Citation: Kazunobu Baba, Tadaaki Miyazaki. Critical function of Siah2 in tumorigenesis[J]. AIMS Molecular Science, 2017, 4(4): 415-423. doi: 10.3934/molsci.2017.4.415
[1] | Patrizia Pucci, Letizia Temperini . On the concentration–compactness principle for Folland–Stein spaces and for fractional horizontal Sobolev spaces. Mathematics in Engineering, 2023, 5(1): 1-21. doi: 10.3934/mine.2023007 |
[2] | Ko-Shin Chen, Cyrill Muratov, Xiaodong Yan . Layered solutions for a nonlocal Ginzburg-Landau model with periodic modulation. Mathematics in Engineering, 2023, 5(5): 1-52. doi: 10.3934/mine.2023090 |
[3] | Boumediene Abdellaoui, Ireneo Peral, Ana Primo . A note on the Fujita exponent in fractional heat equation involving the Hardy potential. Mathematics in Engineering, 2020, 2(4): 639-656. doi: 10.3934/mine.2020029 |
[4] | María Ángeles García-Ferrero, Angkana Rüland . Strong unique continuation for the higher order fractional Laplacian. Mathematics in Engineering, 2019, 1(4): 715-774. doi: 10.3934/mine.2019.4.715 |
[5] | Luigi Montoro, Berardino Sciunzi . Qualitative properties of solutions to the Dirichlet problem for a Laplace equation involving the Hardy potential with possibly boundary singularity. Mathematics in Engineering, 2023, 5(1): 1-16. doi: 10.3934/mine.2023017 |
[6] | Huyuan Chen, Laurent Véron . Weak solutions of semilinear elliptic equations with Leray-Hardy potentials and measure data. Mathematics in Engineering, 2019, 1(3): 391-418. doi: 10.3934/mine.2019.3.391 |
[7] | Marco Cirant, Kevin R. Payne . Comparison principles for viscosity solutions of elliptic branches of fully nonlinear equations independent of the gradient. Mathematics in Engineering, 2021, 3(4): 1-45. doi: 10.3934/mine.2021030 |
[8] | Edgard A. Pimentel, Miguel Walker . Potential estimates for fully nonlinear elliptic equations with bounded ingredients. Mathematics in Engineering, 2023, 5(3): 1-16. doi: 10.3934/mine.2023063 |
[9] | Antonio Vitolo . Singular elliptic equations with directional diffusion. Mathematics in Engineering, 2021, 3(3): 1-16. doi: 10.3934/mine.2021027 |
[10] | David Arcoya, Lucio Boccardo, Luigi Orsina . Hardy potential versus lower order terms in Dirichlet problems: regularizing effects. Mathematics in Engineering, 2023, 5(1): 1-14. doi: 10.3934/mine.2023004 |
This work is dedicated to Ireneo in memoriam, for his teachings, friendship and enthusiasm.
This work deals with the following problem:
{(−Δ)su=λu|x|2s+(F(u)(x))p+ρf in Ω,u>0 in Ω,u=0 in (RN∖Ω), | (1) |
where λ>0, ρ>0, s∈(0,1), 2s<N, 1<p<∞, Ω⊂RN is a bounded regular domain containing the origin and f is a measurable non-negative function satisfying suitable hypotheses.
By (−Δ)s we denote the fractional Laplacian of order 2s introduced by M. Riesz in [39], that is,
(−Δ)su(x):=aN,s P.V. ∫RNu(x)−u(y)|x−y|N+2sdy,s∈(0,1), |
where
aN,s=22s−1π−N2Γ(N+2s2)|Γ(−s)|, |
is the normalizing constant that gives the Fourier multiplier identity
F((−Δ)su)(ξ)=|ξ|2sF(u)(ξ), for u∈S(RN). |
See [26] for details. Our goal is to find natural conditions on p and f (related to the value of λ), in order to get the existence of positive solutions.
If λ=0, the problem (1) can be seen as a Kardar-Parisi-Zhang stationary equation with fractional diffusion and nonlocal gradient term. We refer to [30] for the main model and additional properties of the local case.
The nonlocal case s∈(0,1), but still with the local gradient term, was used recently in order to describe the growing surface in the presence of self-similar hopping surface diffusion. We refer the reader to the papers [29,32,33,35] for a physical rigorous justification.
Existence results for the corresponding problem were obtained in [23] and [11] under suitable hypotheses on f and p. As it was shown in [11], if p>11−s, then the corresponding problem does not have positive solutions with global regularity of the gradient, even in the case of regular datum f. Existence of a solution, in the viscosity sense, is proved in [9,16,17] for some particular cases.
The case λ=0, under the presence of a nonlocal gradient term, was analyzed recently in [7]. Without any limitation on the value of p and under suitable hypothesis of f, the author proved the existence of a solution using a priori estimates and fixed point arguments.
The case λ>0 with a local gradient term was considered in [10] and [12]. Here the authors showed the existence of a critical exponent related to the existence of solutions. Our work can be seen as the non-local counterpart of [12]. However, the non-local gradient term makes the problem more difficult and fine analysis is needed to determine the existence or non-existence scheme.
Notice that for λ>0, problem (1) is related to the Hardy inequality proved in [28], (see also [18] and [38] for equivalent forms.) Namely, for ϕ∈C∞0(IRN), we have
∫IRN|ξ|2s|ˆϕ|2dξ≥ΛN,s∫IRN|x|−2sϕ2dx, | (2) |
where
ΛN,s:=22sΓ2(N+2s4)Γ2(N−2s4) | (3) |
is optimal and not attained.
It is clear that
lims→1ΛN,s=(N−22)2, |
the Hardy constant in the local case.
Inequality (2) can be also formulated in the following way
aN,s2∫IRN∫IRN|ϕ(x)−ϕ(y)|2|x−y|N+2sdxdy⩾ΛN,s∫IRNϕ2|x|2sdx,∀ϕ∈C∞0(IRN). |
If λ>ΛN,s, then we can prove that problem (1) has no positive supersolution. Hence, we assume throughout this paper that 0<λ<ΛN,s.
The presence of the Hardy potential forces the solution to enjoy a singular behavior near the singular point zero and then a loss of regularity is generated.
The paper is organized as follows. In Section 2 we present the functional setting used in order to study our problem. More precisely we describe some related spaces, as the Bessel potential space, and their relationship with the fractional Sobolev space. We introduce also the different forms of the fractional gradient that will be used throughout the paper. In Subsection 2.1 we recall the global regularity results for the Poisson fractional problem proved in [6]. This will be the key in order to show the fractional regularity in our problem.
The analysis of the problem under the presence of the Hardy potential, without the nonlocal gradient term, is considered in Section 3. More precisely, we will consider the semilinear problem
{(−Δ)su=λu|x|2s+f in Ω,u=0 in RN∖Ω, |
where f∈Lm(Ω) with m⩾1. Some partial regularity results are known in the case where λ<Js,m≡ΛN,s4N(m−1)(N−2ms)m2(N−2s)2.
However for Js,m⩽λ⩽ΛN,s, using a different approach based on weighted spaces, we are able to complete the full picture of regularity. As a consequence, we get a complete classification of the fractional regularity of the solution to the above problem.
The first analysis of the KPZ problem (1) is done in Section 4. We begin by considering the case where F(u)(x)=|(−Δ)s2u(x)|. Using suitable radial computations in the whole space, we derive the existence of a critical exponent p+(λ,s) such that if p>p+(λ,s), then for all ρ>0, the problem (1) has no positive solution in a weak sense. Some other non existence results are proved for ρ large under technical condition on p.
The case p<p+(λ,s) is analyzed in Subsection 4.2. Under the hypothesis that f is bounded, we are able to show the existence of a supersolution for ρ small. Moreover, for p<NN−s, and for all f∈L1(Ω) that satisfy a suitable integral condition near the origin, we are able to show the existence of a weak solution for ρ<ρ∗.
In Section 5 we treat the KPZ problem, namely equation (1), under the presence of another version of the non local gradient.
More precisely, we consider the case where F(u)(x)=(aN,s2∫IRN|u(x)−u(y)|2|x−y|N+2sdy)12. Then, also in this case, we are able to show the existence of a critical exponent p+(λ,s) such that non existence holds if p>p+(λ). The proof of the non existence in this case is more technical and need some additional estimates.
Finally, at the end of the section we formulate some interesting open problems that may describe a full picture for the existence in our problem.
The goal of this section is to establish some useful tools and definitions that will play an important role in what follows.
Definition 2.1. Let Ω⊂IRN be a bounded domain and s∈(0,1). For p∈[1,∞), the fractional Sobolev space Ws,p(Ω) is defined by
Ws,p(Ω):={u∈Lp(Ω):∬Ω×Ω|u(x)−u(y)|p|x−y|N+spdxdy<∞}. |
Ws,p(Ω) is a Banach space endowed with the norm
‖u‖Ws,p(Ω):=(‖u‖pLp(Ω)+∬Ω×Ω|u(x)−u(y)|p|x−y|N+spdxdy)1p. |
The space Ws,p0(Ω) is defined as follows:
Ws,p0(Ω):={u∈Ws,p(IRN):u=0 in IRN∖Ω}. |
This is a Banach space endowed with the norm
‖u‖Ws,p0(Ω):=(∬DΩ|u(x)−u(y)|p|x−y|N+spdxdy)1/p, |
where
DΩ:=(IRN×IRN)∖(CΩ×CΩ)=(Ω×IRN)∪(CΩ×Ω). |
Now, for s∈(0,1) and 1≤p<+∞ we define the Bessel potential space by setting
Ls,p(RN) := ¯{u∈C∞0(RN)}||.||Ls,p(RN), |
where
||u||Ls,p(RN)=‖(1−Δ)s2u‖Lp(RN) and (1−Δ)s2u=F−1((1+|⋅|2)s2Fu),∀ u∈C∞c(RN). |
Let us stress that, in the case where s∈(0,1) and 1<p<+∞,
‖u‖Ls,p(RN):=‖u‖Lp(RN)+‖(−Δ)s2u‖Lp(RN) |
is an equivalent norm for Ls,p(RN) (see e.g., [1,page 5] for a precise explanation of this fact). Let us as well recall that, for all 0<ϵ<s<1 and all 1<p<+∞, by [13,Theorem 7.63, (g)], we have
Ls+ϵ,p(RN)⊂Ws,p(RN)⊂Ls−ϵ,p(RN). |
For ϕ∈C∞0(IRN) we define the fractional gradient of order s of ϕ by
∇sϕ(x):=∫IRNϕ(x)−ϕ(y)|x−y|sx−y|x−y|dy|x−y|N,∀ x∈IRN. | (4) |
Notice that, as it was proved in [46,Theorem 2] and [42,Theorem 1.7], we have
Ls,p(IRN):={u∈Lp(IRN) such that |∇su|∈Lp(IRN)}={u∈Lp(IRN) such that |(−Δ)s2u|∈Lp(IRN)} |
with the equivalent norms
‖|u‖|Ls,p(IRN):=‖u‖Lp(IRN)+‖∇su‖Lp(IRN)≃‖u‖Lp(IRN)+‖(−Δ)s2u‖Lp(IRN). |
Another type of "nonlocal gradient" can be defined also by
Ds(u)(x)=(aN,s2∫IRN|u(x)−u(y)|2|x−y|N+2sdy)12. | (5) |
We refer to [20] and [36] for some motivation of this non local version of the gradient.
In this case one has
lims→1−(1−s)D2s(u(x))=|∇u(x)|2,∀ u∈C∞0(RN). | (6) |
If p>2NN+2s, it was proved in [46] that the Bessel potential space Ls,p(IRN) can be defined also as the set of functions u∈Lp(IRN) such that Ds(u)∈Lp(IRN). The space Ls,p(IRN) can be equipped with the equivalent norms
|||u|||Ls,p(IRN)=‖u‖Lp(IRN)+‖Ds(u)‖Lp(IRN). |
The next Sobolev inequality in Ls,p(IRN) is proved in [13], see also [25].
Theorem 2.2. Let 1<p<∞ and s∈(0,1) be such that sp<N. Then there exist two positive constants S1:=S2(N,p,s) and S2:=S1(N,p,s) such that for all u∈Ls,p(IRN), we have
S1||u||Lp∗s(IRN)⩽‖∇su‖Lp(IRN), |
and
S2||u||Lp∗s(IRN)⩽‖(−Δ)s2u‖Lp(IRN), |
with p∗s=pNN−ps.
If Ω⊂IRN, we define the space Ls,p0(Ω) as the set of functions u∈Ls,p(IRN) with u=0 in IRN∖Ω.
From Lemma 1 in [46], if p>2NN+2s and Ω is a bounded domain, then there exist C1:=C1(Ω,N,p,s) and C2:=C2(Ω,N,p,s), two positive constants, such that for all u∈Ls,p0(Ω)
C1‖|u‖|Ls,p(IRN)⩽‖Ds(u)‖Lp(IRN)≤C2‖|u‖|Ls,p(IRN). |
Notice that if Ω is a bounded domain, we can endow Ls,p0(Ω) with the equivalent norms ‖∇su‖Lp(IRN) or ‖(−Δ)s2u‖Lp(IRN). In the same way, by assuming in addition that p>2NN+2s, then we can equip Ls,p0(Ω) also with the equivalent norms ‖Ds(u)‖Lp(IRN). We refer to [47] for more details about the properties of the Bessel potential space and its relation with the fractional Sobolev space.
The next Hardy inequality will be useful in order to prove the non existence result above the critical exponent. See [7] for the proof.
Proposition 2.3. Let Ω⊂RN be a regular domain with 0∈Ω and 0<s<1. Suppose that p>2NN+2s with ps<N and define
L(Ω):=inf{∫IRN(Ds(ϕ)(x))pdx∫Ω|ϕ(x)|p|x|psdx:ϕ∈C∞0(Ω)∖{0}}. | (7) |
Then L(Ω)>0 and L(Ω)=L does not depends on Ω. Moreover, the weight |x|−ps is optimal in the sense that, for all ε>0 we have
inf{∫IRN(Ds(ϕ)(x))pdx∫Ω|ϕ(x)|p|x|ps+εdx:ϕ∈C∞0(Ω)∖{0}}=0. |
Finally, we recall the next standard result from harmonic analysis. See for instance [45,Theorem I, Section 1.2, Chapter V].
Theorem 2.4. Let 0<ν<N and 1≤p<ℓ<∞ be such that 1ℓ+1=1p+νN. For g∈Lp(IRN), we define
Jν(g)(x)=∫IRNg(y)|x−y|νdy. |
Then, it follows that:
a) Jν is well defined in the sense that the integralconverges absolutely for almost all x∈RN.
b) If p>1, then ‖Jν(g)‖Lℓ(IRN)≤cp,l‖g‖Lp(IRN). c) If p=1, then |{x∈RN|Jν(g)(x)>σ}|⩽(A‖g‖L1(IRN)σ)ℓ.
The goal of this section is to state some well known results about the regularity of the Poisson equation
{(−Δ)su=g in Ω,u=0 in RN∖Ω, | (8) |
where Ω is a bounded regular domain of IRN and g∈Lm(Ω) with m⩾1. We begin by the sense for which solutions are defined.
Definition 2.5. We define the class of test functions
T(Ω)={ϕ|(−Δ)s(ϕ)=ψ in Ω,ϕ=0 in RN∖Ω,ψ∈C∞0(Ω)}. | (9) |
Notice that if v∈T(Ω) then, using the results in [34], v∈Hs0(Ω)∩L∞(Ω). Moreover, according to the regularity theory developed in [43], if Ω is smooth enough, there exists a constant β>0 (that depends only on the structural constants) such that v∈Cβ(Ω) (see also [31]).
Definition 2.6. We say that u∈L1(Ω) is a weak solution to (8) if for g∈L1(Ω) we have that
∫Ωuψdx=∫Ωgϕdx, |
for any ϕ∈T(Ω) with ψ∈C∞0(Ω).
Recall also the definition of the truncation operator Tk,
Tk(σ)=max{−k;min{k,σ}} and Gk(σ)=σ−Tk(σ). | (10) |
From [2,22,34] we have the next existence result.
Theorem 2.7. Suppose that g∈L1(Ω), then problem (8) has a unique weak solution u obtained as the limit of {un}n∈N, the sequence of unique solutions to the approximating problems
{(−Δ)sun=gn(x) in Ω,un=0 in IRN∖Ω, | (11) |
with gn=Tn(g). Moreover,
Tk(un)→Tk(u) strongly in Hs0(Ω),∀k>0, | (12) |
u∈Lq(Ω),∀ q∈[1,NN−2s) | (13) |
and
|(−Δ)s2u|∈Lr(Ω),∀ r∈[1,NN−s). | (14) |
In addition, if s>12, then u∈W1,q0(Ω) for all 1≤q<NN−(2s−1) and un→u strongly in W1,q0(Ω).
In what follows we denote Gs the Green function associated to the fractional laplacian (−Δ)s.
Notice that Gs(x,y) solves the problem
{(−Δ)syGs(x,y)=δx(y) if y∈Ω,Gs(x,y)=0 if y∈IRN∖Ω, | (15) |
where x∈Ω is fixed and δx is Dirac's delta function.
It is clear that if u is the unique weak solution to problem (8), then
u(x)=∫ΩGs(x,y)g(y)dy. |
We collect in the next Proposition some useful properties of the Green function Gs (See [21] and [19] for the proof).
Proposition 2.8. Assume that s∈(0,1). Then, for almost every x, y ∈Ω, we have
Gs(x,y)≃1|x−y|N−2s(δs(x)|x−y|s∧1)(δs(y)|x−y|s∧1)≃1|x−y|N−2s(δs(x)δs(y)|x−y|2s∧1). | (16) |
In particular, we have
Gs(x,y)≤C1min{1|x−y|N−2s,δs(x)|x−y|N−s,δs(y)|x−y|N−s} for a.e. x,y∈Ω. | (17) |
In the case where g∈Lm(Ω), we can improve the regularity results of Theorem 2.7. More precisely from [11], we have the next theorem.
Theorem 2.9. Assume that g∈Lm(Ω) with m>1 and let u be the uniquesolution to problem (8), then there exists a positiveconstant C:=C(N,s,m,Ω) (that can change from a line to anotherone), such that
1) If 1<m<N2s, then u∈LmNN−2ms(Ω), uδs∈LmNN−ms(Ω) and
||u||LmNN−2ms(Ω)+||uδs||LmNN−ms(Ω)≤C||g||Lm(Ω). |
2) If m=N2s, then u∈Lr(Ω) for all r<∞, uδs∈LmNN−ms(Ω) and
||u||Lr(Ω)+||uδs||LmNN−ms(Ω)≤C||g||Lm(Ω). |
3) If N2s<m<Ns, then u∈L∞(Ω), uδs∈LmNN−ms(Ω) and
||u||L∞(Ω)+||uδs||LmNN−ms(Ω)≤C||g||Lm(Ω). |
4) If m=Ns, then u∈L∞(Ω), uδs∈Lp(Ω) for all p<∞ and
||u||L∞(Ω)+||uδs||Lp(Ω)≤C||g||Lm(Ω). |
5) If m>Ns, then u∈L∞(Ω), uδs∈L∞(Ω) and
||u||L∞(Ω)+||uδs||L∞(Ω)≤C||g||Lm(Ω). |
Related to the fractional regularity of the solution to problem (8), a global fractional Calderon-Zygmund regularity result was obtained recently in [6].
Theorem 2.10. Let s∈(0,1) and consider u to be the (unique) weak solution to problem (8) with f∈Lm(Ω). Then we have
1) If m⩾Ns, then for all 1≤p<∞, there exists a positive constant C=C(N,s,p,m,Ω) such that
‖(−Δ)s2u‖Lp(RN)≤C‖g‖Lm(Ω). |
Moreover u∈Ls,p(RN) for all 1≤p<∞ and
‖u‖Ls,p(RN)≤C‖g‖Lm(Ω). |
2) 1≤m<Ns, then, for all 1≤p<mNN−ms, there exists a positive constant C=C(N,s,p,m,Ω) such that
‖(−Δ)s2u‖Lp(RN)≤C‖g‖Lm(Ω). |
Hence u∈Ls,p(RN) for all 1≤p<mNN−msand
‖u‖Ls,p(RN)≤C‖g‖Lm(Ω). |
As a direct consequence of the relation between the fractional Sobolev space Ws,p(RN) and the Bessel potential space Ls,p(RN), we get the next result.
Corollary 2.11. Let s∈(0,1). Consider u to be the unique solution of problem (8) with g∈Lm(Ω). Then
1) If 1≤m<Ns,
we have, for all 1<p<mNN−ms, that there exists C=C(N,s,m,p,Ω) such that
‖u‖Ws,p(RN)≤C‖g‖Lm(Ω). |
2) If m⩾Ns then, for all 1<p<∞, there exists C=C(N,s,m,p,Ω) such that
‖u‖Ws,p(RN)≤C‖g‖Lm(Ω). |
Let us recall that another version of the nonlocal gradient is given by
Ds(u)(x)=(aN,s2∫IRN|u(x)−u(y)|2|x−y|N+2sdy)12. |
Taking into consideration the result of [46], we get the following corollary.
Corollary 2.12. Assume that the conditions of Theorem 2.10 hold. Then we have
1) If m>Ns, then for all 2NN+2s<p<∞, there exists C=C(N,s,m,p,Ω) such that
‖Ds(u)‖Lp(RN)≤C‖g‖Lm(Ω). |
2) If 2NN+4s<m⩽Ns, then for all 2NN+2s<p<mNN−ms, there exists C=C(N,s,m,p,Ω) such that
‖Ds(u)‖Lp(RN)≤C‖g‖Lm(Ω). |
In this subsection we analyze the question of regularity of the solution to the problem
{(−Δ)su=λu|x|2s+f in Ω,u=0 in RN∖Ω, | (18) |
in Lebesgue spaces and fractional Sobolev spaces according to the regularity of the datum f. Here Ω⊂IRN is a bounded regular domain containing the origin and s∈(0,1). We will suppose that f∈Lm(Ω) with m⩾1 and 0<λ<ΛN,s.
If f=0, we define the radial potential v±αλ(x)=|x|−N−2s2±αλ with αλ given by
λ=λ(αλ)=λ(−αλ)=22sΓ(N+2s+2αλ4)Γ(N+2s−2αλ4)Γ(N−2s+2αλ4)Γ(N−2s−2αλ4). | (19) |
From [8], we obtain that v±αλ solves the homogeneous equation
(−Δ)su=λu|x|2s in RN∖{0}. | (20) |
It is clear that λ(α)=λ(−α)=mαλm−αλ, with mαλ=2αλ+sΓ(N+2s+2αλ4)Γ(N−2s−2αλ4).
Notice that
0<λ(αλ)=λ(−αλ)⩽ΛN,s if and only if 0⩽αλ<N−2s2. |
Define
μ(λ)=N−2s2−αλ and ˉμ(λ)=N−2s2+αλ. | (21) |
For 0<λ<ΛN,s, then 0<μ(λ)<N−2s2<ˉμ<(N−2s). Since N−2μ(λ)−2s=2αλ>0 and N−2ˉμ(λ)−2s=−2αλ<0, then (−Δ)s/2(|x|−μ(λ))∈L2(Ω), but (−Δ)s/2(|x|−ˉμ(λ)) does not.
As it was proved in [8], if f∈L1(Ω), then the existence of a solution to problem (18) is guaranteed under the necessary and sufficient condition ∫Br(0)f|x|−μ(λ)dx<∞. Hence, throughout this section this condition will be assumed.
The first result concerning the behavior in the neighborhood of zero is given by the next Proposition proved in [8].
Proposition 3.1. Let u∈L1loc(IRN) be such that u⩾0 in IRN and (−Δ)su∈L1loc(Ω). Assume that
(−Δ)su⩾λu|x|2s in Ω,0<λ<ΛN,s. |
Then, there exists r>0 and a positive constant C≡C(r,N,λ) such that
u(x)≥C|x|−μ(λ)=C|x|−N−2s2+αλ in Br(0)⊂⊂Ω. |
We are now in position to prove the first regularity results, in fractional Sobolev space, to the solution of problem (18).
Theorem 3.2. Assume that f∈Lm(Ω) with m>1 satisfying the condition ∫Br(0)f|x|−μ(λ)dx<∞. Let u∈L1(Ω) tobe the unique weak solution to (18) with λ<ΛN,s. Then there exists a positive constant C=C(N,m,p,s,Ω) such that
1) If m⩾N2s, then u|x|2s∈Lσ(Ω) for all 1⩽σ<Nμ(λ)+2s and |(−Δ)s2u|∈Lp(IRN) for all 1≤p<Nμ(λ)+s. Moreover we have
||(−Δ)s2u||Lp(IRN)≤C||f||Lm(Ω). |
2) If 1<m<N2s and λ<Js,m≡ΛN,s4N(m−1)(N−2ms)m2(N−2s)2, then |(−Δ)s2u|∈Lp(IRN) forall 1≤p<NmN−ms. Moreover we have
||(−Δ)s2u||Lp(RN)≤C||f||Lm(Ω). |
Proof. We begin by analyzing the first case. Assume that f∈Lm(Ω) with m>N2s. From Theorem 4.1 in [8], we obtain that u(x)≤C|x|−μ(λ)χΩ. Hence u|x|2s≤C|x|−μ(λ)−2sχΩ. As a consequence, we deduce that u|x|2s∈Lσ(Ω) for all 1⩽σ<Nμ(λ)+2s.
Setting g≡u|x|2s+f, it follows that g∈Lσ(Ω) for all σ<Nμ(λ)+2s. Using the regularity result in Theorem 2.9, we conclude that u∈Lt(Ω) for all t<Nμ(λ). Now by Theorem 2.10, it holds that |(−Δ)s2u|∈Lp(IRN) for all 1≤p<Nμ(λ)+s and
||(−Δ)s2u||Lp(IRN)≤C||f||Lm(Ω). |
Hence we conclude.
We treat now the case 1<m<N2s and 0<λ<Js,m≡ΛN,s4N(m−1)(N−2ms)m2(N−2s)2.
Recall that u solves problem (18). Then by Theorem 4.2 of [8], we get the existence of positive constant C(N,s,m) such that
||u||Lm∗∗s(Ω)≤C||f||Lm(Ω) where m∗∗s=mNN−2sm. | (22) |
Since p<NmN−ms, then we get the existence of m1<m such that p<Nm1N−m1s. Fixed m1<m, using Hölder inequality we deduce that
∫Ωum1|x|2sm1dx≤C. |
Since m1<m, it follows that g≡λu|x|2s+f∈Lm1(Ω).
On the other hand m1<m<N2s<Ns, therefore using the regularity result in Theorem 2.10, we deduce that
||(−Δ)s2u||Lp(IRN)≤C||g||Lm1(Ω) for all p<Nm1N−m1s. |
Thus
||(−Δ)s2u||Lp(IRN)≤C||f||Lm(Ω) for all p<NmN−ms, |
and the result follows in this case.
In order to treat the general case Js,m≡ΛN,s4N(m−1)(N−2ms)m2(N−2s)2⩽λ<ΛN,s, we need to develop a new approach.
Let u be the unique weak solution to problem (18). Setting v(x):=|x|μ(λ)u(x), it follows that v solves the problem
{Lμ(λ)v=|x|−μ(λ)f(x)=:˜f(x) in Ω,v=0 in RN∖Ω, | (23) |
with
Lγv:=aN,s P.V. ∫RNv(x)−v(y)|x−y|N+2sdy|x|γ|y|γ. | (24) |
Since ∫Br(0)f|x|−μ(λ)dx<∞, then ˜f∈L1(Ω). Thus v can be seen as the unique entropy solution to problem (23) as defined in [2]. Following closely the argument used in [4], we get the next general regularity result.
Theorem 3.3. Let s∈(0,1) and 0<λ<ΛN,s. Assume that ˜f∈Lq(Ω,|x|β(q−1)dx) with q>1 and 2Nμ(λ)N−2s⩽β⩽2(μ(λ)+s). Let us denoteby C:=C(N,β,λ,s,q,Ω) a positive constant that may change from line to other.
Then if v solves problem (23), we have
1) If β<2(μ(λ)+s) and q>(N−β)2(μ(λ)+s)−β, then v∈L∞(Ω). Moreover,
‖v‖L∞(Ω)≤C‖˜f‖Lq(Ω,|x|β(q−1)dx). |
2) If β<2(μ(λ)+s) and q=(N−β)2(μ(λ)+s)−β, then v∈Lr(Ω,|x|−βdx), for all 1≤r<+∞. Moreover
(∫Ω|v|r |x|−βdx)1r≤C‖˜f‖Lq(Ω,|x|β(q−1)dx). |
3) If either β=2(μ(λ)+s) or β<2(μ(λ)+s) and 1<q<N−β2(μ(λ)+s)−β, then |v|r∈L1(Ω,|x|−βdx), for all 1≤r≤r∗=(N−β)qN−β−q(2(μ(λ)+s)−β). Moreover
(∫Ω|v|r |x|−βdx)1r≤C‖˜f‖Lq(Ω,|x|β(q−1)dx). |
Before proving the previous Theorem, we recall the following weighted fractional Caffarelli-Khon-Nirenberg inequality, whose proof can be found in [3,15,37].
Theorem 3.4. Assume that s∈(0,1) and −2s<γ<N−2s2. Let θ∈[γ,γ+s], then there exists a positive constant C:=C(N, s, γ, θ), such that for all ϕ∈C∞0(IRN), we have
C(∫IRN|ϕ|ˆσ|x|ˆσθdx)2ˆσ⩽∫IRN∫IRN|ϕ(x)−ϕ(y)|2|x−y|N+2s|x|γ|y|γdxdy, |
with ˆσ=2NN−2s+2(θ−γ).
Setting β=ˆσθ, we obtain that 2NγN−2s⩽β⩽2(γ+s) and
C(∫IRN|ϕ|ˆσ|x|βdx)2ˆσ⩽∫IRN∫IRN|ϕ(x)−ϕ(y)|2|x−y|N+2s|x|γ|y|γdxdy. | (25) |
Notice that by substituting the value of θ in the formula of ˆσ, we reach that ˆσ=2(N−β)N−2(γ+s).
Proof of Theorem 3.3. Notice that, using the notation of Theorem 3.4, then, in our case, we have γ=μ(λ)∈(0,N−2s2).
The main idea of the proof is to use a suitable test function and an approximation argument. To make the paper self contained as possible, we include here all the details.
Without loss of generality we can assume that q>1 and ˜f≩0. Thus v≩0 in IRN.
Consider the following approximating problem
{Lμ(λ)vn=˜fn(x) in Ω,vn=0 in RN∖Ω, | (26) |
where ˜fn(x)=Tn(˜f(x)) is the truncation of ˜fn as defined in (10).
Since v is the unique solution to problem (23), at least in the entropy sense, then
vn↑v a.e. in IRN and vn↑v strongly in L1(IRN). |
In the rest of the proof, we denote by C any positive constant that depends only on N,s,q,r,Ω, and is independent of n,˜f, v, that may change from line to other.
It is not difficult to show that vn is bounded. Thus, for α>0 fixed, to be chosen later, using vαn as a test function in (26), it holds that
12∬DΩ(vn(x)−vn(y))(vαn(x)−vαn(y))|x−y|N+2s|x|μ(λ)|y|μ(λ)dxdy=∫Ω˜fnvαn(x)dx. |
By the algebraic inequality
(a−b)(aα−bα)≥C(aα+12−bα+12)2, |
we reach that
C∬DΩ(vα+12n(x)−vα+12n(y))2|x−y|N+2s|x|μ(λ)|y|μ(λ)dxdy⩽∫Ω˜fn(x)vαn(x)dx. |
Using the weighted fractional Caffarelli-Khon-Nirenberg inequality in Theorem 3.4, we get
∬DΩ(vα+12n(x)−vα+12n(y))2|x−y|N+2s|x|μ(λ)|y|μ(λ)dxdy≥C(∫Ωv(α+1)ˆσ2n|x|βdx)2ˆσ. |
Now by using Hölder's inequality, it holds that
∫Ω˜fnvαn(x)dx⩽(∫Ω˜fqn(x)|x|β(q−1)dx)1q(∫Ωvαq′n(x)|x|βdx)1q′. | (27) |
Hence
C(∫Ω(vn(x))(α+1)ˆσ2|x|βdx)2ˆσ≤(∫Ω˜fqn(x)|x|β(q−1)(x)dx)1q(∫Ωvαq′n(x)|x|βdx)1q′. | (28) |
∙ If β<2(μ(λ)+s) and ˆσ2>q′, namely q>N−β2(μ(λ)+s)−β, in this case we can prove that v∈L∞(Ω). The proof follows using the classical Stampacchia argument as in [44]. Let us give some details. Using Gk(vn) as a test function (26), it follows that
12∬DΩ(vn(x)−vn(y))(Gk(vn(x))−Gk(vn(y)))|x−y|N+2s|x|μ(λ)|y|μ(λ)dxdy=∫Ω˜fn(x)Gk(vn(x))dx. |
Since ˆσ2>q′, then 1ˆσ+1q<1−12q′. Thus Using the Hölder inequality, we get
C∬DΩ(Gk(vn(x))−Gk(vn(y)))2|x−y|N+2s|x|μ(λ)|y|μ(λ)dxdy⩽(∫Ω˜fq(x)|x|β(q−1)dx)1q(∫Ω(Gk(vn(x)))ˆσ|x|βdx)1ˆσ|{x∈Ω:Gk(vn(x))>0}|1−1ˆσ−1q|x|−βdx. |
Now, by the Caffarelli-Kohn-Nirenberg inequality in (25), we deduce that
(∫Ω(Gk(vn(x)))ˆσ|x|βdx)1ˆσ⩽(∫Ω˜fqn(x)|x|β(q−1)dx)1q|{x∈Ω:Gk(vn(x))>0}|1−1ˆσ−1q|x|−βdx. |
Hence
|{x∈Ω:vn(x)>k}|1ˆσ|x|−βdx≤C|{x∈Ω:vn(x)>k}|1−1ˆσ−1q|x|−βdx. |
Thus using the standard Stampacchia argument, see [44], we get the existence of k0>0, independents of n such that
|{x∈Ω:vn(x)>k0}|=0 for all n. |
Hence |{x∈Ω:v(x)>k0}| and then v∈L∞(Ω).
∙ If β<2(μ(λ)+s) and ˆσ2=q′, since (28) holds for all α⩾1, then using Hölder's inequality, we reach that for all n⩾1, vrn|x|−β∈L1(Ω), for all r<∞ and
(∫Ωvrn |x|−βdx)1r≤C‖˜fn‖Lq(Ω,|x|β(q−1)dx),for all 1≤r<+∞. |
Now using Fatou's Lemma we deduce that
(∫Ωvr |x|−βdx)1r≤C‖˜f‖Lq(Ω,|x|β(q−1)dx),for all 1≤r<+∞ |
as requested.
∙ Now, if β<2(μ(λ)+s) and ˆσ2<q′, that is q<N−β2(μ+s)−β, and choosing α=ˆσ2q′−ˆσ, then (α+1)ˆσ2=qˆσ2q−(q−1)ˆσ=q(N−β)N−β−q(2(s+μ(λ))−β):=r∗. Going back to (28), it follows that, for all n⩾1, vr∗n|x|−β∈L1(Ω) and
(∫Ωvr∗n |x|−βdx)1r∗≤C‖˜fn‖Lq(Ω,|x|β(q−1)dx). |
As above, using Fatou's lemma, we get
(∫Ωvr∗ |x|−βdx)1r∗≤C‖˜f‖Lq(Ω,|x|β(q−1)dx). |
∙ If β=2(μ(λ)+s), then ˆσ=2. Again from (28) and choosing α=1q′−1, it follows that r∗=q and vr∗n|x|−β∈L1(Ω) for all n⩾1 with
(∫Ωvqn |x|−βdx)1q≤C‖˜fn‖Lq(Ω,|x|β(q−1)dx). |
Thus
(∫Ωvq |x|−βdx)1q≤C‖˜f‖Lq(Ω,|x|β(q−1)dx). |
As a consequence, we get the next corollary that improves the regularity results obtained in [8].
Corollary 3.5. Let s∈(0,1), 0<λ<λN,s and u be theunique weak solution to problem (18) with f|x|−μ(λ)∈L1(Ω). Suppose in addition that f|x|β−μ(λ)∈Lq(Ω,|x|−βdx) where q>1 and 2Nμ(λ)N−2s⩽β⩽2(μ(λ)+s). Then
1) If β<2(μ(λ)+s) and q>(N−β)2(μ(λ)+s)−β, then u|x|μ(λ)∈L∞(Ω). Moreover, there exists a positive constant C:=C(N,β,λ,s,q,Ω) such that
‖u|x|μ(λ)‖L∞(Ω)≤C‖f|x|β−μ(λ)‖Lq(Ω,|x|−βdx). |
2) If β<2(μ(λ)+s) and q=(N−β)2(μ(λ)+s)−β, then u|x|μ(λ)∈Lr(Ω,|x|−βdx), for all 1≤r<+∞. Moreover, there exists a positive constant C:=C(N,β,λ,s,q,r,Ω) such that
(∫Ωur|x|rμ(λ)−βdx)1r≤C‖f|x|β−μ(λ)‖Lq(Ω,|x|−βdx). |
3) If either β=2(μ(λ)+s) or β<2(μ(λ)+s) and 1<q<N−β2(μ(λ)+s)−β, then
u|x|μ(λ)∈Lr∗(Ω,|x|−βdx) with r∗=(N−β)qN−β−q(2(μ(λ)+s)−β). Moreover, thereexists a positive constant C:=C(N,β,λ,s,q,Ω) such that
(∫Ωur∗ |x|r∗μ(λ)−βdx)1r∗≤C‖f|x|β−μ(λ)‖Lq(Ω,|x|−βdx). |
As a consequence we get the next fractional regularity.
Theorem 3.6. Suppose that f satisfies the same condition as in Corollary 3.5. Let u∈L1(Ω) be the unique weak solution to (18) with λ<ΛN,s. Then
1) If β<2(μ(λ)+s) and q>(N−β)2(μ(λ)+s)−β, then |(−Δ)s2u|∈Lp(IRN) forall 1≤p<Nμ(λ)+s. In particular, thereexists a positive constant C:=C(N,β,λ,s,q,p,Ω) such that
||(−Δ)s2u||Lp(IRN)⩽C‖f|x|β−μ(λ)‖Lq(Ω,|x|−βdx). |
2) If β<2(μ(λ)+s) and q=(N−β)2(μ(λ)+s)−β, then |(−Δ)s2u|∈Lp(IRN) forall 1≤p<Nμ(λ)+s. In particular, thereexists a positive constant C:=C(N,β,λ,s,q,p,Ω) such that
||(−Δ)s2u||Lp(IRN)⩽C‖f|x|β−μ(λ)‖Lq(Ω,|x|−βdx). |
3) If either β=2(μ(λ)+s) or β<2(μ(λ)+s) and ββ−μ(λ)≤q<N−β2(μ(λ)+s)−β, then |(−Δ)s2u|∈Lp(IRN) forall 1≤p<qNN−β−q(μ+s−β). Inparticular, there exists a positive constant C:=C(N,β,λ,s,q,p,Ω) such that
||(−Δ)s2u||Lp(IRN)≤C‖f|x|β−μ(λ)‖Lq(Ω,|x|−βdx). |
4) If either β=2(μ(λ)+s) or β<2(μ(λ)+s) and 1<q≤ββ−μ(λ), then |(−Δ)s2u|∈Lp(IRN) for all 1≤p<qNN−qs. In particular, there exists apositive constant C:=C(N,β,λ,s,q,p,Ω) such that
||(−Δ)s2u||Lp(IRN)≤C‖f|x|β−μ(λ)‖Lq(Ω,|x|−βdx). |
Proof. We start with the first case. Since β<2(μ(λ)+s) and q>(N−β)2(μ(λ)+s)−β, then by Corollary 3.5, we obtain that u(x)≤C|x|−μ(λ).
Hence u|x|2s≤C|x|−μ(λ)−2s∈Lσ(Ω) for all 1⩽σ<Nμ(λ)+2s. Since q>(N−β)2(μ(λ)+s)−β, then using Hölder inequality we can show the existence of a>Nμ(λ)+2s such that f∈La(Ω). Thus g:=u|x|2s+f∈Lσ(Ω) for all 1⩽σ<Nμ(λ)+2s. Using now the regularity result in Theorem 2.10, it holds that |(−Δ)s2u|∈Lp(IRN) for all 1≤p<Nμ(λ)+s and
||(−Δ)s2u||Lp(IRN)≤C‖f|x|β−μ(λ)‖Lq(Ω,|x|−βdx). |
The second case follows as the first case using the fact that u|x|μ(λ)∈Lr(Ω,|x|−βdx), for all 1≤r<+∞.
We consider the third case which is more involved. Assume that β<2(μ(λ)+s) and 1≤q<N−β2(μ(λ)+s)−β, then by Corollary 3.5, we reach that u|x|μ(λ)∈Lr∗(Ω,|x|−βdx) with r∗=(N−β)qN−β−q(2(μ(λ)+s)−β). We claim that u|x|2s∈Lθ(Ω) for all 1≤θ<qNN+q(β−μ(λ))−β. To see this we will use Hölder's inequality. More precisely, for 1≤θ<r∗, we have
∫Ω(u|x|2s)θdx=∫Ω(u|x|μ(λ))θ(|x|β−θ(μ(λ)+2s))|x|−βdx⩽(∫Ω(u|x|μ(λ))r∗|x|−βdx)θr∗(∫Ω(|x|β−θ(μ(λ)+2s))r∗r∗−θ|x|−βdx)r∗−θr∗⩽C(Ω)‖f|x|β−μ(λ)‖θLq(Ω,|x|−βdx)(∫Ω(|x|β−θ(μ(λ)+2s))r∗r∗−θ|x|−βdx)r∗−θr∗. |
The last integral is finite if and only if (β−θ(μ(λ)+2s))r∗r∗−θ−β>−N. This is equivalent to the fact that θ<qNN+q(β−μ(λ))−β. Notice that in this case we have qNN+q(β−μ(λ))−β<r∗. Then the claim follows.
In the same way and taking into consideration that qNN+q(β−μ(λ))−β<q, we can prove that f∈Lθ(Ω) for all 1≤θ<qNN+q(β−μ(λ))−β. As in the previous cases, setting g:=u|x|2s+f, then g∈Lθ(Ω) for all 1≤θ<qNN+q(β−μ(λ))−β. Thus by the regularity result in Theorem 2.10, we obtain that |(−Δ)s2u|∈Lp(IRN) for all 1≤p<θNN−θs. Hence |(−Δ)s2u|∈Lp(IRN) for all 1≤p<qNN−β−q(μ+s−β) and
||(−Δ)s2u||Lp(IRN)≤C‖f|x|β−μ(λ)‖Lq(Ω,|x|−βdx). |
Finally, the fourth case follows easily, using the approach of the previous case.
To end this section we give the next weighted estimate for the fractional gradient if additional assumptions on f are satisfied. This will be used in order to show the existence of a solution to problem (1).
Suppose that f∈L1(|x|−μ(λ)−a0dx,Ω) for some a0>0. Hence there exists λ1∈(λ,ΛN,s) such that μ(λ1)=μ(λ)+a0. Define ψ to be the unique solution to problem
{(−Δ)sψ=λ1ψ|x|2s+1 in Ω,ψ=0 in RN∖Ω, | (29) |
then ψ≃|x|−μ(λ)−a0 near the origin. It is clear also that ψ∈L∞(Ω∖Br(0)).
Using ψ as a test function in problem (18), it holds that
(λ1−λ)∫Ωuψ|x|2sdx⩽∫Ωfψdx. |
Hence
∫Ωu|x|2s+μ(λ)dx≤C(Ωλ,a0)||f||L1(|x|−μ(λ)−a0dx,Ω). | (30) |
The next proposition will be the crucial key in order to show a priori estimates when dealing with problem (1) with general datum f.
Proposition 3.7. Assume that f∈L1(|x|−μ(λ)−a0dx,Ω) for some a0>0. Let v be the unique weak solution to problem (18), then
‖(−Δ)s2u‖Lα(|x|−μ(λ)dx,Ω)≤C(Ω,λ,a0)||f||L1(|x|−μ(λ)−a0dx,Ω) for all 1⩽α<NN−s. | (31) |
To prove Proposition 3.7, we need the following lemma proved in [27].
Lemma 3.8. Let N⩾1, R>0 and α,β∈(−∞,N). There exists C:=C(N,R,α,β)>0 such that:
∙ If (N−α−β)≠0, then
∫BR(0)dz|x−z|α|y−z|β⩽C(1+|x−y|N−α−β), for all x,y∈BR(0) with x≠y. |
∙ If (N−α−β)=0, then
∫BR(0)dz|x−z|α|y−z|β⩽C(1+|ln|x−y||), for all x,y∈BR(0) with x≠y. |
Proof of Proposition 3.7. Since ∫Ωf|x|−μ(λ)dx<∞, then by Theorem 2.10, we know that
‖(−Δ)s2u‖Lα(Ω)≤C(Ω,λ,a0)||f||L1(|x|−μ(λ)−a0dx,Ω) for all 1⩽α<NN−s. |
Thus, to prove the claim we just need to show that
∫Br(0)|(−Δ)s2u|α|x|−μ(λ)dx≤C(Ω,λ,a0)||f||αL1(|x|−μ(λ)−a0dx,Ω) for all 1≤α<NN−s, |
where Br(0)⊂⊂Ω.
We set g(x):=λu|x|2s+μf, then u(x)=∫ΩGs(x,y)g(y)dy. Hence, for a.e. x∈Br(0),
|(−Δ)s2u(x)|⩽∫Ω|(−Δ)s2Gs(x,y)|g(y)dy. | (32) |
Notice that from [6], we know that
|(−Δ)s2xGs(x,y)|⩽C|x−y|N−s(|ln1|x−y||+ln(1δ(x))), for a.e. x,y∈Ω. | (33) |
Since Br(0)⊂⊂Ω, one has
|(−Δ)s2xGs(x,y)|⩽C|x−y|N−sln(C|x−y|), for a.e. (x,y)∈Br(0)×Ω. | (34) |
For the remaining part of this proof, we will use systematically this estimate for a.e. (x,y)∈Br(0)×Ω.
Thus we conclude that
|(−Δ)s2xGs(x,y)|⩽Gs(x,y)h(x,y), for a.e. (x,y)∈Br(0)×Ω, | (35) |
with
h(x,y)={1|x−y|sln(C|x−y|) if |x−y|<δ(y)1δs(y)ln(C|x−y|) if |x−y|≥δ(y). |
Fix 1<α<p∗=NN−s. Going back to (32), we deduce that, for a.e. x∈Br(0), we have
|(−Δ)s2u(x)|⩽∫Ω|(−Δ)s2Gs(x,y)|g(y)dy≤λ∫Ω|(−Δ)s2Gs(x,y)|u(y)|y|2sdy+∫Ω|(−Δ)s2Gs(x,y)|f(y)dy. |
Hence
|(−Δ)s2u(x)||x|−μ(λ)α⩽λ|x|−μ(λ)α∫Ω|(−Δ)s2Gs(x,y)|u(y)|y|2sdy+|x|−μ(λ)α∫Ω|(−Δ)s2Gs(x,y)|f(y)dy. |
We set
K1(x)=∫Ω|(−Δ)s2Gs(x,y)|u(y)|y|2sdy |
and
K2(x)=∫Ω|(−Δ)s2Gs(x,y)|f(y)dy. |
We begin estimating K1. We have
Kα1(x)⩽(∫Ω|(−Δ)s2Gs(x,y)|u(y)|y|2sdy)α≤(∫Ωh(x,y)Gs(x,y)u(y)|y|2sdy)α⩽(∫Ω(h(x,y))αGs(x,y)u(y)|y|2sdy)(∫ΩGs(x,y)g(y)dy)α−1⩽∫Ω(hα(x,y)Gs(x,y)u(y)|y|2sdy)uα−1(x). |
Thus, using Fubini's theorem, it holds that
∫Br(0)Kα1(x)|x|−μ(λ)dx⩽λ∫Ωu(y)|y|2s(∫Br(0)hα(x,y)Gs(x,y)uα−1(x)|x|−μ(λ)dx)dy. |
Recall that, by (30), we have
∫Ωu(y)|y|2s+μ(λ)dy≤C∫Ωf(y)|y|μ(λ)+a0dy. |
Therefore we obtain that
∫Br(0)Kα1(x)|x|−μ(λ)dx⩽C2∫Ωu(y)|y|2s(∫Br(0)∩{|x−y|<δ(y)}uα−1(x)Gs(x,y)|x|μ(λ)1|x−y|sαln(C|x−y|)αdx)dy+C2∫Ωu(y)|y|2sδsα(y)(∫Br(0)∩{|x−y|≥δ(y)}uα−1(x)Gs(x,y)|x|μ(λ)ln(C|x−y|)αdx)dy=J1+J2. |
Respect to J1, using the fact that for all η>0,
1|x−y|sαln(C|x−y|)α⩽αηCη|x−y|sα+η, |
and by Proposition 2.8, we reach that
J1≤C2∫Ωu(y)|y|2s(∫Br(0)uα−1(x)|x|μ(λ)|x−y|N−(2s−sα−η)dx)dy⩽C∫Ωu(y)|y|2s(∫Br(0)∩{|x|⩾12|y|}uα−1(x)|x|μ(λ)|x−y|N−(2s−sα−η)dx)dy+C∫Ωu(y)|y|2s(∫Br(0)∩{|x|<12|y|}uα−1(x)|x|μ(λ)|x−y|N−(2s−sα−η)dx)dy⩽I1+I2. |
To estimate I1, we observe that
I1⩽C∫Ωu(y)|y|2s+μ(λ)(∫Br(0)uα−1(x)|x−y|N−(2s−sα−η)dx)dy. |
Recall that u∈Lσ(Ω) for all σ<NN−2s. Since α<NN−s, fixing σ0<NN−2s and using Hölder inequality, we get
∫Br(0)uα−1(x)|x−y|N−(2s−sα−η)dx⩽(∫Br(0)uσ0dx)α−1σ0(∫Br(0)1|x−y|(N−(2s−sα−η))σ0σ0−(α−1)dx)σ0−(α−1)σ0. |
Since α<NN−s, then we can chose σ0 close to NN−2s and η small enough such that (N−(2s−sα−η))σ0σ0−(α−1)<N. Thus
∫Br(0)1|x−y|(N−(2s−sα−η))σ0σ0−(α−1)dx≤C(r,Ω), |
and then
I1≤C(∫Br(0)uσ0dx)α−1σ0(∫Ωu(y)|y|2s+μ(λ)dy)≤C(∫Ωf(y)|y|μ(λ)+a0dy)α. | (36) |
We deal now with I2. Notice that {|x|⩽12|y|}⊂{|x−y|⩾12|y|}. Thus
I2⩽C∫Ωu(y)|y|μ(λ)+2s(∫Br(0)uα−1(x)|x|μ(λ)|x−y|N−(2s+μ(λ)−sα−η)dx)dy. |
As in the estimate of I1, setting θ=σ0σ0−(α−1), we get
∫Br(0)uα−1(x)|x|μ(λ)|x−y|N−(2s+μ(λ)−sα−η)dx⩽(∫Br(0)uσ0dx)α−1σ0(∫Br(0)1|x|μ(λ)θ|x−y|(N−(2s+μ(λ)−sα−η))θdx)1θ. |
For α<NN−s fixed, we can chose η small enough and σ0 close to p2 such that
N−(N−(2s+μ(λ)−sα−η))θ−μ(λ)θ<N. |
Hence using Lemma 3.8, it holds that
∫Br(0)1|x|μ(λ)θ|x−y|(N−(2s+μ(λ)−sα−η))θdx≤C(r0). |
Therefore we conclude that
I2⩽C∫Ωu(y)|y|μ(λ)+2s(∫Br(0)uσ0dx)α−1σ0⩽C(∫Ωf(y)|y|μ(λ)+a0dy)α. | (37) |
As a consequence, we have
J1≤C(∫Ωf(y)|y|μ(λ)+a0dy)α. | (38) |
We deal now with J2. Let c1>0 be a positive constant to be chosen later, then
J2⩽∫Ωu(y)|y|2sδsα(y)(∫Br(0)∩{|x−y|≥δ(y)}uα−1(x)Gs(x,y)|x|μ(λ)ln(C|x−y|)αdx)dy⩽∫Ω∩{δ(y)>c1}u(y)|y|2sδsα(y)(∫Br(0)∩{|x−y|≥δ(y)}uα−1(x)Gs(x,y)|x|μ(λ)ln(C|x−y|)αdx)dy+∫Ω∩{δ(y)<c1}u(y)|y|2sδsα(y)(∫Br(0)∩{|x−y|≥δ(y)}uα−1(x)Gs(x,y)|x|μ(λ)ln(C|x−y|)αdx)dy⩽C1J1+∫Ω∩{δ(y)<c1}u(y)|y|2sδsα(y)(∫Br(0)∩{|x−y|≥δ(y)}uα−1(x)Gs(x,y)|x|μ(λ)ln(C|x−y|)αdx)dy. |
We set
A=∫Ω∩{δ(y)<c1}u(y)|y|2sδsα(y)(∫Br(0)∩{|x−y|≥δ(y)}uα−1(x)Gs(x,y)|x|μ(λ)ln(C|x−y|)αdx)dy. |
Choosing c1 small, we get the existence of a positive constant c2 such that for δ(y)<c1 and x∈Br(0), we have |x−y|>c2>0. Hence using again Proposition 2.8, we deduce that
A⩽∫Ω∩{δ(y)<c1}u(y)δs(α−1)(y)(∫Br(0)∩{|x−y|≥δ(y)}uα−1(x)|x|μ(λ)dx)dy. |
As above, for α0<NN−2s, we have
∫Br(0)uα−1(x)|x|μ(λ)dx⩽(∫Br(0)uσ0dx)α−1σ0(∫Br(0)1|x|μ(λ)θdx)1θ⩽C(∫Br(0)uσ0dx)α−1σ0≤C(∫Ωf(y)|y|μ(λ)+a0dy)α−1. |
On the other hand, we have
∫Ω∩{δ(y)<c2}u(y)δs(α−1)(y)dy⩽C∫Ωu(y)δs(y)dy⩽C∫Ωf(y)|y|μ(λ)+a0dy. |
Hence
J2≤C(∫Ωf(y)|y|μ(λ)+a0dy)α. |
As a consequence we deduce that
∫Br(0)Kα1(x)|x|−μ(λ)dx≤C(∫Ωf(y)|y|μ(λ)+a0dy)α. |
We treat now the term K2. Recall that
K2(x)=∫Ω|(−Δ)s2Gs(x,y)|f(y)dy. |
Notice that for η>0, small enough, to be chosen later, we have
|(−Δ)s2Gs(x,y)|⩽C|x−y|N−sln(C|x−y|)≤C|x−y|N−(s−η) for a.e. (x,y)∈Br(0)×Ω. |
Thus
K2(x)|x|−μ(λ)α⩽C|x|μ(λ)α∫Ωf(y)|x−y|N−(s−η)dy⩽C|x|μ(λ)α∫Ω∩{|y|⩽4|x|}f(y)|x−y|N−(s−η)dy+C|x|μ(λ)α∫Ω∩{|y|⩾4|x|}f(y)|x−y|N−(s−η)dy⩽C∫Ω∩{|y|⩽4|x|}f(y)|y|μ(λ)1|x−y|N−(s−η)dy+C|x|μ(λ)α∫Ω∩{|y|⩾4|x|}f(y)|x−y|N−(s−η)dy=L1(x)+L2(x). |
We start with the estimate of term L1. Since f(y)|y|μ(λ)∈L1(Ω), then by Theorem 2.4, we deduce that L1∈Lσ(Br(0)) for all 1≤σ<NN−(s−η). Thus L1∈Lα(Ω) and
∫Br(0)Lα1(x)dx⩽C(∫Ωf(y)|y|μ(λ)+a0dy)α. |
We consider now L2. Since |y|⩾4|x|, then |x−y|⩾34|y| and |x−y|⩾3|x|. Hence
L2(x)⩽C|x|μ(λ)α+N−(μ(λ)+s+a0−η)∫Ωf(y)|y|μ(λ)+a0dy. |
Since (μ(λ)α+N−(μ(λ)+s+a0−η))α<N, then we conclude that I2∈Lα(Br(0)) and
∫Br(0)Lα2(x)dx⩽C(∫Ωf(y)|y|μ(λ)+a0dy)α. |
As a consequence, we have proved that
∫Br(0)Kα2(x)|x|−μ(λ)dx≤C(∫Ωf(y)|y|μ(λ)+a0dy)α. |
Therefore we conclude that
∫Br(0)|(−Δs2)u|α|x|−μ(λ)dx≤C(∫Ωf(y)|y|μ(λ)+a0dy)α. |
Hence the main estimate follows and this finishes the proof of our proposition.
In this section we consider the question of existence and non existence of a positive solution to problem (1) with F(u)≡|(−Δ)s2u|. Namely we will treat the problem
{(−Δ)su=λu|x|2s+|(−Δ)s2u|p+ρf in Ω,u>0 in Ω,u=0 in (IRN∖Ω), | (39) |
where Ω⊂IRN is a bounded regular domain containing the origin, s∈(0,1), λ⩽ΛN,s, ρ>0, 1<p<∞ and f is positive measurable function satisfies some hypothesis that will be precised later.
Let us begin with the next definition.
Definition 4.1. Assume that f∈L1(Ω) is a nonnegative function. We say that u is a weak solution to problem (39) if |(−Δ)s2u|p∈L1(Ω),u|x|2s∈L1(Ω) and, setting g≡λu|x|2s+|(−Δ)s2u|p+ρf, then u is a weak solution to problem (8) in the sense of Definition 2.6.
The existence of a solution in the case λ=0 was proved in [7] without any limitation on p under suitable hypotheses on f. However, if λ>0, taking into consideration the singularity generated by the Hardy potential, it is possible to show a non existence result for p large. In the next computation we will find the exact critical exponent for the non existence.
Recall that we are considering the case F(u)(x)=|(−Δ)s2u(x)|. We begin by analyzing the radial case in the whole space as in [8]. Consider the equation
(−Δ)sw−λw|x|2s=|(−Δ)s2w|p in IRN, | (40) |
then we search radial positive solution in the form w=A|x|β−N−2s2, with A>0. By a direct computation, it follows that
Aγβ,s|x|−2s−N−2s2+β−λA|x|β−2s−N−2s2=Ap|γβ,s2|p|x|(N−2s2−β+s)p, |
with
γβ,t:=γ−β,t:=22tΓ(N+2t+2β4)Γ(N+2t−2β4)Γ(N−2t−2β4)Γ(N−2t+2β4) | (41) |
and t∈{s2,s}.
Hence, by homogeneity, we need to have
p=N−2s2−β+2sN−2s2−β+s, |
which means that β=N−2s2+psp−1−2sp−1. Hence the constant A satisfies
γβ,s−λ=Ap−1|γβ,s2|p. |
Using the fact that A>0, it holds that γβ−λ>0. Define the application
Υ:(−N−2s2,N−2s2)↦(0,ΛN,s)β↦γβ |
Then Υ is even and the restriction of Υ to the set [0,N−2s2) is decreasing, see [24] and [26]. So there exists a unique αλ∈(0,ΛN,s] such that γαλ=γ−αλ=λ.
Let β0=−β1=αλ. Setting
p+(λ,s):=N−2s2−β0+2sN−2s2−β0+s=N+2s−2αλN−2αλ, | (42) |
and
p−(λ,s):=N−2s2−β1+2sN−2s2−β1+s=N+2s+2αλN+2αλ, | (43) |
it holds that p−(λ,s)<p+(λ,s) and γβ−λ>0 if and only if
p−(λ,s)<p<p+(λ,s). |
It is easy to check that p+(λ,s) and p−(λ,s) are respectively an increasing and a decreasing function in αλ and, therefore, are respectively a decreasing and an increasing function in the variable λ. Thus
NN−s<p−(λ,s)<N+2sN<p+(λ,s)<2, for 0<λ<ΛN,s. |
Notice that
p+(λ,s)=μ(λ)+2sμ(λ)+s and p−(λ,s)=ˉμ(λ)+2sˉμ(λ)+s, |
where μ(λ) and ˉμ(λ) are defined by (21).
Hence, for p−(λ,s)<p<p+(λ,s) fixed, using the fact that Ω⊂BR(0) for R large, we get the existence of a positive constant C1>0 such that w(x)=C1|x|β−N−2s2 is a radial supersolution for the Dirichlet problem (39) if f(x)⩽1|x|N−2s2+2s−β with ρ small.
To show that p+(λ,s) is critical, we prove the next non existence result.
Theorem 4.2. Let s∈(0,1) and suppose that p>p+(λ,s). Then for all ρ⩾0, problem (39) has no positive weak solution u in the sense of Definition 4.1.
Proof. We argue by contradiction. Assume that problem (39) has a positive solution u in the sense of Definition 4.1, then (−Δ)s2u∈Lp(Ω) and u|x|2s∈L1(Ω). By Lemma 3.1, it follows that
u(x)≥C|x|−μ(λ) in Br(0)⊂⊂Ω. |
Since |(−Δ)s2u|p+λu|x|2s+ρf∈L1(Ω), then from the regularity result in Theorem 2.10, we deduce that (−Δ)s2u∈Lt(IRN) for all 1≤t<NN−s.
Let θ∈L∞(Ω) be a nonnegative function such that Suppθ⊂⊂Br2(0)⊂Br(0)⊂⊂Ω and define ϕθ∈Hs0(Ω)∩L∞(Ω) to be the unique solution of the problem
{(−Δ)s2ϕθ=θ, in Ω,ϕθ=0, in IRN∖Ω. | (44) |
From [40], it holds that ϕθ≃δs2 near the boundary of Ω. Using ϕθ as test function in (39), we get
λ∫Ωuϕθ|x|2sdx+∫Ω|(−Δ)s2u|pϕθdx+ρ∫Ωfϕθ=∫Ωu(−Δ)sϕθdx=∫IRN(−Δ)s2u(−Δ)s2ϕθdx=∫Ω(−Δ)s2u(−Δ)s2ϕθdx+∫IRN∖Ω(−Δ)s2u(−Δ)s2ϕθdx. | (45) |
We treat separately each term in the right hand of the above identity.
Since Suppθ⊂⊂Br(0)⊂⊂Ω, using the fact that ϕθ>0 in Ω and then Hölder inequality, we reach that
|∫Ω(−Δ)s2u(−Δ)s2ϕθdx|=|∫Ω(−Δ)s2uθdx|⩽∫Ω|(−Δ)s2u|ϕθθϕθdx. | (46) |
Next, applying Young's inequality, it holds that
|∫Ω(−Δ)s2u(−Δ)s2ϕθdx|⩽ε∫Ω|(−Δ)s2u|pϕθdx+C(ε)∫Ωθp′ϕp′−1θdx, | (47) |
where ϵ>0 will be chosen later.
Now we deal with the term ∫IRN∖Ω(−Δ)s2u(−Δ)s2ϕθdx.
Since x∈IRN∖Ω, using again Hölder inequality, it follows that
|∫IRN∖Ω(−Δ)s2u(−Δ)s2ϕθdx|⩽||(−Δ)s2u||Lt(IRN∖Ω)||(−Δ)s2ϕθ||Lt′(IRN∖Ω) |
with t<NN−s. By hypothesis ||(−Δ)s2u||Lt(IRN∖Ω)<∞. Now respect to ||(−Δ)s2ϕθ||Lt′(IRN∖Ω), we have
||(−Δ)s2ϕθ||t′Lt′(IRN∖Ω)=∫IRN∖Ω(∫Ωϕθ(y)|y−x|N+sdy)t′dx=∫IRN∖Ω(∫Br(0)ϕθ(y)|y−x|N+sdy)t′dx+∫IRN∖Ω(∫Ω∖Br(0)ϕθ(y)|y−x|N+sdy)t′dx. |
Recalling that Suppθ⊂Br2(0)⊂Br(0)⊂⊂Ω, then we can prove that
||ϕθ||L∞(Ω∖Br(0))≤C||θ||L1(Br(0)), |
where C depends only on the N,s. Thus
∫IRN∖Ω(∫Ω∖Br(0)ϕθ(y)|y−x|N+sdy)t′dx≤C∫IRN∖Ω(∫Ω∖Br(0)1|y−x|N+sdy)t′dx≤C(r,Ω,s,N,t). |
Now, choosing r small enough, we obtain that, for x∈IRN∖Ω and y∈Br(0), |x−y|≥c(|x|+1). Hence
∫IRN∖Ω(∫Br(0)ϕθ(y)|y−x|N+sdy)t′dx≤∫IRN∖ΩC(|x|+1)t′(N+s)(∫Br(0)ϕθ(y)dy)t′dx≤C||ϕθ||t′L1(Ω). |
Now, going back to (45), choosing ε<<1 in estimate (47), we obtain that
λ∫Ωuϕθ|x|2sdx≤C(ε)∫Ωθp′ϕp′−1θdx+C||ϕθ||t′L1(Ω)+C||(−Δ)s2u||Lt(IRN)+C. | (48) |
Recall that p>p+(λ,s)=2s+μ(λ)s+μ(λ), hence p′<2s+μ(λ)s. Using an approximating argument we can take θ=1|x|βχBr4(0) with N−(μ(λ)+s)⩽β<N−(p′−1)s. In this case ϕθ⋍1|x|β−s near the origin and ϕθ∈L∞(B∖Br(0). Therefore ϕ∈L1(Ω). From (48), it holds that
C∫Br41|x|β+μ+sdx≤C(ε)∫Ωθp′ϕp′−1θdx+C||ϕθ||t′L1(Ω)+C||(−Δ)s2u||Lt(IRN)+C. |
Since (β+μ+s)≥N, then in order to conclude we have just to show that ∫Ωθp′ϕp′−1θdx<∞. Notice that
∫Ωθp′ϕp′−1θdx⩽∫Br4(0)1|x|p′β−(p′−1)(β−s)dx=∫Br4(0)1|x|(p′−1)s+βdx. |
Taking into consideration that p′<2s+μ(λ)s, it follows that (p′−1)s+β<N and so we are done.
Remarks 4.3. Following the same arguments as above, we can prove that problem(39) has no positive supersolutions u in the following sense: u=0a.e in IRN∖Ω, |(−Δ)s2u|∈Lr(IRN) for some r>1, g:=|(−Δ)s2u|p+λu|x|2s+ρf∈L1(Ω) and for all nonnegative ϕ∈T(defined in (9)), we have
∫IRN(−Δ)s2u(−Δ)s2ϕdx⩾∫Ωgϕdx. |
For ρ large, we are also able to prove another non existence result.
Theorem 4.4. Assume that f≩0 and p>2s+2s+2, then there exists ρ∗>0 such that problem (39) has non positive solution for ρ>ρ∗.
Proof. Without loss of generality we assume that f∈L∞(Ω).
Assume that u is a positive solution to problem (39). For θ∈C∞0(Ω) with θ≩0, we define ϕθ to be the unique solution to the problem
{(−Δ)sϕθ=θ, in Ω,ϕθ=0, in IRN∖Ω. |
Notice that ϕθ≃δs(x) where δ(x)≡dist(x,∂Ω), see for instance [41].
Using ϕθ as a test function in (39), it holds that
∫Ω(−Δ)sϕθudx⩾∫Ω|(−Δ)s2u|pϕθdx+ρ∫Ωfϕθdx. |
Hence
∫Ωθudx⩾∫Ω|(−Δ)s2u|pϕθdx+ρ∫Ωfϕθdx. | (49) |
Let ψθ to be the unique solution to the problem
{(−Δ)s2ψθ=θ, in Ω,ψθ=0, in IRN∖Ω. |
Thus
∫Ω(−Δ)s2ψθudx⩾∫Ω|(−Δ)s2u|pϕθdx+ρ∫Ωfϕθdx. |
Then
∫Ω(−Δ)s2uψθdx⩾∫Ω|(−Δ)s2u|pϕθdx+ρ∫Ωfϕθdx. |
Notice that
∫Ω(−Δ)s2uψθdx⩽∫Ω|(−Δ)s2u|ϕθψθϕθdx. |
Hence, using Young's inequality, for any ε>0, we get the existence of a positive constant C(ε) such that
∫Ω(−Δ)s2uψθdx⩽ε∫Ω|(−Δ)s2u|pϕθdx+C(ε)∫Ωϕθ(ψθϕθ)p′dx. |
Since θ is bounded, according with [40], then ψθ⋍δs2 and ϕθ⋍δs, it follows that ∫Ωϕθ(ψθϕθ)p′dx<∞ if, p>2s+2s+2. Therefore, in this case, we deduce that
ρ∫Ωfϕdx≤C(ε)∫Ωϕθ(ψθϕθ)p′dx, |
which implies that
ρ≤C(ε)∫Ωϕθ(ψθϕθ)p′dx∫Ωfϕθdx=:ρ∗. |
Hence the result follows in this case.
Remarks 4.5. The condition p>2s+2s+2 in Theorem 4.4 seems to be technical. We conjecture that the non existence result in Theorem 4.4 holds for all p>1. However the above arguments does not hold if p⩽2s+2s+2.
To show the optimality of the exponent p+(λ,s), we show the existence of a supersolution to problem (39). Notice that, in some cases, under suitable conditions on the datum f and the exponent p, we are able to prove the existence of a weak solution to problem (39).
Fix p−(λ,s)<p<p+(λ,s)<2 and let w1(x)=A|x|θ0, with θ0=N−2s2−β, be the solution to the Eq (40) obtained in the previous section. Recall that
(−Δ)sw1(x)−λw1|x|2s=A(γβ,s−λ)|x|θ0+2s=Ap|γβ,s2|p|x|(θ0+s)p=|(−Δ)s2w1|p. |
Taking into consideration the definition of γβ,t given in (41) (with t∈{s2,s}), it holds that (γβ,t−λ)>0 if and only if θ0∈(μ(λ),ˉμ(λ)). Now, if f⩽1|x|2s+θ0, using the fact that Ω is bounded, we can choose C1>0 such that ˆw1=C1w is a supersolution to problem (39) for ρ<ρ∗. In this way we have obtained the following result.
Theorem 4.6. Let Ω be a bounded domain containing the origin. Suppose that p−(λ,s)<p<p+(λ,s). If f⩽1|x|2s+θ, with θ given as above, then problem (39) has asupersolution w such that w,w|x|2s,|(−Δ)s2w1|p∈L1(Ω).
Notice that in order to show the existence of a solution under the presence of a supersolution, we need a comparison principle in the spirit of the work of [14] for the fractional gradient. This is missing at the present time but will be investigated in a forthcoming paper. However, using the compactness approach developed in [7] we are able to show the existence of a solution in some particular cases. More precisely, we have:
Theorem 4.7. Let s∈(0,1), 0<λ<ΛN,s and f∈L1(Ω) be a nonnegative function such that ∫Ωf|x|−μ(λ)−a0dx<∞, for some a0>0. Assume that 1<p<p∗=NN−s. Then, there exists ρ∗:=ρ∗(N,p,s,f,λ,Ω)>0 such that if ρ<ρ∗, problem (39) has asolution u∈Ls,σ0(Ω), for all 1<σ<NN−s.Moreover ∫IRN|(−Δ)s2u|p|x|−μ(λ)dx<∞.
Proof. We follow again the arguments used in [11]. Fix 1<p<p∗ and let f∈L1(Ω) be a nonnegative function with ∫Ωf|x|−μ(λ)−a0dx<∞.
Fix r>1 be fixed such that 1<p<r<p∗. Then, we get the existence of ρ∗>0 such that for some l>0, we have
C0(l+ρ∗||f||L1(|x|−μ(λ)−a0dx,Ω))=l1p, |
where C0 is a positive constant depending only on Ω,λ and the regularity constant in Theorems (2.10).
Let ρ<ρ∗ be fixed and define the set
E={v∈Ls,10(Ω):v∈Ls,r0(|x|−μ(λ)dx,Ω) and ||(−Δ)s2v||Lr(|x|−μ(λ)dx,Ω)≤l1p}. | (50) |
It is clear that E is a closed convex set of Ls,10(Ω). Consider the operator
T:E→Ls,10(Ω)v→T(v)=u, |
where u is the unique solution to problem
{(−Δ)su=λu|x|2s+|(−Δ)s2v|p+ρf in Ω,u=0 in RN∖Ω,u>0 in Ω. | (51) |
Setting
g(x) = |(-\Delta )^{\frac{s}{2}}v|^{p}+\gamma f, |
then taking into consideration the definition of E , it holds that g\in L^1(|x|^{-\mu(\lambda)}dx, \Omega) . Hence the existence and the uniqueness of u follows using the result of [8] with u\in L^{s, \sigma}_0(\Omega) for all {1 < \sigma} < \frac{N}{N-s} . Thus T is well defined.
We claim that T(E)\subset E . Since r > p , using Hölder inequality we get the existence of \hat{a}_0 > 0 such that
\int\limits_ \Omega|(-\Delta )^{\frac{s}{2}}v|^{p}|x|^{-\mu(\lambda )-\hat{a}_0}dx\le C(\Omega)\bigg( \int\limits_ \Omega|(-\Delta )^{\frac{s}{2}}v|^{r}|x|^{-\mu(\lambda )}dx\bigg)^{\frac{p}{r}} < \infty. |
Setting \bar{a}_0 = \min\{a_0, \hat{a}_0\} , it holds that g\in L^1(|x|^{-\mu(\lambda)-\bar{a}}dx, \Omega) . Thus by Proposition 3.7, we reach that, for all 1\le\sigma < \frac{N}{N-s} ,
\bigg( \int\limits_ \Omega|(-\Delta )^{\frac{s}{2}}u|^{\sigma}|x|^{-\mu(\lambda )}dx\bigg)^{\frac{1}{\sigma}}\le C(N,p,\bar{a})\bigg\||(-\Delta )^{\frac{s}{2}}v|^p +\rho f\bigg\|_{L^1(|x|^{-\mu(\lambda )-\bar{a}}dx, \Omega)}. |
Since v\in E , we conclude that
\begin{eqnarray*} \bigg( \int\limits_ \Omega|(-\Delta )^{\frac{s}{2}} u|^{\sigma}|x|^{-\mu(\lambda )}dx\bigg)^{\frac{1}{\sigma}} & \leqslant& C(N,p,\bar{a})\big( \bigg( \int\limits_ \Omega|(-\Delta )^{\frac{s}{2}}v|^{r}|x|^{-\mu(\lambda )}dx\bigg)^{\frac{p}{r}} +\rho ||f||_{L^1(|x|^{-\mu(\lambda )-a_0}dx, \Omega)}\bigg)\\ & \leqslant& C(l+\rho^* ||f||_{L^1(|x|^{-\mu(\lambda )-a_0}dx, \Omega)})\le l. \end{eqnarray*} |
Choosing \sigma = r , it holds that u\in E .
The continuity and the compactness of T follow using closely the same arguments as in [11].
As a conclusion and using the Schauder Fixed Point Theorem as in [11], there exists u\in E such that T(u) = u , u\in L^{s, p}_0(\Omega) and
||(-\Delta )^{\frac{s}{2}}u||_{{L^r(|x|^{-\mu(\lambda )}dx,\Omega)}}\le C. |
Therefore, u solves (39).
Let us consider now the case where \mathfrak{F}(u)\equiv (\mathbb{D}_s (u)) . Then problem (1) takes the form
\begin{equation} \left\{ \begin{array}{rcll} (-\Delta )^s u & = &\lambda \dfrac{u}{|x|^{2s}}+ (\mathbb{D}_s (u))^p+ \rho f & \text{ in } \Omega,\\ u& > &0 & \text{ in }\Omega,\\ u& = &0 & \text{ in }(\mathbb{R}^N\setminus\Omega). \end{array}\right. \end{equation} | (52) |
Recall that \mathbb{D}_s (u)(x) = \bigg(\frac{a_{N, s}}{2} \int_{ {I\!\!R}^N} \frac{|u(x)-u(y)|^2}{|x-y|^{N+2s}} dy \, \bigg)^{\frac 12}. If we consider the equation
\begin{equation} (-\Delta)^s w-\lambda \frac{w}{|x|^{2s}} = (\mathbb{D}_s (w))^p \text{ in } {I\!\!R}^N, \end{equation} | (53) |
then, using the same radial computation as in the previous section, searching for a radial solution in the form w = A|x|^{\beta-\frac{N-2s}{2}} , one sees that we need p = \dfrac{\frac{N-2s}{2}-\beta+2s}{\frac{N-2s}{2}-\beta+s} , which means that \beta = \frac{N-2s}{2}+ \frac{ps}{p-1}-\frac{2s}{p-1} . Hence, as in the previous case, we obtain that w is a solution to (53) if p_-(\lambda, s) < p < p_+(\lambda, s) where p_-(\lambda, s), p_+(\lambda, s) are defined by (42) and (43) respectively.
Notice that if f \leqslant\dfrac{1}{|x|^{2s+\theta}} with \theta = \frac{N-2s}{2}-\beta , then we can chose C_1 > 0 such that C_1w is a supersolution to problem (52) for \rho small enough.
Let us show that p_+(\lambda, s) is the critical exponent for the existence of a weak solution. More precisely we have the next non existence result.
Theorem 5.1. Assume that s\in (0, 1) and p > p_+(\lambda, s) . For \lambda > 0 , problem (52) has no positive solution u in the sense of Definition 4.1.
Proof. We follow closely the arguments in [5]. Without loss of generality we assume that f\in L^\infty(\Omega) .
According to the value of p , we will divide the proof in two parts.
The case p_+(\lambda, s) < p < 2^*_s . In this case p' > \frac{2N}{N+2s} . Assume by contradiction that problem (52) has a weak positive u . Let \phi \in \mathcal{C}_0^{\infty}(\Omega) be a nonnegative function such that \text{Supp}\subset B_{\frac{r}{2}}(0)\subset B_{r}(0)\subset \subset \Omega to be chosen later. Using \phi^{p'} as test function in (52), it holds that
\begin{equation} \int_{\Omega} (-\Delta )^s u \, \phi^{p'}(x) dx \geqslant\int_{\Omega}(\mathbb{D}_s(u)(x))^p\phi^{p'}(x)dx + \lambda \int\limits_ \Omega\frac{u \phi^{p'}}{|x|^{2s}} dx +\rho \int_{\Omega} f(x) \phi^{q'}(x)dx. \end{equation} | (54) |
Using the algebraic inequality, for a, b\ge0, m > 1 ,
(a^m-b^m)\simeq (a-b)(a^{m-1}+b^{m-1}), |
it holds that
\begin{eqnarray*} & & \int_{\Omega} (-\Delta )^s u\, \phi^{p'}(x)dx = \iint_{D_{\Omega}} \frac{(u(x)-u(y))(\phi^{p'}(x) - \phi^{p'}(y))}{|x-y|^{N+2s}} dy dx\\ & \leqslant& C\iint_{D_{\Omega}} \frac{|u(x)-u(y)||\phi(x)-\phi(y)|(\phi^{p'-1}(x)+\phi^{p'-1}(y))}{|x-y|^{N+2s}} dydx \\ & \leqslant& C\iint_{D_{\Omega}} \frac{|u(x)-u(y)| |\phi(x)-\phi(y)|}{|x-y|^{N+2s}}\phi^{p'-1}(x)dydx\\ & + & C\iint_{D_{\Omega}} \frac{|u(x)-u(y)| |\phi(x)-\phi(y)|}{|x-y|^{ N+2s}}\phi^{p'-1}(y)dydx\\ & \leqslant& 2C \iint_{D_{\Omega}} \frac{|u(x)-u(y)| |\phi(x)-\phi(y)|}{|x-y|^{N+2s}}\phi^{p'-1}(x)dydx\\ & \leqslant& C \int_{{I\!\!R}^N} \bigg( \int_{{I\!\!R}^N}\frac{|u(x)-u(y)|^2}{|x-y|^{N+2s}}dy\bigg)^{\frac{1}{2}} \bigg( \int_{{I\!\!R}^N}\frac{|\phi(x)-\phi(y)|^2}{|x-y|^{N+2s}}dy\bigg)^{\frac 12 }\phi^{p'-1}(x)dx\\ & \leqslant& C \int\limits_ \Omega\mathbb{D}_s(u) \mathbb{D}_s(\phi)\phi^{p'-1}(x)dx. \end{eqnarray*} |
Therefore, using Young's inequality, we deduce that for any \varepsilon > 0 , we get the existence of a positive constant C(\varepsilon) such that
\begin{eqnarray*} \int_{\Omega} (-\Delta )^s u\, \phi^{p'}(x)dx & \leqslant& \varepsilon \int\limits_\Omega (\mathbb{D}_s(u))^p\phi^{p'}(x)dx +C( \varepsilon ) \int\limits_\Omega (\mathbb{D}_s(\phi))^{p'}(x)dx. \end{eqnarray*} |
Choosing \varepsilon small enough and going back to (54), we get
\begin{equation} (1- \varepsilon )\int_{\Omega}(\mathbb{D}_s(u)(x))^p\phi^{p'}(x)dx + \lambda \int\limits_ \Omega\frac{u \phi^{p'}}{|x|^{2s}} dx +\rho \int_{\Omega} f(x) \phi^{q'}(x)dx\le C( \varepsilon ) \int\limits_ \Omega(\mathbb{D}_s(\phi)(x))^{p'}dx. \end{equation} | (55) |
Recall that u(x)\ge C|x|^{-\mu(\lambda)} in B_r(0)\subset \subset \Omega . Hence fixed \phi\in \mathcal{C}_0^{\infty}(B_{\frac{r}{2}}(0)) , we have that
\lambda C\int_{B_{\frac{r}{2}}(0)}\frac{\phi^{p'}}{|x|^{2s+\mu(\lambda )}} dx\le C( \varepsilon ) \int\limits_ \Omega(\mathbb{D}_s(\phi)(x))^{p'}dx. |
Replacing \phi by |\phi| in the above estimate, we deduce that
\begin{equation} \lambda C\int_{B_{\frac{r}{2}}(0)}\frac{|\phi|^{p'}}{|x|^{2s+\mu(\lambda )}} dx\le C( \varepsilon ) \int\limits_ \Omega(\mathbb{D}_s(\phi)(x))^{p'}dx \leqslant C( \varepsilon ) \int_{{I\!\!R}^N} (\mathbb{D}_s(\phi)(x))^{p'}dx. \end{equation} | (56) |
Since p > p_+(\lambda, s) , then sp' < 2s+\mu(\lambda) . Recall that p' > \frac{2N}{N+2s} . Hence (56) is in contradiction with the Hardy inequality in Proposition 2.3. Thus we conclude.
The case p > 2^*_s > p_+(\lambda, s) . Notice that for all \lambda < \Lambda_{N, s} , we have p_+(\lambda, s) < 2 < 2^*_s . By a continuity argument we get the existence of \lambda_1 < \lambda and p_1 < 2^*_s such that p_1 > p_+(\lambda_1, s) . Assume that u is a weak solution to problem (52), then
(-\Delta )^s u \geqslant\lambda \dfrac{u}{|x|^{2s}}+ (\mathbb{D}_s (u))^{p_1}-C(p_1) \text{ in } B_r(0). |
Notice that u(x)\ge C_1|x|^{-\mu(\lambda)} \text{ in } B_r(0) . Hence
(-\Delta )^s u \geqslant\lambda_1 \dfrac{u}{|x|^{2s}}+ (\mathbb{D}_s (u))^{p_1}+\frac{C}{|x|^{2s+\mu(\lambda )}}-C(p_1) \text{ in } B_r(0). |
Choosing r small, it holds that
(-\Delta )^s u \geqslant\lambda_1 \dfrac{u}{|x|^{2s}}+ (\mathbb{D}_s (u))^{p_1} \text{ in } B_r(0). |
Since p_+(\lambda_1, s) < p_1 < \frac{2N}{N+2s} , repeating the same argument as in the first case, we reach the same contradiction. Hence we conclude.
Taking advantage of the previous estimate, we can show that the problem (52) has no solution for large value of \rho .
Theorem 5.2. Assume that f\gneqq 0 and p > 1 , then there exists \rho^* > 0 such that problem (52) does not have a positive solutionfor \rho > \rho^* .
Proof. Suppose that u is a nonnegative weak solution to problem (52). Let \phi \in \mathcal{C}_0^{\infty}(\Omega) be a nonnegative function such that
\int_{\Omega} f(x) \phi^{q'}(x) dx > 0. |
From estimate (55), fixing \varepsilon \in(0, 1) , we obtain that
\rho \int_{\Omega} f(x) \phi^{q'}(x)dx\le C( \varepsilon ) \int\limits_ \Omega(\mathbb{D}_s(\phi)(x))^{p'}dx. |
In particular
\rho \leqslant \inf\limits_{\{\phi\in \mathcal{C}^\infty(\Omega), \phi\gneqq 0\}}\frac{ C( \varepsilon ) \int\limits_\Omega(\mathbb{D}_s(\phi))^{p'}(x)dx}{ \int_{\Omega} f(x) \phi^{q'}(x) dx}: = \rho^*, |
and this is in contradiction with our initial assumption.
Remarks 5.3. 1) As in Theorem 4.6, if 1 < p < p_+(\lambda, s) and f\le\dfrac{1}{|x|^{2s+\theta}} , with \theta given as above, thenproblem (39) has a supersolution w such that w, \dfrac{w}{|x|^{ 2s}}, \mathbb{D}_s (u)\in L^p(\Omega) .
2) Using the same compactness approach, we can also treat thecase (\mathfrak{F}(u)(x)) = |\nabla^s u(x)| , where \nabla^s u(x) isdefined in (4).
1) In the local case s = 1 or in the nonlocal case under the existence of a local gradient term, an interesting maximum principle is obtained in the sense that if w\in W^{s, 1}_0(\Omega) is a subsolution to the problem
\begin{equation} \left\{ \begin{array}{rcll} (-\Delta )^s w & = &a(x)|\nabla w| & \text{ in } \Omega,\\ w& = &0 & \text{ in }(\mathbb{R}^N\setminus\Omega), \end{array}\right. \end{equation} | (57) |
with a\in L^{\sigma}(\Omega), \sigma > \frac{N}{s} , then w \leqslant0 in \Omega (see for instance [14] and [11]). It would be very interesting to get a similar result replacing the gradient term |\nabla w| by the nonlocal fractional gradient |(-\Delta)^{\frac{s}{2}}w| , namely for the problem
\begin{equation} \left\{ \begin{array}{rcll} (-\Delta )^s w & = &a(x)|(-\Delta )^{\frac{s}{2}}w| & \text{ in } \Omega,\\ w& = &0 & \text{ in }(\mathbb{R}^N\setminus\Omega). \end{array}\right. \end{equation} | (58) |
2) Since non comparison principle is known for problem (58), then to get general existence result to problem (1), under natural integrability conditions for f , it is necessary to prove a new class of weighted CKN inequalities as in [3], using the norm \| |(-\Delta)^{\frac{s}{2}}u| |x|^{\beta} \|_{L^p(\mathbb{R}^N)} . This will be considered in a forthcoming work.
The work was partially supported by AEI Research Grant PID2019-110712GB-I00 and grant 1001150189 by PRICIT, Spain. Authors 1 and 2 were also supported by a research project from DGRSDT, Algeria.
The authors would like to thank the anonymous reviewer for his/her careful reading of the paper and his/her many insightful comments and suggestions.
The authors declare no conflict of interest.
[1] |
Carthew RW, Rubin GM (1990) Seven in absentia, a gene required for specification of R7 cell fate in the Drosophila eye. Cell 63: 561-577. doi: 10.1016/0092-8674(90)90452-K
![]() |
[2] | Della NG, Senior PV, Bowtell DL (1993) Isolation and characterization of murine homologues of the Drosophila seven in absentia gene (Sina). Development 117: 1333-1343. |
[3] |
Holloway AJ, Della NG, Fletcher CF, et al. (1997) Chromosomal mapping of five highly conserved murine homologues of the Drosophila RING finger gene seven-in-absentia. Genomics 41: 160-168. doi: 10.1006/geno.1997.4642
![]() |
[4] |
House CM, Hancock NC, Moller A, et al. (2006) Elucidation of the substrate binding site of Siah ubiquitin ligase. Structure 14: 695-701. doi: 10.1016/j.str.2005.12.013
![]() |
[5] |
House CM, Frew IJ, Huang HL, et al. (2003) A binding motif for Siah ubiquitin ligase. Proc Natl Acad Sci USA 100: 3101-3106. doi: 10.1073/pnas.0534783100
![]() |
[6] |
Khurana A, Nakayama K, Williams S, et al. (2006) Regulation of the ring finger E3 ligase Siah2 by p38 MAPK. J Biol Chem 281: 35316-35326. doi: 10.1074/jbc.M606568200
![]() |
[7] |
Scortegagna M, Subtil T, Qi J, et al. (2011) USP13 enzyme regulates Siah2 ligase stability and activity via noncatalytic ubiquitin-binding domains. J Biol Chem 286: 27333-27341. doi: 10.1074/jbc.M111.218214
![]() |
[8] |
Matsuzawa SI, Reed JC (2001) Siah-1, SIP, and Ebi Collaborate in a Novel Pathway for b-Catenin Degradation Linked to p53 Responses. Mol Cell 7: 915-926. doi: 10.1016/S1097-2765(01)00242-8
![]() |
[9] |
Winter M, Sombroek D, Dauth I, et al. (2008) Control of HIPK2 stability by ubiquitin ligase Siah-1 and checkpoint kinases ATM and ATR. Nat Cell Biol 10: 812-824. doi: 10.1038/ncb1743
![]() |
[10] | Zundel W, Schindler C, Haas-Kogan D, et al. (2000) Loss of PTEN facilitates HIF-1-mediated gene expression. Genes Dev 14: 391-396. |
[11] |
Carthew RW, Neufeld TP, Rubin GM (1994) Identification of genes that interact with the sina gene in Drosophila eye development. Proc Natl Acad Sci USA 91: 11689-11693. doi: 10.1073/pnas.91.24.11689
![]() |
[12] |
Tang AH, Neufeld TP, Kwan E, et al. (1997) PHYL acts to down-regulate TTK88, a transcriptional repressor of neuronal cell fates, by a SINA-dependent mechanism. Cell 90: 459-467. doi: 10.1016/S0092-8674(00)80506-1
![]() |
[13] | Bogdan S, Senkel S, Esser F, et al. (2001) Misexpression of Xsiah-2 induces a small eye phenotype in Xenopus. Mech Dev 103: 61-69. |
[14] |
Zhang J, Guenther MG, Carthew RW, et al. (1998) Proteasomal regulation of nuclear receptor corepressor-mediated repression. Genes Dev 12: 1775-1780. doi: 10.1101/gad.12.12.1775
![]() |
[15] |
D'Orazi G, Cecchinelli B, Bruno T, et al. (2002) Homeodomain-interacting protein kinase-2 phosphorylates p53 at Ser 46 and mediates apoptosis. Nat Cell Biol 4: 11-19. doi: 10.1038/ncb714
![]() |
[16] |
Johnsen SA, Subramaniam M, Monroe DG, et al. (2002) Modulation of transforming growth factor beta (TGFbeta)/Smad transcriptional responses through targeted degradation of TGFbeta-inducible early gene-1 by human seven in absentia homologue. J Biol Chem 277: 30754-30759. doi: 10.1074/jbc.M204812200
![]() |
[17] | Hu G, Fearon ER (1999) Siah-1 N-terminal RING domain is required for proteolysis function, and C-terminal sequences regulate oligomerization and binding to target proteins. Mol Cell Biol 1: 19724-19732. |
[18] |
Maxwell PH, Ratciffe PJ (2002) Oxygen sensors and angiogenesis. Semin Cell Dev Biol 13: 29-37. doi: 10.1006/scdb.2001.0287
![]() |
[19] |
McNeill LA, Hewitson KS, Gleadle JM (2002) The use of dioxygen by HIF prolyl hydroxylase (PHD1). Bioorg Med Chem Lett 12: 1547-1550. doi: 10.1016/S0960-894X(02)00219-6
![]() |
[20] |
Qi J, Nakayama K, Gaitonde S, et al. (2008) Siah2-dependent concerted activity of HIF and FoxA2 regulates formation of neuroendocrine phenotype and neuroendocrine prostate tumors. Proc Natl Acad Sci USA 105: 16713-16718. doi: 10.1073/pnas.0804063105
![]() |
[21] |
Baba K, Morimoto H, Imaoka S (2013) Seven in absentia homolog 2 (Siah2) protein is a regulator of NF-E2-related factor 2 (Nrf2). J Biol Chem 288: 18393-18405. doi: 10.1074/jbc.M112.438762
![]() |
[22] |
Sajja RK, Green KN, Cucullo L (2015) Altered Nrf2 signaling mediates hypoglycemia-induced blood-brain barrier endothelial dysfunction in vitro. PLos One 10: e0122358. doi: 10.1371/journal.pone.0122358
![]() |
[23] |
Itoh K, Wakabayashi N, Katoh Y, et al. (1999) Keap1 represses nuclear activation of antioxidant responsive elements by Nrf2 through binding to the amino-terminal Neh2 domain. Genes Dev 13: 76-86. doi: 10.1101/gad.13.1.76
![]() |
[24] |
Kim H, Scimia MC, Wilkinson D (2011) Fine-tuning of Drp1/Fis1 availability by AKAP121/Siah2 regulates mitochondrial adaptation to hypoxia. Mol Cell 44: 532-544 doi: 10.1016/j.molcel.2011.08.045
![]() |
[25] |
Li Q, Wang P, Ye K, et al. (2015) Central role of SIAH inhibition in DCC-dependent cardioprotection provoked by netrin-1/NO. Proc Natl Acad USA 112: 899-904. doi: 10.1073/pnas.1420695112
![]() |
[26] |
Schmidt RL, Park CH, Ahmed AU, et al. (2007) Inhibition of RAS-mediated transformation and tumorigenesis by targeting the downstream E3 ubiquitin ligase seven in absentia homologue. Cancer Res 67: 11798-11810. doi: 10.1158/0008-5472.CAN-06-4471
![]() |
[27] |
Kim CJ, Cho YG, Park CH, et al. (2004) Inactivating mutations of the Siah-1 gene in gastric cancer. Oncogene 23: 8591-8596. doi: 10.1038/sj.onc.1208113
![]() |
[28] |
Ahmed AU, Sxhmidt RL, Park CH, et al. (2008) Effect of disrupting seven-in-absentia homolog 2 function on lung cancer cell growth. J Natl Cancer Inst 100: 1606-29. doi: 10.1093/jnci/djn365
![]() |
[29] |
Wong CS, Sceneay J, House CM, et al. (2012) Vascular normalization by loss of Siah2 results an increased chemotherapeutic efficacy. Cancer Res 72: 1694-1704. doi: 10.1158/0008-5472.CAN-11-3310
![]() |
[30] |
Shah M, Stebbins JL, Dewing A, et al. (2009) Inhibition of Siah2 ubiquitin ligase by vitamin K3 (menadione) attenuates hypoxia and MAPK signaling and blocks melanoma tumorigenesis. Pigment Cell Melanoma Res 22: 799-808. doi: 10.1111/j.1755-148X.2009.00628.x
![]() |
[31] |
Qi J, Nakayama K, Cardiff RD, et al. (2010) Siah2-dependent concerted activity of HIF and FoxA2 regulates formation of neuroendocrine phenotype and neuroendocrine prostate tumors. Cancer Cell 18: 23-38. doi: 10.1016/j.ccr.2010.05.024
![]() |
[32] | Yoshibayashi H, Okabe H, Satoh S, et al. (2007) SIAH1 causes growth arrst and apoptosis in hepatoma cells through beta-catenin degradation-dependent and –independent mechanisms. Oncol Rep 17: 549-556. |
[33] |
Roperch J, Lethrone F, Prieur S, et al. (1999) SIAH-1 promotes apoptosis and tumor suppression through a network involving the regulation of protein folding, unfolding, and trafficking: identification of common effectors with p53 and p21(Waf1). Proc Natl Acad USA 96: 8070-8073. doi: 10.1073/pnas.96.14.8070
![]() |
[34] |
Malz M, Aulmann A, Samarin J, et al. (2012) Nuclear accumulation of seven in absentia homologue-2 supports motility and proliferation of liver cancer cells. Int J Cancer 131: 2016-2026. doi: 10.1002/ijc.27473
![]() |
[35] |
Wong CS, Moller A (2013) Siah: a promising anticancer target. Cancer Res 73: 2400-2406. doi: 10.1158/0008-5472.CAN-12-4348
![]() |
[36] |
DeBruyne JP, Baggs JE, Sato TK (2015) Ubiquitin ligase Siah2 regulates RevErba degradation and the mammalian circadian clock. Proc Natl Acad Sci USA 112: 12420-12425. doi: 10.1073/pnas.1501204112
![]() |
[37] |
Adam MG, Matt S, Christian S (2015) SIAH ubiquitin ligases regulate breast cancer cell migration and invasion independent of the oxygen status. Cell Cycle 14: 3734-3747. doi: 10.1080/15384101.2015.1104441
![]() |
[38] |
Frasor J, Danes JM, Funk CC, et al. (2005) Estrogen down-regulation of the corepressor N-CoR: mechanism and implications for estrogen derepression of N-CoR-regulated genes. Proc Natl Acad Sci USA 102: 13153-13157. doi: 10.1073/pnas.0502782102
![]() |