R E S E A R C H
Open Access
BMO and Lipschitz norm estimates for the
composition of Green’s operator and the
potential operator
Zhimin Dai
1, Yuming Xing
1*and Shusen Ding
2*Correspondence:
1Department of Mathematics,
Harbin Institute of Technology, Harbin, China
Full list of author information is available at the end of the article
Abstract
In this paper, we establish BMO and Lipschitz norm inequalities for the composition of Green’s operator and the potential operator. We also investigate the relationship among the Lipschitz norm, the BMO norm and theLp-norm. Finally, we display some examples for applications.
MSC: Primary 35J60; secondary 31B05; 58A10; 46E35
Keywords: differential forms; Lipschitz norm; BMO norm; Green’s operator; potential operator
1 Introduction
Differential forms are extensions of functions and can be used to describe various systems in partial differential equations (or PDEs), physics, theory of elasticity, quasiconformal analysis,etc. Differential forms have become invaluable tools for many fields of sciences and engineering; see [, ] for more details.
Now we introduce some notations and definitions. Letbe an open subset ofRn(n≥)
andObe a ball inRn. LetρOdenote the ball with the same center asOanddiam(ρO) =
ρdiam(O),ρ> . A weightw(x) is a nonnegative locally integrable function inRn.|D|is
used to denote the Lebesgue measure of a setD⊂Rn. Let∧=∧(Rn),= , , . . . ,n, be
the linear space of all-forms(x) =JJ(x)dxJ=
Jjj···j(x)dxj∧dxj· · ·∧dxjinR n,
whereJ= (j,j, . . . ,j), ≤j<j<· · ·<j≤n, are the ordered-tuples. Moreover, if each
of the coefficientJ(x) of(x) is differential on, then we call(x) a differential-form
onand useD(,∧) to denote the space of all differential-forms on.C∞(,∧)
de-notes the space of smooth-forms on. We denote the exterior derivative byd, and the Hodge codifferential operatordis defined asd= (–)n+d:D(,+)→D(,),
where is the Hodge star operator. For ≤p<∞, Lp(,∧) is a Banach space with
the norm p,= (
|(x)|
pdx)/p= ( (
J|J(x)|)p/dx)/p<∞. For a weightw(x), we
write p,,w= (
||pw(x)dx)/p. Similarly, the notationsL p
loc(,∧
) andW,p
loc(,∧
) are
self-explanatory.
From [], ifis a differential form in a bounded convex domain, then there is a de-composition
=d(T) +T(d), (.)
whereTis called a homotopy operator. For the homotopy operator, we know that
T p,O≤C|O|diam(O) p,O (.)
holds for any differential form∈Lploc(O,∧),= , , . . . ,n, <p<∞. Furthermore, we
can define the-form∈D(,∧) by
=
⎧ ⎨ ⎩
||–
(y)dy, = ;
d(T), = , . . . ,n (.)
for all∈Lp(,∧), ≤p<∞.
In this paper, we focus on a class of differential forms satisfying the well-known nonho-mogeneous A-harmonic equation
dA(x,d) =B(x,d), (.)
whereA:× ∧(Rn)→ ∧(Rn) andB:× ∧(Rn)→ ∧–(Rn) satisfy the conditions: |A(x,η)| ≤a|η|s–,A(x,η)·η≥ |η|sand|B(x,η)| ≤b|η|s–for almost everyx∈and all η∈ ∧(Rn). Herea,b> are some constants and <s<∞is a fixed exponent associated
with (.). A solution to (.) is an element of the Sobolev spaceWloc,s(,∧–) such that
A(x,d)·dψ+B(x,d)·ψ= (.)
for all ψ ∈Wloc,s(,∧–) with compact support. The various deformations of (.) are
shown in [].
Recently, Bi extended the definition of a potential operator to the set of all differential forms in []. For any differential -form(x) =JJ(x)dxJ, the potential operatorPis
defined by
P(x) =P J
J(x)dxJ
=
J
PJ(x)
dxJ= J
K(x,y)J(y)dy dxJ, (.)
where the kernelK(x,y) is a non-negative measurable function defined forx=y,J(x) is
defined on⊂Rn and the summation is over all ordered-tuplesJ. For more results related to the potential operatorP, see [–].
Green’s operator and the potential operator are of quite importance in the study of po-tential theory and nonlinear elasticity; see [, , , –] for more properties of these two operators. In many situations, the process of studying solutions of PDEs involves estimat-ing the various norms of the operators. However, the study on the composition of the potential operator and other operators is yet to be fully developed. Hence, we are mo-tivated to establish some norm inequalities for the composite operatorG◦Papplied to differential forms.
Let∈L loc(,∧
),= , , . . . ,n. We write∈locLip
k(,∧), ≤k≤, if locLipk,= sup
ρO⊂|
O|–(n+k)/n –O ,O<∞ (.)
for someρ≥. Further, we writeLipk(,∧) for those forms whose coefficients are in
the usual Lipschitz space with exponentkand write Lipk,for this norm. Similarly, for ∈L
loc(,∧
),= , , . . . ,n, we write∈BMO(,∧) if ,= sup
ρO⊂
|O|– –O ,O<∞ (.)
for someρ≥. Whenis a -form, equation (.) reduces to the classical definition of
BMO(). As to the definitions of the weighted Lipschitz and BMO norms, we will present them in Section .
The purpose of this paper is to derive the Lipschitz and BMO norm inequalities for the composition of Green’s operator Gand the potential operatorP applied to differential forms.
2 Estimates for Lipschitz and BMO norms
In this section, we establish the estimates for Lipschitz and BMO norms for the composite operatorG◦P. We need the following lemmas and definition.
The following inequality is the well-known Hölder inequality and gets proved with the Cauchy-Schwarz inequality in [].
Lemma . Let (,μ) be a measure space and Lp(μ) = Lp(,μ) ={f : ⊂ Rn →
C; f p,,μ<∞}be a Lebesgue space with the Lp-norm
f p,,μ=
⎧ ⎨ ⎩
( f pdμ)/p, ≤p<∞; ess supx∈|f(x)|, p=∞.
(.)
If p,q≥with/p+ /q= ,and if f ∈Lp(μ)and g∈Lq(μ),then fg∈L(μ)and
fg ,,μ≤ f p,,μ g q,,μ. (.)
Remark Ifμis a Lebesgue measure, that is,dμ=dx, then (.) reduces to the inequality
fg ,≤ f p, g q,. (.)
Lemma .[] Let∈D(,∧)be a solution to the nonhomogeneous A-harmonic
equa-tion(.)onandρ> be a constant.Then there exists a constant C,independent of,
such that
d p,O≤Cdiam(O)– –c p,ρO (.)
Lemma .[] Letbe a solution of the nonhomogeneous A-harmonic equation(.)in a domainand <s,t<∞.Then there exists a constant C,independent of,such that
s,O≤C|O|(t–s)/ts t,ρO (.)
for all balls O withρO⊂,whereρ> is a constant. The following definition is introduced in [].
Definition . A kernelKonRn×Rn(n≥) is said to satisfy the standard estimates if
there existα, <α≤, and a constantCsuch that for all distinct pointsxandyinRnand allzwith|x–z|<
|x–y|,
() K(x,y)≤C|x–y|–n; () K(x,y) –K(z,y)≤Cx–z
x–y
α|x–y|–n;
() K(z,y) –K(x,y)≤Cx–z x–y
α|x–y|–n. (.)
The followingLp-norm and Lipschitz norm inequalities for the compositionG◦P of
Green’s operator and the potential operator appear in [].
Lemma . Let∈Lp(,∧),= , , . . . ,n, <p<∞,be a differential form in a bounded convex domain⊂Rn,P be the potential operator defined in(.)with the kernel K(x,y)
satisfying the condition()of the standard estimates(.)and G be Green’s operator.Then there exists a constant C,independent of,such that
GP()–GP()Op,O≤C|O|diam(O) p,O (.)
for all balls O with O⊂.
Lemma . Let∈Lp(,∧),= , , . . . ,n, <p<∞,be a differential form in a bounded domain,P be the potential operator defined in(.)with the kernel K(x,y)satisfying the condition()of the standard estimates(.)and G be Green’s operator.Then there exists a constant C,independent of,such that
GP()locLip
k,≤C p,, (.)
where k is a constant with≤k≤.
Lemma .[] Letϕbe a strictly increasing convex function on[,∞)withϕ() = (that is,ϕis a Young function),and D be a bounded domain inRn.Assume thatis a smooth differential form in D such thatϕ(k(||+|D|))∈L(D;μ)for any real number k> and
μ({x∈D:|–D|> }) > ,whereμis a Radon measure defined by dμ=w(x)dx for a weight w(x).Then,for any positive constant a,we have
D
ϕa||dμ≤C
D
ϕa|–D|
dμ, (.)
Using Lemma . withϕ(t) =tpandw(x) = over the ballO, we obtain
p,O≤C –O p,O, (.)
whereCis a constant.
Theorem . Let∈Lp(,∧),= , , . . . ,n, <p<∞,be a solution of the nonhomo-geneous A-harmonic equation (.)in a bounded convex domain, P be the potential operator defined in(.)with the kernel K(x,y)satisfying the condition()of the standard estimates(.)and G be Green’s operator.Then there exists a constant C,independent of,
such that
GP()locLip
k,≤C ,, (.)
where k is a constant with≤k≤.
Proof From Lemma . and (.), we obtain
GP()–GP()Op,O≤C|O|diam(O) p,O
≤C|O|diam(O) –O p,O. (.)
From the decomposition (.), (.) and (.), we have
–O p,O= Td p,O≤C|O|diam(O) d p,O≤C|O||O|/n d p,O. (.)
Combining (.) with (.) yields
GP()–GP()Op,O≤C|O|diam(O) –O p,O
≤C|O|diam(O)
C|O||O|/n d p,O
≤C|O|+/ndiam(O) d p,O
≤C|O|+/n d p,O. (.)
Using the definition of the Lipschitz norm, (.) with = /p+ (p– )/pand (.), for any ballOwithO⊂, it follows that
GP()–GP()O,O=
O
GP()–GP()Odx
≤
O
G
P()–GP()Opdx
/p
O
p p–dx
(p–)/p
=|O|(p–)/pGP()–GP()Op,O
≤ |O|–/pC|O|+/n d p,O
From Lemma ., we have
d p,O≤Cdiam(O)– –c p,ρO≤C|O|–/n –c p,ρO (.)
for any closed formcand any ballOwithρO⊂, whereρ> is a constant.
Sinceis a solution of equation (.) andcis a closed form,–cis also a solution of equation (.). By Lemma ., we obtain
–c p,ρO≤C|O|(–p)/p –c ,ρO (.)
for some constantρ>ρ> withρO⊂.
Combining (.), (.) and (.), we obtain
GP()–GP()O,O≤C|O|–/p+/n d p,O ≤C|O|–/p+/n
C|O|–/n –c p,ρO
≤C|O|–/p+/n –c p,ρO ≤C|O|–/p+/n
C|O|(–p)/p –c ,ρO
≤C|O|–/p+/n+/p– –c ,ρO
=C|O|+/n –c ,ρO (.)
for any closed formc.
Sincecis any closed form in (.), we may choosec=ρOin (.). By the definitions
of the Lipschitz norms, and noticing ≤k≤, we find that
GP()locLip
k,=ρOsup⊂|O|
–(n+k)/nGP()–GP() O,O
≤ sup
ρO⊂
|O|––k/nC|O|+/n –ρO ,ρO
=C sup
ρO⊂|
O|+/n–k/n –ρO ,ρO ≤C sup
ρO⊂
|O|+/n–k/n|ρO|– –ρO ,ρO ≤C sup
ρO⊂|
|+/n–k/n|ρO|– –ρO ,ρO ≤C||+/n–k/n sup
ρO⊂
|ρO|– –ρO ,ρO
≤C ,, (.)
whereρ>ρ>ρwithρO⊂.
The proof of Theorem . has been completed. We have developed some estimates for the Lipschitz norm · locLipk,. Now, we establish
Lemma .[] If a differential form∈locLipk(,∧), = , , . . . ,n, ≤k≤,in a bounded convex domain,then∈BMO(,∧)and
,≤C locLipk,, (.)
where C is a constant.
SinceG(P()) is a differential form whenis a differential form, we have the following theorem.
Theorem . If a differential form G(P())∈locLipk(,∧),= , . . . ,n, ≤k≤,in a bounded convex domain,then there exists a constant C,independent of,such that G(P())∈BMO(,∧)and
GP(),≤CGP()locLipk,, (.)
where the definitions of G and P are the same as in the preceding theorem.
Based on the above results, we estimate the BMO norm · ,of compositionG◦Pin
terms ofLpnorm.
Theorem . Let∈Lp(,∧), <p<∞,be a differential form in a bounded convex domain,G be Green’s operator and P be the potential operator defined in equation(.)
with the kernel K(x,y)satisfying the condition()of the standard estimates(.).Then there exists a constant C,independent of,such that
GP(),≤C p,. (.)
Proof From Lemma ., we have
GP()locLip
k,≤C p,. (.)
Using Theorem . and (.), it follows that
G
P(),≤CG
P()locLip
k,≤C p,. (.)
The proof of Theorem . has been completed. Similar to the proof of Theorem ., using Theorems . and ., we can prove the following theorem.
Theorem . Let ∈ Lp(,∧), <p <∞, be a solution of the nonhomogeneous
A-harmonic equation (.)in a bounded convex domain,G be Green’s operator and P be the potential operator defined in equation(.)with the kernel K(x,y)satisfying the condition()of the standard estimates(.).Then there exists a constant C,independent of,such that
3 Two weight estimates
In this section, we discuss the weighted Lipschitz and BMO norms []. For∈L loc(,
∧,wα),= , , . . . ,n, we write∈locLip
k(,∧,wα), ≤k≤, if locLipk,,wα= sup
ρO⊂
μ(O)–(n+k)/n –O ,O,wα<∞ (.)
for someρ> , whereis a bounded domain. The measureμis defined bydμ=w(x)αdx, wis a weight andαis a real number. For convenience, we will write the simple notation
locLipk(,∧) for locLip
k(,∧,wα). Similarly, for∈Lloc(,∧
,wα),= , , . . . ,n, we
write∈BMO(,∧,wα) if ,,wα= sup
ρO⊂
μ(O)– –O ,O,wα<∞ (.)
for someρ> , where the measureμis defined bydμ=w(x)αdx,wis a weight andαis a
real number. Again, we will writeBMO(,∧) to replaceBMO(,∧,wα) when it is clear
that the integral is weighted.
Definition .[] A pair of weights (w(x),w(x)) satisfies theAr,λ()-condition in a set
⊂Rn. Write (w(x),w(x))∈Ar,λ() for someλ≥ and <r<∞with r+r = if
sup O⊂
|O|
O wλ
dx
λr
|O|
O
w λr
r
dx
λr
<∞. (.)
Lemma .[] Let∈Lp(,∧,ν),= , . . . ,n, <p<∞,be a solution of the nonho-mogeneous A-harmonic equation(.)in a bounded convex domain,P be the potential operator defined in(.)with the kernel K(x,y)satisfying the condition()of the standard estimates(.)and G be Green’s operator.Assume that(w(x),w(x))∈Ar,λ()for some
λ≥and <r<∞.Then there exists a constant C,independent of,such that
GP()–GP()Op,O,wα
≤C|O|diam(O) p,ρO,w α
(.)
for all balls O withρO⊂,whereρ> andαare two constants with≤α<λ.
Theorem . Let∈Lp(,∧,ν),= , . . . ,n, <p<∞,be a solution of the nonhomo-geneous A-harmonic equation (.)in a bounded convex domain, P be the potential operator defined in(.)with the kernel K(x,y)satisfying the condition()of the standard estimates(.)and G be Green’s operator.The measuresμandνare defined by dμ=wα
dx,
dν=wα
dx,and(w(x),w(x))∈Ar,λ()for someλ≥and <r<∞with w(x)≥> for
any x∈.Then there exists a constant C,independent of,such that
GP()locLip
k,,wα ≤C p,,w α
, (.)
where k andαare constants with≤k≤and <α≤.
Proof Sinceμ(O) =Owα
dx≥
O
αdx=C
|O|, we have
for any ballO. Using (.) and Lemma . with = /p+ (p– )/p, it follows that
GP()–GP()O,O,wα
=
O
GP()–GP()Odμ
≤
O
GP()–GP()Opdμ /p
O
p/(p–)dμ (p–)/p
=μ(O)(p–)/pGP()–GP()Op,O,wα
=μ(O)–/pGP()–GP()Op,O,wα
≤μ(O)–/pC|O|diam(O) p,ρO,wα
≤C
μ(O)–/p|O|+/n p,ρO,wα
. (.)
Notice that – /p+ /n–k/n> and||<∞, from (.), (.) and (.), we obtain
GP()locLip
k,,wα = sup ρO⊂
μ(O)–(n+k)/nGP()–GP()O,O,wα
≤ sup
ρO⊂
μ(O)––k/nC
μ(O)–/p|O|+/n p,ρO,wα
=C sup
ρO⊂
μ(O)–/p–k/n|O|+/n p,ρO,wα
≤C sup
ρO⊂
C|O|– /p+k/n
|O|+/n p,ρO,wα
≤C sup
ρO⊂
|O|–/p+/n–k/n p,ρO,wα
≤C sup
ρO⊂|
|–/p+/n–k/n p,ρO,wα
≤C||–/p+/n–k/n sup
ρO⊂
p,ρO,wα
≤C p,,wα (.)
The proof of Theorem . has been completed. We now estimate the · ,,wα norm in terms of theLp-norm.
Theorem . Let∈Lp(,∧,ν),= , . . . ,n, <p<∞,be a solution of the nonhomo-geneous A-harmonic equation (.)in a bounded convex domain, P be the potential operator defined in(.)with the kernel K(x,y)satisfying the condition()of the standard estimates(.)and G be Green’s operator.The measuresμandνare defined by dμ=wαdx,
dν=wα
dx,and(w(x),w(x))∈Ar,λ()for someλ≥and <r<∞with w(x)≥> for
any x∈.Then there exists a constant C,independent of,such that
GP(),,wα
≤C p,,w α
, (.)
Proof From (.) and (.), it follows that
,,wα = sup
ρO⊂
μ(O)– –O ,O,wα
= sup ρO⊂
μ(O)k/nμ(O)–(n+k)/n –O ,O,wα
≤ sup
ρO⊂
μ()k/nμ(O)–(n+k)/n –O ,O,wα ≤μ()k/n sup
ρO⊂
μ(O)–(n+k)/n –O ,O,wα
≤C sup
ρO⊂
μ(O)–(n+k)/n –O ,O,wα ≤C locLipk,,wα
, (.)
whereCis a positive constant. ReplacingbyG(P()) in (.), we find that GP(),,wα
≤C
GP()locLipk,,wα
, (.)
wherekis a constant with ≤k≤. From Theorem ., we obtain
GP()locLip
k,,wα ≤C p,,w α
, (.)
Substituting (.) into (.), we have
GP(),,wα
≤C p,,w α
. (.)
We have completed the proof of Theorem ..
4 Applications
If we chooseA,Bto be a special operator, for example,A(x,d) =d|d|s–,B= , then (.) reduces to the followings-harmonic equation:
dd|d|s–= . (.)
In particular, we may lets= , then (.) reduces to
d(d) = . (.)
We may make use of the following two specific examples to conform the convenience of the inequality (.) in evaluating the upper bound for the BMO norm ofG(P()). Ob-viously, we may take advantage of (.) to make this estimating process easy, without calculating G(P()) ,in a complicated way.
Example . Let = (x +x + xx+ xx +xx)dx + (x +x+ xx + xx+
xx)dx+ (x+x+ xx+ xx+ xx)dx,={x= (x,x,x) :|≤x+x+x< },
P be the potential operator defined in (.) with the kernelK(x,y) satisfying the condi-tion () of the standard estimates (.) andGbe Green’s operator.
First, by simple computation, we have
d=xdx∧dx+ xdx∧dx+xdx∧dx, (.) (d) =xdx– xdx+xdx,d
(d)= . (.)
Sinced= (–)×+d= –d,
d(d) = –d(d)= . (.)
This implies thatsatisfies (.). Observe that
p, =
||pdx
/p
=
x+x+ xx+ xx+xx
+x+x+ xx+ xx+ xx
+ x+x+ xx+ xx+ xx p/
dx
/p
≤
( + + + + )
+ ( + + + + ) + ( + + + + )p/dx
/p
≤√||/p
=√
π
/p
. (.)
Applying (.), we obtain
GP(),≤C p,≤C √
π
/p
. (.)
Example . Let us assume, in addition to the definitions ofGandPof Example .,
=exsinx
Similarly, to begin with, we observe that
∂ ∂x
=∂
∂x
=exsinx, (.) ∂
∂x
=excosx,
∂ ∂x = –e
xsinx
. (.)
Thus,
= , (.)
which implies the functionis harmonic. Observe that
p,=
||pdx
/p
=
π
dθ
ercosθsin(rsinθ)pr dr
/p
≤e(π·/)/p
=π/pe. (.)
Applying (.), we obtain
G
P(),≤C p,≤Cπ/pe. (.)
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
ZD finished the proof and the writing work. YX gave ZD some excellent advice on the proof and writing. SD gave ZD lots of help in revising the paper. All authors read and approved the final manuscript.
Author details
1Department of Mathematics, Harbin Institute of Technology, Harbin, China.2Department of Mathematics, Seattle
University, Seattle, WA 98122, USA.
Acknowledgements
The authors wish to thank the anonymous referees for their time and thoughtful suggestions.
Received: 23 October 2012 Accepted: 8 January 2013 Published: 22 January 2013
References
1. Agarwal, RP, Ding, S, Nolder, CA: Inequalities for Differential Forms. Springer, New York (2009)
2. Ding, S: Norm estimates for the maximal operator and Green’s operator. Dyn. Contin. Discrete Impuls. Syst., Ser. A Math. Anal.16, 72-78 (2009)
3. Iwaniec, T, Lutoborski, A: Integral estimates for null Lagrangians. Arch. Ration. Mech. Anal.125, 25-79 (1993) 4. Bi, H: Weighted inequalities for potential operators on differential forms. J. Inequal. Appl.2010, Article ID 713625
(2010)
5. Martell, JM: Fractional integrals, potential operators and two-weight, weak type norm inequalities on spaces of homogeneous type. J. Math. Anal. Appl.294, 223-236 (2004)
6. Uribe, DC, Pérez, C: Two-weight, weak-type norm inequalities for fractional integrals, Calderón-Zygmund operators and commutators. Indiana Univ. Math. J.49(2), 697-721 (2000)
9. Agarwal, RP, Ding, S: Inequalities for Green’s operator applied to the minimizers. J. Inequal. Appl.2011, Article ID 66 (2011)
10. Dai, Z, Xing, Y, Ding, S, Wang, Y: Inequalities for the composition of Green’s operator and the potential operator. J. Inequal. Appl.2012, Article ID 271 (2012)
11. Nolder, CA: Hardy-Littlewood theorems for A-harmonic tensors. Ill. J. Math.43, 613-631 (1999)
12. Li, Y, Shaw, S: A proof of Hölder’s inequality using the Cauchy-Schwarz inequality. J. Inequal. Pure Appl. Math.7(2), Article ID 62 (2006)
13. Xing, Y, Ding, S: Norm comparison inequalities for the composite operator. J. Inequal. Appl.2009, Article ID 212915 (2009)
14. Gilbarg, D, Trudinger, NS: Elliptic Partial Differential Equations of Second Order. Springer, Berlin (1997)
doi:10.1186/1029-242X-2013-26