2q
A
q
2
_
_

_
0
f(x)
p
v(x)dx
_
_
q
p
. (5.15)
Here we in the last step apply Theorem 1.1 with q = p, f (x) replaced by
f (x) v (x), v (x) replaced by v
1−p
(x) and u(x) replaced by v (x) V
−p
(x) x
p
so that the condition (1.11) A
PS
= 1. By combining the estimates (5.9),
(5.10), (5.13) and (5.15) we find that (5.1) holds with a constant C satisfying
C ≤
_
1 +p
1/q
p

_
A
0
+
_
p

_
2
A
2
),
i.e. the upper estimate in (5.6) holds.
For the necessity, assume that the inequality (5.1) holds for all f ↓≥ 0.
Consider, for fixed y > 0, the decreasing testfunction
f
y
(s) =
_
_
y
_
s
t
p

V (t)
−p

−1
v(t)dt
_
_
1
p
χ
[0,y]
(s).
A new characterization of the Hardy inequality and its
limit Pólya-Knopp inequality for decreasing functions
65
Applying this function to the right hand side of (5.1) and changing order of
integration we obtain that
_
_

_
0
f
y
(s)
p
v(s)ds
_
_
1/p
=
_
_
y
_
0
_
_
y
_
s
t
p

V (t)
−p

−1
v(t)dt
_
_
v(s)ds
_
_
1
p
=
_
_
y
_
0
t
p

V (t)
−p

−1
_
_
t
_
0
v(s)ds
_
_
dt
_
_
1
p
=
_
_
y
_
0
t
p

V (t)
−p

v(t)dt
_
_
1
p
. (5.16)
For the left hand side of (5.1) we have that
_
_

_
0
_
_
1
x
x
_
0
f(s)ds
_
_
q
u(x)dx
_
_
1
q
=
_
_
_

_
0
_
_
_
1
x
x
_
0
_
_
y
_
s
t
p

V (t)
−p

−1
v(t)dt
_
_
1
p
ds
_
_
_
q
u(x)dx
_
_
_
1
q

_
_
_
y
_
0
_
_
_
x
_
0
_
_
y
_
s
t
p

V (t)
−p

−1
v(t)dt
_
_
1
p
ds
_
_
_
q
u(x)x
−q
dx
_
_
_
1
q
. (5.17)
For the inner integral it yields that
x
_
0
_
_
y
_
s
t
p

V (t)
−p

−1
v(t)dt
_
_
1
p
ds

x
_
0
s
p

p
_
_
y
_
s
V (t)
−p

−1
v(t)dt
_
_
1
p
ds

1
p
x
_
0
s
p

p
_
_
_
x
_
s
_
_
y
_
t
V (z)
−p

−1
v(z)dz
_
_

1
p

V (t)
−p

−1
v(t)dt
_
_
_
ds := I
66 Carleman type inequalities and Hardy type inequalities
for monotone functions
Now, by changing the order of integration, we find that
I =
1
p
x
_
0
V (t)
−p

−1
v(t)
_
_
t
_
0
s
p

p
ds
_
_
_
_
y
_
t
V (z)
−p

−1
v(z)dz
_
_

1
p

dt
=
1
pp

x
_
0
V (t)
−p

−1
v(t)t
p

_
_
y
_
t
V (z)
−p

−1
v(z)dz
_
_

1
p

dt

1
p (p

)
1
p
x
_
0
V (t)
−p

v(t)t
p

dt. (5.18)
Therefore, by combining (5.16) and (5.17) with (5.18), we obtain that
_
_
y
_
0
t
p

V (t)
−p

v(t)dt
_
_

1
p
_
_
y
_
0
_
_
x
_
0
t
p

V
−p

(t)v(t)dt
_
_
q
u(x)x
−q
dx
_
_
1
q
≤ p
_
p

_1
p
C
Hence C ≥
A
2
p(p

)
1
p
. Moreover, since f(x) ≤
1
x
x
_
0
f(t)dt when f is decreasing,
it follows that A
0
≤ C, by choosing f
t
(x) = χ
[0,t]
(x) and taking supremum.
Hence also the lower estimate of (5.6) holds and the proof is complete.
Remark 5.1. The crucial part of the proof above is to use the results in
[84], [99] and the technique in [100] combined with finding explicitly the cor-
responding test functions. For the estimates C ≈ max (A
0
, A
2
) we have found
that
max
_
A
0
, A
2
1
p (p

)
1/p
_
≤ C ≤
_
1 +p
1/q
p

_
A
0
+
_
p

_
2
A
2
) (5.19)
but we strongly believe that these estimates can be improved.
5.3 A characterization of the Pólya-Knopp inequal-
ity for decreasing functions
In this Section we will give a weight characterization of the Pólya-Knopp
inequality (1.14) for decreasing functions. In fact, the following Pólya-Knopp
type inequality may be regarded as a limit result of our Theorem 5.2:
A new characterization of the Hardy inequality and its
limit Pólya-Knopp inequality for decreasing functions
67
Theorem 5.4. Let 0 < p ≤ q < ∞. Let u(x) and v (x) be weight functions
and assume that v (x) is decreasing. Then the inequality
_
_

_
0
_
_
exp
_
_
1
x
x
_
0
ln f(t)dt
_
_
_
_
q
u(x)dx
_
_
1
q
≤ C
_
_

_
0
f
p
(x)v(x)dx
_
_
1
p
(5.20)
holds for all f ↓≥ 0 if and only if
D
PS
= D
PS
(p, q, w) := sup
t>0
t

1
p
_
_
t
_
0
w(x)dx
_
_
1
q
< ∞, (5.21)
holds. Moreover, if C is the best possible constant in (5.20), then
D
PS
≤ C ≤ inf
r>1
k (r) D
PS
, (5.22)
where
k (r) =
_
1 +r
p
qr
r

+
_
r

_
2
_r
p
,
Proof. Assume that (5.21) holds and let g(x) = f(x)v(x)
1
p
. Then as v(x) is
decreasing and f(x) is decreasing so is g(x). Moreover, if w(x) be defined as
in (1.13), then (5.20) is equivalent to
_
_

_
0
_
_
exp
_
_
1
x
x
_
0
ln g(t)dt
_
_
_
_
q
w(x)dx
_
_
1
q
≤ C
_
_

_
0
g
p
(x)dx
_
_
1
p
, (5.23)
and with g(t) = h(t)
r
p
equivalent to
_
_
_

_
0
_
_
exp
_
_
1
x
x
_
0
ln h(t)dt
_
_
_
_
qr
p
w(x)dx
_
_
_
1
q
≤ C
_
_

_
0
h
r
(x)dx
_
_
1
p
. (5.24)
Note that if r > 1 and 1 < p ≤ q < ∞, then we have 1 < r ≤
qr
p
< ∞. By
applying Jensen’s inequality we find that if the inequality
_
_
_

_
0
_
_
x
_
0
h(t)dt
_
_
qr
p
w(x)x

qp
s
dx
_
_
_
1
q
≤ C
_
_

_
0
h
r
(x)dx
_
_
1
p
(5.25)
68 Carleman type inequalities and Hardy type inequalities
for monotone functions
holds for all decreasing functions h(x) , then also (5.24) holds with the same
constant C. Now we replace p by r and q by
qr
p
, u(x) with w(x) and let
v(x) = 1 in the inequality (5.1) and apply Theorem 5.2. This means that
the inequality (5.25) holds if and only if
_
A
0
(r,
qr
p
, w, 1)
_r
p
= D
PS
(p, q, w) = sup
t>0
t

1
p
_
_
t
_
0
w(x)dx
_
_
1
q
< ∞,
and
_
A
2
(r,
qr
p
, w, 1)
_r
p
= D
PS
(p, q, w) = sup
t>0
t

1
p
_
_
t
_
0
w(x)dx
_
_
1
q
< ∞,
where w(x) is defined by (1.13). We conclude that if (5.21) holds, then
(5.24) and, thus, (5.20) holds for all f ↓≥ 0. Moreover, by using the upper
bound in (5.19) we have that for the best constant C in (5.20), we have the
estimate
C ≤
_
1 +r
p
qr
r

+
_
r

_
2
_r
p
D
PS
= k (r) D
PS
.
By taking infimum the upper bound in (5.22) is proved.
The neccesity part and the proof of the lower bound in (5.22) follows by
applying
g(x) = t

1
p
χ
[0,t]
(x)
to (5.23). The proof is complete.
Remark 5.2. We have obtained the same condition for characterizing the
Pólya-Knopp inequality for decreasing functions as for all positive functions
in [84] but our upper estimate of the best constant C in (5.22) is different.
5.4 Another approach via an equivalence theorem
In [93] E. Sawyer showed that to characterize the inequality (5.1) for de-
creasing functions f is equivalent to characterize the following inequalities:
_
_
_

_
0
_
_
x
_
0
f(t)dt
_
_
p

V (x)
−p

v(x)dx
_
_
_
1
p

≤ C
_
_

_
0
f
q

(x)u(x)
1−q

dx
_
_
1
q

,
(5.26)
A new characterization of the Hardy inequality and its
limit Pólya-Knopp inequality for decreasing functions
69
and
_
_
_

_
0
_
_

_
x
f(t)
t
dt
_
_
p

x
p

V (x)
−p

v(x)dx
_
_
_
1
p

≤ C
_
_

_
0
f
q

(x)u(x)
1−q

dx
_
_
1
q

,
(5.27)
where V is defined by (5.4). Then, by using the Hardy inequality and the
so called Muckenhoupt condition to characterize the inequalities (5.26) and
(5.27), E. Sawyer received the before mentioned result in Theorem 5.1.
Recently the weighted Hardy inequality has been characterized with some
new conditions (see e.g. [64] and [102]). More generally, nowadays we know
that the Hardy inequality (5.1) for 1 < p ≤ q < ∞ holds for all f ≥ 0 if
and only if just one of the following (infinite many equivalent) conditions is
satisfied (see [33]):
Theorem 5.5. Let 1 < p ≤ q < ∞, 0 < s < ∞ and define
/
1
(s) := sup
0<x<∞
_
_
_
_

_
x
u(t)
_
_
t
_
0
v
1−p

(y)dy
_
_
q
_
1
p

−s
_
dt
_
_
_
_
1
q
_
_
x
_
0
v
1−p

(t)dt
_
_
s
,
/
2
(s) := sup
0<x<∞
_
_
_
_
x
_
0
u(t)
_
_
t
_
0
v
1−p

(y)dy
_
_
q
_
1
p

+s
_
dt
_
_
_
_
1
q
_
_
x
_
0
v
1−p

(t)dt
_
_
−s
,
/
3
(s) := sup
0<x<∞
_
_
_
_
x
_
0
v
1−p

(t)
_
_

_
t
u(y)dy
_
_
p

_
1
q
−s
_
dt
_
_
_
_
1
p

_
_

_
t
u(y)dy
_
_
s
,
/
4
(s) := sup
0<x<∞
_
_
_
_

_
x
v
1−p

(t)
_
_

_
t
u(y)dy
_
_
p

_
1
q
+s
_
dt
_
_
_
_
1
p

_
_

_
t
u(y)dy
_
_
−s
.
(5.28)
Then the Hardy inequality
_
_

_
0
_
_
x
_
0
f (t) dt
_
_
q
u(x) dx
_
_
1
q
≤ C
_
_

_
0
f
p
(x) v (x) dx
_
_
1
p
(5.29)
70 Carleman type inequalities and Hardy type inequalities
for monotone functions
holds for all measurable functions f ≥ 0 if and only if any of the quantities
/
i
(s) is finite. Moreover, for the best constant C in (5.29) we have C ≈
/
i
(s), i = 1, 2, 3, 4.
This gives us the possibility to characterize the weighted Hardy inequal-
ity for decreasing functions with some new conditions, by just using the
technique of E. Sawyer described above. We can therefore formulate the
following generalization of our Theorem 5.2:
Theorem 5.6. Let 1 < p ≤ q < ∞. Then the inequality (5.1) holds for all
f ↓≥ 0 if and only if for any s, r, 0 < s, r < ∞, one of the quantities
B
1
(s) := sup
0<x<∞
_
_
_
_

_
x
V (t)
−p

v (t)
_
_
t
_
0
u(y)dy
_
_
p

_
1
q
−s
_
dt
_
_
_
_
1
p

_
_
x
_
0
u(t)dt
_
_
s
,
B
2
(s) := sup
0<x<∞
_
_
_
_
x
_
0
V (t)
−p

v (t)
_
_
t
_
0
u(y)dy
_
_
p

_
1
q
+s
_
dt
_
_
_
_
1
p

_
_
x
_
0
u(t)dt
_
_
−s
,
B
3
(s) := sup
0<x<∞
_
_
_
_
x
_
0
u(t)
_
_

_
t
V (y)
−p

v (y) dy
_
_
q
_
1
p

−s
_
dt
_
_
_
_
1
q

_
_

_
t
V (y)
−p

v (y) dy
_
_
s
,
B
4
(s) := sup
0<x<∞
_
_
_
_

_
x
u(t)
_
_

_
t
V (y)
−p

v (y) dy
_
_
q
_
1
p

+s
_
dt
_
_
_
_
1
q

_
_

_
t
V (y)
−p

v (y) dy
_
_
−s
and one of the quantities
c
1
(r) := sup
0<x<∞
_
_
_
_

_
x
u(t) t
−q
_
_
t
_
0
Ψ
1−p

(y)dy
_
_
q
_
1
p

−r
_
dt
_
_
_
_
1
q
_
_
x
_
0
Ψ
1−p

(t)dt
_
_
r
,
A new characterization of the Hardy inequality and its
limit Pólya-Knopp inequality for decreasing functions
71
c
2
(r) := sup
0<x<∞
_
_
_
_
x
_
0
u(t) t
−q
_
_
t
_
0
Ψ
1−p

(y)dy
_
_
q
_
1
p

+r
_
dt
_
_
_
_
1
q

_
_
x
_
0
Ψ
1−p

(t)dt
_
_
−r
,
c
3
(r) := sup
0<x<∞
_
_
_
_
x
_
0
Ψ
1−p

(t)
_
_

_
t
u(y)y
−q
dy
_
_
p

_
1
q
−r
_
dt
_
_
_
_
1
p

_
_

_
t
u(y)y
−q
dy
_
_
r
,
c
4
(r) := sup
0<x<∞
_
_
_
_

_
x
Ψ
1−p

(t)
_
_

_
t
u(y)y
−q
dy
_
_
p

_
1
q
+r
_
dt
_
_
_
_
1
p

_
_

_
t
u(y)y
−q
dy
_
_
−r
are finite, where V is defined by (5.4) and Ψ is defined by (5.14).
Remark 5.3. By using B
1
with s =
1
q
and V (∞) = ∞ (this is no real restic-
tion see e.g. [93, p. 148] and [63, p. 316]) and integrating, and also using
c
1
with r =
1
p

we get the result due to Sawyer (Theorem 5.1). Furthermore,
if we again use B
1
as but with s =
1
q
and also c
2
with r =
1
p
we get our result
(Theorem 5.2).
Proof. (Theorem 5.6). We apply the above mentioned technique by E.
Sawyer [93]. First we give weight characterizations for the inequality (5.26).
If we use Theorem 5.5 with p, q, u(x) , v (x) replaced by q

, p

,V (x)
−p

v (x)
and u(x)
1−q

,respectively, then we obtain the first bunch of conditions in
Theorem 5.6.
Now, we characterize the inequality (5.27) and first we note that, by
duality (see e.g. [63, p. 314]), (5.27) is equivalent to the Hardy inequality
_
_

_
0
_
_
x
_
0
f(t)dt
_
_
q
u(x)x
−q
dx
_
_
1
q
≤ C
_
_

_
0
f
p
(x)Ψ(x)dx
_
_
1
p
,
72 Carleman type inequalities and Hardy type inequalities
for monotone functions
where Ψ(x) is defined by (5.14) so we can again use Theorem 5.5 now with
s, u(x) and v (x) replaced by r, u(x) x
−q
and Ψ(x) respectively, and we get
the second bunch of conditions in Theorem 5.6. The proof is complete.
Remark 5.4. By using Theorem 5.6 and arguing as in the proof of Theo-
rem 5.4 we can get a number of other characterizations of the Pólya-Knopp
inequality (5.20) then that is stated in Theorem 5.4.
Chapter 6
A Unified approach to Sawyer and
Sinnamon characterizations of the
Hardy inequality for decreasing
functions
6.1 Introduction
Let us first recall that in the Sawyer characterization of Hardy’s inequality
for decreasing functions two independent conditions were used (see Theorem
1.3 = Theorem 5.1 and c.f. also our alternative descriptions in Theorem 5.2
and 5.6).
On the contrary, in the next result by G. Sinnamon [97] only one condi-
tion is necessary to characterize the corresponding Hardy type inequality.
Theorem 6.1. Let 1 < p ≤ q < ∞and let u(x) and v (x) be weight functions
on (0, ∞). Then the inequality
_
_

_
0
_
_
x
_
0
f (t) v (t) dt
_
_
q
u(x) dx
_
_
1
q
≤ C
_
_

_
0
f
p
(x) v (x) dx
_
_
1
p
(6.1)
73
74 Carleman type inequalities and Hardy type inequalities
for monotone functions
holds for all decreasing f ≥ 0 if and only if
A
3
:= sup
t>0
_
_
t
_
0
u(x)
_
_
x
_
0
v (y) dy
_
_
q
dx
_
_
1
q
_
_
t
_
0
v (x) dx
_
_

1
p
< ∞. (6.2)
Remark 6.1. The main argument of Sinnamon was to prove (via so called
level functions) that (6.1) is in fact equivalent to that
_
_

_
0
f
q
(x) V
q
(x) u(x) dx
_
_
1
q
≤ C
_
_

_
0
f
p
(x) v (x) dx
_
_
1
p
(6.3)
holds for all decreasing f ≥ 0 and use the fact that a characterization for such
continuous embeddings between Lorentz spaces is known. (Here, as usual,
V (x) =
x
_
0
v (t) dt.)
At a first glance the two inequalities (5.1) and (6.1) seems to be similar,
but to characterize (5.1) we need two conditions while (6.1) can be charac-
terized by only one condition. Moreover, in [93] E. Sawyer gives an example
for the fact that neither (5.2) nor (5.3) is sufficient for (5.1) to hold namely
the following:
Example: If u(x) = x
q−1
and v (x) = x
p−1
, then (5.2) holds but not
(5.3), while if u(x) = x
q
(x + 1)
−q−1
and v (x) = χ
(0,1)
(x) , then (5.3) holds
but not (5.2).
Let us now look at another example:
Example: If v(x) ≡ 1, then
A
0
= sup
0<t<∞
_
_
t
_
0
u(x)dx
_
_
1
q
t

1
p
and
A
1
= sup
0<t<∞
_
_

_
t
x
−q
u(x)dx
_
_
1
q
t
1
p

.
We also note that if u(x) is replaced by u(x) x
−q
in (6.1)-(6.2) and v (x) ≡ 1,
then (5.1) coincides with (6.1) and (5.2) coincides with (6.2). In particular,
this means that in this case the conditions A
0
< ∞ and A
1
< ∞ must be
equivalent. This can be proved directly but it is also a consequence of the
general equivalence result proved in [33] (see Theorem 5.5 and Remark 6.2):
A Unified approach to Sawyer and Sinnamon characteri-
zations of the Hardy inequality for decreasing functions
75
Remark 6.2. The equivalence of the conditions A
0
< ∞ and A
1
< ∞ for
v (x) ≡ 1 follows by using Theorem 5.5 with u(x) = u(x) x
−q
and /
1
with
s =
1
p

and /
2
with s =
1
p
.
In Section 6.2 we state and prove the main result of this Chapter (The-
orem 6.2), which generalize the above mentioned result of Sinnamon (see
Corollary 6.3) and also put some new light on the result of Sawyer for the
case when the weight v (x) is increasing. Section 6.3 is reserved for some
concluding remarks.
6.2 The main result
Our main result reads:
Theorem 6.2. Let 1 < p ≤ q < ∞ and let u(x) , v (x) and ϕ(x) be weight
functions on (0, ∞), where ϕ(x) is decreasing. Then the inequality
_
_

_
0
_
_
x
_
0
f (t) ϕ(t) v (t) dt
_
_
q
u(x) dx
_
_
1
q
≤ C
_
_

_
0
f
p
(x) v (x) dx
_
_
1
p
(6.4)
holds for all C < ∞ and decreasing f if and only if one of the following
conditions holds for some s > 0:
D
ϕ
(s) := sup
t>0
_
_
_
_

_
t
u(x)
_
_
x
_
0
v (y) ϕ
p

(y) dy
_
_
q
_
1
p

−s
_
dx
_
_
_
_
1
q
(6.5)
_
_
t
_
0
v (x) ϕ
p

(x) dx
_
_
s
< ∞.
D

ϕ
(s) := sup
t>0
_
_
_
_
t
_
0
v (x) ϕ
p

(x)
_
_

_
x
u(y) dy
_
_
p

_
1
q
−s
_
dx
_
_
_
_
1
p

(6.6)
_
_

_
t
u(x) dx
_
_
s
< ∞.
76 Carleman type inequalities and Hardy type inequalities
for monotone functions
E
ϕ
(s) := sup
t>0
_
_
_
_
t
_
0
u(x)
_
_
x
_
0
v (y) ϕ
p

(y) dy
_
_
q
_
1
p

+s
_
dx
_
_
_
_
1
q
(6.7)
_
_
t
_
0
v (x) ϕ
p

(x) dx
_
_
−s
< ∞.
E

ϕ
(s) := sup
t>0
_
_
_
_

_
t
v (x) ϕ
p

(x)
_
_

_
x
u(y) dy
_
_
p

_
1
q
+s
_
dx
_
_
_
_
1
p

(6.8)
_
_

_
t
u(x) dx
_
_
−s
< ∞.
Proof. First we use Theorem 5.5 with u(x), v (x) = ϕ
−p

(x) to conclude
that, for any s > 0, the conditions (6.5) - (6.8) are equivalent. In particular,
now we assume that (6.5) holds for some s, 0 < s <
1
p

. Since f is decreasing
we can write
f
p
(x) =
_

x
h(y) dy, where h ≥ 0.
Thus, by using the Fubini theorem, we find that (6.4) can equivalently be
written as
_
_
_

_
0
_
_
_
x
_
0
_
_

_
t
h(y) dy
_
_
1
p
v (t) ϕ(t) dt
_
_
_
q
u(x) dx
_
_
_
1
q
(6.9)
≤ C
_
_

_
0
h(y)
_
_
y
_
0
v (x) dx
_
_
dy
_
_
1
p
Now, we define
˜
V
ϕ
(t) =
x
_
0
v (t) ϕ
p

(t) dt and C
0
=
_
1
1 −sp

_ 1
p

.
A Unified approach to Sawyer and Sinnamon characteri-
zations of the Hardy inequality for decreasing functions
77
By using Hölder’s inequality, Minkowski’s generalized integral inequality and
changing order of integration we find that
_
_

_
0
_
_
x
0
__

t
h(y) dy
_1
p
v (t) ϕ(t) dt
_
q
u(x) dx
_
_
1
q
=
_
_

_
0
_
_
x
0
__

t
h(y) dy
_1
p
ϕ(t)
˜
V
−s
ϕ
(t)
˜
V
s
ϕ
(t) v
1
p
(t) v
1
p

(t) dt
_
q
u(x) dx
_
_
1
q

_
_

_
0
__
x
0
__

t
h(y) dy
_
˜
V
−sp
ϕ
(t) v (t)
_
q
p

__
x
0
˜
V
−sp

ϕ
(t) ϕ
p

(t) v(t)dt
_
q
p

u(x) dx
_1
q
≤ C
0
_
_

_
0
__
x
0
__

t
h(y) dy
_
˜
V
sp
ϕ
(t) v (t) dt
_
q
p
˜
V
q
_
1
p

−s
_
ϕ
(x) u(x) dx
_
_
1
q
≤ C
0
_
_
_

_
0
__

t
h(y) dy
_
˜
V
sp
ϕ
(t) v (t)
_
_

_
t
˜
V
q
_
1
p

−s
_
ϕ
(x) u(x) dx
_
_
p
q
dt
_
_
_
1
p
≤ C
0
_
_
_

_
0
h(y)
_
_
_
_
y
0
˜
V
sp
ϕ
(t) v (t)
_
_

_
t
˜
V
q
_
1
p

−s
_
ϕ
(x) u(x) dx
_
_
p
q
dt
_
_
_
dy
_
_
_
1
p
≤ C
0
sup
t≥0
˜
V
s
ϕ
(t)
_
_

_
t
˜
V
q
_
1
p

−s
_
ϕ
(x) u(x) dx
_
_
1
q
_
_

_
0
h(y)
__
y
0
v (t) dt
_
dy
_
_
1
p
=
_
1
1 −sp

_ 1
p

D
1
(s)
_
_

_
0
h(y)
__
y
0
v (t) dt
_
dy
_
_
1
p
=
_
1
1 −sp

_ 1
p

D
ϕ
(s)
_
_

_
0
f
p
(t) v (t) dt
_
_
1
p
.
78 Carleman type inequalities and Hardy type inequalities
for monotone functions
Hence (6.9) and, thus, (6.4) holds.
Now we assume that (6.4) holds for some C < ∞ and choose the test-
function
f (x) =
_
p
p −(ps + 1)
_
˜
V

_
1
p
+s
_
ϕ
(t) ϕ
p

p
(x) χ
(0,t)
(x) +
˜
V

_
1
p
+s
_
ϕ
(x) ϕ
p

p
(x) χ
(t,∞)
(x)
where t > 0 is fixed. Then the integral on the right hand side of (6.4) is
equal to
=
_
_
t
_
0
_
p
p (1 −s) −1
_
p
˜
V
−(1+ps)
ϕ
(t) ϕ
p

(x) v (x) dx (6.10)
+

_
t
˜
V
−(1+ps)
ϕ
(x) ϕ
p

(t) v (x) dx
_
_
1
p
=
__
p
p (1 −s) −1
_
p
˜
V
−ps
ϕ
(t) +
1
ps
˜
V
−ps
ϕ
(t)
_
1
p
=
__
p
p (1 −s) −1
_
p
+
1
ps
_
1
p
˜
V
−s
ϕ
(t) .
Moreover, the left hand side of (6.4) is greater than
_
_

_
t
_
_
x
_
0
f (t) ϕ(t) v (t) dt
_
_
q
u(x) dx
_
_
1
q
=
_
since
p

p
+ 1 = p

_
=
_
_

_
t
_
_
t
_
0
_
p
p (1 −s) −1
_
˜
V

_
1
p
+s
_
ϕ
(t) ϕ
p

(y) v (y) dy +
x
_
t
˜
V

_
1
p
+s
_
ϕ
(y) ϕ
p

(y) v (y) dy
_
_
q
u(x) dx
_
_
1
q
=
_
_

_
t
_
_
p
p (1 −s) −1
_
˜
V
_
1
p

−s
_
ϕ
(t) +
_
p
p (1 −s) −1
_
˜
V
_
1
p

−s
_
ϕ
(x) −
A Unified approach to Sawyer and Sinnamon characteri-
zations of the Hardy inequality for decreasing functions
79
_
p
p (1 −s) −1
_
˜
V
_
1
p

−s
_
ϕ
(t)
_
q
u(x) dx
_1
q
=
_
p
p (1 −s) −1
_
_
_

_
t
˜
V
q
_
1
p

−s
_
ϕ
(x) u(x) dx
_
_
1
q
.
By combining this estimate with (6.10) we find that, in view of (6.4),
_
p
p (1 −s) −1
_
_
_

_
t
˜
V
q
_
1
p

−s
_
ϕ
(x) u(x) φ(x) dx
_
_
1
q
≤ C ≤
__
p
p (1 −s) −1
_
p
+
1
ps
_
1
p
˜
V
−s
ϕ
(t)
i.e. that
_
_
_
p
p(1−s)−1
_
p
_
p
p(1−s)−1
_
p
+
1
ps
_
_
1
p
_
_

_
t
˜
V
q
_
1
p

−s
_
ϕ
(x) u(x) dx
_
_
1
q
˜
V
s
ϕ
(t) ≤ C < ∞.
By taking supremum over all t > 0 we find that (6.5) holds for 0 < s <
1
p
.
Hence, according to Theorem 5.5, all conditions (6.5) - (6.8) holds for all
s > 0. The proof is complete.
By applying Theorem 6.2 with ϕ ≡ 1 we obtain the following version of
the before mentioned result by G. Sinnamon:
Corollary 6.3. Let 1 < p ≤ q < ∞ and let u(x) and v (x) be weight
functions on (0, ∞). Then the inequality
_
_

_
0
_
_
x
_
0
f (t) v (t) dt
_
_
q
u(x) dx
_
_
1
q
≤ C
_
_

_
0
f
p
(x) v (x) dx
_
_
1
p
holds for some C < ∞ and for all decreasing f ≥ 0 if and only if one of the
following conditions holds for some s > 0 :
D
1
(s) = sup
t>0
_
_
_
_

_
t
u(x)
_
_
x
_
0
v (y) dy
_
_
q
_
1
p

−s
_
dx
_
_
_
_
1
q _
_
x
_
0
v (t) dt
_
_
s
< ∞.
80 Carleman type inequalities and Hardy type inequalities
for monotone functions
D

1
(s) = sup
t>0
_
_
_
_
t
_
0
v (x)
_
_

_
x
u(y) dy
_
_
p

_
1
q
−s
_
dx
_
_
_
_
1
p

_
_

_
t
u(x) dx
_
_
s
< ∞.
E
1
(s) = sup
t>0
_
_
_
_
t
_
0
u(x)
_
_
x
_
0
v (y) dy
_
_
q
_
1
p

+s
_
dx
_
_
_
_
1
q _
_
x
_
0
v (t) dt
_
_
−s
< ∞.
E

1
(s) = sup
t>0
_
_
_
_

_
t
v (x)
_
_

_
x
u(y) dy
_
_
p

_
1
q
+s
_
dx
_
_
_
_
1
p

_
_

_
t
u(x) dx
_
_
−s
< ∞.
Remark 6.3. Theorem 6.1 is obtained by just noting that condition (6.2)
coincides with condition E
1
_
1
p
_
< ∞.
Moreover, by applying Theorem 6.2 with ϕ(t) =
1
v(t)
and u(x) replaced
by u(x) x
−q
we obtain the following alternative to the Sawyer result for the
case when v (t) is increasing.
Corollary 6.4. Let 1 < p ≤ q < ∞ and let u(x) and v (x) be weight
functions on (0, ∞), where v (x) is increasing. Then the inequality
_
_

_
0
_
_
1
x
x
_
0
f (t) dt
_
_
q
u(x) dx
_
_
1
q
≤ C
_
_

_
0
f
p
(x) v (x) dx
_
_
1
p
(6.11)
holds for some C < ∞ and for all decreasing f ≥ 0 if and only if one of the
following conditions holds for some s > 0 :
F
1
(s) = sup
t>0
_
_
_
_

_
t
u(x) x
−q
_
_
x
_
0
v
1−p

(y) dy
_
_
q
_
1
p

−s
_
dx
_
_
_
_
1
q

_
_
t
_
0
v
1−p

(x) dx
_
_
s
< ∞.
A Unified approach to Sawyer and Sinnamon characteri-
zations of the Hardy inequality for decreasing functions
81
F

1
(s) = sup
t>0
_
_
_
_
t
_
0
v
1−p

(x)
_
_

_
x
u(y) y
−q
dy
_
_
p

_
1
q
−s
_
dx
_
_
_
_
1
p

_
_

_
t
u(x) x
−q
dx
_
_
s
< ∞.
F
2
(s) = sup
t>0
_
_
_
_
t
_
0
u(x) x
−q
_
_
x
_
0
v
1−p

(y) dy
_
_
q
_
1
p

+s
_
dx
_
_
_
_
1
q

_
_
t
_
0
v
1−p

(x) dx
_
_
−s
< ∞.
F

2
(s) = sup
t>0
_
_
_
_

_
t
v
1−p

(x)
_
_

_
x
u(y) y
−q
dy
_
_
p

_
1
q
+s
_
dx
_
_
_
_
1
p

_
_

_
t
u(x) x
−q
dx
_
_
−s
< ∞.
Remark 6.4. According to Theorem 5.1 two conditions are necessary to
characterize (6.11) for the general case but, in view of Corollary 6.4, for the
special case when v (x) is increasing only one condition is required (but there
are infinite many such equivalent conditions). Moreover, in this case each of
the conditions above are also equivalent.
6.3 Concluding remarks
Remark 6.5. Our proof of Theorem 6.2 shows that if C is the best possible
constant in (6.4), then
sup
s∈
_
0,
1
p

_
_
_
_
p
p(1−s)−1
_
p
_
p
p(1−s)
_
p
+
1
ps
_
_
1
p
D
ϕ
(s) ≤ C ≤ inf
s∈
_
0,
1
p

_
_
1
1 −sp

_ 1
p

D
ϕ
(s) .
82 Carleman type inequalities and Hardy type inequalities
for monotone functions
Remark 6.6. According to our proof of Theorem 6.2 we see that the suf-
ficient part of the Theorem holds in general i.e. without the additional as-
sumption that ϕ(x) is decreasing.
Remark 6.7. Note that condition F
1
(s) < ∞, with 0 < s <
1
p

coincides
with that obtained by Wedestig in [102] to characterize (6.11) for all functions
f ≥ 0. The explanation of this phenomenon is that among all rearrangements
of the function f the integral

_
0
f
p
(x) v (x) dx with v (x) non-decreasing, will
be the smallest possible when f is rearranged in decreasing order. It follows
that if v (x) is increasing, then (6.11) for all functions is equivalent to (6.11)
restricted to the cone of decreasing functions.
Remark 6.8. In view of Sinnamon’s result that (6.1) and (6.3) are in fact
equivalent our result also gives directly new results concerning embedding be-
tween Lorentz spaces. More exactly, let f

denote the decreasing rearrange-
ment of function f from the measurespace Ω. Then, for 1 < p ≤ q < ∞, the
inequality
_
_

_
0
(f

(t) dt)
q
u(x) dx
_
_
1
q
≤ C
_
_

_
0
(f

(x))
p
v (x) dx
_
_
1
p
holds for some finite constant C if and only if any of the conditions in Corol-
lary 6.3 holds for any s > 0 and with u(x) replaced by
u(x)
V
q
(x)
. Except for the
case E
1
_
1
p
_
< ∞ used by Sinnamon we think this fact is not pointed out
before in the literature.
Remark 6.9. A generalization of Sinnamon’s results in Theorem 6.1 can be
found in [85]. In fact, they prove a similar result for the case with a (Oinarov
type) kernel involved in the Hardy operator in the left hand side in (6.1). It
is an open question if the results in this Chapter can also be generalized to
this case in this generality.
Chapter 7
Hardy inequality with three
measures on monotone functions
7.1 Introduction
Let M
+
be the class consisting of all Borel functions f : [0, ∞) → [0, +∞]
and M ↓ (M ↑) be a subclass of M
+
, which consists of all decreasing
(increasing) functions f ∈ M
+
. Suppose that λ, µ and ν are positive Borel
σ-finite measures on [0, ∞) and u, v, w ∈ M
+
are weight functions.
In this chapter we study for 0 < p, q, s < ∞ the problem when the Hardy
inequality of the form
_
_
[0,∞)
(H
s
f)
q
vdµ
_1
q
≤ C
_
_
[0,∞)
f
p
wdν
_1
p
, (7.1)
holds for all f ∈ M↓ or for all f ∈ M↑, where
(H
s
f) (x) :=
_
_
[0,x]
f
s
udλ
_1
s
. (7.2)
Since by the substitution f
s
→ f the inequality (7.1) can be reduced to
the equivalent inequality with new parameters p and q of the form
_
_
[0,∞)
(Hf)
q
vdµ
_1
q
≤ C
_
_
[0,∞)
f
p
wdν
_1
p
, (7.3)
83
84 Carleman type inequalities and Hardy type inequalities
for monotone functions
where
(Hf) (x) := (H
1
f) (x) =
_
[0,x]
fudλ (7.4)
we may and shall restrict our studies to the inequality (7.3). All the charac-
terizations of (7.1) can be easily reproduced from the results for (7.3).
The weighted inequality (7.3) for f ∈ M ↓, when λ = µ = ν is the
Lebesgue measure, was essentially characterized in [93] and [100] with the
complement for the case 0 < q < 1 = p in [98] and recent contribution in
[13] for the case 0 < q < p < 1. In fact, [93], [100], [98] and [13] deal with
the case u(x) = 1, but a weight u can be incorporated with no change in
the arguments. A piece of historical remarks and the literature can be found
in ( [62] and [63, Chapter 6]). We summarize these results in the following
theorems:
Theorem 7.1. Let λ = µ = ν be the Lebesgue measure. Then the inequality
(7.3) holds for all f ∈ M↓ if and only if:
(a) 1 < p ≤ q < ∞, max (A
0
, A
1
) < ∞, where
A
0
:= sup
t>0
__
t
0
__
s
0
u
_
q
v(s)ds
_
1
q
__
t
0
w
_

1
p
,
and
A
1
:= sup
t>0
__

t
v
_1
q
_
_
t
0
__
s
0
u
_
p
__
s
0
w
_
−p

w(s)ds
_ 1
p

and C ≈ A
0
+A
1
.
(b) 0 < q < p < ∞, 1 < p < ∞,
1
r
:=
1
q

1
p
, max (B
0
, B
1
) < ∞, where
B
0
:=
_
_

0
__
t
0
w
_

r
p
__
t
0
__
s
0
u
_
q
v(s)ds
_
r
p
__
t
0
u
_
q
v (t) dt
_
1
r
,
and
B
1
:=
_
_
_

0
__

t
v
_r
p
_
_
t
0
__
s
0
u
_
p
__
s
0
w
_
−p

w(s)ds
_ r
p

v(t)dt
_
_
1
r
and C ≈ B
0
+B
1
.
Hardy inequality with three measures on monotone func-
tions
85
(c) 0 < q < p ≤ 1, max (B
0
, B
1
) < ∞, where
B
1
:=
_
_
_

0
_
esssup
s∈[0,t]
__
s
0
u
_
p
__
s
0
w
_
−1
_r
p
__

t
v
_r
p
v(t)dt
_
_
1
r
and C ≈ B
0
+B
1
.
(d) 0 < p ≤ q < ∞, 0 < p ≤ 1, max (A
0
, /
1
) < ∞, where
/
1
:= sup
t>0
__
t
0
u
___

t
v
_1
q
__
t
0
w
_

1
p
and C ≈ A
0
+/
1
.
It is important to note, that the weighted case of (7.3) for 1 < p, q < ∞
was solved in [93] by proving the principle of duality which allows to reduce
an inequality with a positive operator on monotone functions to an inequality
with modified operator on non-negative functions. The other cases, when
p, q ,∈ (1, ∞) were studied by different methods.
Our aim is twofold. First we study the inequality (7.3) in the case 0 < p ≤
1 proving a complete analog of the parts (c) and (d) of Theorem 7.1 (Section
7.3). In the case 0 < q < p ≤ 1 our method is based on the characterization
of the Hardy inequality on nonnegative functions in the case 0 < q < 1 = p,
which we establish in Section 7.3 (Theorem 7.12). This approach is direct
and different from discretization methods of [13] and [37].
Hardy inequality (7.3) on monotone functions with two different measures
was recently investigated by G. Sinnamon [97]. Namely, for 1 < p < ∞ and
0 < q < ∞ the author established the equivalence of (7.3) with u ≡ v ≡ w ≡
1 and dλ = dν for f ∈ M
+
to the same inequality restricted to f ∈ M ↓ .
Moreover, such equivalence takes place also for more general operator than
(7.4), that is for the operator
(Kf) (x) =
_
[0,x]
k (x, y) f (y) dλ(y)
with a kernel k (x, y) ≥ 0, which is monotone in the variable y (see [85,
Theorem 2.3]). Moreover, G. Sinnamon [97] extended the Sawyer principle
of duality for measures. We apply this extension to characterize (7.3) in
case 1 < p, q < ∞ (Section 4) combining with the recent results by D.V.
Prokhorov [87] for the inequality (7.3) on f ∈ M
+
with 1 < p < ∞ and
0 < q < ∞ extended by the same author for the Hardy operator with
Oinarov kernel [88].
86 Carleman type inequalities and Hardy type inequalities
for monotone functions
7.2 Preliminary remarks
Denote
Λ
f
(x) :=
_
[0,x]
fdλ, and
¯
Λ
f
(x) :=
_
[x,∞)
fdλ. (7.5)
We need the following statements.
Lemma 7.2. ([87], Lemma 1) If γ > 0, then
Λ
f
(∞)
γ+1
max ¦1, γ + 1¦

_
[0,∞)
f (x) Λ
f
(x)
γ
dλ(x) ≤
Λ
f
(∞)
γ+1
min ¦1, γ + 1¦
(7.6)
holds. If γ ∈ (−1, 0) and Λ
f
(∞) < +∞, then (7.6) holds.
Lemma 7.3. ([87], Lemma 2) If γ > 0, then
¯
Λ
f
(0)
γ+1
max ¦1, γ + 1¦

_
[0,∞)
f (x)
¯
Λ
f
(x)
γ
dλ(x) ≤
¯
Λ
f
(0)
γ+1
min ¦1, γ + 1¦
(7.7)
holds. If γ ∈ (−1, 0) and
¯
Λ
f
(0) < +∞, then (7.7) holds.
The following two statements can be obtained from [[96], Lemma 2.1]
(see also [[97], Proposition 1.5]).
Lemma 7.4. Let f ∈ M ↑ with f(0) = 0 and let η be a Borel measure on
[0, ∞). Then there exist f
0
∈ M ↑ and the sequence ¦h
n
¦
n≥1
⊂ M
+
such
that
(1) f
0
(x) ≤ f(x) for all x ∈ [0, ∞).
(2) f
0
(x) = f(x) for η-a.e. x ∈ [0, ∞).
(3) f
n
(x) :=
_
[0,x])
h
n
dη ≤ f
0
(x) for all x ∈ [0, ∞).
(4) For all x ∈ [0, ∞) the sequence ¦f
n
(x)¦
n≥1
is nondecreasing in n and
f
0
(x) = lim
n→∞
f
n
(x) η-a.e. x ∈ [0, ∞).
Lemma 7.5. Let f ∈ M ↓ with f(+∞) = 0 and let η be a Borel measure
on [0, ∞). Then there exist f
0
∈ M↓ and the sequence ¦h
n
¦
n≥1
⊂ M
+
such
that
(1) f
0
(x) ≤ f(x) for all x ∈ [0, ∞).
(2) f
0
(x) = f(x) for η-a.e. x ∈ [0, ∞).
(3) f
n
(x) :=
_
[x,∞)
h
n
dη ≤ f
0
(x) for all x ∈ [0, ∞).
(4) For all x ∈ [0, ∞) the sequence ¦f
n
(x)¦
n≥1
is nondecreasing in n and
f
0
(x) = lim
n→∞
f
n
(x) η-a.e. x ∈ [0, ∞).
Hardy inequality with three measures on monotone func-
tions
87
Remark 7.1. Two similar lemmas are valid for the approximation from
above.
The following statements are taken from [96] and concern the weighted
L
p
λ
[0, ∞) −L
q
µ
[0, ∞) inequality with the operator of the form
(K
u
f) (x) =
_
[0,x]
k (x, y) u(y) f (y) dλ(y) .
Here the kernel k (x, y) ≥ 0 is µ λ - measurable on [0, ∞) [0, ∞) and
satisfies the following Oinarov condition. There is a constant D ≥ 1 such
that
D
−1
k (x, y) ≤ k (x, z) +k (z, y) ≤ Dk (x, y) , 0 ≤ y ≤ z ≤ x. (7.8)
Theorem 7.6. Let 1 < p ≤ q < ∞. Then the inequality
_
_
[0,∞)
(K
u
f)
q
vdµ
_1
q
≤ C
_
_
[0,∞)
f
p

_1
p
(7.9)
holds for all f ∈ M
+
if and only if A : =max (A
0,1
, A
0,2
) < ∞, where
A
0,1
:= sup
t∈[0,∞)
_
_
[t,∞)
v(x)k (x, t)
q
dµ(x)
_1
q
_
_
[0,t]
u
p


_ 1
p

,
A
0,2
:= sup
t∈[0,∞)
_
_
[t,∞)
vdµ
_1
q
_
_
[0,t]
k (t, y)
p

u(y)
p

dλ(y)
_ 1
p

.
If 1 < q < p < ∞ and
1
r
=
1
q

1
p
, then inequality (7.9) holds for all f ∈ M
+
if and only if B : =max (B
0,1
, B
0,2
) < ∞, where
B
0,1
:=
_
_
_
[0,∞)
_
_
[t,∞)
v(x)k (x, t)
q
dµ(x)
_r
q
_
_
[0,t]
u
p


_ r
q

u(t)
p

dλ(t)
_
_
1
r
,
B
0,2
:=
_
_
_
[0,∞)
_
_
[t,∞)
vdµ
_r
p
_
_
[0,t]
k (t, y)
p

u(y)
p

dλ(y)
_ r
p

v (t) dµ(t)
_
_
1
r
.
88 Carleman type inequalities and Hardy type inequalities
for monotone functions
The next statement is an analog of the previous theorem for the operator
K

u
of the dual form
(K

u
f) (x) =
_
[x,∞)
k (y, x) u(y) f (y) dλ(y)
with a kernel satisfying Oinarov’s condition (7.8).
Theorem 7.7. Let 1 < p ≤ q < ∞. Then the inequality
_
_
[0,∞)
(K

u
f)
q
vdµ
_1
q
≤ C
_
_
[0,∞)
f
p

_1
p
(7.10)
holds for all f ∈ M
+
if and only if A

: =max
_
A

0,1
, A

0,2
_
< ∞, where
A

0,1
:= sup
t∈[0,∞)
_
_
[0,t])
v(x)k (t, x)
q
dµ(x)
_1
q
_
_
[t,∞)
u
p


_ 1
p

,
A

0,2
:= sup
t∈[0,∞)
_
_
[0,t]
vdµ
_1
q
_
_
[t,∞)
k (y, t)
p

u(y)
p

dλ(y)
_1
p

.
If 1 < q < p < ∞ and
1
r
=
1
q

1
p
, then inequality (7.10) holds for all f ∈ M
+
if and only if B

: =max
_
B

0,1
, B

0,2
_
< ∞, where
B

0,1
:=
_
_
_
[0,∞)
_
_
[0,t]
v(x)k (t, x)
q
dµ(x)
_r
q
_
_
[t,∞)
u
p


_r
q

u(t)
p

dλ(t)
_
_
1
r
,
B

0,2
:=
_
_
_
[0,∞)
_
_
[0,t]
vdµ
_r
p
_
_
[t,∞)
k (y, t)
p

u(y)
p

dλ(y)
_ r
p

v (t) dµ(t)
_
_
1
r
.
In the following theorems we collect weight versions of the results ob-
tained by G. Sinnamon in [97] for embeddings the cones of monotone func-
tions. Put
W (t) :=
_
[0,t]
wdν, and
¯
W (x) :=
_
[x,∞)
wdν. (7.11)
Hardy inequality with three measures on monotone func-
tions
89
Theorem 7.8. If 0 < p ≤ q < ∞, then
sup
F∈M↓
_
_
[0,∞)
F
q
vdµ
_1
q
_
_
[0,∞)
F
p
wdν
_1
p
= sup
x∈[0,∞)
_
_
[0,x]
vdµ
_1
q
_
_
[0,x]
wdν
_1
p
. (7.12)
Theorem 7.9. If 0 < q < p < ∞, and
1
r
=
1
q

1
p
then
sup
F∈M↓
_
_
[0,∞)
F
q
vdµ
_1
q
_
_
[0,∞)
F
p
wdν
_1
p

_
_
_
[0,∞)
w(y)
_
_
[y,∞)
W
−1
vdµ
_r
q
dν(y)
_
_
1
r
.
(7.13)
Analogous results take place for F ∈ M↑ .
Theorem 7.10. If 0 < p ≤ q < ∞, then
sup
F∈M↑
_
_
[0,∞)
F
q
vdµ
_1
q
_
_
[0,∞)
F
p
wdν
_1
p
= sup
x∈[0,∞)
_
_
[x,∞)
vdµ
_1
q
_
_
[x,∞)
wdν
_1
p
. (7.14)
Theorem 7.11. If 0 < q < p < ∞ and
1
r
=
1
q

1
p
, then
sup
F∈M↑
_
_
[0,∞)
F
q
vdµ
_1
q
_
_
[0,∞)
F
p
wdν
_1
p

_
_
_
[0,∞)
w(y)
_
_
[0,y]
¯
W
−1
vdµ
_r
q
dν(y)
_
_
1
r
.
(7.15)
Note that Theorems 7.9 and 7.11 with q = 1 give analogs of Sawyer’s
principle of duality with general Borel measures.
7.3 The case 0 < p ≤ 1
We need the following extension of ([98], Theorem 3.3) from the weighted
case to the case of measures.
Theorem 7.12. Let 0 < q < 1, ν = ν
a

s
, ν
a
=

a

dλ and ν
s
⊥ λ . Then
_
_
[0,∞)
_
_
[0,x]
fudλ
_
q
v (x) dµ(x)
_
1
q
≤ C
_
[0,∞)
fwdν (7.16)
90 Carleman type inequalities and Hardy type inequalities
for monotone functions
holds for all f ∈ M
+
if and only if
B :=
_
_
_
[0,∞)
_
_
[0,y]
v (z) dµ(z)
˜ w

(z)
_
q
1−q
v (y) dµ(y)
_
_
1−q
q
< ∞,
where
˜ w :=
w
u

a

and ˜ w(x)

:= ess inf
t∈[0,x]
λ
˜ w(t) . (7.17)
Moreover, C ≈ B.
Proof. Let us start with proving that (7.16) is equivalent to the following
inequality
_
_
[0,∞)
_
_
[0,x]
fudλ
_
q
v (x) dµ(x)
_
1
q
≤ C
_
[0,∞)
fw

a

dλ. (7.18)
Obviously, (7.18) implies (7.16). Let (7.16) hold and f ∈ M
+
. If ν
s
⊥ λ,
then there exists A ⊂ [0, ∞) such that λ(A) = 0, suppν
s
= A and suppν
a
=
[0, ∞) ¸A . Let
˜
f = fχ
[0,∞)\A
. Then
_
_
[0,∞)
_
_
[0,x]
fudλ
_
q
v (x) dµ(x)
_1
q
=
_
_
[0,∞)
_
_
[0,x]
˜
fudλ
_
q
v (x) dµ(x)
_
1
q
≤ C
_
[0,∞)
˜
fwdν
= C
_
_
[0,∞)
˜
fwdν
a
+
_
[0,∞)
˜
fwdν
s
_
= C
_
[0,∞)
fwdν
a
.
Now if we use (7.17), then (7.18) is equivalent to
_
_
[0,∞)
_
_
[0,x]
fudλ
_
q
v (x) dµ(x)
_
1
q
≤ C
_
[0,∞)
fu ˜ wdλ. (7.19)
Hardy inequality with three measures on monotone func-
tions
91
Then, by [96, Theorem 3.1] and changing fu to f, we get that (7.19) is
equivalent to
_
_
[0,∞)
_
_
[0,x]
fdλ
_
q
v (x) dµ(x)
_
1
q
≤ C
_
[0,∞)
f ˜ w

dλ. (7.20)
Now we follow the proof of [98, Theorem 3.3]. First let ˜ w

(x) =
_
[x,∞)
bdλ
λ-a.e. x ∈ [0, ∞) ,
_
[0,∞)
bdλ = ∞ and
_
[x,∞]
bdλ < ∞. Then, by changing
order of integration the right hand side of (7.20) is equal to
C
_
[0,∞)
_
_
[0,x]
fdλ
_
bdλ
and hence (7.20) is equivalent to
_
_
[0,∞)
_
_
[0,x]
fdλ
_
q
v (x) dµ(x)
_
1
q
≤ C
_
[0,∞)
_
_
[0,x]
fdλ
_
bdλ. (7.21)
Since
_
[0,x]
fdλ is increasing we can replace it with F and so (7.21) is equiv-
alent to
_
_
[0,∞)
F
q
v (x) dµ(x)
_1
q
≤ C
_
[0,∞)
Fbdλ with F ∈ M↑ . (7.22)
By [97, Theorem 2.5] and using Lemma 7.3 we get that
C ≈
_
_
_
[0,∞)
_
_
[0,x]
v (y) dµ(y)
˜ w

(y)
_ 1
1−q
b (x) dλ(x)
_
_
q
1−q

_
_
_
[0,∞)
_
[0,x]
v (y) dµ(y)
˜ w

(y)
_
_
[0,y]
v (z) dµ(z)
˜ w

(z)
_
q
1−q
b (x) dλ(x)
_
_
1−q
q
=
_
_
_
[0,∞)
_
_
[0,y]
v (z) dµ(z)
˜ w

(z)
_
q
1−q
v (y) dµ(y)
_
_
1−q
q
.
For a general ˜ w

we may and shall suppose that ˜ w

(x) < ∞ for all x > 0.
Let N ∈ N and
w
N
(x) := χ
[0,N]
(x) ˜ w

(x) .
92 Carleman type inequalities and Hardy type inequalities
for monotone functions
Then w
N
(+∞) = 0 and similar to Lemma 7.5 we find that w
(0)
N
∈ M↓ and
h
n
∈ M
+
(n ∈ N) such that
(1) w
N
(x) ≤ w
(0)
N
(x) for all x ∈ [0, ∞).
(2) w
N
(x) = w
(0)
N
(x) for λ-a.e. x ∈ [0, ∞).
(3) w
N,k
(x) :=
_
[x,∞)
h
k
dλ ≥ w
(0)
N
(x) for all x ∈ [0, ∞).
(4) The sequence ¦w
N,k
(x)¦
k≥1
is decreasing in k for all x ∈ [0, ∞) and
w
(0)
N
(x) = lim
k→∞
w
N,k
(x) λ-a.e. x ∈ [0, ∞). Then, by the previous part of the
proof, for any f ∈ M
+
we have that
_
_
[0,N]
_
_
[0,x]
fdλ
_
q
v (x) dµ(x)
_
1
q
¸
_
_
_
[0,N]
_
_
[0,x]
v (z) dµ(z)
w
N,k
(z)
_
q
1−q
v (x) dµ(x)
_
_
1−q
q
_
[0,N]
fw
N,k
dλ.
According to [87, Lemma 5] this is equivalent to
_
_
[0,N]
_
_
[0,x]
f
w
N,k

_
q
v (x) dµ(x)
_1
q
¸
_
_
_
[0,N]
_
_
[0,x]
v (z) dµ(z)
w
N,k
(z)
_
q
1−q
v (x) dµ(x)
_
_
1−q
q
_
[0,N]
fdλ.
By (3) and (1) we have that
1
w
N,k
(z)

1
w
N
(z)
and by (4), (2) and the Mono-
tone Convergence Theorem, it yields that
lim
k→∞
_
[0,x]
f
w
N,k
dλ =
_
[0,x]
f
w
(0)
N
dλ =
_
[0,x]
f
w
N
dλ.
Making the reverse change
f
w
N
→ f we find that
_
_
[0,N]
_
_
[0,x]
fdλ
_
q
v (x) dµ(x)
_
1
q
¸
_
_
_
[0,N]
_
_
[0,x]
v (z) dµ(z)
w
N
(z)
_
q
1−q
v (x) dµ(x)
_
_
1−q
q
_
[0,N]
fw
N

Hardy inequality with three measures on monotone func-
tions
93
=
_
_
_
[0,N]
_
_
[0,x]
v (z) dµ(z)
˜ w

(z)
_
q
1−q
v (x) dµ(x)
_
_
1−q
q
_
[0,N]
f ˜ w


≤ B
_
[0,∞)
f ˜ w

dλ.
Letting N −→ ∞ we arrive at C ¸ B. To show the reverse inequality
we again approximate ˜ w

from above by a monotone sequence of functions
w
k
(x) :=
_
[x,∞)
b
k
dλ ↓ ˜ w

. Then, by applying (7.21), (7.22) and [97, Theo-
rem 2.5], we find that
_
_
_
[0,∞)
_
_
[0,y]
v (z) dµ(z)
w
k
(z)
_
q
1−q
v (y) dµ(y)
_
_
1−q
q
¸ C
and since w
−1
k
↑ ˜ w
−1

the result follows.
Definition 7.1. Let w ∈ M ↓ and be continuos on the left. It is known (
[92, Chapter 12, §3]), that there exists a Borel measure, say η
w
, such that
w(x) =
_
[x,∞)

w
+w(+∞) . We say that w ∈ 1
2
(0) if there exist a constant
C ≥ 1 such that
1
w(x)

1
w(0)
≤ C
_
[0,x]

w
w
2
, x > 0.
Corollary 7.13. Let 0 < q < 1, w ∈ M↓ and w ∈ 1
2
(0). Then
_
_
[0,∞)
_
_
[0,x]
hdλ
_
q
v (x) dµ(x)
_
1
q
≤ C
_
[0,∞)
hwdλ
holds for all h ∈ M
+
if and only if
B :=
_
_
_
[0,∞)
_
_
[0,x]
vdµ
w
_
q
1−q
v (x) dµ(x)
_
_
1−q
q
< ∞.
Moreover, C ≈ B ≈ B
0
+B
1
, where
B
0
:=
_
_
[0,x]
vdµ
_1
q
w(0)

1
p
94 Carleman type inequalities and Hardy type inequalities
for monotone functions
B
1
:=
_
_
_
[0,∞)
w

q
1−q
_
_
[x,∞)
vdµ
_
q
1−q
v (x) dµ(x)
_
_
1−q
q
.
Proof. It follows from Theorem 7.12, Lemma 7.3 and [97, Theorem 2.6].
Denote
Λ(t) := Λ
u
(t) =
_
[0,t]
udλ (7.23)
and observe that by the change f
p
→ f in the inequality (7.3) we get the
following equivalent inequality:
_
_
[0,∞)
_
Hf
1
p
_
q
vdµ
_
p
q
≤ C
p
_
_
[0,∞)
fwdν
_
, f ∈ M↓ . (7.24)
Theorem 7.14. (a) Let 0 < p ≤ q < ∞ and 0 < p ≤ 1. Then (7.3) holds
for all f ∈ M↓ if and only if
A
0
:= sup
t∈[0,∞)
_
_
[0,t]
wdν
_

1
p
_
_
[0,t]
Λ
q
vdµ
_1
q
< ∞,
and
/
1
:= sup
t∈[0,∞)
Λ(t)
_
_
[0,t]
wdν
_

1
p
_
_
[t,∞)
vdµ
_1
q
< ∞
and C ≈ A
0
+/
1
(b) Let 0 < q < 1 = p. Then (7.3) holds for all f ∈ M↓ if and only if
B
0
:=
_
_
_
[0,∞)
w(y)
_
_
[y,∞)
W
−1
Λ
q
vdµ
_ 1
1−q
dν (y)
_
_
1−q
q
< ∞,
and
B
1
:=
_
_
_
[0,∞)
_
_
[0,x)
ess sup
s∈[0,t]
λ
Λ(s)
W (s)
v(t)dµ(t)
_
q
1−q
v (t) dµ(t)
_
_
1−q
q
< ∞
and C ≈ B
0
+B
1
Hardy inequality with three measures on monotone func-
tions
95
(c) Let 0 < q < p < 1, 1
p
(t) := ess sup
s∈[0,t]
λ
Λ
p
(s)
W(s)
. Then (7.3) holds for all
f ∈ M↓ if
B
0
:=
_
_
_
[0,∞)
w(y)
_
_
[y,∞)
W
−1
Λ
q
vdµ
_
p
p−q
dν (y)
_
_
p−q
pq
< ∞,
and
B
1
:=
_
_
_
[0,∞)
_
_
[0,x)
1
p
(t) v(t)dµ(t)
_
q
p−q
v (t) dµ(t)
_
_
p−q
pq
< ∞
and if and only if B
0
+B
1
< ∞, provided 1
p
(t) is continuous on (0, ∞) and
1
V
p
(t)
∈ 1
2
(0). Then C ≈ B
0
+B
1
Proof. (a) Since f ∈ M↓, then (H
u
f) (x) ≥ f (x) Λ(x) and (7.3) implies
_
_
[0,∞)
f
q
Λ
q
vdµ
_1
q
≤ C
_
_
[0,∞)
f
p
wdν
_1
p
, f ∈ M↓ .
It is known (see Theorem 7.8) that C = A
0
for 0 < p ≤ q < ∞.
Now, if f
t
= χ
[0,t]
in (7.3), then
C
_
_
[0,t]
wdν
_1
p

_
_
[t,∞)
(H
u
f
t
)
q
vdµ
_1
q
= Λ(t)
_
_
[t,∞)
vdµ
_1
q
,
which implies that C ≥ /
1
. Consequently, A
0
+/
1
≤ 2C.
For the sufficiency we suppose first that f ∈ M↓, f (x) =
_
[x,∞)
hudλ for
λ-a.e. x ∈ [0, ∞), where h ∈ M
+
and f (x) ≥
_
[x,∞)
hudλ for all x ∈ [0, ∞).
96 Carleman type inequalities and Hardy type inequalities
for monotone functions
Let 0 < p < 1. We have, by Lemma 7.3,
_
[0,x]
_
_
[s,∞)
hudλ
_
u(s) dλ(s)

_
[0,x]
_
_
_
[s,∞)
_
_
[y,∞)
hudλ
_
p−1
h(y)u(y)dλ(y)
_
_
1
p
u(s) dλ(s)
¸
_
[0,x]
_
_
_
[s,x]
_
_
[y,∞)
hudλ
_
p−1
h(y)u(y)dλ
_
_
1
p
u(s) dλ(s) + Λ(x) f (x)
[by the Minkowski inequality] (7.25)

_
_
_
[0,x]
_
_
[y,∞)
hudλ
_
p−1
h(y)u(y)Λ(y)
p
dλ(y)
_
_
1
p
+ Λ(x) f (x) .
Applying (7.25) we obtain that
_
_
[0,∞)
(Hf)
q
vdµ
_1
q
¸
_
_
[0,∞)
f
q
Λ
q
vdµ
_1
q
+J, (7.26)
where
J :=
_
_
_
_
[0,∞)
_
_
_
[0,x]
_
_
[y,∞)
hudλ
_
p−1
h(y)u(y)Λ(y)
p
dλ(y)
_
_
q
p
v (x) dµ(x)
_
_
_
1
q
.
For the first term on the right hand side of (7.26), according to Theorem
7.8, we have that
_
_
[0,∞)
f
q
Λ
q
vdµ
_1
q
≤ A
0
_
_
[0,∞)
f
p
wdν
_1
p
. (7.27)
For the second term on the right hand side of (7.26), by the Minkowski
inequality with
q
p
≥ 1 and Lemma 7.3, we find that
J ≤
_
_
_
[0,∞)
_
_
[y,∞)
hudλ
_
p−1
h(y)u(y)Λ(y)
p
_
_
[y,∞)
vdµ
_
p
q
dλ(y)
_
_
1
p
Hardy inequality with three measures on monotone func-
tions
97
≤ /
1
_
_
_
[0,∞)
_
_
[y,∞)
hudλ
_
p−1
h(y)u(y)
_
_
[0,y]
wdν
_
dλy
_
_
1
p
≈ /
1
_
_
[0,∞)
_
_
[s,∞)
hudλ
_
p
w(s)dν(s)
_
1
p
≤ /
1
_
_
[0,∞)
f
p
wdν
_1
p
and the inequality
_
_
[0,∞)
(Hf)
q
vdµ
_1
q
¸ (A
0
+/
1
)
_
_
[0,∞)
f
p
wdν
_1
p
(7.28)
in this case follows. For an arbitrary f ∈ M↓ without loss of generality we
may suppose that f(+∞) = 0 and find, by Lemma 7.5, that f
0
∈ M↓ and
a sequence ¦h
n
¦
n≥1
⊂ M
+
such that
(1) f
0
(x) ≤ f(x) for all x ∈ [0, ∞).
(2) f
0
(x) = f(x) for λ-a.e. x ∈ [0, ∞).
(3) f
n
(x) :=
_
[x,∞)
h
n
udλ ≤ f
0
(x) for all x ∈ [0, ∞).
(4) For all x ∈ [0, ∞) the sequence ¦f
n
(x)¦
n≥1
is nondecreasing in n and
f
0
(x) = lim
n→∞
f
n
(x) λ-a.e. x ∈ [0, ∞).
Then, by the Monotone Convergence Theorem and (7.28), it yields that
_
_
[0,∞)
(Hf)
q
vdµ
_1
q
(2)
=
_
_
[0,∞)
(Hf
0
)
q
vdµ
_1
q
(4)
= lim
n→∞
_
_
[0,∞)
(Hf
n
)
q
vdµ
_1
q
(7.28)
¸ (A
0
+/
1
) lim
n→∞
_
_
[0,∞)
f
p
n
wdν
_1
p
(3)
≤ (A
0
+/
1
)
_
_
[0,∞)
f
p
0
wdν
_1
p
(1)
≤ (A
0
+/
1
)
_
_
[0,∞)
f
p
wdν
_1
p
98 Carleman type inequalities and Hardy type inequalities
for monotone functions
and the upper bound C ¸ A
0
+/
1
is proved. The case p = 1 is treated by
the same method, but even simpler.
(b) Necessity. It follows from the inequality
_
_
[0,∞)
(Hf)
q
vdµ
_1
q
≤ C
_
[0,∞)
fwdν, f ∈ M↓, (7.29)
that
_
_
[0,∞)
f
q
Λ
q
vdµ
_1
q
≤ C
_
[0,∞)
fwdν, f ∈ M↓ . (7.30)
The last inequality is characterized by B
0
(see Theorem 7.9 with p = 1.)
Hence, B
0
≤ C. Now, suppose h ∈ M
+
and f(x) =
_
[x,∞)
hudλ. Then
f ∈ M↓ and (7.29) gives
_
_
[0,∞)
_
_
[0,x]
_
_
[s,∞)
hudλ
_
u(s) dλ(s)
_
q
v (x) dµ(x)
_
1
q
≤ C
_
[0,∞)
_
_
[s,∞)
hudλ
_
w(x) dν (x) .
This implies that
_
_
[0,∞)
_
_
[0,x]
hΛudλ
_
q
v (x) dµ(x)
_
1
q
≤ C
_
[0,∞)
hWudλ.
Changing the variable hΛu → h, we obtain that
_
_
[0,∞)
_
_
[0,x]
hdλ
_
q
v (x) dµ(x)
_
1
q
≤ C
_
[0,∞)
h
W
Λ
udλ.
The last inequality is characterized by Theorem 7.12. Consequently,
B
1
¸ C.
Sufficiency. Again, suppose first that f ∈ M ↓, f (x) =
_
[x,∞)
hudλ for
λ-a.e. x ∈ [0, ∞), where h ∈ M and f (x) ≥
_
[x,∞)
hudλ for all x ∈ [0, ∞).
Hardy inequality with three measures on monotone func-
tions
99
Then, we have that
_
_
[0,∞)
(Hf)
q
vdµ
_1
q
=
_
_
[0,∞)
_
_
[0,x]
_
_
[s,∞)
hudλ
_
u(s)dλ(s)
_
q
vdµ
_1
q
¸
_
_
[0,∞)
_
_
[0,x]
_
_
[s,x)
hudλ
_
u(s)dλ(s)
_
q
vdµ
_
1
q
+
_
_
[0,∞)
_
_
[x,∞)
hudλ
_
q
Λ
q
(x)v(x)dµ(x)
_
1
q

_
_
[0,∞)
_
_
[0,x]
hΛudλ
_
q
v(x)dµ(x)
_
1
q
+
_
_
[0,∞)
f
q
Λ
q
vdµ
_1
q
[applying Theorem 7.12 and Theorem 2.8]
¸B
1
_
_
[0,∞)
_
_
[x,∞]
hudλ
_
w(x)dν(x)
_
+B
0
_
_
[0,∞)
fwdν
_
≤ (B
0
+B
1
)
_
[0,∞)
fwdν.
For an arbitrary f ∈ M↓ we use the arguments from the end of the part
(a).
(c) Sufficiency. To prove (7.3) we again, suppose first that f ∈ M ↓
, f (x) =
_
[x,∞)
hudλ for λ-a.e. x ∈ [0, ∞), where h ∈ M
+
and f (x) ≥
_
[x,∞)
hudλ for all x ∈ [0, ∞). Then, arguing as before and applying Minkowskii’s
inequality, we find that
_
_
[0,∞)
_
Hf
1
p
_
q
v(x)dµ(x)
_
p
q
=
_
_
_
[0,∞)
_
_
_
[0,x]
_
_
[s,∞)
hudλ
_1
p
u(s)dλ(s)
_
_
q
v(x)dµ(x)
_
_
p
q
100 Carleman type inequalities and Hardy type inequalities
for monotone functions
¸
_
_
_
[0,∞)
_
_
_
[0,x]
_
_
[s,x)
hudλ
_1
p
u(s)dλ(s)
_
_
q
vdµ
_
_
p
q
+
_
_
_
[0,∞)
_
_
[x,∞)
hudλ
_
q
p
Λ
q
(x)v(x)dµ(x)
_
_
p
q

_
_
_
[0,∞)
_
_
[0,x]

p
udλ
_
q
p
v(x)dµ(x)
_
_
p
q
+
_
_
[0,∞)
f
q
p
Λ
q
vdµ
_
p
q
[applying Theorem 7.12 and Theorem 7.8]
¸ B
p
1
_
_
[0,∞)
_
_
[x,∞]
hudλ
_
w(x)dν(x)
_
+B
p
0
_
_
[0,∞)
fwdν
_
≤ (B
p
0
+B
p
1
)
_
[0,∞)
fwdν.
For an arbitrary f ∈ M ↓ we again use the arguments from the end of the
part (a).
Necessity. The inequality B
0
≤ C follows by using similar arguments as
in the proof of A
0
≤ C and B
0
≤ C in the parts (a) and (b).
For the rest it is sufficient to show that (7.24) implies the inequality
C ¸ B
1
.
Suppose for simplicity, that 1
p
(0) = 0. Let
g (t) := max
_
2
m
, m ∈ Z: 2
m
≤ 1
r
p
p
(t)
_
and
τ
m
:= inf
_
y ∈ [0, ∞) : 2
m
≤ 1
r
p
p
(y)
_
.
Since 1
p
(t) is continuous, then τ
m
exists for all m ∈ Z, τ
m
↑ and
Λ(τ
m
)
r
W (τ
m
)
r
p
= 2
m
= 1
r
p
p

m
) ≤ 1
r
p
p
(t) ≤ 2
m+1
, t ∈ [τ
m
, τ
m+1
] ,
g (τ
m
) = 2
m
, g (s) ≤ 2
m−1
for all s ∈ [0, τ
m
) .
We note that
g (t) =

m∈Z
2
m
χ

m

m+1
)
(t) ≤ 1
r
p
p
(t) (7.31)
Hardy inequality with three measures on monotone func-
tions
101
and define
f (t) :=
_
[t,∞)
_
_
[x,∞)
vdµ
_r
q
W (x)
dg (x) .
Then f ∈ M↓ and, by Lemma 7.3, it yields that
_
[0,∞)
fwdν =
_
[0,∞)
_
_
[x,∞)
vdµ
_r
q
dg (x)

_
[0,∞)
g (x)
_
_
[x,∞)
vdµ
_r
p
v (x) dµ(x)

_
[0,∞)
1
r
p
p
(x)
_
_
[x,∞)
vdµ
_r
p
v (x) dµ(x) := B
r
2,1
.
On the other hand
_
_
[0,∞)
_
_
[0,x]
f
1
p
(y)dΛ(y)
_
q
v (x) dµ(x)
_
1
q

_

m
_

m

m+1
)
v (x)

_
_
_
_
_
[0,τ
m
]
_
_
_
_
[y,τ
m
]
_
_
[s,∞)
vdµ
_r
q
W (s)
dg (s)
_
_
_
1
p
dΛ(y)
_
_
_
_
q
dµ(x)
_
_
_
_
1
q

_
_

m
_
_

m

m+1
)
vdµ
__
_

m
,∞)
vdµ
_r
p

_
W (τ
m
)

1
p
_
[0,τ
m
]
(g (τ
m
) −g (y))
1
p
dΛ(y)
_
q
_
1
q
¸
_
_

m
_
_

m

m+1
)
vdµ
__
_

m
,∞)
vdµ
_r
p
_
2
m
p
Λ(τ
m
)
W (τ
m
)
1
p
_
q
_
_
1
q
102 Carleman type inequalities and Hardy type inequalities
for monotone functions

_
_

m
2
m
_

m

m+1
)
_
_
[s,∞)
vdµ
_r
p
v (s) dµ(s)
_
_
1
q
¸
_
_
_
[0,∞)
1
r
p
p
(s)
_
_
[s,∞)
vdµ
_r
p
v (s) dµ(s)
_
_
1
q
=: B
r
q
2,1
With such f (x) the inequality (7.24) gives that: C
p
B
r
2,1
¸ B
pr
q
2,1
⇒ C ¸ B
2,1
.
Now, if we put f = χ
{0}
in (7.24), then we find that
C ≥
_
_
[0,∞)
vdµ
_1
q
_
W (0)
Λ
p
(0)
_

1
p
=
_
_
[0,∞)
vdµ
_1
q
_
1
1
p
(0)
_

1
p
=: B
2,0
.
It follows from Corollary 7.13 that B
2,1
+ B
2,0
¸ B
1
. Hence, C ¸ B
1
and
the proof is complete.
In conclusion of this section we give an analog of part (a) of the previous
theorem for non-decreasing functions.
Theorem 7.15. Let 0 < p ≤ q < ∞ and 0 < p ≤ 1. Then, (7.3) holds for
all f ∈ M↑ if and only if
¯
A
1
:= sup
t∈[0,∞)
_
_
[t,∞)
Λ
q
(x, t) v (x) dµ(x)
_1
q
¯
W

1
p
(x) < ∞,
where
Λ(x, t) :=
_
[t,x]
udλ,
and C ≈
¯
A
1
.
Proof. Replacing f in (7.3) by f
t
:= χ
[t,∞)
we find that
¯
A
1
≤ C. For suffi-
ciency we suppose that
f (x) =
_
[0,x]
hudλ, h ∈ M
+
and let 0 < p < 1. Then, by Minkowskii inequality and Lemma 7.2, we find
Hardy inequality with three measures on monotone func-
tions
103
that
_
[0,x]
_
_
[0,s]
hudλ
_
u(s) dλ(s)

_
[0,x]
_
_
_
[0,s]
_
_
[0,y]
hudλ
_
p−1
h(y) u(y) dλ(y)
_
_
1
p
u(s) dλ(s)

_
_
_
[0,x]
_
_
[0,y]
hudλ
_
p−1
h(y) u(y) Λ
p
(x, y) dλ(y)
_
_
1
p
.
Thus, again by the Minkowskii inequality
_
_
[0,∞)
(Hf)
q
vdµ
_1
q

_
_
_
_
[0,∞)
_
_
_
[0,x]
_
_
[0,y]
hudλ
_
p−1
h(y) u(y) Λ
p
(x, y) dλ(y)
_
_
q
p
v (x) dµ(x))
1
q

_
_
_
[0,∞)
_
_
[0,y]
hudλ
_
p−1
h(y) u(y)

_
_
[y,∞)
Λ(x, y) v (x) dµ(x)
_
p
q
dλ(y)
_
_
1
p

¯
A
1
_
_
_
[0,∞)
_
_
[0,y]
hudλ
_
p−1
h(y) u(y)
_
_
[y,∞)
wdν
_
dλ(y)
_
_
1
p

¯
A
1
_
_
[0,∞)
f
p
wdν
_1
p
.
A general case f ∈ M↑ follows by Lemma 7.4 similar to the proof of Theorem
7.14.
104 Carleman type inequalities and Hardy type inequalities
for monotone functions
7.4 The case 1 < p, q < ∞
The result of this section is based on the following statement, which follows
from Theorems 7.9 and 7.11 with q = 1.
Corollary 7.16. Let (Tf)(x) =
_
[0,∞)
k(x, y)f(y)u(y)dλ(y), where k(x, y)
is a defined on [0, ∞) [0, ∞), non-negative, µ λ-measurable kernel.
(a) The inequality
_
_
[0,∞)
(Tf)
q
vdµ
_1
q
≤ C
_
_
[0,∞)
f
p
wdν
_1
p
(7.32)
for f ∈ M↓, holds if and only if the inequality
_
_
_
[0,∞)
w(y)
_
_
[y,∞)
W
−1
(T

g)udλ
_
p

dν(y)
_
_
1
p

(7.33)
≤ C
_
_
[0,∞)
q
q

vdµ
_ 1
q

, g ∈ M
+
,
holds with (T

g)(z) =
_
[0,∞)
k(z, x)g(z)v(z)dµ(z).
(b) The inequality (7.32) for f ∈ M ↑ holds if and only if the following
inequality holds:
_
_
_
[0,∞)
w(y)
_
_
[0,y]
¯
W
−1
(T

g)udλ
_
p

dν(y)
_
_
1
p

≤ C
_
_
[0,∞)
q
q

vdµ
_1
q

, g ∈ M
+
.
Now let us present our result for the case 1 < p, q < ∞ .
Theorem 7.17. Let k(x, y) =
_
[y,x]
W
−1
udλ and f ∈ M↓ . The inequality
(7.3) holds for 1 < p ≤ q < ∞ if and only if / = max ¦/
0,1
+/
0,2
¦ < ∞,
where
/
0,1
: = sup
t∈[0,∞)
_
_
[0,t]
w(y)k(t, y)
p

dν(y)
_ 1
p

_
_
[t,∞)
vdµ
_1
q
,
Hardy inequality with three measures on monotone func-
tions
105
/
0,2
: = sup
t∈[0,∞)
_
_
[0,t]
wdν
_ 1
p

_
_
[t,∞)
v(x)k(x, t)
q
dµ(x)
_1
q
.
Moreover, if C is the best constant in (7.3), then C = /.
In the case 1 < q < p < ∞ the inequality (7.3) holds if and only if
B = max ¦B
0,1
+B
0,2
¦ < ∞, where
B
0,1
: =
_
_
_
[0,∞)
_
_
[0,t]
w(y)k(t, y)
p

dν(y)
_ r
p

_
_
[t,∞)
vdµ
_r
p
v(t)dµ(t)
_
_
1
r
,
B
0,2
: =
_
_
_
[0,∞)
_
_
[0,t]
wdν
_ r
q

_
_
[t,∞)
v(x)k(x, t)
q
dµ(x)
_r
q
w(t)dν(t)
_
_
1
r
and
1
r
=
1
q

1
p
. Moreover, if C is the best constant in (7.3), then C = B.
Proof. Because of Corollary 7.16 (a) the inequality (7.3) is equivalent to
_
_
_
[0,∞)
w(y)
_
_
[y,∞)
W(x)
−1
_
_
[x,∞)
gvdµ
_
u(x)dλ(x)
_
p

dν(y)
_
_
1
p

≤ C
_
_
[0,∞)
q
q

vdµ
_1
q

, g ∈ M
+
. (7.34)
By changing the order of integration in the left hand side of (7.34) we obtain
the Hardy inequality with Oinarov kernel of the form
_
_
_
[0,∞)
w(y)
_
_
[y,∞)
g(z)k(z, y)v(z)dµ(z)
_
p

dν(y)
_
_
1
p

≤ C
_
_
[0,∞)
q
q

vdµ
_1
q

.
By substitution f = g
q

v and according to Lemma 7 in [88] the last inequality
is equivalent to
_
_
_
[0,∞)
w(y)
_
_
[y,∞)
f(z)k(z, y)v(z)
1/q
dµ(z)
_
p

dν(y)
_
_
1
p

≤ C
_
_
[0,∞)
f
q


_1
q

.
106 Carleman type inequalities and Hardy type inequalities
for monotone functions
Thus the proof follows by applying Theorem 7.6.
Similarly we can obtain the result for non-decreasing functions as follows.
Theorem 7.18. Let
¯
k(y, x) =
_
[x,y]
¯
W
−1
udλ and f ∈ M↑ . The inequality
(7.3) holds for 1 < p ≤ q < ∞ if and only if
¯
/ = max
_
¯
/
0,1
+
¯
/
0,2
_
< ∞,
where
¯
/
0,1
: = sup
t∈[0,∞)
_
_
[t,∞)
w(y)
¯
k(y, t)
p

dν(y)
_ 1
p

_
_
[0,t]
vdµ
_1
q
,
¯
/
0,2
: = sup
t∈[0,∞)
_
_
[t,∞)
wdν
_ 1
p

_
_
[0,t]
v(x)
¯
k(t, x)
q
dµ(x)
_1
q
.
Moreover, if C is the best constant in (7.3), then C =
¯
/. In the case 1 < q <
p < ∞ the inequality (7.3) holds if and only if
¯
B = max
_
¯
B
0,1
+
¯
B
0,2
_
< ∞,
where
¯
B
0,1
: =
_
_
_
[0,∞)
_
_
[t,∞)
w(y)
¯
k(y, t)
p

dν(y)
_ r
p

_
_
[0,t]
vdµ
_r
p
v(t)dµ(t)
_
_
1
r
,
¯
B
0,2
: =
_
_
_
[0,∞)
_
_
[t,∞)
wdν
_r
q

_
_
[0,t]
v(x)
¯
k(t, x)
q
dµ(x)
_r
q
w(t)dν(t)
_
_
1
r
and
1
r
=
1
q

1
p
. Moreover, if C is the best constant in (7.3), then C = B

.
Chapter 8
Characterizations of embeddings in
Lorentz spaces
8.1 Introduction
Let (1, µ) be a totally σ-finite measure space with a non-atomic measure
µ, and let /(1, µ) be the set of all extended complex-valued µ-measurable
functions on 1. For f ∈ /(1, µ), let
f

(t) = µ(¦x ∈ 1; [f (x)[ > t¦) , t ∈ (0, ∞) ,
be a distribution function of f. The non-increasing rearrangement f

of f
is defined by
f

(t) = inf ¦x > 0; f

(x) ≤ t¦ , t ∈ (0, ∞) .
We will assume that µ(1) = ∞. Everywhere in this chapter, we assume
that u, v, and w are weights, that is, locally integrable non-negative functions
on (0, ∞) and we denote
V (s) =
_
s
0
v (t) dt and U (s) =
_
s
0
u(t) dt.
We assume that u is such that U(t) > 0 for every t ∈ (0, ∞) and denote
f
∗∗
u
(t) =
1
U(t)
_
t
0
f

(s) u(s)ds.
107
108 Carleman type inequalities and Hardy type inequalities
for monotone functions
Definition 8.1. Let p ∈ (0, ∞) and let u and v be weight functions. Then
we define the following four spaces (of Lorentz and Marcinkiewicz type):
Λ
p
(v) =
_
f ∈ /(1, µ) ; |f|
Λ
p
(v)
=
__

0
f
∗p
(t) v (t) dt
_1
p
< ∞
_
;
Λ

(v) =
_
f ∈ /(1, µ) ; |f|
Λ

(v)
= ess sup
t∈(0,∞)
f

(t)v(t) < ∞
_
;
Γ
p
u
(v) =
_
f ∈ /(1, µ) ; |f|
Γ
p
u
(v)
=
__

0
f
∗∗p
u
(t) v (t) dt
_1
p
< ∞
_
;
Γ

u
(v) =
_
f ∈ /(1, µ) ; |f|
Γ

u
(v)
= ess sup
t∈(0,∞)
f
∗∗
u
(t)v (t) < ∞
_
.
When u ≡ 1 (hence U(t) = t), we will omit the subscript u. We thus
write, in such a case f
∗∗
, Γ
p
(v) and Γ

(v) in place of f
∗∗
u
, Γ
p
u
(v) and Γ

u
(v),
respectively. The spaces Λ
p
(v) and Γ
p
(v) are called classical Lorentz spaces.
The spaces Λ
p
(v) were introduced by G. G. Lorentz in 1951 in [66]. The
spaces Γ
p
(v) were first used by E. Sawyer in [93] and the weak classical
Lorentz spaces were introduced by M. Carro and J. Soria in [19]. Weak
Lorentz spaces were further investigated in [17], [18], [20] and [21]. For
definitions and a detailed study of rearrangement-invariant Banach function
spaces see e.g. [12].
During the last two decades many authors have spent enormous ef-
forts in order to find necessary and sufficient conditions on the parameters
p, q ∈ [0, ∞] and weights u, w such that the embeddings Λ
p
(v) → Γ
q
(w)
hold. Such embeddings proved to be indispensable in several important ar-
eas in analysis, including the theory of interpolation and modern study of
Sobolev spaces. This research brought plenty of deep results, see in e.g. [5],
[19], [18], [20], [34], [35], [36], [37], [38], [93], [94], [95], [96], [98], [100] and
[99]. A summary of results of embeddings of classical Lorentz spaces known
by the end the of 1990’s, as well as some references, can be found in [21].
In this survey paper there were some cases of parameters which were not
known. In the paper [37] all cases have been established, but the resulting
characterization were expressed in a difficult way, not very satisfactory from
a practical point of view.
Our main goal in this chapter is to give some new necessary and sufficient
conditions on the weight functions u, v and w such that the inequality
|f|
Γ
q
u
(w)
≤ C |f|
Λ
p
(v)
(8.1)
Characterizations of embeddings in Lorentz spaces 109
holds for every f ∈ /(1, µ).
Our approach is based on discretization and anti-discretization methods
developed in [35] and [36]. Let as outline the structure of the chapter. In
Section 8.2 we give some preliminaries including some definitions and basic
facts concerning discretization and anti-disctertization from the papers [35]
and [36]. In Section 8.3 we present the main results and finally in Section
8.4 we give the proofs.
Throughout this chapter a b, (b a), means that a ≤ λb, where λ > 0
depends on inessential parameters. If b a b, then we write a ≈ b.
8.2 Preliminaries
Let us now recall some definitions and basic facts concerning discretization
and anti-discretization which can be found in [35] and [36].
Definition 8.2. Let ¦a
k
¦ be a sequence of positive real numbers. We say
that ¦a
k
¦ is strongly increasing or strongly decreasing and write a
k
↑↑
or a
k
↓↓ when
inf
k∈Z
a
k+1
a
k
> 1 or sup
k∈Z
a
k+1
a
k
< 1,
respectively.
Definition 8.3. Let U be a continuous strictly increasing function on [0, ∞)
such that U (0) = 0 and lim
t→∞
U (t) = ∞. Then we say that U is admissible.
Let U be an admissible function. We say that a function ϕ is U-quasiconcave
if ϕ is equivalent to an increasing function on [0, ∞) and
ϕ
U
is equivalent to
a decreasing function on (0, ∞). We say that U-quasiconcave function ϕ is
non-degenerated if
lim
t→0+
ϕ(t) = lim
t→∞
1
ϕ(t)
= lim
t→∞
ϕ(t)
U (t)
= lim
t→0+
U (t)
ϕ(t)
= 0.
The family of non-degenerated U-quasiconcave functions will be denoted by

U
. We say that ϕ is quasiconcave when ϕ ∈ Ω
U
with U (t) = t.
Definition 8.4. Assume that U is admissible and ϕ ∈ Ω
U
. We say that
¦x
k
¦
k∈Z
is a discretizing sequence for ϕ with respect to U if
(i) x
0
= 1 and U (x
k
) ;
(ii) ϕ(x
k
) and
ϕ(x
k
)
U(x
k
)
;
110 Carleman type inequalities and Hardy type inequalities
for monotone functions
(iii) there is a decomposition Z = Z
1
∪Z
2
such that Z
1
∩Z
2
= ∅ and, for
every t ∈ [x
k
, x
k+1
] ,
ϕ(x
k
) ≈ ϕ(t) if k ∈ Z
1
,
ϕ(x
k
)
U (x
k
)

ϕ(t)
U (t)
if k ∈ Z
2
.
Let us recall [35, Lemma 2.7] that if ϕ ∈ Ω
U
, then there always exists a
discretizing sequence for ϕ with respect to U.
Definition 8.5. Let U be an admissible function and let ν be a non-negative
Borel measure on [0, ∞). We say that the function ϕ defined by
ϕ(t) = U(t)
_
[0,∞)
dν(s)
U(s) +U(t)
, t ∈ (0, ∞),
is the fundamental function of the measure ν with respect to U. We will also
say that ν is a representation measure of ϕ with respect to U.
We say that ν is non-degenerate if the following conditions are satisfied
for every t ∈ (0, ∞):
_
[0,∞)
dν(s)
U(s) +U(t)
< ∞, t ∈ (0, ∞),
_
[0,1]
dν(s)
U(s)
=
_
[1,∞)
dν(s) = ∞.
We recall from Remark 2.10 of [35] that
ϕ(t) ≈
_
[0,t]
dν(s) +U(t)
q
_
[t,∞)
U(s)
−q
dν(s), t ∈ (0, ∞).
We shall further assume that ν is non-degenerate.
Lemma 8.1. Let v be a weight function. Then
Λ

(v) = Λ

(ess sup
0<s<t
v(s)),
with identical norms.
The proof follows from the following simple observation: If F and G are
non-negative functions on (0, ∞) and F is non-increasing, then
ess sup
0<t<∞
F(t)G(t) = ess sup
0<t<∞
F(t) ess sup
0<s<t
G(s).
Characterizations of embeddings in Lorentz spaces 111
Lemma 8.2 ([36], Lemma 1.5). Let u, v be weights and let ϕ be defined
by
ϕ(t) = ess sup
s∈(0,t)
U(s) ess sup
τ∈(s,∞)
v(τ)
U(τ)
, t ∈ (0, ∞). (8.2)
Then ϕ is the least U-quasiconcave majorant of v and
Γ

u
(v) = Γ

u
(ϕ),
with identical norms. Further, for t ∈ (0, ∞);
ϕ(t) = ess sup
τ∈(0,∞)
v(τ) min
_
1,
U(t)
U(τ)
_
= U(t) ess sup
s∈(t,∞)
1
U(s)
ess sup
τ∈(0,s)
v(τ),
ϕ(t) ≈ ess sup
s∈(0,∞)
v(s)U(t)
U(s) +U(t)
.
Theorem 8.3 ([35], Theorem 2.11). Let p, q, r ∈ (0, ∞). Assume that
U is an admissible function, let ν be a non-negative non-degenerate Borel
measure on [0, ∞) and let ϕ be the fundamental function of ν with respect to
U
q
. Moreover, let σ ∈ Ω
U
and let ¦x
k
¦ be a discretizing sequence for ϕ with
respect to U
q
. Then
_
[0,∞)
ϕ(t)
r
q
−1
σ(t)
r
p
dν(t) ≈

k∈Z
ϕ(x
k
)
r
q
σ(x
k
)
r
p
.
Lemma 8.4 ([35], Corollary 2.13). Let q ∈ (0, ∞), let U be an admissible
function, let f ∈ Ω
U
, let ν be a positive non-degenerated Borel measure
on [0, ∞) and let ϕ be the fundamental function of ν with respect to U
q
.
Moreover, assume that ¦x
k
¦ is a discretizing sequence for ϕ with respect to
U
q
. Then
_
_
[0,∞)
_
f (t)
U (t)
_
q
dν (t)
_1
q

_

k∈Z
_
f (x
k
)
U (x
k
)
_
q
ϕ(x
k
)
_1
q
.
Lemma 8.5 ([35], Lemma 3.5 ). Let U be an admissible function, let
f, ϕ ∈ Ω
U
and let ¦x
k
¦ be a discretizing sequence for ϕ with respect to U.
Then
ess sup
0<t<∞
f (t)
U (t)
ϕ(t) ≈ sup
k∈Z
f (x
k
)
U (x
k
)
ϕ(x
k
) .
By using Lemma 3.1 and Lemma 3.2 from [35] we get the following result:
112 Carleman type inequalities and Hardy type inequalities
for monotone functions
Lemma 8.6. If τ
k
, then, for any q > 0,

k∈Z
__
x
k
0
h
_
q
τ
k

k∈Z
_
_
x
k
x
k−1
h
_
q
τ
k
,
sup
k∈Z
__
x
k
0
h
_
q
τ
k
≈ sup
k∈Z
_
_
x
k
x
k−1
h
_
q
τ
k
,

k∈Z
__

x
k
h
_
q
τ
−1
k

k∈Z
__
x
k+1
x
k
h
_
q
τ
−1
k
,
and
sup
k∈Z
__

x
k
h
_
q
τ
−1
k
≈ sup
k∈Z
__
x
k+1
x
k
h
_
q
τ
−1
k
.
Lemma 8.7 ([35], Lemma 3.6). Let q ∈ (0, ∞), let U be an admissible
function, let ν be a positive non-degenerated Borel measure on [0, ∞) and let
ϕ be the fundamental function of ν with respect to U
q
. Assume that ¦x
k
¦ is
a discretizing sequence for ϕ with respect to U
q
. Then
_

0
__

0
[f(t)[dt
U(t) +U(x)
_
q
dν(x) ≈

k∈Z
__

0
[f(t)[dt
U(t) +U(x
k
)
_
q
ϕ(x
k
)

k∈Z
_
U
−1
(x
k
)
_
x
k
x
k−1
[f(y)[ dy +
_
x
k+1
x
k
[f (y)[ U
−1
(y) dy
_
q
ϕ(x
k
)

k∈Z
__
x
k+1
x
k
[f (y)[ U
−1
(y) ϕ(y)
1
q
dy
_
q
.
Lemma 8.8 ([35], Lemma 3.7). Let q ∈ (0, ∞), let U be an admissible
function and let ν be a non-degenerate positive Borel measure. Moreover, let
ϕ be the fundamental function of ν with respect to U
q
and let f be a measur-
able function on [0, ∞). If ¦x
k
¦ is a discretizing sequence for ϕ with respect
to U
q
, then
_
[0,∞)
_
sup
y∈(0,∞)
[f(y)[
U(x) +U(y)
_
q
dν(x) ≈

k∈Z
_
sup
y∈(0,∞)
[f(y)[
U(x
k
) +U(y)
_
q
ϕ(x
k
)

k∈Z
_
U
−1
(x
k
) sup
x
k−1
≤y<x
k
[f(y)[ + sup
x
k
≤y<x
k+1
[f(y)[U(y)
−1
_
q
ϕ(x
k
)

k∈Z
sup
x
k
≤y<x
k+1
[f(y)[
q
U(y)
−q
ϕ(y).
Characterizations of embeddings in Lorentz spaces 113
Lemma 8.9 ([35], Lemma 3.8). Let v be a weight function, let U be
an admissible function and ϕ be defined by (8.2). Moreover, let ¦x
k
¦ be a
discretizing sequence for ϕ with respect to U. Then
ess sup
x∈(0,∞)
__

0
[f(t)[dt
U(t) +U(x)
_
v(x) ≈ sup
k∈Z
__

0
[f(t)[dt
U(t) +U(x
k
)
_
ϕ(x
k
)
≈sup
k∈Z
_
U
−1
(x
k
)
_
x
k
x
k−1
[f (y)[ dy +
_
x
k+1
x
k
[f (y)[ U
−1
(y) dy
_
ϕ(x
k
)
≈ sup
k∈Z
__
x
k+1
x
k
[f (y)[ U
−1
(y) ϕ(y)dy
_
.
Lemma 8.10 ([35], Lemma 3.9). Let U be an admissible function and let
ϕ ∈ Ω
U
. Let ¦x
k
¦ be a discretizing sequence for ϕ with respect to U. Then
sup
x∈(0,∞)
ϕ(x) sup
0<y<∞
[f(y)[
U(x) +U(y)
≈ sup
k∈Z
ϕ(x
k
) sup
0<y<∞
[f(y)[
U(x
k
) +U(y)
≈ sup
k∈Z
ϕ(x
k
)U(x
k
)
−1
sup
x
k−1
<y<x
k
[f(y)[ + sup
k∈Z
ϕ(x
k
) sup
x
k
<y<x
k+1
[f(y)[U(y)
−1
≈ sup
k∈Z
sup
x
k
≤y<x
k+1
[f(y)[U(y)
−1
ϕ(y).
Proposition 8.11 ([35], Proposition 4.1). Let ¦w
k
¦ and ¦v
k
¦ , k ∈ Z,
be two sequences of positive real numbers. Let p, q ∈ (0, ∞) and assume that
the inequality
_

k∈Z
a
q
k
v
k
_1
q
≤ C
_

k∈Z
a
p
k
w
k
_1
p
is satisfied for every sequence ¦a
k
¦ of positive real numbers.
(i) If p ≤ q, then
sup
k∈Z
w

q
p
k
v
k
< ∞.
(ii) If p > q and r =
pq
p−q
, then
_

k∈Z
w

r
p
k
v
r
q
k
_1
r
< ∞.
Lemma 8.12 ([80] and [98]). We have the following Hardy type inequali-
ties:
114 Carleman type inequalities and Hardy type inequalities
for monotone functions
(a) For the case 0 < p ≤ 1,
_
_
x
k
x
k−1
__
x
k
s
h(z) dz
_1
p
u(s)ds
_
p
≤ C
_
_
x
k
x
k−1
h(s) V (s)ds
_
; (8.3)
the best constant C ≈ sup
x
k−1
<t<x
k
_
_
t
x
k−1
u(s)ds
_
p
V (t)
−1
and the equiva-
lent constant is independent of x
k
.
(b) For 1 < p < ∞,
_
_
x
k
x
k−1
__
x
k
s
h(z) dz
_1
p
u(s)ds
_
p
≤ C
_
_
x
k
x
k−1
h(s) V (s)ds
_
; (8.4)
the best constant C ≈
_
_
x
k
x
k−1
_
_
t
x
k−1
u(s)ds
_
p

p
V (t)

p

p
u(t)dt
_
p
p

and the
equivalent constant is independent of x
k
.
We will also use the following trivial inequality
_
x
k+1
x
k
h(z)dz ≤ CV (x
k
)
−1
_
x
k+1
x
k
h(s)V (s)ds. (8.5)
8.3 The main result
Our main result reads:
Theorem 8.13. Let p, q ∈ (0, ∞] and let u, v, w be weights. Assume that u
is such that U
q
is admissible and the measure w(t) dt is non-degenerate with
respect to U
q
. Then the inequality (8.1) holds for every f ∈ /(1, µ) if and
only if
(i) 0 < p ≤ 1, p ≤ q < ∞
A
1
:= sup
x>0
__

0
_
U(x)
U(t) +U(x)
_
q
w(t) dt
_
1
q
V

1
p
(x) < ∞.
Moreover, the best constant C in (8.1) satisfies C ≈ A
1
.
(ii) 0 < q < p ≤ 1
A
2
:=
__

0
__

0
_
U(x)
U(t) +U(x)
_
q
w(t) dt
_
r−q
q
w(x) U(x)
−r

_
sup
0<t<x
U(t)V (t)

1
p
_
r
dx
_
1
r
< ∞.
Characterizations of embeddings in Lorentz spaces 115
Moreover, the best constant C in (8.1) satisfies C ≈ A
2
.
(iii) 1 < p ≤ q < ∞,
A
3
:= sup
x>0
__

0
_
U(x)
U(t) +U(x)
_
q
w(t) dt
_
1
q

_
_

0
_
U(t)
U(t) +U(x)
_
p

v (t)
V
p

(t)
dt
_ 1
p

< ∞.
Moreover, the best constant C in (8.1) satisfies C ≈ A
3
.
(iv) 1 < p < ∞, 0 < q < p, r =
pq
p−q
A
4
:=
_
_

0
__

0
_
U(x)
U(t) +U(x)
_
q
w(t)dt
_
r−q
q
w(x)
_
_

0
_
U(t)
U(t) +U(x)
_
p

v(t)dt
V
p

(t)
_ r
p

dx
_
_
1
r
< ∞.
Moreover, the best constant C in (8.1) satisfies C ≈ A
4
.
(v) 0 < p ≤ 1, q = ∞
A
5
:= sup
x>0
_
ess sup
t∈(0,∞)
_
w(t)U(x)
U(t) +U(x)
_
_
V

1
p
(x) < ∞.
Moreover, the best constant C in (8.1) satisfies C ≈ A
5
.
(vi) 1 < p < ∞, q = ∞
A
6
:= sup
x>0
_
ess sup
t∈(0,∞)
w(t)U(x)
U(t) +U(x)
__
_

0
_
U(t)
U(t) +U(x)
_
p

v (t)
V
p

(t)
dt
_ 1
p

< ∞.
Moreover, the best constant C in (8.1) satisfies C ≈ A
6
.
(vii) p = q = ∞
A
7
:= ess sup
x∈(0,∞)
w(x)
U(x)
_
x
0
u(t)dt
ess sup
0<s<t
v (s)
< ∞.
Moreover, the best constant C in (8.1) satisfies C ≈ A
7
.
(viii) p = ∞, 0 < q < ∞
A
8
:=
_
_
_

0
_
_
1
U(x)
_
x
0
u(t)dt
ess sup
0<s<t
v (s)
_
_
q
w(x)dx
_
_
1
q
< ∞.
116 Carleman type inequalities and Hardy type inequalities
for monotone functions
Moreover, the best constant C in (8.1) satisfies C ≈ A
8
.
Our next goal is to formulate Theorem 8.13 in a more compact and nice
form. We need the following definition of the associate space (also called the
Köthe dual) Λ
p
(v)

u
of Λ
p
(v) .
Definition 8.6. Let 0 < p ≤ ∞ and let u, v be weight functions. The
u-associate or Köthe dual space Λ
p
(v)

u
of Λ
p
(v) is defined by
|f|
Λ
p
(v)

u
=: sup
g∈Λ
p
(v)
_

0
f

(t) g

(t) u(t)dt
|g|
Λ
p
(v)
< ∞. (8.6)
In the classical situation (that is when u = 1) the associate norm Λ
p
(v)

was
characterised by E. Sawyer [93] for 1 < p < ∞ and by V. D. Stepanov [100]
for 0 < p ≤ 1. In the case p = ∞ it was considered in [21] (see also [34]).
In fact we can describe Λ
p
(v)

in the following way:
Λ
p
(v)

=
_
¸
¸
¸
¸
¸
¸
_
¸
¸
¸
¸
¸
¸
_
Γ

_
t
V (t)
1/p
_
, 0 < p ≤ 1
Γ
p

_
t
p

v(t)
V (t)
p

_
, 1 < p < ∞
Λ
1
_
1
ess sup
0<s<t
v(s)
_
, p = ∞.
(8.7)
Remark 8.1. By letting t → U
−1
(t) (where U
−1
is the inverse function of
U) in (8.6) and after using (8.7) we can obtain the following characterization
of Λ
p
(v)

u
:
Λ
p
(v)

u
=
_
¸
¸
¸
¸
¸
¸
_
¸
¸
¸
¸
¸
¸
_
Γ

u
_
U(t)
V (t)
1/p
_
, 0 < p ≤ 1
Γ
p

u
_
U(t)
p

v(t)
V (t)
p

_
, 1 < p < ∞
Λ
1
_
u(t)
ess sup
0<s<t
v(s)
_
, p = ∞.
(8.8)
Using Γ
q
u
(w) and Λ
p
(v)

u
we can now formulate Theorem 8.13 in the
following compact form:
Theorem 8.14. Let p, q ∈ (0, ∞] and let u, v, w be weights. When q ∈
(0, ∞) we assume that u is such that U
q
is admissible and the measure w(t) dt
is non-degenerate with respect to U
q
. Then the inequality (8.1) holds for
every f ∈ /(1, µ) if and only if
Characterizations of embeddings in Lorentz spaces 117
(i) 0 < p ≤ q ≤ ∞
sup
t
_
_
χ
(0,t)
_
_
Γ
q
u
(w)
_
_
χ
(0,t)
_
_
Λ
p
(v)

u
U(t)
−1
< ∞;
(ii) 0 < q < p ≤ ∞, r =
pq
p−q
__

0
_
_
χ
(0,t)
_
_
r−q
Γ
q
u
(w)
w(t) U(t)
−r
_
_
χ
(0,t)
_
_
r
Λ
p
(v)

u
dt
_
1/r
< ∞.
Remark 8.2. Theorem 8.13 and Theorem 8.14 are in fact equivalent, the
only difference is that Theorem 8.14 is stated in a more compact form. All
the eight conditions A
1
− A
8
in Theorem 8.13 can be obtained from the two
conditions in Theorem 8.14.
8.4 Proofs
Proof of Theorem 8.13: We start with the upper bounds (sufficiency).
Let 0 < p < ∞. First we use the fact that f is non-increasing if and only
if f
p
is non-increasing and a standard argument shows that it is enough to
prove (8.1) for f satisfying f

(t)
p
=
_

t
h(s)ds, where h is some positive
measurable function on (0, ∞). That is, when 0 < q < ∞ we only have to
prove that
_
_

0
_
1
U(t)
_
t
0
__

s
h(z)dz
_1
p
u(s)ds
_
q
w(t) dt
_
1
q
≤ C
__

0
__

t
h(s)ds
_
v (t) dt
_1
p
(8.9)
holds for every non-negative function h, and for q = ∞, by Lemma 8.2 we
have to prove that the inequality
sup
t∈(0,∞)
ϕ(t)
U(t)
_
t
0
__

s
h(z)dz
_1
p
u(s)ds ≤ C
__

0
__

t
h(s)ds
_
v (t) dt
_1
p
(8.10)
holds for every non-negative function h, where the function ϕ is defined by
(8.2).
According to Fubini’s theorem the inequalities (8.9) and (8.10) are equiv-
alent to
118 Carleman type inequalities and Hardy type inequalities
for monotone functions
_
_

0
_
1
U(t)
_
t
0
__

s
h(z)dz
_1
p
u(s)ds
_
q
w(t)dt
_
1
q
≤ C
__

0
h(t)V (t)dt
_1
p
,
and
sup
t∈(0,∞)
ϕ(t)
U(t)
_
t
0
__

s
h(z)dz
_1
p
u(s)ds ≤ C
__

0
h(t)V (t)dt
_1
p
, (8.11)
respectively.
Assume first that 0 < q < ∞. Define
ϕ(x) = U
q
(x)
_

0
w(s) ds
(U (s) +U (x))
q
. (8.12)
Then ϕ ∈ Ω
U
q and therefore there exists a discretizing sequence for ϕ with
respect to U
q
. Let ¦x
k
¦ be one such sequence. Then ϕ(x
k
) and
ϕ(x
k
)
U
q
(x
k
)
.
Furthermore, there is a decomposition Z = Z
1
∪Z
2
, Z
1
∩Z
2
= ∅ such that for
every k ∈ Z
1
and t ∈ [x
k
, x
k+1
], ϕ(x
k
) ≈ ϕ(t) and for every k ∈ Z
2
and
t ∈ [x
k
, x
k+1
] ,
ϕ(x
k
)
U
q
(x
k
)

ϕ(t)
U
q
(t)
.
For the left-hand side of (8.9), using Lemma 8.4 with dυ(t) = w(t) dt
and f (t) =
_
t
0
__

s
h(z) dz
_1
p
u(s)ds we get that
__

0
_
1
U (t)
_
t
0
__

s
h(z) dz
_1
p
u(s)ds
_
q
w(t) dt
_
1
q

_

k∈Z
_
_
x
k
0
__

s
h(z) dz
_1
p
u(s)ds
_
q
ϕ(x
k
)
U (x
k
)
q
_
1
q
[using Lemma8.6]

_

k∈Z
_
_
x
k
x
k−1
__

s
h(z) dz
_1
p
u(s)ds
_
q
ϕ(x
k
)
U (x
k
)
q
_
1
q
=
_

k∈Z
_
_
x
k
x
k−1
__
x
k
s
h(z) dz +
_

x
k
h(z) dz
_1
p
u(s)ds
_
q
ϕ(x
k
)
U (x
k
)
q
_
1
q
Characterizations of embeddings in Lorentz spaces 119

_

k∈Z
_
_
x
k
x
k−1
__
x
k
s
h(z) dz
_1
p
u(s)ds
_
q
ϕ(x
k
)
U (x
k
)
q
_
1
q
+
_

k∈Z
_
_
x
k
x
k−1
__

x
k
h(z) dz
_1
p
u(s)ds
_
q
ϕ(x
k
)
U (x
k
)
q
_
1
q

_

k∈Z
_
_
x
k
x
k−1
__
x
k
s
h(z) dz
_1
p
u(s)ds
_
q
ϕ(x
k
)
U (x
k
)
q
_
1
q
+
_

k∈Z
__

x
k
h(z) dz
_
q
p
_
_
x
k
x
k−1
u(s)ds
_
q
ϕ(x
k
)
U (x
k
)
q
_
1
q
[using the fact that
_
x
k
x
k−1
u(s)ds = U(x
k
) −U(x
k−1
) ≈ U (x
k
)]

_

k∈Z
_
_
x
k
x
k−1
__
x
k
s
h(z) dz
_1
p
u(s)ds
_
q
ϕ(x
k
)
U (x
k
)
q
_
1
q
+
_

k∈Z
__

x
k
h(z) dz
_
q
p
ϕ(x
k
)
_1
q
[using Lemma 8.6 on the second term]

_

k∈Z
_
_
x
k
x
k−1
__
x
k
s
h(z) dz
_1
p
u(s)ds
_
q
ϕ(x
k
)
U (x
k
)
q
_
1
q
+
_

k∈Z
__
x
k+1
x
k
h(z) dz
_
q
p
ϕ(x
k
)
_1
q
= I +II. (8.13)
Now we shall distinguish several cases. Starting with the case 0 < p ≤ 1.
Using (8.3) for I and (8.5) for II we find that
I =:
_

k∈Z
_
_
x
k
x
k−1
__
x
k
s
h(z) dz
_1
p
u(s)ds
_
q
ϕ(x
k
)
U (x
k
)
q
_
1
q
120 Carleman type inequalities and Hardy type inequalities
for monotone functions

_

k∈Z
ϕ(x
k
)
U(x
k
)
q
_
sup
x
k−1
<t<x
k
_
_
t
x
k−1
u(s) ds
_
V (t)

1
p
_
q

_
_
x
k
x
k−1
h(s) V (s) ds
_
q
p
_
_
1
q

_
_

k∈Z
ϕ(x
k
)
U(x
k
)
q
_
sup
x
k−1
<t<x
k
U (t)
q
V (t)

q
p
__
_
x
k
x
k−1
h(s) V (s) ds
_
q
p
_
_
1
q
(8.14)
and
II = :
_

k∈Z
__
x
k+1
x
k
h(z) dz
_
q
p
ϕ(x
k
)
_1
q

_

k∈Z
ϕ(x
k
) V (x
k
)

q
p
__
x
k+1
x
k
h(s) V (s) ds
_
q
p
_1
q
. (8.15)
(i) In the case 0 < p ≤ 1, p ≤ q < ∞ (i.e.
q
p
≥ 1) we have according to
(8.14) that
I ≤ sup
k
ϕ(x
k
)
U(x
k
)
1
q
sup
x
k−1
<t<x
k
U (t) V (t)

1
p
__

0
h(s) V (s) ds
_1
p
,
and, similarly, if
q
p
≥ 1 we have according to (8.15) that
II ≤ sup
k
ϕ(x
k
)
1
q
V (x
k
)

1
p
__

0
h(s) V (s) ds
_1
p
and, finally, according to (8.12) and Lemma 8.10, we get that
I +II ≤ sup
k
ϕ(x
k
)
1
q
U (x
k
)
_
sup
x
k−1
<t<x
k
U (t) V (t)
−1/p
_
__

0
h(s) V (s) ds
_1
p
= A
1
__

0
h(s) V (s) ds
_1
p
.
(ii) For the case 0 < q < p ≤ 1, r =
pq
p−q
, by applying Hölder’s inequality
for sums to the right-hand side of (8.14) and the right hand side of (8.15)
Characterizations of embeddings in Lorentz spaces 121
with exponents
p
q
and
r
q
, we find that
I
_

k∈Z
ϕ(x
k
)
r
q
U (x
k
)
r
_
sup
x
k−1
<t<x
k
U (t) V (t)

1
p
_
r
_
1
r
__

0
h(s) V (s) ds
_1
p
II
_

k∈Z
ϕ(x
k
)
r
q
V (x
k
)

r
p
_1
r
__

0
h(s) V (s) ds
_1
p
.
Therefore, we get that
I +II
_

k∈Z
ϕ(x
k
)
r
q
U (x
k
)
r
_
sup
x
k−1
<t<x
k
U (t) V (t)

1
p
_
r
_
1
r

__

0
h(s) V (s) ds
_
q
p
. (8.16)
By using Lemma 8.8 we obtain from (8.16) that
I +II A
2
__

0
h(s) V (s) ds
_1
p
.
Now let us assume that 1 < p < ∞. Using (8.4) for I and (8.5) for II, we
get that
I
_
_

k∈Z
ϕ(x
k
)
U(x
k
)
q
_
_
x
k
x
k−1
U (t)
p

p
V (t)

p

p
u(t)dt
_
q
p

_
_
x
k
x
k−1
h(s) V (s) ds
_
q
p
_
_
1
q
(8.17)
II
_

k∈Z
ϕ(x
k
) V (x
k
)

q
p
__
x
k+1
x
k
h(s) V (s) ds
_
q
p
_1
q
. (8.18)
(iii) Suppose that 1 < p ≤ q < ∞. Then, according to (8.17) and (8.18),
we obtain that
I sup
k>0
ϕ(x
k
)
1
q
U (x
k
)
_
_
x
k
x
k−1
U (t)
p

p
V (t)

p

p
u(t)dt
_ 1
p

__

0
h(s) V (s) ds
_1
p
,
II sup
k
ϕ(x
k
)
1
q
V (x
k
)

1
p
__

0
h(s) V (s) ds
_1
p
.
122 Carleman type inequalities and Hardy type inequalities
for monotone functions
Hence, using integration by part and Lemma 8.9, we find that
I +II sup
k>0
ϕ(x
k
)
1
q
U (x
k
)
_
_
x
k
x
k−1
V (t)

p

p
U (t)
p

p
u(t)dt +U (x
k
)
p

V (x
k
)

p

p
_ 1
p

__

0
h(s) V (s) ds
_1
p
sup
k>0
ϕ(x
k
)
1
q
U (x
k
)
_
_
x
k
x
k−1
V (t)
−p

U (t)
p

v(t)dt +V (x
k
)

p

p
U (x
k
)
p

_ 1
p

__

0
h(s) V (s) ds
_1
p
sup
k>0
ϕ(x
k
)
1
q
_
_

0
U (t)
p

V (t)
−p

v(t)dt
(U(t) +U (x
k
))
p

_ 1
p

__

0
h(s) V (s) ds
_1
p
= A
3
__

0
h(s) V (s) ds
_1
p
.
(iv) Next we consider the case 1 < p < ∞, 0 < q < p, r =
pq
p−q
. By using
Hölder’s inequality for sums to the right-hand side of (8.17) and the right
hand side of (8.18) with exponents
p
q
and
r
q
, we find that
I
_
_

k∈Z
ϕ
r
q
(x
k
)
U (x
k
)
r
_
_
x
k
x
k−1
U (t)
p

p
V (x
k
)

p

p
u(t)dt
_ r
p

_
_
1
r __

0
h(s) V (s) ds
_1
p
and
II
_

k∈Z
ϕ
r
q
(x
k
) V (x
k
)

r
p
_1
r
__

0
h(s) V (s) ds
_1
p
.
Therefore, using integration by part and Lemma 8.7, we see that
I +II

_
_

k∈Z
ϕ
r
q
(x
k
)
U (x
k
)
r
_
_
x
k
x
k−1
V (t)

p

p
U (t)
p

p
u(t)dt +V (x
k
)

p

p
U (x
k
)
p

_ r
p

_
_
1
r

__

0
h(s) V (s) ds
_1
p
Characterizations of embeddings in Lorentz spaces 123

_
_

k∈Z
ϕ
r
q
(x
k
)
U (x
k
)
r
_
_
x
k
x
k−1
V (t)
−p

U (t)
p

v(t)dt +V (x
k
)

p

p
U (x
k
)
p

_ r
p

_
_
1
r

__

0
h(s) V (s) ds
_1
p

_
_

k∈Z
ϕ
r
q
(x
k
)
_
_

0
U (t)
p

V (t)
−p

v(t)dt
(U(t) +U(x
k
))
p

_ r
p

_
_
1
r __

0
h(s) V (s) ds
_1
p
≤ A
4
__

0
h(s) V (s) ds
_1
p
.
Let now q = ∞. In the same way, by using Lemma 8.2, Lemma 8.5 and
Lemma 8.6 we obtain for the left-hand side of (8.11) that
sup
t∈(0,∞)
ϕ(t)
U(t)
_
t
0
__

s
h(z)dz
_1
p
u(s)ds
≈sup
k∈Z
ϕ(x
k
)
U(x
k
)
_
x
k
0
__

s
h(z)dz
_1
p
u(s)ds
≈sup
k∈Z
ϕ(x
k
)
U(x
k
)
_
x
k
x
k−1
__
x
k
s
h(z)dz
_1
p
u(s)ds
+sup
k∈Z
ϕ(x
k
)
__
x
k+1
x
k
h(z)dz
_1
p
:= III +IV, (8.19)
where ϕ is defined by (8.2).
(v) Let 0 < p ≤ 1, q = ∞. By using the inequality (8.3) for III and
(8.5) for IV , we arrive at
III sup
k
ϕ(x
k
)
U(x
k
)
_
sup
x
k−1
<t<x
k
U (t) V (t)

1
p
_
__
x
k+1
x
k
h(s) V (s) ds
_1
p
,
IV sup
k
ϕ(x
k
) V (x
k
)

q
p
__
x
k+1
x
k
h(s) V (s) ds
_1
p
124 Carleman type inequalities and Hardy type inequalities
for monotone functions
and, finally, by using Lemma 8.10, we get that
III +IV sup
k
ϕ(x
k
)
U (x
k
)
_
sup
x
k−1
<t<x
k
U (t) V (t)
−1/p
_
__

0
h(s) V (s) ds
_1
p
A
5
__

0
h(s) V (s) ds
_1
p
.
(vi) Let now 1 < p < ∞, q = ∞. By using (8.4) for III and (8.5) for IV ,
we obtain that
III sup
k∈Z
ϕ(x
k
)
U(x
k
)
_
_
x
k
x
k−1
U (t)
p

p
V (t)

p

p
u(t)dt
_ 1
p

_
_
x
k
x
k−1
h(s) V (s) ds
_1
p
and
IV

k∈Z
ϕ(x
k
) V (x
k
)

1
p
__
x
k+1
x
k
h(s) V (s) ds
_1
p
.
Using integration by part and Lemma 8.9, we find that
III +IV sup
k>0
ϕ(x
k
)
U (x
k
)
_
_
x
k
x
k−1
V (t)

p

p
U (t)
p

p
u(t)dt +V (x
k
)

p

p
U (x
k
)
p

_ 1
p

__

0
h(s) V (s) ds
_1
p
sup
k>0
ϕ(x
k
)
U (x
k
)
_
_
x
k
x
k−1
V (t)
−p

U (t)
p

v(t)dt +V (x
k
)

p

p
U (x
k
)
p

_1
p

__

0
h(s) V (s) ds
_1
p
sup
k>0
ϕ(x
k
)
_
_

0
U (t)
p

V (t)
−p

v(t)dt
U(t)
p

+U (x
k
)
p

_ 1
p

__

0
h(s) V (s) ds
_1
p
= A
6
__

0
h(s) V (s) ds
_1
p
.
Characterizations of embeddings in Lorentz spaces 125
(vii) For the case p = q = ∞, we use Lemma 8.1 and get that
|f|
Γ

u
(w)
= ess sup
x∈(0,∞)
w(x)
U(x)
_
x
0
f

(t)u(t)dt
≤ ess sup
x∈(0,∞)
w(x)
U(x)
_
x
0
u(t)dt
ess sup
0<s<t
v (s)
ess sup
x∈(0,∞)
f

(x)ess sup
0<s<x
v (s)
≤ A
7
|f|
Λ

(v)
.
(viii) Finally, for p = ∞, 0 < q < ∞, by Lemma 8.1 we obtain that
|f|
Γ
q
u
(w)

_
_
_

0
_
_
1
U(x)
_
x
0
u(t)dt
ess sup
0<s<t
v (s)
_
_
q
w(x)dx
_
_
1
q
ess sup
x∈(0,∞)
f

(x)ess sup
0<s<x
v (s)
≤ A
8
|f|
Λ

(v)
.
Next we prove the lower bounds (Necessity). Let 0 < q < ∞ and let ¦x
k
¦
be a discretizing sequence for ϕ from (8.12). Then by (8.13) we have that
_

k∈Z
_
_
x
k
x
k−1
__
x
k
s
h(z) dz
_1
p
u(s)ds
_
q
ϕ(x
k
)
U (x
k
)
q
_
1
q
+
_

k∈Z
__
x
k+1
x
k
h(z) dz
_
q
p
ϕ(x
k
)
_1
q

_

k∈Z
_
x
k
x
k−1
h(t) V (t) dt
_1
p
. (8.20)
Let 0 < p ≤ 1. For k ∈ Z, let h
k
be functions that satisfy the Hardy in-
equality (8.3) and the Hölder inequality (8.5), that is, functions h
k
satisfying
supp h
k
⊂ [x
k−1
, x
k
] ,
_
x
k
x
k−1
h
k
(s) V (s)ds = 1,
_
_
x
k
x
k−1
__
x
k
s
h
k
(t) dt
_1
p
u(s)ds
_
p
sup
x
k−1
<t<x
k
(U (t) −U (x
k−1
))
p
V (t)
−1
,
_
x
k
x
k−1
h
k
(s) ds V (x
k−1
)
−1
.
126 Carleman type inequalities and Hardy type inequalities
for monotone functions
Now we define the test function
h(s) =

k∈Z
a
k
h
k
(s) , (8.21)
where ¦a
k
¦ is a sequence of positive real numbers. Then, by using this test
function in (8.20), we get that
_

k∈Z
a
q
p
k
ϕ(x
k
)
U (x
k
)
q
sup
x
k−1
<t<x
k
(U (t) −U (x
k−1
))
q
V (t)

q
p
_1
q
+
_

k∈Z
a
q
p
k
ϕ(x
k−1
) V (x
k−1
)

q
p
_1
q

_

k∈Z
a
k
_1
p
. (8.22)
Now, by using Proposition 8.11 for the case 0 < p ≤ 1, p ≤ q, we find that
sup
k∈Z
ϕ(x
k
)
1
q
V (x
k
)

1
p
+
sup
k∈Z
_
sup
x
k−1
<t<x
k
(U (t) −U (x
k−1
)) V (t)
−1/p
_
ϕ(x
k
)
1
q
U (x
k
)
< ∞.
Moreover, according to Lemma 8.10, we see that
A
1
sup
k∈Z
sup
x
k−1
<t<x
k
V (t)
−1/p
ϕ(t)
1
q
sup
k∈Z
ϕ(x
k
)
1
q
V (x
k
)

1
p
+ sup
k∈Z
_
sup
x
k−1
<t<x
k
U (t) V (t)
−1/p
_
ϕ(x
k
)
1
q
U (x
k
)
sup
k∈Z
ϕ(x
k
)
1
q
V (x
k
)

1
p
+ sup
k∈Z
_
sup
x
k−1
<t<x
k
(U (t) −U(x
k−1
)) V (t)
−1/p
_
ϕ(x
k
)
1
q
U (x
k
)
< ∞.
Let 0 < q < p ≤ 1. We obtain from (8.22) and Proposition 8.11 that
_

k∈Z
ϕ(x
k
)
r
q
V (x
k
)

r
p
_1
r
+
_

k∈Z
_
sup
x
k−1
<t<x
k
(U (t) −U (x
k−1
))
r
V (t)
−r/p
_
ϕ(x
k
)
r
q
U (x
k
)
r
_1
r
< ∞.
Characterizations of embeddings in Lorentz spaces 127
Therefore, in view of Theorem 8.3, Lemma 8.6 and Lemma 8.8, we have that
A
2

_

k∈Z
_
sup
x
k−1
<t<x
k
U(t)
r
V (t)
−r/p
_
ϕ(x
k
)
r
q
U(x
k
)
r
_1
r

_

k∈Z
ϕ(x
k
)
r
q
V (x
k
)
−r/p
_1
r
+
_

k∈Z
_
sup
x
k−1
<t<x
k
(U (t) −U(x
k−1
))
r
V (t)
−r/p
_
ϕ(x
k
)
r
q
U (x
k
)
r
_1
r
< ∞.
Let 1 < p < ∞. For k ∈ Z, let h
k
be functions that satisfy the Hardy in-
equality (8.4) and the Hölder inequality (8.5), that is, functions h
k
satisfying
supp h
k
⊂ [x
k−1
, x
k
] ,
_
x
k
x
k−1
h
k
(s) V (s)ds = 1,
_
_
x
k
x
k−1
__
x
k
s
h
k
(t) dt
_1
p
u(s)ds
_
p

_
_
x
k
x
k−1
(U (t) −U (x
k−1
))
p

p
V (t)

p

p
u(t)dt
_
p
p

_
x
k
x
k−1
h
k
(s) ds V (x
k−1
)
−1
.
Now we define the test function
h(s) =

k∈Z
a
k
h
k
(s) , (8.23)
where ¦a
k
¦ is a sequence of positive real numbers. Then, by using this test
function in (8.20) we get that
_
_

k∈Z
a
q
p
k
ϕ(x
k
)
U (x
k
)
q
_
_
x
k
x
k−1
(U (t) −U (x
k−1
))
p

p
V (t)

p

p
u(t)dt
_
q
p

_
_
1
q
+
_

k∈Z
a
q
p
k
ϕ(x
k−1
) V (x
k−1
)

q
p
_1
q

_

k∈Z
a
k
_1
p
. (8.24)
128 Carleman type inequalities and Hardy type inequalities
for monotone functions
Now, by using Proposition 8.11 for the case 1 < p < q < ∞, we find that
sup
k∈Z
_
_
x
k
x
k−1
(U (t) −U (x
k−1
))
p

/p
V (t)
−p

/p
u(t)dt
_
1/p

ϕ(x
k
)
1
q
U (x
k
)
+ sup
k∈Z
V (x
k
)

1
p
ϕ(x
k
)
1
q
< ∞.
Hence, by using Lemma 8.5 and Lemma 8.9 and performing integration by
part, we have that
A
3
sup
k∈Z
ϕ(x
k
)
1
q
U (x
k
)
_
_
x
k
x
k−1
V (t)
−p

U (t)
p

v(t)dt
_
1/p

+ sup
k∈Z
ϕ(x
k
)
1
q
V (x
k
)

1
p
sup
k∈Z
ϕ(x
k
)
1
q
U (x
k
)
_
_
x
k
x
k−1
V (t)
−p

(U (t) −U (x
k−1
))
p

v(t)dt
_
1/p

+sup
k∈Z
ϕ(x
k
)
1
q
V (x
k
)

1
p
sup
k∈Z
ϕ(x
k
)
1
q
U (x
k
)
_
_
x
k
x
k−1
V (t)

p

p
(U (t) −U (x
k−1
))
p

p
u(t)dt
_
1/p

+sup
k∈Z
ϕ(x
k
)
1
q
V (x
k
)

1
p
< ∞.
Let now 1 < p < ∞, 0 < q < p. Then, by using (8.24) and Proposition 8.11,
we obtain that
_
_

k∈Z
ϕ
r/q
(x
k
)
U (x
k
)
r
_
_
x
k
x
k−1
V (t)

p

p
(U (t) −U(x
k−1
))
p

p
u(t)dt
_ r
p

_
_
1
r
+
_

k∈Z
ϕ
r
q
(x
k
) V x
k

r
p
_1
r
< ∞.
Therefore, by using Theorem 8.3 and Lemma 8.7 and integration by part,
we get that
A
4

_
_

k∈Z
ϕ(x
k
)
r
q
U (x
k
)
r
_
_
x
k
x
k−1
V (t)
−p

U (t)
p

v(t)dt
_ r
p

_
_
1
r
+
_

k∈Z
ϕ(x
k
)
r
q
V (x
k
)

r
p
_1
r
Characterizations of embeddings in Lorentz spaces 129

_
_

k∈Z
ϕ(x
k
)
r
q
U (x
k
)
r
_
_
x
k
x
k−1
V (t)
−p

(U (t) −U (x
k−1
))
p

v(t)dt
_ r
p

_
_
1
r
+
_

k∈Z
ϕ(x
k
)
r
q
V (x
k
)

r
p
_1
r

_
_

k∈Z
ϕ(x
k
)
r
q
U (x
k
)
r
_
_
x
k
x
k−1
V (t)

p

p
(U (t) −U (x
k−1
))
p

p
u(t)dt
_ r
p

_
_
1
r
+
_

k∈Z
ϕ(x
k
)
r
q
V (x
k
)

r
p
_1
r
< ∞.
Let q = ∞ and let ¦x
k
¦ be a discretizing sequence for ϕ defined by (8.2).
Then, according to (8.19), we have that
sup
k∈Z
ϕ(x
k
)
U(x
k
)
_
x
k
x
k−1
__
x
k
s
h(z)dz
_1
p
u(s)ds
+ sup
k∈Z
ϕ(x
k
)
__
x
k+1
x
k
h(z)dz
_1
p

_

k∈Z
_
x
k
x
k−1
h(t) V (t) dt
_1
p
. (8.25)
Let 0 < p ≤ 1. If we use in (8.25) the test function defined by (8.21), then
we obtain that
sup
k∈Z
a
1
p
k
ϕ(x
k
)
U (x
k
)
sup
x
k−1
<t<x
k
(U (t) −U (x
k−1
)) V (t)

1
p
+sup
k∈Z
a
1
p
k
ϕ(x
k
) V (x
k
)

1
p

_

k∈Z
a
k
_1
p
.
Therefore, by Proposition 8.11, we have that
sup
k∈Z
ϕ(x
k
)
U (x
k
)
sup
x
k−1
<t<x
k
(U (t) −U (x
k−1
)) V (t)

1
p
+sup
k∈Z
ϕ(x
k
) V (x
k
)

1
p
< ∞.
130 Carleman type inequalities and Hardy type inequalities
for monotone functions
Thus, by Lemma 8.10, we find that
A
5
sup
k∈Z
sup
x
k−1
<t<x
k
V (t)
−1/p
ϕ(t)
sup
k∈Z
ϕ(x
k
)V (x
k
)

1
p
+ sup
k∈Z
_
sup
x
k−1
<t<x
k
U (t) V (t)
−1/p
_
ϕ(x
k
)
U (x
k
)
sup
k∈Z
ϕ(x
k
)V (x
k
)

1
p
+ sup
k∈Z
_
sup
x
k−1
<t<x
k
(U (t) −U(x
k−1
)) V (t)
−1/p
_
ϕ(x
k
)
U (x
k
)
< ∞.
Now let 1 < p < ∞. Using in (8.25) the test function defined by (8.23) we
obtain that
sup
k∈Z
a
1
p
k
ϕ(x
k
)
U (x
k
)
_
_
x
k
x
k−1
(U (t) −U (x
k−1
))
p

p
V (t)

p

p
u(t)dt
_ 1
p

+ sup
k∈Z
a
1
p
k
ϕ(x
k
) V (x
k
)

1
p

_

k∈Z
a
k
_1
p
.
Moreover, by using Proposition 8.11, we get that
sup
k∈Z
_
_
x
k
x
k−1
(U (t) −U (x
k−1
))
p

/p
V (t)
−p

/p
u(t)dt
_
1/p

ϕ(x
k
)
U (x
k
)
+ sup
k∈Z
V (x
k
)

1
p
ϕ(x
k
) < ∞.
Hence, by using Lemma 8.5, Lemma 8.9 and performing integration by part,
Characterizations of embeddings in Lorentz spaces 131
we find that
A
6
sup
k∈Z
ϕ(x
k
)
U (x
k
)
_
_
x
k
x
k−1
V (t)
−p

U (t)
p

v(t)dt
_
1/p

+ sup
k∈Z
ϕ(x
k
) V (x
k
)

1
p
sup
k∈Z
ϕ(x
k
)
U (x
k
)
_
_
x
k
x
k−1
V (t)
−p

(U (t) −U (x
k−1
))
p

v(t)dt
_
1/p

+ sup
k∈Z
ϕ(x
k
) V (x
k
)

1
p
sup
k∈Z
ϕ(x
k
)
U (x
k
)
_
_
x
k
x
k−1
V (t)

p

p
(U (t) −U (x
k−1
))
p

p
u(t)dt
_
1/p

+ sup
k∈Z
ϕ(x
k
) V (x
k
)

1
p
< ∞.
The corresponding estimates for A
7
and A
8
are trivial, by using Lemma 8.1
and taking the test function f(t) = (ess sup
0<s<t
v(s))
−1
. The proof is com-
plete.
Proof of Theorem 8.14:
It is easy to see that
_
χ
(0,t)
_
∗∗
u
(s) ≈
U (t)
U (s) +U (t)
. (8.26)
Using (8.26) we calculate the following norms in Theorem 8.14:
_
_
_
χ
(0,t)
__
_
Γ
q
u
(w)

_
_

_
0
U (t)
q
(U (t) +U (s))
q
w(s) ds
_
_
1
q
(8.27)
and
_
_
_
χ
(0,t)
__
_
Γ

u
(w)
≈ ess sup
0<s<∞
U (t) w(s)
U (t) +U (s)
, (8.28)
in each of the eight cases considered in Theorem 8.13.
1. For the case 1 < p < ∞, 0 < p ≤ q < ∞, by using (8.8) and (8.27),
132 Carleman type inequalities and Hardy type inequalities
for monotone functions
we find that
_
_
_
χ
(0,t)
__
_
Λ
p
(v)

u

_
_
_
χ
(0,t)
__
_
Γ
p

u
_
U(t)
p

v(t)
V (t)
p

_
≈ U (t)
_
_

0
_
U (s)
U (t) +U (s)
_
p

v (s)
V (s)
p

ds
_ 1
p

. (8.29)
Thus, using (8.27) and (8.29)
sup
t>0
_
_
χ
(0,t)
_
_
Γ
q
u
(w)
_
_
χ
(0,t)
_
_
Λ
p
(v)

u
U(t)
−1
≈ sup
t>0
_
_

_
0
U (t)
q
(U (t) +U (s))
q
w(s) ds
_
_
1
q
_
_

_
0
_
U (s)
U (t) +U (s)
_
p

v (s)
V (s)
p

ds
_
_
1
p

= A
3
.
2. For the case 1 < p < ∞, q = ∞, by using (8.8) and (8.28), we have
that
sup
t>0
_
_
χ
(0,t)
_
_
Γ

u
(w)
_
_
χ
(0,t)
_
_
Λ
p
(v)

u
U(t)
−1
≈ sup
t>0
_
ess sup
s∈(0,∞)
U (t) w(s)
U (t) +U (s)
_
_
_

_
0
_
U (s)
U (t) +U (s)
_
p

v (s)
V (s)
p

ds
_
_
1
p

= A
6
.
For the case 0 < p ≤ 1, by using (8.8) and (8.26), we obtain that
_
_
χ
(0,t)
_
_
Λ
p
(v)

u

_
_
χ
(0,t)
_
_
Γ

u
_
U(t)
V (t)
1/p
_
≈ U(t) ess sup
t∈(0,∞)
U (s)
U (s) +U (t)
V (s)
−1/p
. (8.30)
3. For the case, 0 < p ≤ 1, p < q < ∞, by using Lemma 8.2 and (8.30), we
Characterizations of embeddings in Lorentz spaces 133
get that
sup
t>0
_
_
χ
(0,t)
_
_
Γ

u
(w)
_
_
χ
(0,t)
_
_
Λ
p
(v)

u
U(t)
−1
≈ sup
t>0
_
_
χ
(0,t)
_
_
Γ
q
u
(w)
_
_
χ
(0,t)
_
_
Γ

u
_
U(t)
V (t)
1/p
_
U(t)
−1
≈ sup
t>0
_
_

_
0
_
U (t)
U (t) +U (s)
_
q
w(s) ds
_
_
1
q
ess sup
s∈(0,∞)
U (s)
U (s) +U (t)
V (s)
−1/p
≈ sup
t>0
_
_

_
0
_
U (t)
U (t) +U (s)
_
q
w(s) ds
_
_
1
q
V (t)
−1/p
= A
1
.
4. For the case 0 < p ≤ 1, q = ∞, by using (8.28), (8.30) and Lemma 8.2,
we find that
sup
t>0
_
_
χ
(0,t)
_
_
Γ

u
(w)
_
_
χ
(0,t)
_
_
Λ
p
(v)

u
U(t)
−1
≈ sup
t>0
_
_
χ
(0,t)
_
_
Γ

u
(w)
_
_
χ
(0,t)
_
_
Γ

u
_
U(t)
V (t)
1/p
_
U(t)
−1
≈ sup
t>0
ess sup
s∈(0,∞)
U (t) w(s)
U (s) +U (t)
ess sup
t∈(0,∞)
U (s)
U (s) +U (t)
V (s)
−1/p
≈ sup
t>0
_
ess sup
s∈(0,∞)
U (s) w(s)
U (s) +U (t)
_
V (t)
−1/p
= A
5
.
5. Now consider the case 1 < p < ∞, 0 < q < p, r =
pq
p−q
. By using (8.8)
(8.27), and (8.29), we have that
_
_

_
0
_
_
_
χ
(0,t)
__
_
r−q
Γ
q
u
(w)
w(t) U (t)
−r
_
_
χ
(0,t)
_
_
r
Λ
p
(v)

u
dt
_
_
1/r

_
_

_
0
_
_
_
χ
(0,t)
__
_
r−q
Γ
q
u
(w)
w(t) U (t)
−r
_
_
χ
(0,t)
_
_
r
Γ
p

u
_
U(t)
p

v(t)
V (t)
p

_
dt
_
_
1/r
134 Carleman type inequalities and Hardy type inequalities
for monotone functions

_
_
_

_
0
_
_

_
0
U (t)
q
(U (t) +U (s))
q
w(s) ds
_
_
r−q
q
w(t)
_
_

_
0
_
U (t)
U (s) +U (s)
_
p

v (s)
V (s)
p

ds
_
_
r
p

dt
_
_
_
1/r
= A
4
.
6. For the case 0 < q < p ≤ 1, by using (8.8), (8.27) and (8.30), we obtain
that
_
_

_
0
_
_
_
χ
(0,t)
__
_
r−q
Γ
q
u
(w)
w(t)U(t)
−r
_
_
χ
(0,t)
_
_
r
Γ

u
_
U(t)
V (t)
1/p
_
dt
_
_
1/r

_
_
_

_
0
_
_

_
0
U(t)
q
(U(t) +U(s))
q
w(s)ds
_
_
r−q
q
w(t)U(t)
−r

_
ess sup
s∈(0,∞)
U(t)U(s)V (s)

1
p
U(t) +U(s)
_
r
dt
_
1
r

_
_
_

_
0
_
_

_
0
U(t)
q
(U(t) +U(s))
q
w(s)ds
_
_
r−q
q
w(t)U(t)
−r
_
sup
s∈(0,t)
U(s)V (s)

1
p
_
r
dt
_
_
_
1
r
= A
2
.
7. For the case p = q = ∞, by using (8.8), (8.28) and Lemma 8.2, we
have that
sup
t>0
_
_
_
χ
(0,t)
__
_
Γ

u
(w)
_
_
χ
(0,t)
_
_
Λ

(v)

u
≈ sup
t>0
_
_
χ
(0,t)
_
_
Γ

u
(w)
_
_
_
χ
(0,t)
__
_
Λ
1
_
u(τ)
ess sup
s∈(0,τ)
v(s)
_
U(t)
−1
≈ sup
t>0
_
ess sup
s∈(0,∞)
U (t) w(s)
U (t) +U (s)
_
1
U (t)
t
_
0
u(τ)
ess sup
s∈(0,τ)
v(s)

Characterizations of embeddings in Lorentz spaces 135
≈ ess sup
s∈(0,∞)
w(s)
U (s)
t
_
0
u(τ)
ess sup
s∈(0,τ)
v(s)

= A
7
.
8. For the case 0 < q < p = ∞, we have that r = q and by using (8.8),
we find that
_
_

_
0
_
_
_
χ
(0,t)
__
_
r−q
Γ
q
u
(w)
w(t) U (t)
−r
_
_
χ
(0,t)
_
_
r
Γ

u
_
U(t)
V (t)
1/p
_
dt
_
_
1/r

_
_

_
0
w(t) U (t)
−q
_
_
_
χ
(0,t)
__
_
q
Λ
1
_
u(t)
ess sup
s∈(0,t)
v(s)
_
dt
_
_
1/q
=
_
_

_
0
_
_
1
U (t)
t
_
0
u(τ)
ess sup
s∈(0,τ)
v(s)

_
_
w(t) dt
_
_
1/q
= A
8
.
The proof is complete.
136 Carleman type inequalities and Hardy type inequalities
for monotone functions
Bibliography
[1] Svenskt Biografiskt Lexikon, Sjunde Bandet. Bonniers förlag, 1927.
[2] H. Alzer. On Carleman’s inequality. Portugal. Math., 50(3):331–334,
1993.
[3] H. Alzer. Refinement of a Carleman-type inequality. Studia Sci. Math.
Hungar., 32(3-4):361–366, 1996.
[4] H. Alzer. A refinement of Carleman’s inequality. J. Approx. Theory,
95(3):497–499, 1998.
[5] M. Ari˜ no and B. Muckenhoupt. Maximal functions on classical Lorentz
spaces and Hardy’s inequality with weights for non-increasing func-
tions. Trans. Amer. Math. Soc., 320, 1990.
[6] I. Asekritova, N. Krugljak, L. Maligranda, and L.-E. Persson. Dis-
tribution and rearrangement estimates of the maximal function and
interpolation. Studia Math., 124:107–132, 1997.
[7] S. Barza. Weighted Multidimensional Integral Inequalities and Appli-
cations. Ph.D. Thesis, Luleå University of Technology, 1999.
[8] S. Barza, H. Heinig, and L.-E. Persson. Duality theorem over the cone
of monotone functions and sequences in higher dimensions. J. Inequal.
Appl., 7(1):79–108, 2002.
[9] S. Barza, M. Johansson, and L.-E. Persson. A Sawyer duality principle
for radially monotone functions in R
n
. J. Inequal. Pure Appl. Math.,
6(2):13, 2005.
137
138 Carleman type inequalities and Hardy type inequalities
for monotone functions
[10] S. Barza, L.-E. Persson, and J. Soria. Sharp weighted multidimensional
integral inequalities for monotone functions. Math. Nachr., 210:43–58,
2000.
[11] S. Barza, L.-E. Persson, and V. D. Stepanov. On weighted multidimen-
sional embeddings for monotone functions. Math. Scand., 88(2):303–
319, 2001.
[12] C. Bennett and R. Sharpley. Interpolation of Operators. Vol. 129.
Academic Press, Pure and Applied Mathematics, Boston, 1988.
[13] G. Bennett and K.-G. Grosse-Erdmann. Weighted Hardy inequalities
for decreasing sequences and functions. Math. Ann., 334:489–531, 2006.
[14] T. Carleman. Sur les fonctions quasi-analytiques. Helsingfors: Aka-
dem. Buchh. (5. Kongress Skandinav. Mathematiker in Helsingfors
vom 4. bis 7. Juli 1922), 181-196, 1923.
[15] T. Carleman and G. H. Hardy. Fourier series and analytic functions.
Proc. Royal Soc. A, 101:124–133, 1922.
[16] L. Carleson. A proof of an inequality of Carleman. Proc. Amer. Math.
Soc., 5:932–933, 1954.
[17] M. Carro, A. García del Amo, and J. Soria. Weak-type weights and
normable Lorentz spaces. Proc. Amer. Math. Soc., 124:849–857, 1996.
[18] M. Carro and J. Soria. Boundedness of some integral operators. Canad.
J. Math., 45:1155–1166, 1993.
[19] M. Carro and J. Soria. Weighted Lorentz spaces and the Hardy oper-
ator. J. Funct. Anal., 112:480–494, 1993.
[20] M. Carro and J. Soria. The Hardy-Littlewood maximal function and
weighted Lorentz spaces. J. London Math. Soc., 55:146–158, 1997.
[21] M. Carro, J. Soria, Pick L., and V.D. Stepanov. On embeddings be-
tween classical Lorentz spaces. Math. Ineq. Appl., 4:397–428, 2001.
[22] A.
ˇ
Ciˇ zmeˇ sija and J. Peˇ cari´ c. Mixed means and Hardy’s inequality.
Math. Inequal. Appl., 1(4):491–506, 1998.
[23] A.
ˇ
Ciˇ zmeˇ sija, J. Peˇ cari´ c, and L.-E. Persson. On strengthened Hardy
and Pólya-Knopp inequalities. J. Approx. Theory, 125(1):74–84, 2003.
Bibliography 139
[24] A.
ˇ
Ciˇ zmeˇ sija, L.-E. Persson, and A. Wedestig. Weighted integral in-
equalities for Hardy and geometric mean operators with kernals over
cones in R
n
. Ital. J. Pure Appl. Math., 18:89–118, 2005.
[25]
ˇ
Ciˇ zmeˇ sija A. and Peˇ cari´ c J. Classical Hardy’s and Carleman’s Inequal-
ities and Mixed Means. Survey on Classical Inequalities (Ed: T.M.
Rassias) Kluwer Acad. Publ., Dordrecht-Boston-London, 2000.
[26] M. Cwikel, A. Kaminska, L. Maligranda, and L. Pick. Are generalized
Lorentz "spaces" really spaces? Proc. Amer. Math. Soc., 132(12):3615–
3625, 2004.
[27] P. Drábek, H. P. Heinig, and A. Kufner. Higher dimensional Hardy
inequality. General Inequalities, (Oberwolfach, 1995), Internat. Ser.
Numer. Math., Birkhäuser, Basel, 7(123):3–16, 1997.
[28] D. E. Edmunds and W. D. Evans. Spectral Theory and Differential
Operators. Oxford Mathematical Monographs. Oxford Science Pub-
lications. The Clarendon Press, Oxford University Press, New York,
1987.
[29] D. E. Edmunds, V. Kokilashvili, and A. Meskhi. Bounded and Compact
Integral Operators. Mathematics and its Applications, 543. Kluwer
Academic Publishers, Dordrecht, 2002.
[30] L. Gårding. Mathematics and Mathematicians. Mathematics in Swe-
den before 1950. Amer. Math. Soc. Providence, 1998.
[31] I. Genebashvili, A. Gogatishvili, V. Kokilashvili, and M. Krbec. Weight
Theory for Integral Transforms on Spaces of Homogeneous Type. Pit-
man Monographs and Surveys in Pure and Applied Mathematics, 92.
Longman, Harlow, 1998.
[32] A. Gogatishvili, M. Johansson, C. A. Okpoti, and L.-E. Persson. Char-
acterizations of embeddings in Lorentz spaces. Bull. Austral. Math.
Soc., 76:69–92, 2007.
[33] A. Gogatishvili, A. Kufner, L.-E. Persson, and A. Wedestig. An equiv-
alence theorem for integral conditions related to Hardy’s inequality.
Real Anal. Exchange, 29(2):867–880, 2003/04.
[34] A. Gogatishvili and L. Pick. Duality principles and reduction theorems.
Math. Ineq. Appl., 3:539–558, 2000.
140 Carleman type inequalities and Hardy type inequalities
for monotone functions
[35] A. Gogatishvili and L. Pick. Discretization and anti-discretization of
rearrangement-invariant norms. Publ. Mat., 47:311–358, 2003.
[36] A. Gogatishvili and L. Pick. Embeddings and duality theory for weak
classical Lorentz spaces. Canad. Math. Bull., 49:82–95, 2006.
[37] M. L. Goldman. Sharp estimates for the norms of Hardy-type opera-
tors on cones of quasimonotone functions. Proc. Steklov Inst. Math.,
232(1):109–137, 2001.
[38] M. L. Gol’dman, Heinig H. P., and V. D. Stepanov. On the principle
of duality in Lorentz spaces. Canad. J. Math., 48:959–979, 1996.
[39] M. Gyllenberg and P. Yan. On a conjecture by Yang. J. Math. Anal.
Appl., 264(2):687–690, 2001.
[40] G. H. Hardy. Notes on some points in the integral calculus. Messenger
of Math., 41(45):163–166, 1915.
[41] G. H. Hardy. Notes on some points in the integral calculus. Messenger
of Math., 51(48):107–112, 1919.
[42] G. H. Hardy. Notes on a theorem of Hilbert. Math. Z, 6:314–317, 1920.
[43] G. H. Hardy. Notes on some points in the integral calculus. Messenger
of Math., 60(54):150–156, 1925.
[44] G. H. Hardy, J. E. Littlewood, and G. Pólya. Inequalities. 2nd ed.,
Cambridge University Press, 1952, 1934.
[45] H. Heinig and L. Maligranda. Weighted inequalities for monotone and
concave functions. Studia Math., 116(2):133–165, 1995.
[46] H. P. Heinig, R. Kerman, and M. Krbec. Weighted exponential in-
equalities. Georgian Math. J, 8(1):69–86, 2001.
[47] L. Hörmander. The Analysis of Linear Partial Differential Operators.
2nd ed. Springer Verlag (1983), 1989.
[48] M. Johansson. A new characterization of the Hardy and it’s limit
Polya-Knopp inequality for decreasing functions. Research report 7,
Department of Mathematics, Luleå University of Technology, submit-
ted, page 14, 2007.
Bibliography 141
[49] M. Johansson. A unified approach to the Sawyer and Sinnamon char-
acterizations of Hardy’s inequality for decreasing functions. Research
report 6, Department of Mathematics, Luleå University of Technology,
page 11, 2007.
[50] M. Johansson, L.-E. Persson, and A. Wedestig. Carleman’s inequality:
History, proofs and some new generalizations. J. Inequal. Pure Appl.
Math., 4(3), 2003.
[51] M. Johansson, L.-E. Persson, and A. Wedestig. Carlemans olikhet-
historik, skärpningar och generaliseringar. Normat, 51(3):89–108, 2003.
[52] M. Johansson, V. D. Stepanov, and E. P. Ushakova. Hardy inequal-
ity with three measures on monotone functions. to appear in Math.
Inequal. Appl., page 20.
[53] S. Kaijser, L.-E. Persson, and A. Öberg. On Carleman’s inequality. J.
Approx. Theory, 117(1):140–151, 2002.
[54] T. Kaluza and G. Szegö. Über Reihen mit Lauter Positivern Gliedern.
J. London Math. Soc, 2:266–272, 1927.
[55] A.A. Kalybay. A new development of Nikol’skii - Lizorkin and Hardy
type inequalities with applications. Ph.D. Thesis, Luleå University of
Technology, 2006.
[56] A. Kamińska and L. Maligranda. On Lorentz spaces Gamma
p,w
. Israel
J. Math., 140:285–318, 2004.
[57] A. Kamińska and L. Maligranda. Order convexity and concavity of
lorentz spaces Lambda
p,w
, 0 < p < ∞. Studia Math., 160(3):267–286,
2004.
[58] B. Kjellberg. Matematiker i Uppsala - Några Minnen. Festschift in hon-
our of Lennart Carleson and Yngve Domar, Proc. Conf. Dept. Math.
Uppsala Univ, 1993.
[59] K. Knopp. Über Reihen mit Positivern Gliedern. J. London Math.
Soc, 3:205–212, 1928.
[60] V. Kokilashvili. On Hardy’s inequalities in weighted spaces. (Russian)
Soobshch. Akad. Nauk Gruzin. SSR, 96(1):37–40, 1979.
[61] A. Kufner, L. Maligranda, and L.-E. Persson. The pre-history of the
Hardy inequality. Amer. Math. Monthly, 113(8):715–732, 2006.
142 Carleman type inequalities and Hardy type inequalities
for monotone functions
[62] A. Kufner, L. Maligranda, and L.-E. Persson. The Hardy Inequality -
About its History and some Related Results. Vydavatelsky Sevis Pub-
lishing House, Pilsen, 2007.
[63] A. Kufner and L.-E. Persson. Weighted Inequalities of Hardy Type.
World Scientific Publishing Co, Singapore/ New Jersey/ London/
Hong Kong, 2003.
[64] A. Kufner, L.-E. Persson, and A. Wedestig. A study of some con-
stants characterizing the weighted Hardy inequality. Orlicz Centenary
Volume, Banach Center Publ., 64, Polish Acad. Sci., Warsaw, pages
135–146, 2004.
[65] E. H. Lieb and M. Loss. Analysis. Graduate Studies in Mathematics.
14. AMS, 1997.
[66] G. G. Lorentz. On the theory of spaces λ. Pacific J. Math., 1:411–429,
1951.
[67] L. Maligranda. Torsten Carleman. at: http://www-history.mcs.st-
and.ac.uk/Biographies/Carleman.html.
[68] L. Maligranda. Generalized hardy inequalities in rearrangement invari-
ant spaces. J. Math. Pures Appl., no. 4, 59(9):405–415, 1980.
[69] L. Maligranda. Weighted estimates of integral operators for decreas-
ing functions. In: Boundary Value Problems, Special Functions and
Fractional Calculus (Minsk, 1996), Belorus. Gos. Univ., Minsk, pages
226–236, 1996.
[70] L. Maligranda. Weighted inequalities for monotone functions. Fourth
International Conference on Function Spaces (Zielona Góra, 1995).
Collect. Math., 48(4-6):687–700, 1997.
[71] L. Maligranda. Weighted inequalities for quasi-monotone functions. J.
London Math. Soc., 57(2):363–370, 1998.
[72] D. S. Mitrinovi, J. Peˇ cari´ c, and A. M. Fink. Inequalities Involving
Functions and their Integrals and Derivatives. Kluwer Academic Pub-
lishers, 1991.
[73] B. Mond and J. Peˇ cari´ c. A mixed means inequality. Austral. Math.
Soc. Gaz, 23(2):67–70, 1996.
Bibliography 143
[74] B. Muckenhoupt. Hardy’s inequality with weights. Studia Math.,
44:31–38, 1972.
[75] M. Nassyrova. Weighted inequalities involving Hardy-type and limiting
geometric mean operators. Ph.D. Thesis, Luleå University of Technol-
ogy, 2002.
[76] M. Nassyrova, L.-E. Persson, and V. D. Stepanov. On weighted in-
equalities with geometric mean operator by the Hardy-type integral
transform. J. Inequal. Pure Appl. Math., 3(4), 2002.
[77] C. Okpoti, L.-E. Persson, and G. Sinnamon. An equivalence theorem
for some integral conditions with general measures related to Hardy’s
inequality. J. Math. Anal. Appl., 326(1):398–413, 2007.
[78] C. Okpoti, L.-E. Persson, and G. Sinnamon. An equivalence theorem
for some integral conditions with general measures related to Hardy’s
inequality II. To appear, 2007.
[79] C.A. Okpoti. Weight Characterization of Hardy and Carleman Type
Inequalities. Ph.D. Thesis, Luleå University of Technology, 2006.
[80] B. Opic and A. Kufner. Hardy-Type Inequalities. Pitman Research
Notes in Mathematics 219, Longman Scientific & Technical, Harlow,
1990.
[81] B. Opic and L. Pick. On the geometric mean operator. J. Math. Anal.
Appl., 183(3):652–662, 1994.
[82] A. Ostrowski. Über quasi-analytische Funktionen und Bestimmtheit
Asymptotischer Entwicklungen. Acta Math., 53:181–266, 1929.
[83] J. Peˇ cari´ c and K. B. Stolarsky. Carleman’s inequality: History and
new generalizations. Aequationes Math., 61(1-2):49–62, 2001.
[84] L.-E. Persson and V. D. Stepanov. Weighted integral inequalities with
the geometric mean operator. J. Inequal. Appl., 7(5):727–746, 2002.
[85] L.-E. Persson, V. D. Stepanov, and E. P. Ushakova. Equivalence of
Hardy-type inequalities with general measures on the cones of non-
negative respective non-increasing functions. Proc. Amer. Math. Soc.,
134(8):2363–2372, 2006.
[86] G. Pólya. Proof of an inequality. Proc. London Math. Soc, 24(2), 1926.
144 Carleman type inequalities and Hardy type inequalities
for monotone functions
[87] D. V. Prokhorov. Hardy’s inequality with three measures. Proc. Steklov
Inst. Math., 255:233–242, 2007.
[88] D. V. Prokhorov. Inequalities of Hardy type for a class of integral
operators with measures. Anal. Math., 33:199–225, 2007.
[89] D.V. Prokhorov. Weighted Inequalities Involving Riemann-Liouville
and Hardy Type Operators. Ph.D. Thesis, Luleå University of Technol-
ogy, 2003.
[90] R. M. Redheffer. Recurrent inequalities. Proc. London Math. Soc,
17(3):683–699, 1967.
[91] R. M. Redheffer. Easy proofs of hard inequalities. In: General inequal-
ities, Oberwolfach, 3:123–140, 1981.
[92] H. L. Royden. Real Analysis. Third edition. Macmillan Publishing
Company, New York, 1988.
[93] E. Sawyer. Boundedness of classical operators on classical Lorentz
spaces. Studia Math., 96(2):145–158, 1990.
[94] G. Sinnamon. Spaces defined by level functions and their duals. Studia
Math., 111:19–52, 1994.
[95] G. Sinnamon. Embeddings of concave functions and duals of Lorentz
spaces. Publ. Math., 46:489–515, 2002.
[96] G. Sinnamon. Transferring monotonicity in weighted norm inequalities.
Collect. Math., 54:181–216, 2003.
[97] G. Sinnamon. Hardy’s inequality and monotonocity. In: Function
Spaces and Nonlinear Analysis (Eds.: P. Drábec and J. Rákosnik),
Mathematical Institute of the Academy of Sciences of the Czech Re-
public, Prague 292-310, 2005.
[98] G. Sinnamon and V. D. Stepanov. The weighted Hardy inequality:
New proofs and the case p=1. J. London Math. Soc., (2) 54:89–101,
1996.
[99] V. D. Stepanov. Integral operators on the cone of monotone functions.
J. London Math. Soc., 48(3):465–487, 1993.
[100] V. D. Stepanov. The weighted Hardy’s inequality for nonincreasing
functions. Trans. Amer. Math. Soc., 338(1):173–186, 1993.
Bibliography 145
[101] E. Ushakova. Norm inequalities of Hardy and Pólya-Knopp types.
Ph.D. Thesis, Luleå University of Technology, 2006.
[102] A. Wedestig. Some new Hardy type inequalities and their limiting
inequalities. J. Inequal. Pure Appl. Math., 4(3), 2003.
[103] A. Wedestig. Weighted Inequalities of Hardy Type and their Limiting
Inequalities. Ph.D. Thesis, Luleå University of Technology, 2003.
[104] P. Yan and G. Sun. A strengthened Carleman inequality. J. Math.
Anal. Appl., 240(1):290–293, 1999.
[105] X. Yang. On Carleman’s inequality. J. Math. Anal. Appl., 253(2):691–
694, 2001.