Commentationes Mathematicae Universitatis Carolinae

Alberto Fiorenza; Miroslav Krbec

Indices of Orlicz spaces and some applications
Commentationes Mathematicae Universitatis Carolinae, Vol. 38 (1997), No. 3, 433--451

Persistent URL: http://dml.cz/dmlcz/118943

Terms of use:
Charles University in Prague, Faculty of Mathematics and Physics, 1997
Institute of Mathematics of the Academy of Sciences of the Czech Republic provides access to
digitized documents strictly for personal use. Each copy of any part of this document must
contain these Terms of use.
This paper has been digitized, optimized for electronic delivery and stamped
with digital signature within the project DML-CZ: The Czech Digital
Mathematics Library http://project.dml.cz

433

Comment.Math.Univ.Carolin. 38,3 (1997)433451

Indices of Orlicz spaces and some applications

Alberto Fiorenza , Miroslav Krbec

Abstract. We study connections between the Boyd indices in Orlicz spaces and the
growth conditions frequently met in various applications, for instance, in the regularity theory of variational integrals with non-standard growth. We develop a truncation
method for computation of the indices and we also give characterizations of them in
terms of the growth exponents and of the Jensen means. Applications concern variational integrals and extrapolation of integral operators.
Keywords: Boyd indices, Orlicz spaces, Simonenko indices, non-standard growth conditions, variational integrals, interpolation, extrapolation
Classification: Primary 46E30; Secondary 26A12, 35A15, 35B10, 42B20, 46E35

1. Introduction
The aim of this paper is to establish connections between the Boyd indices
of Orlicz spaces and the growth conditions on Young functions appearing in the
theory of non-linear b.v.p. with non-standard growth and in the interpolation and
extrapolation theory in Lebesgue and Orlicz spaces.
From large number of relevant references dealing with various sorts of indices in
Orlicz and also more general r.i. spaces we recall the Matuszewska-Orlicz indices
in [17], the Boyd indices [4], [5], [6], the Zippin indices [28], Maligranda [16]
with many further references. The detailed exposition in the general setting of
r.i. spaces and also in Orlicz spaces can be found in Bennett and Sharpley [3]. The
Boyd indices in Lorentz-Orlicz spaces have been studied in Montgomery-Smith
[19]. It is well known that all the indices mentioned above coincide in Orlicz
spaces so that henceforth we shall speak only about the Boyd indices.
The definition of the Boyd indices is very simple, nevertheless, a particular
computation might be extremely difficult. This paper is intended also as a contribution to development of effective methods of establishing their values.
In Sections 2 and 3 we present estimates for the Boyd indices of a Young
function in terms of the growth conditions
(1.1)

p(t) t (t) q(t),

t R1 ,

This work has been partly performed as a part of a National Research Project and partly
supported by G.N.A.F.A.
Partly supported by Grant No. 201/96/0431 of GA CR.

434

A. Fiorenza, M. Krbec

giving birth to the Simonenko indices. Of crucial importance is the following theorem, linking the reciprocals i() and I() of the Boyd indices of a Young function
with the Simonenko indices p() and q() of equivalent Young functions. We
shall agree that from now on all Young functions and their complementary functions satisfy the 2 -condition.
Theorem 1.1. Let be a Young function. Then
i() = sup p() and I() = inf q().

As observed above, the definition of the Boyd indices does not often represent
an efficient tool for computation and one has to find another way. The following
theorems give an answer in this direction. We develop a truncation technique,
which itself represents another means for dealing with the Boyd indices of Young
functions, particularly in the cases when the limits r0 and r in Theorem 1.3 do
not exist (cf. Example 5.8); in Section 4 we prove the following theorem:
Theorem 1.2. Let be a Young function. Put
t (t)
,
(t)

F (t) =

t > 0,

and let us define

[F ] (t) = max(F (t), )
Then
i() sup{ > 0 ;

I() inf{ > 0 ;

and

If there exist

and [F ] (t) = min(F (t), ),

(1.2)

t (t)
t0 (t)

i() = sup{ > 0 ;

(1.3)

I() = inf{ > 0 ;

([F ] (s) F (s))

ds
< }.
s

t (t)
,
t (t)

and r = lim

and

ds
< }
s

r0 = lim
then

([F ] (s) F (s))

t > 0.

([F ] (s) F (s))

ds
< }
s

([F ] (s) F (s))

ds
< }.
s

When proving Theorem 1.2 we also obtain the following estimates, which have
been established in Fiorenza and Krbec [8], by a different method.

435

Indices of Orlicz spaces and some applications

Theorem 1.3. Let be a Young function. Then



t (t)
t (t)
i() min lim sup
,
, lim sup
(t)
t (t)
t0


t (t)
t (t)
,
, lim inf
i() min lim inf
t (t)
t0
(t)
and



t (t)
t (t)
,
, lim sup
I() max lim sup
(t)
t (t)
t0


t (t)
t (t)
I() max lim inf
.
, lim inf
t (t)
t0
(t)

Hence if there exist

t (t)
t0 (t)

r0 = lim

t (t)
,
t (t)

and r = lim

then i() = min(r0 , r ) and I() = max(r0 , r ).

Besides that we also pursue the problem of relationsRbetween the

indices of
1
and the (Jensen) integral mean M (f ) = 1 ||
(f
)
dx
, studied in

Fiorenza [7] in connection with (1.1) and we get still another characterization
of the indices in Theorem 3.7. A straightforward application of this theorem
yields an alternative proof and actually an improvement of Migliaccios theorem
(see [18]) on extrapolation of reverse H
olders inequality.
In the concluding Section 5 we present several further applications. Since in
general p() i() I() q() and any of these inequalities can be sharp it
often occurs that conditions in terms of the Simonenko indices are more restrictive.
On the other hand, many particular problems have been dealt with, from one
reason or another, with use of the growth exponents. A general common feature
of the approach offered by Theorem 1.1 is use of the Boyd indices in claims while
sticking to the Simonenko indices in proofs. This can be done, for instance, when
studying regularity properties of (Q-quasi)minima of
Z
(Dv)w dx,
(1.4)
I(, v) =

1,1

where is a bounded open subset of Rn , v = (v1 , . . . , vN ) Wloc (, RN ), w is

an Ap weight and satisfies (1.1). Let m0 = inf{m 1; w Am }.
Sbordone [24] proved that the condition
nm0 q
< p q < nm0
(1.5)
nm0 + q
implies that r (|Du|) is locally integrable in with some r > 1. We shall show
that p and q in (1.5) can be replaced by the lower and the upper index of ,

436

A. Fiorenza, M. Krbec

respectively. An analogous approach yields a variant of Harnacks inequality

for non-negative (Q-quasi)minima of a non-weighted version of (1.4) for scalar
functions v, studied under the condition (1.1) by Moscariello [20].
We are also concerned with extrapolation of inequalities for integral operators, we formulate and prove a generalization of Simonenkos extrapolation theorem [26].
At the end of the paper we present examples of particular Young functions,
showing how our results can be used for computations of the Boyd indices when
the definition seems to be of no practical use.
2. Preliminaries
Let us fix notation and recall basic concepts. For our purposes, a Young function will be any non-negative, even, convex function : R1 R1 such that
is (strictly) increasing on [0, ), and lim (t)/t = 0, lim (t)/t = . To avoid
t0

non-important technicalities we shall suppose that exists everywhere in R1 .

A Young function satisfies the 2 -condition ( 2 ) if there is c > 0 such
that (2t) c(t) for all t R1 .
e
Let be a Young function. Then (s)
= sup {|st| (t); t R1 }, s R1 , is
the so called complementary function with respect to . The function is said
to be equivalent to another Young function (we shall write ) if there is
c > 0 such that (c1 t) (t) (ct), t R1 , with some c > 0.
Let 2 be a Young function and let us define
h () = sup
t>0

(t)
,
(t)

> 0.

The numbers
(2.1)

i() = lim

log h ()
log h ()
= sup
log
log
0<<1

I() = lim

log h ()
log h ()
= inf
log
log
1<<

0+

and
(2.2)

are called the lower index of and the upper index of , respectively. Sometimes
these indices are called the fundamental indices of .
The numbers i() and I() are reciprocals of the Boyd indices (see Boyd [4],
[5], [6]). The right wing equalities in (2.1) and (2.2) follow from known properties
of subadditive functions (since log h enjoys this property), see, e.g. [12], and
I() < by virtue of the assumption 2 .
Always 1 i() I() and it is i() > 1 if and only if the complementary
e satisfies the 2 -condition. The couples i()
e and I(), and I()
e and
function

437

Indices of Orlicz spaces and some applications

i() behave similarly as conjugate exponents of power functions (see, e.g. [5], [6],
e = I()/(I() 1) and I()
e = i()/(i() 1).
[12]), namely, i()
Throughout the paper we shall assume that all Young functions under consideration satisfy the 2 -condition together with their complementary functions.
We observe that the Boyd indices turned out to be particularly useful in the
theory of classical operators in Orlicz spaces. The well-known Muckenhoupt theorem [21] gives a characterization of weights w in Rn for which the Hardy maximal
operator takes Lp (w) boundedly into Lp (w), namely, w Ap , the Muckenhoupt
class. This has been extended to the context of reflexive Orlicz spaces (that is,
e 2 ) by Kerman and Torchinsky [13]. They showed that the maximal op,
erator is modularly continuous if and only if w belongs to the Muckenhoupt class
Ai() .
Some more connections between the indices and the theory of classical operators in Orlicz spaces can be found in Kokilashvili and Krbec [14, Chapters 2, 3].
In many applications (calculus of variations, interpolation etc.) it is useful
to assume that the Young function in question is in a certain sense between two
powers tp and tq and an appropriate quantitative analysis is needed. At one hand,
2 is equivalent to existence of p0 , p1 [1, ), p0 p1 , such that
(2.3)

(t) C max (p0 , p1 ) (t),

, t 0,

(see Gustavsson and Peetre [12]) and this in turn gives

(2.4)

(t) C 1 min (p0 , p1 ) (t),

, t 0,

and one can see that sup of those p0 and inf of those p1 such that (2.3) and (2.4)
hold equals to i() and I(), respectively. We observe that the growth estimates
(2.3) and/or (2.4) immediately give the formulas (cf. [16])
i() = sup{p ; inf p
u>0
1

(u)
> 0},
(u)

I() = inf{p ; sup p

u>0
1

(u)
< }.
(u)

On the other hand, a control of a Young function C 1 by power functions

frequently used, for instance, in connections with applications to PDEs (see, e.g.
[9], [24], [25], [20]) has the form (1.1), which can be equivalently written as
(2.5)

(t)

tp

and

(t)

tq

on (0, ).

To distinguish the couples of exponents p, q in (1.1) belonging to different Young

functions we shall sometimes write p = p and q = q . The numbers p and q will
be called growth exponents of .

438

A. Fiorenza, M. Krbec

Let us consider the Simonenko indices, see [26], that is, the best p and q such
that (1.1) holds:
t (t)
t>0 (t)

p() = inf

and q() = sup

t>0

t (t)
.
(t)

e 2 if and only if 1 < p()

It is known ([15, Theorem 5.1]) that ,
q() < .
We conclude this section with a survey of useful properties of the lower, upper
and Simonenko indices. Not all of them are used in the sequel, but they are listed
for completeness. Their straightforward proofs are omitted.
Proposition 2.1. For r > 0 let er (t) = |t|r . Then
= p() ,
= q() ,
q()
p()
q() 1
p() 1
i()
I()

,
I() =
,
i() =
I() 1
i() 1
p(1 ) = 1 ,
q(1 ) = 1 ,
q()
p()
i(1 ) = 1 ,
I(1 ) = 1 ,
I()
i()
p( ) p()p( ),
q( ) q()q( ),
p(er ) = p()r,
q(er ) = q()r,
p( er ) = p()r,
q( er ) = q()r,
i( ) i()i( ),
I( ) I()I( ),
i(er ) = i()r,
I(er ) = I()r,
i( er ) = i()r,
I( er ) = I()r,
p( ) p() + p( ),
q( ) q() + q( ),
p( er ) = p() + r,
q( er ) = q() + r,
i( ) i() + i( ),
I( ) I() + I( ),
i( er ) = i() + r,
I( er ) = I() + r.
3. Characterizations of indices by growth exponents and integral
means
The following results are stated for both the lower and the upper indices.
Nevertheless, the proofs are analogous or could be deduced thanks to a certain
duality between i() and I(), therefore we shall restrict ourselves only to the
proofs of the part concerning the lower indices.
We start with a quantitative relation between (2.3) and (2.5) (see Gustavsson
and Peetre [12] and Persson [22]) that will be useful in the sequel:
Proposition 3.1. Let be a Young function. Then the following statements
are equivalent
(i) There are 1 p0 , p1 < such that (t) C max (p0 , p1 ) (t) for all
, t 0, with C independent of and t;

439

Indices of Orlicz spaces and some applications

(ii) is equivalent to a function such that (t)/tp0 and (t)/tp1 on

(0, ).
Lemma 3.2. Let a Young function satisfy (1.1). Then p p() i()
I() q() q.
Proof: Given (0, 1) we have by Proposition 3.1 and (2.3) that (t)
cp (t) for every p p(). Hence h () cp and
log h ()
log c
p+
,
log
log
implying that i() p.

Example 3.3. In general p() 6= i(). Put

2
if t [0, 1),

t
(t) = 2t 1
if t [1, 2),

2
t /2 + 1 if t [2, ),

and (t) = (t) for t < 0. Then is a Young function with a continuous
derivative. An elementary calculation yields that p() = 4/3 and q() = 2. On
the other hand, by virtue of Theorem 1.3 we get i() = 2. Let us observe that
t2 /2 (t) t2 , t R1 , and 4/3 = p() 6= p(t2 ) = 2.
Remark 3.4. When dealing only with the Boyd indices, then the assumption
C 1 often is not a restriction. Indeed, if 2 , then
Z |t|
(s)
(t) =
ds,
t R1 ,
s
0
is a continuously differentiable Young function equivalent to and such that
e C 1 . Moreover, if (1.1) holds, then the same is true for . Also,
2 , and
e 2 , then
e 2 . This is a consequence of invariance
i() = i(), and, if
of these properties with respect to the equivalence of Young functions, namely, if
2 and , then 2 , and if , then i() = i( ) and I() = I( ).
Remark 3.5. Let us observe, however, that does not generally imply that
p() = p( ). For instance, a simple computation shows that the Simonenko
indices of from Example 3.3 and the corresponding from Remark 3.4 do not
coincide. This indicates that relations of p() and q() at one hand and the Boyd
indices at the other hand are of a rather delicate nature.
Given a Young function , let us consider the class of equivalent Young functions. By Lemma 3.2 we know that
(3.1)

p() i() = i()

and I() = I() q() for all .

440

A. Fiorenza, M. Krbec

We are now in a position to prove Theorem 1.1

Proof of Theorem 1.1: By (3.1) we need only to prove that sup p() i().

Let > 0. Then (see, e.g. [14, Chapter I]) there is C such that


(t) C max i() , I()+ (t),

, t 0.

By Proposition 3.1 there is an even function with growth exponents i()

and I() + , and equivalent to . Moreover, can be chosen in such a way
that it is also a Young function. Indeed, multiplying (1.1), with instead of
by t1 and integrating over (0, s), s > 0, we arrive at
(i() ) (s) s (s) (I() + ) (s),
where
(t) =

|t|

(s)
ds,
s

s > 0,

t R1 ,

(cf. Remark 3.4). Therefore the functions and are such that
(3.2)

(t)

ti()

and

(t)
I()+
t

on (0, ),

and, on the other hand, are equivalent because both are equivalent to . But
(3.2) implies immediately that i() p( ) sup p() and therefore the

proof is complete.

We shall finish this section with another characterization using the Jensen
means


Z
1
M (f ) = 1
(f ) dx ,
||
where is a bounded open subset of Rn . Let
Mr (f ) =

1
||

1/r
|f |r dx
.

In [7] there is proved that the growth conditions (1.1) imply the existence of
positive constants c1 , c2 such that
(3.3)

c1 Mp (f ) M (f ) c2 Mq (f ).

This can be slightly improved as follows:

441

Indices of Orlicz spaces and some applications

Proposition 3.6. The inequality (3.3) holds with every 0 < p < i() and every
I() < q < .
Proof: Let p < i(). By virtue of Theorem 1.1 there exists such that
p() > p. H
olders inequality and the fact that 1 1 yield
Mp (f ) Mp() (f ) c M (f ) c M (f ).

A natural question arises, namely, whether the inequality (3.3) can be used
for another characterization of the indices. We shall show that this is indeed the
case.
Theorem 3.7. Let Rn be a bounded open subset of Rn . Then
(3.4)
(3.5)

i() = sup{r 1 ; M (f ) cMr (f ) for some c > 0},

I() = inf{s 1 ; M (f ) cMs (f ) for some c > 0}.

Proof: By Proposition 3.6 we already know that

i() sup{r 1 ; M (f ) cMr (f ) for some c > 0}.
Let us assume that there are > 0, c > 0 such that
(3.6)

M (f ) cMi()+ (f ).

Put g(t) = (f (t))i()+ , (t) = (t1/(i()+) ), then 1 (s) = (1 (s))i()+ .

From (3.6) we get that there is c1 > 0 such that

1
||

 Z


1
|f | dx c1
(c1 f ) dx ,

that is, according to [14, Lemma 1.1.1] the function is pseudoconvex in the sense
that there is a convex function and a constant c2 > 0 such that (t) (t)
c2 (c2 t) for all t 0. It is easy to see that the respective indices of and must
coincide, in particular, i( ) 1. Our assumptions give i( ) = i()/(i()+ ) < 1
which is a contradiction. Hence (3.4) holds.

The preceding theorem yields an alternative proof and actually a mild improvement of a theorem due to Migliaccio [18] on the extrapolation of the reverse
H
older inequality.

442

A. Fiorenza, M. Krbec

Corollary 3.8. Let p < i(), Rn a bounded open set, and



Z
Z
1
1
(3.7)
(w) dx b
w dx ,
||
||
then

1
||


p
Z
1
w dx c(b, p)
w dx .
||

Proof: According to (3.4) we have

1/p



Z
Z
Z
c(b, p)
1
1
p
1
w dx
(w) dx
w dx
c(p)
||
||
||
and we are done. We observe that the assumption (3.7) is weaker than the original
one, when (3.7) is required for w with every > 0.

4. Proofs of formulas for the Boyd indices
Proofs of Theorems 1.2, 1.3: Let us assume first that r0 and r exist. We will
divide the proof into the following steps:
Step 1. i() min(r0 , r ).
R
Step 2. min(r0 , r ) sup{ > 0 ; 0 ([F ] (s) F (s)) ds/s < }.
R
Step 3. sup{ > 0 ; 0 ([F ] (s) F (s)) ds/s < } i().

Step 1. We shall make use of Theorem 1.1. Let be a Young function. Then
(t) = a(t)(t), where 0 < m a(t) M < for all t > 0 with m and M
independent of t. We have
F (t) =

ta (t)
+ F (t),
a(t)

t > 0,

and let us point out that

(4.1)

lim inf Fa (t) 0.

t0

Indeed, assuming that lim inf Fa (t) > > 0, there is > 0 such that Fa (t) >
t0

for all t (0, ) and therefore

Z
Z
a (t) dt >
M a() a(0) =

a(t)

dt
> m
t

dt
= .
t

By virtue of (4.1) we have

p lim inf F (t) lim inf Fa (t) + lim F (t) r0 ,
t0

t0

i() = sup p r0 .

t0

443

Indices of Orlicz spaces and some applications

The proof of lim inf Fa (t) 0 is analogous. We arrive at i() min(r0 , r ).

t

Step 2. We claim that

< min(r0 , r ) sup{ > 0 ;

([F ] (s) F (s))

ds
< }
s

and therefore that supremum is greater than or equal to min(r0 , r ) = i().

Indeed, given > 0, put = min(r0 , r ) . Then there is 0 < 1 < 2 such
that F (t) for all t (0, 1 ) (2 , ) and we have [F ] (s) F (s) = 0 for
all s (0, 1 ) (2 , ), hence
Z
ds
< .
([F ] (s) F (s))
s
0
Step 3. Let > 0 be such that the integral in (1.3) is finite and put
Z t

ds
a(t) = exp
([F ] (s) F (s))
,
t 0.
s
1
By the definition of [F ] we have a L and therefore the function
Z |t|
a(s)(s)
ds,
t R1 ,
(t) =
s
0
is a Young function equivalent to . Further,






t(a(t)(t))
= inf F (t) + t(log a) (t) = inf [F ] (t)
p( ) inf
t>0
t>0
a(t)(t)
t>0
which yields p( ) i( ) = i().
If one of the above limits does not exist, Theorem 1.2 follows from Step 3, and
Theorem 1.3 follows by replacing r0 , r by lim sup and lim sup F (t), respectively,
t

t0

in the proof of Step 1, and by lim inf F (t) and lim inf F (t), respectively, in the
t0

proof of Step 2.
In accordance with our previous agreement we omit the analogous proof of the
part of the theorems concerning I().

5. Applications
As we have observed in Introduction we are going to present some applications
of the developed theory, following the general pattern, namely replacement of the
growth exponents by the Boyd indices in the respective claims.
Let be a Young function such that
(5.1)

p(t) t (t) q(t),

t R1 ,

444

A. Fiorenza, M. Krbec

with some 1 < p q < . Further, let

(5.2)

I(, v) = I (, v) =

(|Dv|) dx,

where is a bounded open set of Rn and v = (v1 , . . . , vN ).

1,1
A function u Wloc
(, RN ) is a minimizer of I if
I(supp , u) I(supp , u + )
1,1
for every Wloc
(, RN ) supported in .
1,1
Let Q 1. A function u Wloc
(, RN ) is a Q-quasiminimizer of I if

I(supp , u) QI(supp , u + )
1,1

for every Wloc (, RN ) supported in .

Regularity of minimizers for the functionals in (5.2) have been studied in Fusco
and Sbordone [9]. They proved that if u is a minimizer of I then the condition
(5.1) together with
(5.3)

q < p =

np
np

if p < n

guarantee that there is r > 1 such that (|Du|) Lrloc (). Note that they use
the Young function from Example 5.8 to illustrate the result.
The minima of functionals of type (5.2) have been studied by Giaquinta and
Giusti [10], Giaquinta and Modica [11], Sbordone [23] under the condition
(5.4)

c1 tp c2 (t) c3 (1 + tq ),

t > 0,

with p = q. It is easy to see that (5.1) implies (5.4), nevertheless, as observed

in [9] the growth exponents from (5.1) need not be necessarily the best ones for
which (5.4) holds; if p and q are the best possible exponents from (5.4), then
p p q q and in general any of these inequalities can be sharp.
More generally, let w Am with some m > 1 and consider
Z
(5.5)
J (, v) = J (, v) =
(|Dv|)w dx.

Let
(5.6)

m0 = inf{m 1 ; w Am }.

Sbordone [24] has proved that the condition

(5.7)

nm0 q
< p q < nm0
nm0 + q

Indices of Orlicz spaces and some applications

implies that r (|Du|) is locally integrable in with some r > 1.

The condition (5.3) can be rewritten as
1 1
1
<
p q
n

if p < n

so that (5.3) is actually a condition on the distance of the reciprocals of the growth
exponents p and q; the situation is similar as to the left wing inequality in (5.7).
Recalling Lemma 3.2 we see that the conditions
(5.8)

I() < (i()) =

if i() < n, and

(5.9)

ni()
n i()

nm0 I()
< i() I() < nm0
nm0 + I()

are weaker than (5.3) and (5.7), respectively.

As noticed in [9] the regularity statement for (5.2) holds for Q-quasiminimizers,
too. In fact, giving a closer look at the proof in [9] one can see that the original
assumption about u is only used for proving the Caccioppoli type inequality


Z
Z
u uR
dx,

(|Du|) dx c
(5.10)
R
BR/2
BR/2
where BR is any ball of the radius R. Going along the lines of the proof it
becomes clear that the inequality (5.10) with a possibly different constant c can
be proved without any extra effort if u is a Q-quasiminimizer, too.
We can now present several applications.
Theorem 5.1.
1,1
(i) Let be a Young function and let u Wloc (, RN ) be a Q-quasiminimizer
of (5.2). Suppose that (5.8) holds provided i() < n. Then there is r > 1
such that r (|Du|) L1loc ().
1,1
(ii) Let be a Young function and let u Wloc
(, RN ) be a Q-quasiminimizer
of (5.5) with w Am for some m > 1. Let m0 be the critical index from
(5.6). If (5.9) holds, then there is r > 1 such that r (|Du|) L1loc ().
In the paper by Moscariello [20], the condition (5.3) has been also used to prove
a Harnack type inequality for scalar-valued Q-quasiminimizers of (5.2), namely,
if u 0 is a Q-quasiminimizer of (5.2) and BR is a ball, then for every
(0, 1) there exists a constant C = C(p, q, Q, n, ) such that
sup u(x) C

xBR

inf

xBR

u(x).

(See Di Benedetto and Trudinger [2] for the original Lp -setting.)

This result can also be formulated in terms of the Boyd indices:

445

446

A. Fiorenza, M. Krbec

Theorem 5.2. Let n > 1 and suppose that (5.8) holds provided i() < n. Let
1,1
BR be a ball whose closure is contained in . If u Wloc
(, R1 ) is a non-negative
Q-quasiminimizer of (5.2), then for every (0, 1) there is C = C(p, q, Q, n, )
such that
sup u(x) C inf u(x).
xBR

xBR

Proofs of Theorems 5.1, 5.2: We restrict ourselves to the proof of Theorem 5.1
(i). After that it will be clear how to proceed in the remaining cases.
Let u be a Q-quasiminimizer of (5.2) with satisfying the condition (5.8).
Choose > 0 in such a way that
I() + (i() ) .

(5.11)

We know (see the proof of Theorem 1.1) that there is a Young function
such that
(5.12)

p( ) i() ,

Then the functional

I (, v) =

q( ) I() + .
Z

(|Dv|) dx

is of type (5.2) and by virtue of (5.11) and (5.12) the growth exponents of
satisfy inequality (5.3). As Q-quasiminima of are Q-quasiminima of (with
possibly different Q) and vice versa we are done.

Remark 5.3. The calculus of the indices i() and I() can be difficult in particular
cases, for instance, when Theorems 1.2, 1.3 are not sufficient. On the other hand,
the behaviour of near the origin, although relevant for i(), does not in fact
play any role in Theorem 5.1 (cf. the final remarks in [9] and [24]). This suggests
another couple of indices. If we put
pe = pe = lim inf
t

t (t)
,
(t)

qe = qe = lim sup
t

t (t)
,
(t)

then p pe qe q , and therefore the conditions

(5.13)
and
(5.14)

pe qe < (e
p ) =

ne
p
n pe

nm0 qe
< pe qe < nm0
nm0 + qe

are weaker than (5.3) and (5.7), respectively.

It is possible to prove a result better than Theorem 5.1. To this goal we shall
need the following special construction:

Indices of Orlicz spaces and some applications

Lemma 5.4. Let be a Young function such that

1 < lim inf
t

t (t)
.
(t)

Then there is a Young function such that

(5.15)

t R1 ,

c1 (t) c2 (t) c3 (t) + c4 ,

for some c1 , c2 , c3 , c4 > 0 independent of t, and

(5.16)

i( ) = ,

I( ) lim sup
t

t (t)
.
(t)

Proof: Step 1. We show that there is a Young function G and constants

c1 , c2 > 0 such that
c1 (t) c2 G(t) (t),

(5.17)

t R1 ,

and
(5.18)

qG > .

If q > it suffices to put G = and we are done. If this is not the case, let
c > 0 be such that
2 (2)
> ,
(2) c

(2) c
< 1,
(1) + (1)

and define
(2) c

(t)

(1) + (1)
(2) c
G(t) =
[ (1)(|t| 1) + (1)]

(1) + (1)

(t) c

if |t| 1,
if 1 < |t| < 2,
if |t| 2.

It is easy to check that G satisfies (5.17) and (5.18).

Step 2. We shall construct . Choose t0 > 0 such that t0 G (t0 )/G(t0 ) > and
put

G(t0 ) |t| if |t| t0 ,

t0
(t) =

G(t)
if |t| > t0 .

447

448

A. Fiorenza, M. Krbec

Then is a Young function; its convexity follows from

lim (t) =

tt0

G(t0 ) 1 G(t0 )
t0 =
< G (t0 ) lim G (t).
tt0+
t0
t0

Moreover, G(t) G(t0 ) (t) G(t) + G(t0 ), t R1 , hence

(2) c
(t) G(t0 ) (t) G(t) + G(t0 ) (t) + (t0 ),
(1) + (1)

t R1 ,

and (5.15) is proved.

Step 3. We shall prove (5.16). By Theorem 1.3 we have




t (t)
tG (t)
t (t)
i( ) min lim sup
= min , lim sup
, lim sup
(t)
t G(t)
t (t)
t0


t (t)
= .
= min , lim sup
t (t)
On the other hand,




tG (t)
t (t)
t (t)
= min , lim inf
, lim inf
i( ) min lim inf
t G(t)
t (t)
(t)
t0



t (t)
= .
= min , lim inf
t (t)
The estimate for I( ) follows directly from Theorem 1.3 because = G near
infinity, and G and differ by a constant on (2, ).

Similarly, one can prove the dual statement:
Lemma 5.5. Let be a Young function such that
lim sup
t

t (t)
< .
(t)

Then there is a Young function such that

(5.19)

c1 (t) c2 (t) c3 (t) + c4 ,

t R1 ,

for some c1 , c2 , c3 , c4 > 0 independent of t, and

(5.20)

I( ) = ,

i( ) lim inf
t

t (t)
.
(t)

Combining Lemma 5.4 with the foregoing considerations we arrive at

Indices of Orlicz spaces and some applications

Theorem 5.6. The assertions (i) and (ii) in Theorem 5.1 hold also under the
conditions (5.13) with pe < n and (5.14), respectively.

Next we shall pay attention to one of the interesting pioneering results due to
Simonenko [26] on extrapolation of integral operators with a kernel, which enables
very simply and naturally to carry over for instance the theorems on singular
integrals from Lp to L spaces with finite indices. The original assumption in
[26] for the claim in the following theorem reads 1 < p() q() < q.
Z
Theorem 5.7. Let
Kf (x) =
A(x, y)f (y) dy

and let us suppose that K is continuous in Lq (). Let B(x0 , r) denote a ball
centered at x0 and of Zthe radius r. Assume that
sup

(B(x0 ,r)c )

|A(x, y) A(x, y )| dx < ,

where the sup is taken over all y, y B(x0 , 2r), x0 Rn , r > 0, and (B(x0 , r))c
denotes the complement of B(x0 , r). If is a Young function such that 1 <
i() I() < q, then K is continuous in L ().
Proof: It is clear that the function can be substituted by an equivalent Young
function in the claim and at the same time we see that this cannot be done in
the assumption. Nevertheless, arguing as in the proof of Theorem 1.1, we find
an equivalent Young function whose Simonenko index q() is arbitrarily near

to I() and we apply Simonenkos theorem from [26] to K in L .
One can think about possible applications in the above described spirit to other
problems, too, since the growth condition (1.1) appears also in connections with
areas which we have not touched here. For instance, (1.1) plays a major role in
the recent paper by Assaoui [1] on Bessel potentials in Orlicz spaces.
Example 5.8. The following examples illustrate possible behaviour of Young
functions. The verification is a matter of simple calculation.
(1) Let (t)= e|t|3 if |t| <e and (t) = t4+sin log log t if |t| e (cf. [9]). Then
p() = 4 2, q() = 4 + 2, further, (5.4) holds with p = 3 and q = 5. At the
same time, using just the definition, it is likely extremely difficult to find the lower
and the upper indices. Further, the limit r does not exist so that Theorem 1.3
cannot be used. Nevertheless, invoking Theorem 1.2, we simply obtain that i()
and I() coincide
with p() and q(), respectively. Indeed, every truncation of
F by > 4 2 makes the integral in (4.1) infinite; similarly for the upper
index.
q
(2) Put (0) = 0 and (t) = |t|r exp( 1 + s log+ |t|) otherwise, r 1, s > 0
(Talenti [27]). Then i() = I() = r simply by Theorem 1.3. As to the Simonenko
indices we have p() = r and q() = r + s/2. Notice that q() depends on s
while I() does not.

449

450

A. Fiorenza, M. Krbec

Acknowledgment. This paper was finished when the first author visited Institute of Mathematics of the Czech Academy of Sciences in Prague. He would like
to thank for support and hospitality.
References
[1] Assaoui N., Bessel potentials in Orlicz spaces, preprint.
[2] Di Benedetto E., Trudinger N., Harnack inequalities for quasi-minima of variational integrals, Ann. Inst. H. Poincar
e Anal. Non Lin
eaire 1 (1984), 295308.
[3] Bennett C., Sharpley R., Interpolation of Operators, Academic Press, 1988.
[4] Boyd D.W., The Hilbert transform on rearrangement invariant spaces, Canad. J. Math.
19 (1967), 599616.
[5] Boyd D.W., Indices of function spaces and their relationship to interpolation, Canad.
J. Math. 21 (1969), 12451254.
[6] Boyd D.W., Indices for the Orlicz spaces, Pacific J. Math. 38 (1971), 315323.
[7] Fiorenza A., An inequality for Jensen means, Nonlinear Analysis, TMA 16 (1991), 191198.
[8] Fiorenza A., Krbec M., A formula for the Boyd indices in Orlicz spaces, to appear.
[9] Fusco N., Sbordone C., Higher integrability of the gradient of minimizers of functionals
with nonstandard growth conditions, Comm. Pure Appl. Math. 43 (1990), 673683.
[10] Giaquinta M., Giusti E., On the regularity of the minima of variational integrals, Acta
Math. 148 (1982), 3146.
[11] Giaquinta M., Modica G., Regularity results for some classes of higher order non linear
elliptic systems, J. Reine Angew. Math. 311/312 (1979), 145169.
[12] Gustavsson J., Peetre J., Interpolation of Orlicz spaces, Studia Math. 60 (1977), 3359.
[13] Kerman R., Torchinsky A., Integral inequalities with weights for the Hardy maximal function, Studia Math. 71 (1982), 277284.
[14] Kokilashvili V., Krbec M., Weighted inequalities in Lorentz and Orlicz spaces, World Scientific, Singapore-New Jersey-London-Hong Kong, 1991.
[15] Krasnoselski M.A., Rutitski Ya.B., Convex functions and Orlicz spaces, Noordhof, Groningen, 1961; English transl. from the first Russian edition Gos. Izd. Fiz. Mat. Lit., Moskva,
1958
[16] Maligranda L., Indices and interpolation, Dissert. Math. 234 (1984), 149.
[17] Matuszewska W., Orlicz W., On certain properties of functions, Bull. Acad. Polon. Sci.
8 (1960), 439443.
[18] Migliaccio L., Reverse H
older from reverse Jensen inequalities, in Methods of Real Analysis
and Partial Differential Equations, An International Workshop, Capri, Sept. 17-20, Liguori,
1990, pp. 129134.
[19] Montgomery-Smith S.J., Boyd indices of Orlicz-Lorentz spaces, in Function spaces (Edwardsville, IL, 1994), Lecture Notes in Pure and Appl. Math., 172 (Krzysztof Jarosz, ed.),
Marcel Dekker, Inc., New York, 1995, pp. 321334.
[20] Moscariello G., Regularity results for quasiminima of functionals with non-polynomial
growth, J. Math. Anal. Appl. 168 (1992), 500510.
[21] Muckenhoupt B., Weighted norm inequalities for the Hardy maximal function, Trans. Amer.
Math. Soc. 165 (1972), 207226.
[22] Persson L.E., Interpolation with a parameter function, Math. Scand. 59 (1986), 199222.
[23] Sbordone C., On some integral inequalities and their applications to the Calculus of Variations, Boll. Un. Mat. Ital., Ser. VI, Anal. Funz. Appl. V,C,1 (1986), 7394.
[24] Sbordone C., Quasiminima of degenerate functionals with nonpolynomial growth, Rend.
Sem. Mat. Fis. Milano 59 (1989), 173184.

Indices of Orlicz spaces and some applications

[25] Sbordone C., Maximal inequalities and applications to regularity problems in the calculus
of variations, to appear in Calculus of Variations, Applications and Computations, Ponta
`-Mousson 1994, Pitman Research Notes in Math. Series 326, 1995, 230244.
[26] Simonenko I.B., Interpolation and extrapolation in Orlicz spaces (in Russian), Mat. Sb.
(N. S.) 63 (1964), 536553.
[27] Talenti G., Boundedness of Minimizers, Hokkaido Math. J. 19 (1990), no. 2, 259279.
[28] Zippin M., Interpolation of operators of weak type between rearrangement invariant spaces,
J. Funct. Anal. 7 (1971), 267284.
Dipartimento di Matematica e Applicazioni R. Caccioppoli, via Cintia,
80126 Napoli, Italy
E-mail : fiorenza@matna2.dma.unina.it

25, 115 67 Prague 1,

Institute of Mathematics, Czech Academy of Sciences, Zitn
a
Czech Republic
E-mail : krbecm@matsrv.math.cas.cz
(Received October 24, 1996)

451