Processing math: 100%
Research article

Existence of infinitely many solutions for a nonlocal problem

  • Received: 12 May 2020 Accepted: 22 June 2020 Published: 10 July 2020
  • MSC : 35B33, 35J60

  • In this paper, we deal with a class of fractional Hénon equation and by using the Lyapunov-Schmidt reduction method, under some suitable assumptions, we derive the existence of infinitely many solutions, whose energy can be made arbitrarily large. Compared to the previous works, we encounter some new challenges because of the nonlocal property for fractional Laplacian. But by doing some delicate estimates for the nonlocal term we overcome the difficulty and find infinitely many nonradial solutions.

    Citation: Jing Yang. Existence of infinitely many solutions for a nonlocal problem[J]. AIMS Mathematics, 2020, 5(6): 5743-5767. doi: 10.3934/math.2020369

    Related Papers:

    [1] Ramzi Alsaedi . Infinitely many solutions for a class of fractional Robin problems with variable exponents. AIMS Mathematics, 2021, 6(9): 9277-9289. doi: 10.3934/math.2021539
    [2] Zonghu Xiu, Shengjun Li, Zhigang Wang . Existence of infinitely many small solutions for fractional Schrödinger-Poisson systems with sign-changing potential and local nonlinearity. AIMS Mathematics, 2020, 5(6): 6902-6912. doi: 10.3934/math.2020442
    [3] Zhilin Li, Guoping Chen, Weiwei Long, Xinyuan Pan . Variational approach to p-Laplacian fractional differential equations with instantaneous and non-instantaneous impulses. AIMS Mathematics, 2022, 7(9): 16986-17000. doi: 10.3934/math.2022933
    [4] Najla Alghamdi, Abdeljabbar Ghanmi . Multiple solutions for a singular fractional Kirchhoff problem with variable exponents. AIMS Mathematics, 2025, 10(1): 826-838. doi: 10.3934/math.2025039
    [5] Mengyu Wang, Xinmin Qu, Huiqin Lu . Ground state sign-changing solutions for fractional Laplacian equations with critical nonlinearity. AIMS Mathematics, 2021, 6(5): 5028-5039. doi: 10.3934/math.2021297
    [6] Zhiqiang Li . The finite time blow-up for Caputo-Hadamard fractional diffusion equation involving nonlinear memory. AIMS Mathematics, 2022, 7(7): 12913-12934. doi: 10.3934/math.2022715
    [7] Xueqin Peng, Gao Jia, Chen Huang . Multiplicity of solutions for Schrödinger-Poisson system with critical exponent in R3. AIMS Mathematics, 2021, 6(3): 2059-2077. doi: 10.3934/math.2021126
    [8] Khaled Kefi, Abdeljabbar Ghanmi, Abdelhakim Sahbani, Mohammed M. Al-Shomrani . Infinitely many solutions for a critical p(x)-Kirchhoff equation with Steklov boundary value. AIMS Mathematics, 2024, 9(10): 28361-28378. doi: 10.3934/math.20241376
    [9] Lixiong Wang, Ting Liu . Existence and regularity results for critical (p,2)-Laplacian equation. AIMS Mathematics, 2024, 9(11): 30186-30213. doi: 10.3934/math.20241458
    [10] Xianyong Yang, Zhipeng Yang . Existence of multiple non-trivial solutions for a nonlocal problem. AIMS Mathematics, 2019, 4(2): 299-307. doi: 10.3934/math.2018.2.299
  • In this paper, we deal with a class of fractional Hénon equation and by using the Lyapunov-Schmidt reduction method, under some suitable assumptions, we derive the existence of infinitely many solutions, whose energy can be made arbitrarily large. Compared to the previous works, we encounter some new challenges because of the nonlocal property for fractional Laplacian. But by doing some delicate estimates for the nonlocal term we overcome the difficulty and find infinitely many nonradial solutions.


    In this paper, we consider the following nonlocal Hénon equation

    {Asu=|x|αup,   u>0xB1(0),u=0,onB1(0) (1.1)

    with critical growth, where α>0 is a positive constant, p=n+2sn2s,n2+2s, 12<s<1, B1(0) is the unit ball in Rn and As stands for the fractional Laplacian operator in B1(0) with zero Dirichlet boundary values on B1(0).

    Here, to define the fractional Laplacian operator As in B1(0), let {λk,φk} be the eigenvalues and corresponding eigenfunctions of the Laplacian operator Δ in B1(0) with zero Dirichlet boundary values on B1(0), namely, {λk,φk} satisfies

    {Δφk=λkφk,inB1(0),φk=0,onB1(0)

    with φkL2(B1(0))=1. Then we can define the fractional Laplacian operator As: Hs0(B1(0))Hs0(B1(0)) as

    Asu=k=1λskckφk,

    where the fractional Sobolev space Hs0(B1(0))(0<s<1) is given by

    Hs0(B1(0))={u=k=1ckφkL2(B1(0)):k=1λskc2k<}

    and equipped with the following inner product

    k=1ckφk,k=1dkφkHs0(B1(0))=k=1λskckdk.

    Form the above definitions, it immediately follows that for any u,vHs0(B1(0)),

    u,vHs0(B1(0))=B1(0)A12suA12sv=B1(0)Asuv.

    It is well known that the nonlinear fractional equations appear in diverse areas including physics, biological modeling and mathematical finances and have attracted the considerable attention in the recent period. Also in recent years, there have been many investigations for the related fractional problem Asu=f(u), where f:RnR is a certain function. But a complete review of the available results in this context goes beyond the aim of this paper. Here we just mention some very recent papers which study fractional equations involving the critical sobolev exponent (cf. [3,7,19,20,21]).

    On the other hand, our main interest in the present paper is motivated by some works that have appeared in recent years related to the classical local Hénon equation of this kind,

    {Δu=|x|αup,   u>0xB1(0),u=0,onB1(0). (1.2)

    Among pioneer works we mention Ni [13], where the author established a compactness result of H10,rad(B1(0))Lp+1(B1(0)) and thus got the existence of one positive radial solution for (1.2) if p(1,n+2+2αn2). Later, in [18], Smets, Su and Willem established some symmetry breaking phenomenon and obtained the non-radial property of the ground state solution of (1.2) if 1<p<n+2n2 and α is large enough. When n3 and p=n+2n2σ, Cao and Peng [1] verified that the ground state solutions of (1.2) are non-radial and blow up as σ0. Meanwhile, when p=n+2n2, Serra [17] showed that (1.2) has a non-radial solution if n4 and α is large enough. More recently, Wei and Yan [23] proved the existence of infinitely many non-radial solutions for (1.2) for any α>0. For other results related to the Hénon problem (1.2), one can refer to [2,11,14,15] and the references therein.

    Up to our knowledge, not much is obtained for the existence of multiple solutions of equation (1.2) with fractional operator. Motivated by [23] and [12], we want to exploit the finite dimensional reduction method to investigate the existence of infinitely many non-radial solutions for (1.1). To achieve our aim, we will study the following more general problem

    {Asu=Φ(|x|)up,   u>0xB1(0),u=0,onB1(0), (1.3)

    where Φ(r) is a bounded function defined in [0,1]. It is easy to check that a necessary condition for the existence of a solution of (1.3) is that Φ(r) is positive somewhere from Pohozaev identity (see[16]). At this point we call attention to the recent work of [12], where we studied (1.3) in Rn and proved that if n>2+2s,0<s<1 and Φ(|x|) satisfies

    Φ(r)=Φ0arm+O(1rm+θ)asr(r0δ,r0+δ) (1.4)

    with max{2,n2s2(n2s)2n+2s}<m<n2s,Φ0>0,r0>0,θ>0, δ>0, then (1.3) has infinitely many non-radial solutions. It is worth mentioning that from assumption (1.4), r0 is a local maximum point of Φ(r) and then a critical point of Φ(r). Also the function rα achieves its maximum on [0,1] at r0=1 but r0=1 is not a critical point of rα. However we will verify that if Φ(r) is increasing near r0=1, through r0=1 is not a critical point of Φ(r), the zero Dirichlet boundary condition makes it possible to construct infinitely many solutions of (1.3). Now we state our main result as follows:

    Theorem 1.1. Suppose that n2+2s,12<s<1. If Φ(1)>0 and Φ(1)>0, then problem (1.3) has infinitely many non-radial solutions. Particularly, the Hénon equation (1.1) has infinitely many non-radial solutions.

    In the end of this part, let us outline the main idea in the proof of Theorem 1.1.

    Given any ε>0 and y0Rn, let

    Uε,y0(x)=σn,s(εε2+|xy0|2)n2s2

    for xRn and σn,s=2n2s2(Γ(n+2s2)Γ(n2s2))n2s4s. In 1983, Lieb [10] (also see [6,7,8,9]) proved that Uε,y0(x) solves the following critical fractional equation

    Asu=up,  lim|x|u(x)=0,  u>0  in   Rn. (1.5)

    Also, very recently, J. DÁvila, M. del Pino and Y. Sire [5] obtained the non-degeneracy of Uε,y0(x). More precisely, if we define the corresponding functional of (1.5) as

    I0(u)=12Rn|A12su|21pRn|u|p,

    then I0 possesses a finite-dimensional manifold Z of least energy critical points, given by

    Z={Uε,y0:ε>0,y0Rn}

    and

    kerI0(u)=spanR{Uε,y0y01,,Uε,y0y0n,Uε,y0ε},Uε,y0Z.

    Now let us fix a positive integer kk0, where k0 is large, which is to be determined later and set

    ν=kn2s+1n2s,

    to be the scaling parameter. Using the transformation u(x)νn2s2u(xν), (1.3) becomes

    {Asu=Φ(|x|ν)up,   u>0xBν(0),u=0,onBν(0). (1.6)

    Since Uε,y0 is not zero on Bν(0), we define PUε,y0 as the solution of the following problem

    AsPUε,y0=AsUε,y0  in Bν(0),   PUε,y0=0 on Bν(0), (1.7)

    and we will use the solution PUε,y0 to build up the approximate solutions for (1.6).

    For x=(x,x)R2×Rn2, we define

    Hk={u:uHs0(Bν(0)),u is even in xj,j=2,,n,u(rcosθ,rsinθ,x)=u(rcos(θ+2iπk),rsin(θ+2iπk),x)}.

    Also, we denote

    Uε,r(x)=ki=1PUε,xi(x),

    where

    xi=(rcos2(i1)πk,rsin2(i1)πk,0),  i=1,,k

    with 0 is the zero vector in Rn2. And throughout this paper, we always assume that r[ν(1r0k),ν(1r1k)], ε0εε1 for some constants r1>r0>0 and ε1>ε0>0.

    To prove Theorem 1.1, it suffices to verify the following result:

    Theorem 1.2. Under the assumption of Theorem 1.1, there is an integer k0>0, such that for any integer kk0, (1.6) has a solution uk of the form

    uk=Uε,rk(x)+ωk

    where ωkHk, and as k+,ωkL(Bν(0))0, rk[ν(1r0k),ν(1r1k)], ε0εε1.

    We want to point out that compared with [23], due to the fact that the fractional Laplacian operator is nonlocal and very few things on this topic are known about the fractional Laplacian, we have to face much difficulties in the reduction process and need some more delicate estimates in the proof of our results.

    The rest of the paper is organized as follows. In Section 2, we will carry out a reduction procedure and we prove our main result in Section 3. Finally, in Appendix, some basic estimates and an energy expansion for the functional corresponding to problem (1.6) will be established.

    In this section, we perform a finite-dimensional reduction. Let

    u=supxBν(0)(ki=11(1+|xxi|)n2s2+τ)1|u(x)| (2.1)

    and

    f=supxBν(0)(ni=11(1+|xxi|)n+2s2+τ)1|f(x)|, (2.2)

    where τ=n2sn2s+1. For this choice of τ, we find that

    ki=21|xix1|τCkτντki=21iτC.

    Let

    Zi,1=PUε,xir,   Zi,2=PUε,xiε.

    Now we consider

    {AsφkpΦ(|x|ν)Up1ε,rφk=gk+2l=1clki=1Up1ε,xiZi,l,in Bν(0),φkHk,Up1ε,xiZi,l,φk=0,  i=1,...,k,l=1,2 (2.3)

    for some numbers cl, where u,v=Bν(0)uv. Then we have

    Lemma 2.1. Suppose that φk solves (2.3) for g=gk. If gk goes to zero as k goes to infinity, so does φk.

    Proof. We will argue by an indirect method. Suppose by contradiction that there exist k+,g=gk,εk[ε0,ε1],rk[ν(1r0k),ν(1r1k)] and φk solving (2.3) for g=gk,ε=εk,r=rk with gk0 and φkc>0. We may assume that φk=1. Also for simplicity, we drop the subscript k.

    Note that we can rewrite (2.3) as

    φ(x)=pBν(0)1|yx|n2sΦ(|y|ν)Up1ε,r(y)φ(y)dy+Bν(0)1|yx|n2s(g(y)+2l=1clki=1Up1xi,ε(y)Zi,l(y))dy. (2.4)

    Now we estimate each terms in (2.4). Analogous to Lemma A.3, we have

    |Bν(0)1|yx|n2sΦ(|y|ν)Up1ε,r(y)φ(y)dy|CφBν(0)1|yx|n2sUp1ε,r(y)ki=11(1+|yxi|)n2s2+τdyCφki=11(1+|xxi|)n2s2+τ+ϵ. (2.5)

    Meanwhile, by Lemma A.2, we get

    |Bν(0)1|xy|n2sg(y)dy|CgBν(0)1|xy|n2ski=11(1+|yxi|)n+2s2+τdyCgki=11(1+|xxi|)n2s2+τ (2.6)

    and

    |Bν(0)1|xy|n2ski=1Up1ε,xi(y)Zi,l(y)dy|CBν(0)1|xy|n2ski=1Upε,xi(y)dyCBν(0)1|xy|n2ski=11(1+|yxi|)n+2sdyCki=11(1+|xxi|)n2s2+τ. (2.7)

    Next, we estimate cl,l=1,2. Multiplying (2.3) by Z1,t(t=1,2) and integrating, we see that cl satisfies

    2l=1clki=1Up1ε,xiZi,l,Z1,t=AsφpΦ(|y|ν)Up1ε,rφ,Z1,tg,Z1,t. (2.8)

    First, it follows from Lemma A.1 that

    |g,Z1,t|CgBν(0)1(1+|yx1|)n2ski=11(1+|yxi|)n+2s2+τCg.

    On the other hand,

    AsφpΦ(|y|ν)Up1ε,rφ,Z1,t=AsZ1,tpΦ(|y|ν)Up1ε,rZ1,t,φ=pUp1x1,εZ1,tpΦ(|y|ν)Up1ε,rZ1,t,φCφBν(0)|Φ(|y|ν)1|(ki=11(1+|yxi|)n2s)p11(1+|yx1|)n2ski=11(1+|yxi|)n2s2+τ=:J0. (2.9)

    Define

    Ωi={x=(x,x)Bν(0):x|x|,(xi)|(xi)|cosπk},i=1,2,,k.

    Observe that for yΩ1, |yxi||yx1| and then

    ki=21(1+|yxi|)n2sC1(1+|yx1|)n2s2ki=21(1+|yxi|)n2s2C1(1+|yx1|)n2sτki=21|x1xi|τC(kν)τ1(1+|yx1|)n2sτ,

    which implies

    (ki=11(1+|yxi|)n2s)4sn2sC1(1+|yx1|)4s4sτn2s

    and

    ki=11(1+|yxi|)n2s2+τC1(1+|yx1|)n2s2.

    As a result, from (2.9), we have

    J0CkφBν(0)|Φ(|y|ν)1|1(1+|yx1|)3n2+s4sn2s. (2.10)

    Using the same argument used for proving (A.7), it follows from (2.9) and (2.10) that

    AsφpΦ(|y|ν)Up1ε,rφ,Z1,t=o(φ). (2.11)

    Finally, we have

    ki=1Up1ε,xiZi,l,Z1,tCki=1Bν(0)(1(1+|xxi|)n2s)p1(1+|xx1|)n2sdxC.

    Hence using (2.8), we get

    cl=o(φ)+O(g). (2.12)

    So, combining (2.4)-(2.7) and (2.12), one has

    φ(g+ki=11(1+|yxi|)n2s2+τ+ϵki=11(1+|yxi|)n2s2+τ). (2.13)

    Being φ=1, we obtain from (2.13) that there is R>0 such that

    φ(x)L(BR(xi))a>0, (2.14)

    for some i. But, by using (2.3), ˜φ(x)=φ(xxi) converges, uniformly in any compact set, to a solution ϕ of the following equation

    AsϕpUp1ε,0ϕ=0,inRn

    for some ε[ε0,ε1]. Due to the non-degeneracy of Uε,0, we can infer that ϕ=0, which yields a contradiction with (2.14) and then this proof has been proved.

    From Lemma 2.1, arguing as proving Proposition 4.1 in [6] or Proposition 2.2 in [12], we can show the following result.

    Proposition 2.2. There exists k0>0 and a constant C>0, independent of k, such that for all kk0, and gL(Rn), problem (2.3) has a unique solution φ=Lk(g). Also

    Lk(g)Cg

    and

    |cl|Cg.

    To prove our results, we consider

    {As(Uε,r+φ)=Φ(|x|ν)(Uε,r+φ)p+2l=1clki=1Up1ε,xiZi,l,in Bν(0),φHk,Up1ε,xiZi,l,φ=0,  i=1,...,k,l=1,2. (2.15)

    In order to use the contraction mapping theorem to prove that (2.15) is uniquely solvable in the set that φ is small, we rewrite (2.15) as

    {AsφpΦ(|x|ν)Up1ε,rφ=N(φ)+lk+2l=1clki=1Up1ε,xiZi,l,in Bν(0),φHk,Up1ε,xiZi,l,φ=0,  i=1,...,k,l=1,2, (2.16)

    where

    N(φ)=Φ(|x|ν)((Uε,r+φ)pUpε,rpUp1ε,rφ)

    and

    lk=Φ(|x|ν)Upε,rki=1Upε,xi.

    Lemma 2.3. There holds

    N(φ)Cφmin{p,2}.

    Proof. Firstly, we deal with the case p2. By Hölder inequality, we find

    |N(φ)|Cφp(ki=11(1+|xxi|)n2s2+τ)pCφpki=11(1+|xxi|)n+2s2+τ(ki=11(1+|xxi|)τ)4sn2sCφpki=11(1+|yxi|)n+2s2+τ.

    Using the same argument as above, if p>2, we also have

    |N(φ)|Cφ2(ki=11(1+|xxi|)n2s)p2(kj=11(1+|xxj|)n2s2+τ)2+φp(ki=11(1+|xxi|)n2s2+τ)pC(φp+φ2)(ki=11(1+|xxi|)n2s2+τ)pCφ2ki=11(1+|xxi|)n+2s2+τ,

    which completes our proof.

    Lemma 2.4. Assume that r[ν(1r0k),ν(1r1k)]. Then there is a small ϵ>0, such that

    lkC(1ν)12+ϵ.

    Proof. Recall that

    Ωi={x=(x,x)Bν(0):x|x|,(xi)|(xi)|cosπk}

    and

    lk=Φ(|x|ν)(Upε,rki=1(PUε,xi)p)+Φ(|x|ν)(ki=1(PUε,xi)pki=1Upε,xi)+ki=1Upε,xi(Φ(|x|ν)1)=:J1+J2+J3.

    By symmetry, we can suppose that xΩ1 and for any xΩ1,

    |xxi||xx1|.

    Thus

    |J1|C1(1+|xx1|)4ski=21(1+|xxi|)n2s+C(ki=21(1+|xxi|)n2s)p. (2.17)

    By Lemma A.1, taking any 1<θn+2s2, we obtain that for any xΩ1,

    1(1+|xx1|)4ski=21(1+|xxi|)n2s1(1+|xx1|)n+2s2ki=21(1+|xxi|)n+2s2Cki=2[1(1+|xx1|)n+2sθ+1(1+|xxi|)n+2sθ]1|xix1|θC1(1+|xx1|)n+2sθki=21|xix1|θC(1+|xxi|)n+2sθ(kν)θ.

    Choosing θ>n2s+12 with n+2sθn+2s2+τ, we have

    1(1+|xx1|)4ski=21(1+|xxi|)n2sC(1+|xx1|)n+2s2+τ(1ν)12+ϵ. (2.18)

    On the other hand, for xΩ1, by Lemma A.1 again, we get

    1(1+|xxi|)n2s1(1+|xx1|)n2s21(1+|xxi|)n2s2C|xix1|n2s2n2sn+2sτ(1(1+|xx1|)n2s2+n2sn+2sτ+1(1+|xxi|)n2s2+n2sn+2sτ)C|xix1|n2s2n2sn+2sτ1(1+|xx1|)n2s2+n2sn+2sτ.

    So

    ki=21(1+|xxi|)n2sC(kν)n2s2n2sn+2sτ1(1+|xx1|)n2s2+n2sn+2sτ,

    which gives

    (ki=21(1+|xxi|)n2s)pC(kν)n+2s2τ1(1+|xx1|)n+2s2+τ. (2.19)

    Combining (2.17)-(2.19), we have

    J1C(1ν)12+ϵ.

    Now, we estimate J2. Let H(x,y) be the regular part of the Green function for As in B1(0) with the Dirichlet boundary condition (see the definition in Appendix) and let ˉxi be the reflection point of ˉxi with respect to B1(0), where ˉxi=xiν. Then

    H(ˉx,ˉxi)νn2s=Cνn2s|ˉxˉxi|n2sC(1+|xxi|)n2s.

    Take κ=1θ with θ>0 small. By (A.1), we have

    |J2|ki=1C(1+|xxi|)4sH(ˉx,ˉxi)νn2ski=1C(1+|xxi|)4s+κ(n2s)(H(ˉx,ˉxi)νn2s)κC(1ν)κ(n2s)ki=11(1+|xxi|)4s+κ(n2s)C(1ν)12+ϵki=11(1+|xxi|)4s+κ(n2s)C(1ν)12+ϵki=11(1+|xxi|)n+2s2+τ,

    since θn2sn2s+1>12 for n2+2s.

    Finally, we estimate J3. For any xΩ1 and i=2,,k, applying Lemma A.1, we get

    Upε,xi(x)C1(1+|xx1|)n+2s21(1+|xxi|)n+2s2C(1(1+|xx1|)n+2s2+τ+1(1+|xxi|)n+2s2+τ)1|xix1|n+2s2τC1(1+|xx1|)n+2s2+τ1|xix1|n+2s2τ,

    which gives that

    |ki=2(Φ(|x|ν)1)Upε,xi|C1(1+|xxi|)n+2s2+τki=11|xix1|n+2s2τC1(1+|xx1|)n+2s2+τ(kν)n+2s2τC(1ν)12+ϵ1(1+|xx1|)n+2s2+τ. (2.20)

    On the other hand, if xΩ1 and ||x|ν|δν, where δ>0 is a fixed constant, then

    ||x||x1||||x|ν|||x1|ν|12δν

    and thus

    |(Φ(|x|ν)1)Upε,x1|Cνn+2s2τ1(1+|xx1|)n+2s2+τ. (2.21)

    If xΩ1 and ||x|ν|δν, we find

    |Φ(|x|ν)1|C||x|ν1|Cν((|x||x1|)+(|x1|ν))Cν(|x||x1|)+CkCν(|x||x1|)+Cν12+ϵ (2.22)

    and ||x||x1||||x|ν|+|ν|x1||2δν. Thus,

    |x||x1|ν1(1+|xx1|)n+2sCν12+ϵ||x||x1||12ϵ(1+|xx1|)n+2sC(1ν)12+ϵ1(1+|xx1|)n+2s2+τ,

    which, together with (2.20)-(2.22), yields that

    J3C(1ν)12+ϵ.

    This finished this proof.

    Proposition 2.5. There is an integer k0>0, such that for each kk0, ε0εε1, r[ν(1r0k),ν(1r1k)], (2.16) has a unique solution φ=φ(r,ε) satisfying

    φC(1ν)12+ϵ,   |cl|C(1ν)12+ϵ,

    where ϵ>0 is a small constant.

    Proof. Recall that ν=kn2s+1n2s and set

    N={w:wCα0(Bν(0))Hk,wCν12,Bν(0)Up1ε,xiZi,lw=0},

    where 0<α0<s and i=1,2,...,k,l=1,2. From Proposition 2.2, solving (2.16) is equivalent to solving

    φ=B(φ):=Lk(N(φ))+Lk(lk),

    where Lk is defined in Proposition 2.2.

    Firstly, we find

    B(φ)CN(φ)+ClkCφmin{p,2}+C(1ν)12+ϵC(1ν)12(min{p,2})+C(1ν)12+ϵC(1ν)12+ϵCν12,

    which tells that B maps N to N.

    On the other hand, since

    |N(t)|C|t|min{p1,1},

    we get

    |Lk(N(φ1))Lk(N(φ2))|C(|φ1|min{p1,1}+|φ2|min{p1,1})|φ1φ2|C(φ1min{p1,1}+φ2min{p1,1})φ1φ2(ki=11(1+|xxi|)n2s2+τ)min{p,2}.

    Taking into account that

    (ki=11(1+|xxi|)n2s2+τ)min{p,2}ki=1C(1+|xxi|)n+2s2+τ,

    we have

    B(φ1)B(φ2)CN(φ1)N(φ2)C(φ1min{p1,1}+φ2min{p1,1})|φ1φ2|12φ1φ2.

    Thus B is a contraction map.

    Therefore, Applying the contraction mapping theorem, we can find a unique φ=φ(r,ε)N such that

    φ=B(φ)

    and

    φC(1ν)12+ϵ.

    Moreover, we get the estimate of cl from (2.12).

    Let F(,ε)=I(Uε,r+φ), where r=|x1|,=1rν, φ is the function obtained in Proposition 2.5, and

    I(u)=12Bν(0)|A12su|21p+1Bν(0)Φ(|x|ν)|u|p+1.

    Proposition 3.1. We have

    F(,ε)=I(Uε,r)+O(kν1+ϵ)=k(A+A1H(¯x1,¯x1)εn2sνn2s+A2Φ(1)ki=2A1G(¯xi,¯x1)εn2sνn2s+O(1ν1+ϵ))

    where A,A1,A2 are some positive constants, ϵ>0 is a small constant.

    Proof. Since

    I(Uε,r+φ),φ=0,φN,

    there is t(0,1) such that

    F(,ε)=I(Uε,r+φ)=I(Uε,r)12D2I(Uε,r+tφ)(φ,φ)=I(Uε,r)12Bν(0)(|A12sφ|2pΦ(|x|ν)(Uε,r+tφ)p1φ2)=I(Uε,r)12Bν(0)(N(φ)+lk)φ+p2Bν(0)Φ(|x|ν)((Uε,r+tφ)p1Up1ε,r)φ2=I(Uε,r)+O(Bν(0)(|φ|p+1+|N(φ)||φ|+|lk||φ|)).

    Firstly,

    Bν(0)(|N(φ)||φ|+|lk||φ|)C(N(φ)+lk)φBν(0)ki=11(1+|xxi|)n+2s2+τkj=11(1+|xxj|)n2s2+τ.

    From ki=21|xix1|τC and Lemma A.1, we obtain

    ki=11(1+|xxi|)n+2s2+τkj=11(1+|xxj|)n2s2+τ=ki=11(1+|xxi|)n+2τ+kj=1ji1(1+|xxi|)n+2s2+τ1(1+|xxj|)n2s2+τki=11(1+|xxi|)n+2τ+Cki=11(1+|xxi|)n+τkj=21|xjx1|τCki=11(1+|xxi|)n+τ.

    Therefore, we see

    Bν(0)(|N(φ)||φ|+|lk||φ|)C(N(φ)+lk)φBν(0)ki=11(1+|xxi|)n+τCk(N(φ)+lk)φCk(1ν1+ϵ).

    On the other hand, by Hölder inequality, we have

    Bν(0)|φ|p+1Cφp+1Bν(0)(ki=11(1+|xxi|)n2s2+τ)p+1Cφp+1Bν(0)(ki=11(1+|xxi|)n+τ)(ki=11(1+|xxi|)τ)pCφp+1Bν(0)ki=11(1+|xxi|)n+τCkφp+1Ck(1ν1+ϵ).

    Combining the estimates above and applying Proposition A.5, we have proved

    F(,ε)=k(A+A1H(ˉx1,ˉx1)εn2sνn2s+A2Φ(1)ki=2A1G(¯xi,ˉx1)εn2sνn2s+O(1ν1+ϵ)).

    Proposition 3.2. We have

    F(,ε)ε=k(n2s)A1(H(ˉx1,ˉx1)εn+12sνn2s+ki=2G(ˉxi,ˉx1)εn+12sνn2s+O(1ν1+ϵ)),

    and

    F(,ε)=k(A1H(ˉx1,ˉx1)εn2sνn2s+A2Φ(1)A1ki=2G(ˉxi,ˉx1)εn2sνn2s+O(1νϵ)),

    where A1,A2 are the same constants as in Proposition 3.1, and ϵ>0 is a small constant.

    Proof. Note that

    F(,ε)ε=I(Uε,r+φ),Uε,rε+φε=I(Uε,r+φ),Uε,rε+2l=1ki=1clUp1ε,xiZi,l,φε=I(Uε,r)εBν(0)Φ(|x|ν)[(Uε,r+φ)pUpε,r]Uε,rε+2l=1ki=1clUp1ε,xiZi,l,φε.

    Considering

    Up1ε,xiZi,l,φε=(Up1ε,xiZi,l)ε,φ,

    and from Proposition 2.5, we can get

    |ki=1clUp1ε,xiZi,l,φε|C|cl|φBν(0)ki=11(1+|xxi|)n2s2+τkj=11(1+|xxj|)n+2sCkν1+ϵ.

    On the other hand, being φN, we have

    Bν(0)Φ(|x|ν)[(Uε,r+φ)pUpε,r]Uε,rε=Bν(0)pΦ(|x|ν)Up1ε,rUε,rεφ+O(Bν(0)|φ|2)=Bν(0)pΦ(|x|ν)(Up1ε,rUε,rεφki=1Uε,xiUε,xiε)φ+pki=1Bν(0)(Φ(|x|ν)1)Up1ε,xiUε,xiεφ+O(Bν(0)|φ|2)=kΩ1pΦ(|x|ν)(Up1ε,rUε,rεφki=1Uε,xiUε,xiε)φ+kBν(0)p(Φ(|x|ν)1)Up1ε,x1Uε,x1εφ+O(Bν(0)|φ|2),
    |Ω1Φ(|x|ν)(Up1ε,rUε,rεφki=1Uε,xiUε,xiε)φ|CΩ1(Up1ε,r(Uε,x1PUε,x1)+Up1ε,x1ki=2Uε,xi+ki=2Upε,xi)|φ|Cν1+ϵ

    and

    |Bν(0)(Φ(|x|ν)1)Up1ε,x1Uε,x1εφ||||x|ν|ν+||x|ν|ν(Φ(|x|ν)1)Up1ε,x1Uε,x1εφ|Cν1+ϵ.

    Thus, using the estimates above,

    F(,ε)ε=I(Uε,r)ε+O(kν1+ϵ),

    from which and Proposition A.6, the conclusion follows.

    Finally, noting that =νr and arguing as before, we can get the estimate F(,ε).

    To prove Theorem 1.2, now we estimate H(ˉx1,ˉx1) and G(ˉxi,ˉx1),i2. Let ˉx1=(1|ˉx1|,0,,0) be the reflection of ˉx1 with respect to the unit sphere. Then

    H(ˉx1,ˉx1)=C|ˉx1ˉx1|n2s=C2n2sn2s

    for some constant C>0.

    On the other hand,

    |ˉxiˉx1|=|ˉxiˉx1|2+424|ˉxiˉx1|cosθi,

    where θi is the angle between ˉxiˉx1 and (1,0,,0) and then θi=π2+(i1)πk. Thus

    G(ˉxi,ˉx1)=C|ˉxiˉx1|n2sC|ˉxiˉx1|n2s=C|ˉxiˉx1|n2s(111+42+4|ˉxiˉx1|sin(i1)πk|ˉxiˉx1|2)n2s2.

    As

    |x1xi|=2|x1|sin(i1)πk, i=2,,k,

    kc>0 and

    0<c<sin(i1)πk(i1)πkc, i=2,,[k2],

    we find

    c1i242+4|ˉxiˉx1|sin(i1)πk|ˉxiˉx1|2c2i2

    for some constants c2c1>0.

    So, there exists a constant A3>0 such that

    ki=2G(ˉxi,ˉx1)=ki=2C|ˉxiˉx1|n2s(1+O(1(1+ci2)n2s))=ki=2C|ˉxiˉx1|n2s(˜c+O(1i2))=A3kn2s+O(1k).

    Therefore, it follows from Propositions 3.1 and 3.2 that there are constants B1,B2,B3 such that

    F(,ε)=k(A+B1εn2sνn2sn2s+B2B3kn2sεn2sνn2s+O(1ν1+ϵ)), (3.1)
    F(,ε)ε=k(B1(n2s)εn+12sνn2sn2s+B3(n2s)kn2sεn+12sνn2s+O(1ν1+ϵ)) (3.2)

    and

    F(,ε)=k(B1(n2s)εn2sνn2sn+12s+B2+O(1νϵ)). (3.3)

    Proof of Theorem 1.2. Recall that =1rν and ν=kn2s+1n2s. Denote L=k. Then from (3.2) and (3.3), F(,ε)ε=0 and F(,ε)=0 are equivalent to

    B1(n2s)εn+12sLn2s+B3(n2s)εn+12s+O(1νϵ)=0 (3.4)

    and

    B1(n2s)εn2sLn+12s+B2+O(1νϵ)=0 (3.5)

    respectively.

    Denote

    h1(L,ε)=B1(n2s)εn+12sLn2s+B3(n2s)εn+12s

    and

    h2(L,ε)=B1(n2s)εn2sLn+12s+B2.

    Thus h1=0 and h2=0 have a unique solution

    L0=(B1B3)1n2s,ε0=(B1(n2s)B2Ln+12s0)1n2s.

    Moreover, it is easy to verify that

    h1(L0,ε0)ε=0,h2(L0,ε0)L>0

    and

    h1(L0,ε0)L=h2(L0,ε0)ε>0,

    which means that h1=0 and h2=0 at (L0,ε0) is invertible. So, (3.4) and (3.5) have a solution near (L0,ε0).

    This work was partially supported by NSFC (No.11601194) and the author thanks the referee's thoughtful reading of details of the paper and nice suggestions to improve the results.

    The author declares that there are no conflicts of interest.

    In this section, we will give some basic estimates and the energy expansion for the approximate solutions. First, recall that

    xi=(rcos2(i1)πk,rsin2(i1)πk,0), i=1,,k, 

    where 0Rn2, r[ν(1r0k),ν(1r1k)], and

    I(u)=12Bν(0)|A12su|21p+1Bν(0)Φ(|x|ν)|u|p+1.

    Now, we introduce the following two lemmas which are important in this sequel and have been proved in [22] and [12] respectively.

    Lemma A.1. For any constant 0<σ≤</italic><italic>min{α,β}, there exists a constant C>0, such that

    1(1+|yxi|)α1(1+|yxj|)βC|xixj|σ(1(1+|yxi|)α+βσ+1(1+|yxj|)α+βσ).

    Lemma A.2. For any constant 0<κ<</italic><italic>n2s, there is a constant C>0 such that

    Bν(0)1|y|n2s1(1+|xy|)2s+κdyC(1+|x|)κ.

    Note that

    Uε,r(x)=ki=1PUε,xi(x)

    and

    Uε,xi(x)=Cn,sεn2s2(1+ε2|xxi|2)n2s2

    for some suitable Cn,s. We have

    Lemma A.3. There is a small ϵ>0 and some constant C>0 such that

    Bν(0)1|xy|n2sU4sn2sε,r(y)ki=11(1+|yxi|)n2s2+τdyCki=11(1+|xxi|)n2s2+τ+ϵ.

    Proof. We can find this proof in [12]. Here we just need to use

    Uε,r(x)Cki=11(1+|xxi|)n2s.

    Let G(x,y) be the Green function of As in B1(0) with the Dirichlet boundary condition (see [4]), namely, G(x,y) solves

    AsG(,y)=δy   in B1(0),   G(,y)=0  on B1(0),

    and the regular part of G is given by

    H(x,y)=αn,s|xy|n2sG(x,y),

    where αn,s=1|Sn1|212sΓ(n2s2)Γ(n2)Γ(s).

    Let ˉxi=xiν. Then

    Proposition A.4. We have

    Uε,xi(x)PUε,xi(x)=1εn2s2νn2sH(ˉx,ˉxi)+O(1(ν)n+22s), (A.1)

    where =1|ˉxi|=1rν.

    Proof. First, letting φxi,ε=Uε,xiPUε,xi, from the equations satisfied by Uε,xi and PUε,xi, we can get that

    {Asφxi,ε=0   in Bν(0),φxi,ε=Cn,s1ε(n2s)/2|xxi|n2s+O(1εn+22s|xxi|n+22s)   on Bν(0). (A.2)

    Denote by ˜G(x,y) the Green function of As in Bν(0) with the Dirichlet boundary condition and by ˜H(x,y) the regular part of ˜G(x,y). So we can find

    {As˜H(x,xi)=0   in Bν(0),˜H(x,xi)=αn,s|xxi|n2s   on Bν(0),

    which, together with (A.2), yields that

    {As(φxi,ε1ε(n2s)/2˜H(x,xi))=0   in Bν(0),φxi,ε1ε(n2s)/2˜H(x,xi)=O(1εn+22s|xxi|n+22s)   on Bν(0).

    As a result,

    φxi,ε1ε(n2s)/2˜H(x,xi)LCεn+22sd(xi,Bν(0))n+22s=Cεn+22s|ν|xi||n+22s (A.3)

    for some constant C>0.

    Using (A.3), we have proved

    Uε,xi(x)PUε,xi(x)=1εn2s2νn2sH(ˉx,ˉxi)+O(1(ν)n+22s).

    Proposition A.5. There holds

    I(Uε,r)=k(A+A1H(ˉx1,ˉx1)εn2sνn2s+A2Φ(1)ki=2A1G(ˉxi,ˉx1)εn2sνn2s+O(1ν1+ϵ)),

    where A,A1,A2 are some positive constants, and ϵ is a small constant.

    Proof. Note that

    I(Uε,r)=12Bν(0)|A12sUε,r|21p+1Bν(0)Φ(|x|ν)|Uε,r|p+1.

    Using the symmetry and (A.1), we have

    Bν(0)|A12sUε,r|2=ki=1kj=1Bν(0)Upε,xjPUε,xi=kki=1Bν(0)Upε,x1PUε,xi=k(Bν(0)Up+1ε,x1Bν(0)Upε,x1(Uε,x1PUε,x1)+ki=2Bν(0)Upε,x1PUε,xi)=k(RnUp+11,0B0H(ˉx1,ˉx1)εn2sνn2s+ki=2B0G(ˉxi,ˉx1)εn2sνn2s+O(kν)n+22s), (A.4)

    where B0=RnUp1,0.

    Recalling that

    Ωi={x=(x,x)Bν(0):x|x|,(xi)|(xi)|cosπk},i=1,2,,k,

    we can obtain that

    Bν(0)Φ(|x|ν)Up+1ε,r=kΩ1Φ(|x|ν)(PUε,x1)p+1+kpΩ1ki=2(PUε,x1)pPUε,xi+kO(Ω1|Φ(|x|ν)1|ki=2Upε,x1Uε,xi+Ω1U(p+1)/2ε,x1(ki=2Uε,xi)(p+1)/2). (A.5)

    Next, we estimate each terms in (A.5). For xΩ1, |xxi||xx1|, we obtain for any κ(1,n2s),

    ki=2Uε,xiCki=21|x1xi|κ1(1+|xx1|)n2sκ,

    which implies that if κ close to n2s,

    Ω1Up+1/2ε,x1(ki=2Uε,xi)p+1/2=O(kν)κnn2s=O(1ν1+ϵ).

    On the other hand, from (A.1) again, we have

    Ω1ki=2(PUε,x1)pPUε,xi=Ω1ki=2(PCn,sεn2s2(1+ε2|xxi|2)n2s2)p(1εn2s2νn2sG(ˉx,ˉxi)+O(1(ν)n+22s))=εn+2s2˜Ω1ki=2(PCn,s(1+|y|2)n2s2)p(1εn2s2νn2sG(¯yε+xi,ˉxi)+O(1(ν)n+22s))εndy=Bν(0)Up1,0ki=2G(ˉx,ˉxi)εn2sνn2s+O(1ν1+ϵ),

    where ˜Ω1={yε+x1Ω1}.

    Now we estimate Ω1|Φ(|x|ν)1|Upε,x1ki=2Uε,xi. For xΩ1 and ||x|ν|δν, where δ>0 is a fixed constant, we have

    ||x||x1||||x|ν|||x1|ν|12δν

    and from Lemma A.1,

    Ω1|Φ(|x|ν)1|Upε,x1ki=2Uε,xiCΩ1ki=21(1+|xx1|)n+2s1(1+|xxi|)n2sC(1ν)1+ski=21|xix1|n12sΩ11(1+|xx1|)n+sC(1ν)1+ϵ. (A.6)

    If xΩ1 and ||x|ν|δν, we have

    |Φ(|x|ν)1|C||x|ν1|Cν((|x||x1|)+(|x1|ν))Cν(|x||x1|)+Ck,

    and ||x||x1||||x|ν|+|ν|x1||2δν. Choosing κ(1,n2s), we obtain that

    Ω1|Φ(|x|ν)1|Upε,x1ki=2Uε,xiCΩ1(1ν(|x||x1|)+1k)ki=21(1+|xx1|)n+2s1(1+|xx1|)n2sCνΩ1ki=21(1+|xx1|)n+2s11(1+|xx1|)n2s+Ckki=21|xix1|κΩ11(1+|xx1|)2nκC(1ν)1+ϵ.

    This, together with (A.6), tells us that

    Ω1|Φ(|x|ν)1|Upε,x1ki=2Uε,xiC(1ν)1+ϵ. (A.7)

    Finally,

    Ω1Φ(|x|ν)(PUε,x1)p+1=Ω1(PUε,x1)p+1+Ω1(Φ(|x|ν)1)Up+1ε,x1+O(Ω1|Φ(|x|ν)1|Upε,x1H(ˉx,ˉx1)νn2s)=Bν(0)Up+11,0(p+1)B0H(ˉx1,ˉx1)εn2sνn2s+Ω1(Φ(|x|ν)1)Up+1ε,x1+O(1ν1+ϵ).

    But

    Ω1(Φ(|x|ν)1)Up+1ε,x1=Ω1(Φ(|ˉx1|)1)Up+1ε,x1+Ω1(Φ(|x|ν)Φ(|ˉx1|))Up+1ε,x1=Ω1(Φ(|ˉx1|)1)Up+1ε,x1+O(1ν1+ϵ)=Φ(1)Bν(0)Up+11,0+O(1ν1+ϵ).

    Thus, from the estimates above, we have proved

    Bν(0)Φ(|x|ν)Up+1ε,r=k(Bν(0)Up+11,0Φ(1)Bν(0)Up+11,0(p+1)B0H(ˉx1,ˉx1)εn2sνn2s+(p+1)ki=2B0G(ˉxi,ˉx1)εn2sνn2s+O(1ν1+ϵ)),

    which, together with (A.4), implies that our desired result holds.

    Similar to Proposition A.5, we also have

    Proposition A.6.

    I(Uε,r)ε=k(n2s)A1(H(ˉx1,ˉx1)εn+12sνn2s+ki=2G(ˉxi,ˉx1)εn+12sνn2s+O(1ν1+ϵ)),

    and

    I(Uε,r)r=k(A1H(ˉx1,ˉx1)rεn2sνn2sA2Φ(1)1νA1ki=2G(ˉxi,ˉx1)rεn2sνn2s+O(1ν1+ϵ)),

    where A1,A2 are the same constants as in Proposition A.5.



    [1] D. Cao, S. Peng, The asymptotic behavior of the ground state solutions for Hénon equation, J. Math. Anal. Appl., 278 (2003), 1-17. doi: 10.1016/S0022-247X(02)00292-5
    [2] D. Cao, S. Peng, S. Yan, Asymptotic behavior of ground state solutions for Hénon equation, IMA J. Appl. Math., 74 (2009), 468-480. doi: 10.1093/imamat/hxn035
    [3] W. Chen, C. Li, B. Ou, Classification of solutions for an integral equation, Comm. Pure Appl. Math., 59 (2006), 330-343. doi: 10.1002/cpa.20116
    [4] W. Choi, S. Kim, K.-A. Lee, Asymptotic behavior of solutions for nonlinear elliptic problems with the fractional Laplacian, J. Funct. Anal., 266 (2014), 6531-6598. doi: 10.1016/j.jfa.2014.02.029
    [5] J. DÁvia, M. Del Pino, Y. Sire, Non degeneracy of the bubble in the critical case for nonlocal equations, Proc. Amer. Math. Soc., 141 (2013), 3865-3870. doi: 10.1090/S0002-9939-2013-12177-5
    [6] M. del Pino, P. Felmer, M. Musso, Two-bubble solutions in the super-critical Bahri-Coron's problem, Calc. Var. Partial Dif., 16 (2003), 113-145. doi: 10.1007/s005260100142
    [7] S. Dipierro, G. Palatucci, E. Valdinoci, Existence and symmetry results for a Schrödinger type problem involving the fractional Laplacian, Matematiche, 68 (2013), 201-216.
    [8] R.-L. Frank, E.-H. Lieb, Inversion positivity and the sharp Hardy-Littlewood-Sobolev inequality, Calc. Var. Partial Dif., 39 (2010), 85-99. doi: 10.1007/s00526-009-0302-x
    [9] R.-L. Frank, E.-H. Lieb, A new rearrangement-free proof of the sharp Hardy-Littlewood- Sobolev inequality, Spectral Theory, Function Spaces and Inequalities, 219 (2012), 55-67.
    [10] E.-H. Lieb, Sharp constants in the Hardy-Littlewood-Sobolev and related inequalities, Ann. Math., 118 (1983), 349-374. doi: 10.2307/2007032
    [11] N. Hirano, Existence of positive solutions for the Hénon equation involving critical Sobolev exponent, J. Differ. Equations 247 (2009), 1311-1333.
    [12] W. Long, J. Yang, Positive or sign-changing solutions for a critical semilinear non-local equation, Z. Angew. Math. Phys., 67 (2016), 45.
    [13] W.-M. Ni, A nonlionear Dirichlet problem on the unit ball and its application, Indiana Univ. Math. J., 6 (1982), 801-807.
    [14] S. Peng, Multiple boundary concentrating solutions to Dirichlet problem of Hénon equation, Acta Math. Appl. Sin., 22 (2006), 137-162. doi: 10.1007/s10255-005-0293-0
    [15] A. Pistoia, E. Serra, Multi-peak solutions for the Hénon equation with slightly subcritical growth, Math. Z., 256 (2007), 75-97. doi: 10.1007/s00209-006-0060-9
    [16] X. Ros-Oton, J. Serra, The Pohozaev identity for the fractional Laplacian, Arch. Ration. Mech. Anal., 213 (2014), 587-628. doi: 10.1007/s00205-014-0740-2
    [17] E. Serra, Non-radial positive solutions for the Hénon equation with critical growth, Calc. Var. Partial Dif., 23 (2005), 301-326. doi: 10.1007/s00526-004-0302-9
    [18] D. Smets, J. Su, M. Willem, Non-radial ground states for the Hénon equation, Comm. Contemp. Math., 4 (2002), 467-480. doi: 10.1142/S0219199702000725
    [19] R. Servadei, E. Valdinoci, A Brezis-Nirenberg result for non-local critical equations in low dimension, Commun. Pure Appl. Anal., 12 (2013), 2445-2464. doi: 10.3934/cpaa.2013.12.2445
    [20] J. Tan, The Brezis-Nirenberg type problem involving the square root of the Laplacian, Calc. Var. Partial Dif., 42 (2012), 21-41.
    [21] J. Tan, Positive solutions for nonlocal ellipitc problems, Discrete Contin. Dyn. Syst., 33 (2013), 837-859. doi: 10.3934/dcds.2013.33.837
    [22] J. Wei, S. Yan, Infinitely many positive solutions for the prescribed scalar curvature problem on SN, J. Funct. Anal., 258 (2010), 3048-3081.
    [23] J. Wei, S. Yan, Infinitely many nonradial solutions for the Hénon equation with critical growth, Rev. Mat. Iberoam., 29 (2013), 997-1020. doi: 10.4171/RMI/747
  • Reader Comments
  • © 2020 the Author(s), licensee AIMS Press. This is an open access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0)
通讯作者: 陈斌, bchen63@163.com
  • 1. 

    沈阳化工大学材料科学与工程学院 沈阳 110142

  1. 本站搜索
  2. 百度学术搜索
  3. 万方数据库搜索
  4. CNKI搜索

Metrics

Article views(3326) PDF downloads(227) Cited by(0)

Other Articles By Authors

/

DownLoad:  Full-Size Img  PowerPoint
Return
Return

Catalog