Abstract
We deal with the following critical Choquard equation
where \(\varepsilon >0\) is a small parameter, \(0<\mu <3\), \(p\in (4,6)\). Under some conditions on the potential functions V(x), P(x), and Q(x), we obtain the existence of multiple solutions and their asymptotical behavior as \(\varepsilon \rightarrow 0\).
Similar content being viewed by others
Avoid common mistakes on your manuscript.
1 Introduction and Main Results
In this paper, we consider the following critical Choquard equation
where \(\varepsilon >0\) is a parameter, \(0<\mu <3\), \(p\in (4,6)\). The potential functions V(x), P(x), and Q(x) are three bounded and continuous functions in \({\mathbb {R}}^3\) satisfying \(\inf \limits _{x\in {\mathbb {R}}^3} V(x)>0\), \(\inf \limits _{x\in {\mathbb {R}}^3} P(x)>0\) and \(\inf \limits _{x\in {\mathbb {R}}^3} Q(x)>0\).
The Choquard equation
was used by Pekar [17] to describe the quantum theory of polaron at rest. Then it was introduced by Choquard [10] as an approximation to Hartree–Fock theory of one-component plasma. Penrose [18] also derived it as a model of self-gravitating matter, in which quantum state reduction is understood as a gravitational phenomenon. Lieb [10] proved the existence and uniqueness (up to translations) of solutions by using symmetric decreasing rearrangement inequalities. Lions [11] obtained the existence of infinitely many spherically symmetric solutions. Ma and Zhao [14] showed the positive solutions of this equation must be radially symmetric and monotone decreasing about some fixed point by the method of moving planes. Moroz and Van Schaftingen [15] studied the generalized Choquard equation
where \(I_\alpha \) is a Riesz potential and \(p>1\). For an optimal range of parameters, they showed the regularity, positivity, and radial symmetry of the ground states and derived decay property at infinity as well.
Gao and Yang [6] studied the Brezis–Nirenberg type problem of the nonlinear Choquard equation
where \(\Omega \) is a bounded domain and \(\lambda \) is a parameter, \(N\ge 3\) and \(2^*_\mu =\frac{2N-\mu }{N-2}\) is the critical exponent under the sense of Hardy–Littlewood–Sobolev inequality. They established some existence results for this equation. Shen, Gao, and Yang [19] investigated the critical Choquard equation with potential well
where \(\lambda ,\beta >0\), \(0<\mu <N\), \(N\ge 4\), \(2^*_\mu \) is the critical exponent. They proved the existence of ground state solutions which localize near the potential well \(\inf V^{-1}(0)\) and also characterize the asymptotic behavior as \(\lambda \rightarrow \infty \). Furthermore, the multiple solutions were also established by Lusternik–Schnirelmann category theory.
For the semiclassical problem, Liu and Tang [12] studied the following subcritical equation
where \(\varepsilon >0\), \(N>2\), \(\theta \in [2.\frac{N+\theta }{N-2})\). The potential functions V(x), W(x) are bounded positive functions. By using pseudo-index theory, they established the multiplicity of solutions. Alves et al. [1] studied the following critical equation
where \(\varepsilon >0\) is a parameter, \(0<\mu <3\). The potential functions V(x) and Q(x) are two bounded and continuous functions in \({\mathbb {R}}^3\) satisfying \(\inf _{x\in {\mathbb {R}}^3} V(x)>0\) and \(\inf _{x\in {\mathbb {R}}^3} Q(x)>0\). When \(Q(x)\equiv 1\) and V(x) satisfies
they proved the existence of ground state solution and multiple solutions. Moreover, the concentration phenomenon was also considered. Zhang and Zhang [29] considered the following critical Choquard equation
where \(\varepsilon >0\) is a parameter, \(0<\mu <3\). The potential functions V(x) and Q(x) are two bounded and continuous functions. Under the condition,
and
they established a relationship between the category of the set \(\mathcal {V\cap Q}\) and the number of solutions by employing the Lusternik–Schnirelmann category theory.
On the other hand, the reduction methods are also used to study the Choquard equation. Wei and Winter [22] considered
where \(\varepsilon >0\), \(V\in C^2({\mathbb {R}}^3)\) and \(\inf _{x\in {\mathbb {R}}^3} V(x)>0\). They proved that for any given positive integer K, if \(P_1, P_2, ... , P_K\in {\mathbb {R}}^3\) were given nondegenerate critical points of V(x), then for \(\varepsilon \) sufficiently small, there existed a positive solution for the equation and this solution had exactly K local maximum points \(Q^{\varepsilon }_i(i=1,2,...,K)\) with \(Q^{\varepsilon }_i\rightarrow P_i\) as \(\varepsilon \rightarrow 0\). Luo, Peng and Wang [13] also investigated the above problem. For \(\varepsilon \) small enough, by using a local Pohozaev type of identity, blow-up analysis, and the maximum principle, they showed the uniqueness of positive solutions concentrating at the nondegenerate critical points of V(x). For more results about Choquard equations, we refer to [5, 7,8,9, 16, 23, 26, 28, 31] and the references therein.
Motivated by the above works, we are concerned with the existence and concentration behavior of positive solutions for (1.1). We note that (1.1) involves three different potentials. This brings a competition between the potentials V, P, and Q: each one would like to attract ground states to their minimum or maximum points, respectively. It makes difficulties in determining the concentration position of solutions. This kind of problem can be traced back to [20, 21] for the semilinear Schrödinger equation. See also [24, 25, 27, 30] for other related results. We first recall the following famous Hardy–Littlewood–Sobolev inequality.
Proposition 1.1
(Hardy–Littlewood–Sobolev inequality). Let \(t,r>1\) and \(0<\mu <3\) with \(\frac{1}{t}+\frac{\mu }{3}+\frac{1}{r}=2\), \(f\in L^t({\mathbb {R}}^3)\) and \(h\in L^r({\mathbb {R}}^3)\). There exists a sharp constant \(C(t,\mu ,r)\), independent of f, h, such that
Remark 1.2
By Proposition 1.1, the term
is well defined if \(|u|^r\in L^s({\mathbb {R}}^3)\) satisfies \(\frac{2}{s}+\frac{\mu }{3}=2\). Therefore, for \(u\in H^1({\mathbb {R}}^3)\), we will require \(sr\in [2,6]\). Then \(\frac{6-\mu }{3}\le r\le 6-\mu \). Here, \(\frac{6-\mu }{3}\) is called the lower critical exponent and \(6-\mu \) is called the upper critical exponent in the sense of Hardy–Littlewood–Sobolev inequality.
Proposition 1.3
(Optimizers for \(S_{H,L}\)). [6] Define
Then \(S_{H,L}\) is achieved if and only if
where \(C>0\) is a fixed constant, \(a\in {\mathbb {R}}^3\) and \(b>0\) are parameters.
Remark 1.4
[6] In fact,
is a minimizer for S, the best Sobolev constant, and is also the minimizer for \(S_{H,L}\). Moreover,
where \(C(3,\mu )\) is the sharp constant in Proposition 1.1.
To state our main results, some hypotheses about the potential functions are needed as follows:
- \((H_1)\):
-
\(V_{\infty }>V_\mathrm{{min}}\) or \(P_\mathrm{{max}}>P^{\infty }\),
- \((H_2)\):
-
\(Q(x)\le Q^{\infty }\) for \(x\in {\mathbb {R}}^3\),
- \((H_3)\):
-
\(\mathcal {V\cap P\cap Q}=\{x\in {\mathbb {R}}^3: V(x)=V_{\min }, P(x)=P_\mathrm{{max}}, Q(x)=Q_\mathrm{{max}}\}\ne \varnothing \),
where
Obviously, under the assumptions \((H_1)\), the set \(\mathcal {V\cap P\cap Q}\) is bounded.
Our main results are as follows:
Theorem 1.5
Suppose that the potentials V(x), P(x), Q(x) satisfy conditions \((H_1)\), \((H_2)\) and \((H_3)\). Then
-
(i)
For any \(\delta >0\), there exists \(\varepsilon _\delta >0\) such that problem (1.1) has at least \(cat_{({\mathcal {V}}\cap {\mathcal {P}}\cap {\mathcal {Q}})_\delta }({\mathcal {V}}\cap {\mathcal {P}}\cap {\mathcal {Q}})\) solutions for \(\varepsilon \in (0,\varepsilon _\delta )\), where \(({\mathcal {V}}\cap {\mathcal {P}}\cap {\mathcal {Q}})_\delta =\{x\in {\mathbb {R}}^3: dist(x,{\mathcal {V}}\cap {\mathcal {P}}\cap {\mathcal {Q}})\le \delta \}\).
-
(ii)
For \(\varepsilon _n\rightarrow 0\) as \(n\rightarrow \infty ,\) up to a subsequence, there exists \(y_n\) such that \(u_{\varepsilon _n}(x+y_n)\), where \(u_{\varepsilon _n}\) is a solution in (i), converges in \(H^1({\mathbb {R}}^3)\) to a ground state solution u of
$$\begin{aligned} -\Delta u+ V_\mathrm{{min}}u=P_\mathrm{{max}}|u|^{p-2}u+\Big (\int _{{\mathbb {R}}^3}\frac{Q^2_\mathrm{{max}}|u(y)|^{6-\mu }}{|x-y|^\mu }\mathrm{{d}}y\Big )|u|^{4-\mu }u, \ \ x\in {\mathbb {R}}^3,\nonumber \\ \end{aligned}$$(1.2)
The proof of our main results is based on the variational method. The main difficulties lie in two aspects: (i) The unboundedness of the domain \({\mathbb {R}}^3\) and the critical exponent under the sense of Hardy–Littlewood–Sobolev inequality lead to the lack of compactness. Some arguments developed by Brezis and Nirenberg [3] can be applied to prove that the functional associated with (1.1) satisfies the Palais-Smale (PS) condition under some energy level. (ii) When the critical term has a potential Q(x), the proof of the existence of multiple solutions become more complicated. As far as we know, there are no results about this problem. By using Lusternik–Schnirelmann theory, we establish the relationship between the category of the set \({\mathcal {V}}\cap {\mathcal {P}}\cap {\mathcal {Q}}\) and the number of solutions.
This paper is organized as follows. In the forthcoming section, we collect some necessary preliminary lemmas which will be used later. In Sect. 3, we are devoted to the energy functional with constant coefficients. In Sect. 4, the PS condition is given. In Sect. 5, the Lusternik–Schnirelmann theory is applied to prove the existence of multiple solutions.
Notation. In this paper, we make use of the following notations.
-
For any \(R>0\) and \(x\in {\mathbb {R}}^3\), \(B_{R}(x)\) denotes the open ball of radius R centered at x.
-
The letter C stands for positive constants (possibly different from line to line).
-
”\(\rightarrow \)” denotes the strong convergence and "\(\rightharpoonup \)" denotes the weak convergence.
-
\(|u|_q=(\int _{{\mathbb {R}}^3}|u|^q\mathrm{{d}}x)^{\frac{1}{q}}\)denotes the norm of u in \(L^q({\mathbb {R}}^3)\) for \(2\le q\le 6\).
2 Preliminaries
The standard norm of \(E:=H^1({\mathbb {R}}^3)\) is given by
Since V(x) is bounded and \(\inf _{x\in {\mathbb {R}}^3} V(x)>0\), we have the following equivalent norm
For \(f\in L^1_{loc}({\mathbb {R}}^3)\), define
and this integral converges in the classical Lebesgue sense for a.e. \(x\in {\mathbb {R}}^3\) if and only if \(f\in L^1({\mathbb {R}}^3,(1+|x|)^{-\mu }\mathrm{{d}}x).\)
Remark 2.1
By Hardy–Littlewood–Sobolev inequality, \(I_\mu \) defines a linear continuous map from \(L^{\frac{6}{6-\mu }}({\mathbb {R}}^3)\) to \(L^{\frac{6}{\mu }}({\mathbb {R}}^3)\).
Define \(F:E\rightarrow {\mathbb {R}}\) by
To prove the properties about \(F(\cdot )\), for simplicity, we assume that \(Q(x)\equiv 1\) in the following three Lemmas.
Lemma 2.2
Let \(u_n\rightharpoonup u\) in E and \(u_n\rightarrow u\), a.e. in \({\mathbb {R}}^3\). Then
Proof
By Hardy–Littlewood–Sobolev inequality, \(I_\mu *|u_n|^{6-\mu }\in L^{\frac{6}{\mu }}({\mathbb {R}}^3)\). Choose a function \(v\in L^{\frac{6}{6-\mu }}({\mathbb {R}}^3)\) satisfying \(v>0\) in \({\mathbb {R}}^3\). Then
Since \(I_\mu *v\in L^{\frac{6}{\mu }}({\mathbb {R}}^3)\), and \(\Big ||u_n(y)|^{6-\mu }-|u(y)|^{6-\mu }\Big |\rightharpoonup 0\) in \(L^{\frac{6}{6-\mu }}({\mathbb {R}}^3)\), we can obtain
It follows from \(v>0\) that the result holds. \(\square \)
Lemma 2.3
Let \(u_n\rightharpoonup u\) in E and \(u_n\rightarrow u\), a.e. in \({\mathbb {R}}^3\). Then
and
Proof
Let \(v_n=u_n-u\), then
By Hardy–Littlewood–Sobolev inequality, \((I_\mu *|v_n|^{6-\mu })^{\frac{6}{5}}\in L^{\frac{5}{\mu }}({\mathbb {R}}^3)\), and is bounded in \(L^{\frac{5}{\mu }}({\mathbb {R}}^3)\). From Lemma 2.2, \(I_\mu *|v_n|^{6-\mu }\rightarrow 0, \ \text {a.e. in}\ {\mathbb {R}}^3, \ \text {as}\ n\rightarrow \infty \). Then, we have \((I_\mu *|v_n|^{6-\mu })^{\frac{6}{5}}\rightharpoonup 0\) in \(L^{\frac{5}{\mu }}({\mathbb {R}}^3)\). It follows from \(|u|^{\frac{6(5-\mu )}{5}}\in L^{\frac{5}{5-\mu }}({\mathbb {R}}^3)\) that the first result holds. Similarly, the second limit can be obtained. \(\square \)
Lemma 2.4
Let \(u_n\rightharpoonup u\) in E and \(u_n\rightarrow u\), a.e. in \({\mathbb {R}}^3\). Then
-
(i)
\(F(u_n-u)=F(u_n)-F(u)+o_n(1)\);
-
(ii)
\(F^\prime (u_n-u)=F^\prime (u_n)-F^\prime (u)+o_n(1)\), in \((H^1({\mathbb {R}}^3))^{-1}\).
Proof
The first part (i) has been proved in [6]. We just prove the second part (ii). In fact, for any \(\phi \in H^1({\mathbb {R}}^3)\),
Next, we prove that
Let \(v_n=u_n-u\). Then, for any small \(\delta >0\),
where
and
Therefore,
Define
It is easy to see that
By Lemma 2.2, we can obtain
and
Thus,
Then, we have
By the definition of \(G_{\delta ,n(x)}\) and the boundedness of \(f_n\) in \(L^{\frac{6}{5}}({\mathbb {R}}^3)\),
It follows from the arbitrariness of \(\delta \) that
\(\square \)
Making the change of variable \(x\rightarrow \varepsilon x\), we can rewrite problem (1.1) as
Thus, the corresponding energy functional is
It is easy to check that \(I_\varepsilon \) is well defined on E and \(I_\varepsilon \in C^1(E,{\mathbb {R}}).\) Then we can define the Nehari manifold
Lemma 2.5
There exists \(C_0>0\) which is independent of \(\varepsilon \) such that
Proof
For any \(u\in {\mathcal {N}}_\varepsilon \), we have
It follows from the Hardy–Littlewood–Sobolev inequality and Sobolev embedding theorem that
Without loss of generality, we assume that \(\Vert u\Vert _\varepsilon \le 1\). Then
Thus, the first desired result follows. On the other hand, we have
\(\square \)
Lemma 2.6
For any \(u\in E\setminus \{0\}\), there exists a unique \(t(u)>0\) such that \(t(u)u\in ({\mathcal {N}})_\varepsilon \) and
Proof
For any \(u\in E\setminus \left\{ 0\right\} \), define \(g(t)=I_\varepsilon (tu),\ t\in [0,+\infty ).\) Then
It is easy to see that \(g(t)>0\) for \(t>0\) small and \(g(t)<0\) for \(t>0\) large enough, so there exists \(t_0>0\) such that
It follows from \(g'(t_0)=0\) that \(t_0u\in {\mathcal {N}}_\varepsilon \).
If there exist \(0<t_1<t_2\) such that \(t_1u\in {\mathcal {N}}_\varepsilon \) and \(t_2u\in {\mathcal {N}}_\varepsilon \). Then
and
It follows that
which is a contradiction. \(\square \)
Lemma 2.7
For any \(\varepsilon >0\), let
where
Then, \(c_\varepsilon =c_\varepsilon ^*=c_\varepsilon ^{**}.\)
Proof
We divide the proof into three steps.
Step1. \(c_\varepsilon ^*=c_\varepsilon \). By Lemma 2.6, we have
Step2. \(c_\varepsilon ^*\ge c_\varepsilon ^{**}.\) For any \(u\in E\setminus \{0\}\), there exists T large enough, such that \(I_\varepsilon (Tu)<0\). Define \(\gamma (t)=tTu\), \(t\in [0,1]\). Then we have \(\gamma (t)\in \Gamma _\varepsilon \) and, therefore,
It follows that \(c_\varepsilon ^*\ge c_\varepsilon ^{**}\).
Step3. \( c_\varepsilon ^{**}\ge c_\varepsilon \). For any \(u\in E\setminus \{0\}\) with \(\Vert u\Vert _\varepsilon \) small, we know
We claim that every \(\gamma (t)\in \Gamma _\varepsilon \) has to cross \({\mathcal {N}}_\varepsilon \). Otherwise, by the continuity of \(\gamma (t)\), (2.3) still holds when u is replaced by \(\gamma (1)\). Then, we can obtain
which contradicts the definition of \(\gamma (1)\). It follows from the claim that \( c_\varepsilon ^{**}\ge c_\varepsilon \). \(\square \)
One can easily check that the functional \(I_\varepsilon \) satisfies the mountain-pass geometry that is the following lemma holds ( [26]).
Lemma 2.8
\(I_\varepsilon \) has the mountain geometry structure.
-
(i)
There exist \(a_0,r_0>0\) independent of \(\varepsilon \), such that \(I_\varepsilon (u)\ge a_0\), for all \(u\in E\) with \(\Vert u\Vert _\varepsilon =r_0.\)
-
(ii)
For any \(u\in E\setminus \{0\}\), \(\lim _{t\rightarrow \infty }I_\varepsilon (tu)=-\infty .\)
Lemma 2.9
For any \(\varepsilon >0\) and \(Q(x)\equiv q\), we have \(c_\varepsilon <\displaystyle \frac{5-\mu }{2(6-\mu )} S^{\frac{6-\mu }{5-\mu }}_{H,L}q^{\frac{-2}{5-\mu }}\), where q is a positive constant.
Proof
For any \(\epsilon >0\), define
where \(\phi (x)\in C_0^\infty ({\mathbb {R}}^3)\) is such that \(\phi =1\) on \(B_1(0)\) and \(\phi =0\) on \(B^c_2(0)\). From Lemma 2.6 in [1], we know that
and
Then, for \(t>0\),
It is easy to see that \(h(t)\rightarrow -\infty \) as \(t\rightarrow +\infty \), \(h(0)=0\) and \(h(t)>0\) as t is small. Therefore, there exists \(t_\varepsilon >0\) such that h(t) attains its maximum. Then, differentiating h at \(t_\epsilon \), we can obtain
When \(\epsilon \) is small enough, it follows from the above expression that there exist \(t_1,t_2>0\) independent of \(\epsilon \) such that \(t_1<t_\epsilon <t_2\). Noting
attains its maximum at
Then, we have
Since \(p\in (4,6)\), then \(0<\frac{6-p}{2}<1\). Thus, as \(\epsilon \) is small enough, we have
Then, we can get
By Lemma 2.7, the proof is completed. \(\square \)
Lemma 2.10
Any \((PS)_{c}\) sequence \(\left\{ u_n\right\} \) for \(I_\varepsilon \) is bounded, and
Proof
Suppose that \(\left\{ u_n\right\} \) is a \((PS)_{c}\) sequence of \(I_\varepsilon \), we have
Thus
It follows that
Then \(\left\{ u_n\right\} \) is bounded in E, and the second result holds. \(\square \)
Lemma 2.11
If u is a critical point of \(I_\varepsilon \) on \({\mathcal {N}}_\varepsilon \), then u is a critical point of \(I_\varepsilon \) in E.
Proof
Since u is a critical point of \(I_\varepsilon \) on \({\mathcal {N}}_\varepsilon \), there exists \(\theta \in {\mathbb {R}}\) such that
where \(J_\varepsilon (u)=\langle I'_\varepsilon (u), u\rangle \).
It follows from \(u\in {\mathcal {N}}_\varepsilon \) that
Then, by \(0=\langle I'_\varepsilon (u), u\rangle =\theta \langle J'_\varepsilon (u), u\rangle \), we have \(I'_\varepsilon (u)=0\). \(\square \)
3 The Energy Functional with Constant Coefficients
We need some results about Eq. (2.2) with constant coefficients. Consider the following problem
where k, \(\tau \), and \(\nu \) are positive constants. The associated energy functional is
where
By Lemma 2.7, we have
where \({\mathcal {N}}_{k\tau \nu }=\left\{ u\in E\setminus \left\{ 0\right\} | \langle I_{k\tau \nu }'(u), u\rangle =0\right\} \). Especially, \(I_\infty (u)\), \(m_\infty \), and \({\mathcal {N}}_\infty \) mean \(I_{V_\infty P^\infty Q^\infty }(u)\), \(m_{V_\infty P^\infty Q^\infty }\), and \({\mathcal {N}}_{V_\infty P^\infty Q^\infty }\), respectively.
Lemma 3.1
Problem (3.1) has at least one ground state solution.
Proof
By Lemma 2.7 and Lemma 2.8, there exits a sequence \(\left\{ u_n\right\} \) which is a \((PS)_{m_{k\tau \nu }}\) sequence of \(I_{k\tau \nu }\). By Lemma 2.10, we know that \(\left\{ u_n\right\} \) is bounded in E. Hence, up to a subsequence, we have
It is easy to verify that \(I_{k\tau \nu }'(u)=0\) .
Case1. \(u\ne 0\).
For this case, we have \(u\in {\mathcal {N}}_{k\tau \nu }\). Therefore, \(I_{k\tau \nu }(u)\ge m_{k\tau \nu }\). Then we get
Thus, \(I_{k\tau \nu }(u)= m_{k\tau \nu }\). Moreover, we have \(u_n\rightarrow u\) in E.
Case2. \(u=0\).
Since \(\{u_n\}\) is a \((PS)_{m_{k\tau \nu }}\) sequence of \(I_{k\tau \nu }\), we have
Assume that
It is easy to see that \(l\ne 0\). If \(\int _{{\mathbb {R}}^3}|u_n|^p\mathrm{{d}}x\rightarrow 0\), then \(\nu ^2{\widetilde{F}}(u_n)\rightarrow l\). By the definition of \(S_{H,L}\), we can get
Letting \(n\rightarrow \infty \), we have
Then,
Thus,
which is a contradiction. Therefore, \(\int _{{\mathbb {R}}^3}|u_n|^p\mathrm{{d}}x\rightarrow b>0\) as \(n\rightarrow \infty .\) Thus, by Lions’s Lemma, there exists \(\{y_n\}\subset {\mathbb {R}}^3\), \(\rho ,\eta >0\) such that
Let \({\widetilde{u}}_n(x)=u_n(x+y_n)\). Then \(||{\widetilde{u}}_n||\le C\) in E. This implies that there exists \({\widetilde{u}}\in E\) such that \({\widetilde{u}}_n\rightharpoonup {\widetilde{u}}\) in E and \({\widetilde{u}}_n\rightarrow {\widetilde{u}}\) a.e. in \({\mathbb {R}}^3\). By (3.3), we get \({\widetilde{u}}\ne 0\). It is easy to prove that
Thus, we have \(I_{k\tau \nu }'({\widetilde{u}})=0\) and \({\widetilde{u}}\in {\mathcal {N}}_{k\tau \nu }\). Then the proof follows from the argument used in the case of \(u\ne 0.\) \(\square \)
Lemma 3.2
For \(k_i>0,\) \(\tau _i>0\) and \(\nu _i>0,\) \(i=1,2\). If
then \(m_{k_1\tau _1\nu _1}\le m_{k_2\tau _2\nu _2}\). Additionally, if \(\max \left\{ k_2-k_1,\tau _1-\tau _2,\nu _1-\nu _2\right\} > 0,\) then \(m_{k_1\tau _1\nu _1}< m_{k_2\tau _2\nu _2}\).
Proof
By Lemma 3.1, there exists \(u\in E\) satisfying \(I_{k_2\tau _2\nu _2}(v)= m_{k_2\tau _2\nu _2}=\displaystyle \max _{t\ge 0}I_{k_2\tau _2\nu _2}(tu)\). By Lemma 2.6, there exists \(t_0>0\) such that \(I_{k_1\tau _1\nu _1}(t_0u)=\displaystyle \max _{t\ge 0}I_{k_1\tau _1\nu _1}(tu)\). Then
\(\square \)
Lemma 3.3
For any \(\xi \in {\mathbb {R}}^3\), \(\displaystyle \limsup _{\varepsilon \rightarrow 0}c_\varepsilon \le m_{V(\xi )P(\xi )Q(\xi )}\).
Proof
For any \(\xi \in {\mathbb {R}}^3\), by Lemma 3.1, we assume that u is a ground state solution to the equation corresponding to the functional \(I_{V(\xi )P(\xi )Q(\xi )}\). Set \(u_\varepsilon (x)=\varphi (\varepsilon x-\xi ) u(x-\frac{\xi }{\varepsilon })\), where \(\varphi \in C_0^\infty ({\mathbb {R}}^3,[0,1])\) is a cut-off function satisfying \(\varphi =1\), \(|x|<1\) and \(\varphi =0\), \(|x|\ge 2\). Then, there exists T large enough, such that \(I_\varepsilon (Tu_\varepsilon )<0\). Define \(\gamma _\varepsilon (t)=tTu_\varepsilon \), \(t\in [0,1]\). It is easy to see that \(\gamma _\varepsilon (t)\in \Gamma _\varepsilon \) in Lemma 2.7. By direct computation, we have
Therefore,
Thus,
It follows that \(\displaystyle \limsup _{\varepsilon \rightarrow 0}c_\varepsilon \le m_{V(\xi )P(\xi )Q(\xi )}\). \(\square \)
4 The Palais-Smale Condition
Lemma 4.1
Suppose that the condition \((H_2)\) holds. Let \(\{u_n\}\subset E\) be a \((PS)_c\) sequence for \(I_\varepsilon \) with \(c<\displaystyle \frac{5-\mu }{2(6-\mu )} S^{\frac{6-\mu }{5-\mu }}_{H,L}(Q^\infty )^{\frac{-2}{5-\mu }}\) and such that \(u_n\rightharpoonup 0\) in E. Then, one of the following conclusions holds.
-
(i)
\(u_n\rightarrow 0\) in E;
-
(ii)
There exists a sequence \(\{y_n\}\subset {\mathbb {R}}^3\) and constants \(R,\beta >0\) such that
$$\begin{aligned}\liminf _{n\rightarrow \infty }\int _{B_R(y_n)}|u_n|^2\mathrm{{d}}x\ge \beta .\end{aligned}$$
Proof
Suppose that (ii) does not occur. Then, for any \(R>0\), one has
Then, we have
Noting \(o_n(1)=\langle I'_\varepsilon (u_n),u_n\rangle \), we can obtain
By Lemma 2.10, \(\{u_n\}\) is bounded in E. Up to a subsequence, we can assume that
Assume by contradiction that \(l>0\). From condition \((H_2)\),
By the definition of \(S_{H,L}\), we can get
It follows that
Since \(I_\varepsilon (u_n)=c+o_n(1)\), we can deduce that
which is a contradiction with our assumption. Therefore, \(l=0\) and the conclusion follows. \(\square \)
Lemma 4.2
Suppose that the condition \((H_2)\) holds. Let \(\{u_n\}\subset E\) be a \((PS)_c\) sequence for \(I_\varepsilon \) with \(c<m_\infty \) and \(u_n\rightharpoonup 0\) in E. Then \(u_n\rightarrow 0\) in E.
Proof
Assume that \(u_n\nrightarrow 0\) in E. Let \(\{t_n\}\subset (0,+\infty )\) be a sequence such that \(\{t_nu_n\}\subset {\mathcal {N}}_\infty \). Then, we claim that the sequence \(\{t_n\}\) satisfies that \(\displaystyle \limsup _{n\rightarrow \infty }t_n\le 1\).
Assume by contradiction that there exists \(\delta >0\) and a subsequence still denoted by \(\{t_n\}\), such that, for all \(n\in \mathbb {N}\),
Since \(\langle I'_\varepsilon (u_n),u_n\rangle =o_n(1)\), we get
Using \(t_nu_n\in {\mathcal {N}}_\infty \), we have
Then, we can obtain
By the definition of \(V_\infty \) and \(P^\infty \), for any \(\sigma >0\), there exists \(R=R(\sigma )>0\), such that, for \(|\varepsilon x|\ge R\),
and
Moreover, \(\Vert u_n\Vert _\varepsilon \) is bounded and \(u_n\rightarrow 0\ \hbox {in}\ L^{q}_{loc}({\mathbb {R}}^3), \ \hbox {for }\ 2\le q < 6\). Then, we can obtain
and
Therefore,
Since \(t_n>1+\delta \) and \(Q(\varepsilon x)\le Q^\infty \), it follows from the above inequality that
By the arbitrariness of \(\sigma \), we can obtain
Since \(u_n\nrightarrow 0\) in E, by Lemma 4.1, we know that there exists a sequence \(\{y_n\}\subset {\mathbb {R}}^3\) and constants \(R,\beta >0\) such that
Set \(v_n(x)=u_n(x+y_n)\). Then \(\{v_n(x)\}\) is a bounded sequence in E. Therefore, there exists \(v\in E\) such that
By (4.5), \(v\ne 0\) in E. Then, it follows from Fatou Lemma and (4.4) that
which is a contradiction.
We next distinguish the following two cases.
Case 1: \(\displaystyle \limsup _{n\rightarrow \infty }t_n=1\).
In this case, there exists a subsequence, still denoted by \(\{t_n\}\) such that \(t_n\rightarrow 1\) as \(n\rightarrow \infty \). Then,
Therefore,
By (4.2) and \(u_n\rightarrow 0\ \hbox {in}\ L^{2}_{loc}({\mathbb {R}}^3)\),
Similarly,
Then, noting \(Q(\varepsilon x)\le Q^\infty \), we can obtain
By the arbitrariness of \(\sigma \), we have \(c\ge m_\infty \), which is a contradiction.
Case 2: \(\displaystyle \limsup _{n\rightarrow \infty }t_n<1\).
In this case, we may suppose that \(t_n<1\) for all \(n\in \mathbb {N}\). From (4.1), we can deduce that
Using this result, we have
which means that \(c\ge m_\infty \), a contradiction. \(\square \)
Lemma 4.3
Suppose that the condition \((H_2)\) holds. Then \(I_\varepsilon \) satisfies the \((PS)_c\) condition at any level \(c<m_\infty \).
Proof
Let \(\{u_n\}\) be a \((PS)_c\) sequence. By Lemma 2.10, \(\{u_n\}\) is bounded in E. Then there exists \(u\in E\) such that \(u_n\rightharpoonup u\) in E. By standard argument, \(I'_\varepsilon (u)=0\) and \(I_\varepsilon (u)\ge 0\). Set \(w_n=u_n-u\). It follows from Lemma 2.4 and Brezis–Lieb’ Lemma that \(\{w_n\}\) is a \((PS)_{c-I_\varepsilon (u)}\) sequence. Since \(c-I_\varepsilon (u)<m_\infty \), by Lemma 4.2, \(w_n\rightarrow 0\) in E. Therefore, \(u_n\rightarrow u\) in E. \(\square \)
Lemma 4.4
Suppose that the condition \((H_2)\) holds. Let \(\{u_n\}\) be a \((PS)_c\) sequence restricted on \({\mathcal {N}}_\varepsilon \) and assume \(c<m_\infty \). Then \(\{u_n\}\) has a convergent subsequence in E.
Proof
Let \(\{u_n\}\) be a \((PS)_c\) sequence restricted on \({\mathcal {N}}_\varepsilon \). Then, there exist \(\{\theta _n\}\subset {\mathbb {R}}\) such that
where \(J_\varepsilon (u)=\langle I'_\varepsilon (u), u\rangle \).
It follows from \(u_n\in {\mathcal {N}}_\varepsilon \) and Lemma 2.5 that
From \(0=\langle I'_\varepsilon (u_n), u_n\rangle \) and the above inequality, we have \(\theta _n=o_n(1)\). Therefore, \(I'_\varepsilon (u_n)=o_n(1)\). Thus, by Lemma 4.3, the proof is completed. \(\square \)
5 The Existence of Multiple Solutions
We assume that the conditions \((H_1)\), \((H_2)\), and \((H_3)\) hold in this section. Let us consider a cut-off function \(\eta \in C_0^\infty ({\mathbb {R}}^3,[0,1])\) such that \(\eta (x)=1\) if \(|x|<1,\) \(\eta (x)=0\) if \(|x|>2\) and \(|\nabla \eta |\le C\). Choose \(w\in E\) with \(I^\prime _{V_\mathrm{{min}}P_\mathrm{{max}}Q_\mathrm{{max}}}(w)=0\) and \(I_{V_\mathrm{{min}}P_\mathrm{{max}}Q_\mathrm{{max}}}(w)=m_{V_\mathrm{{min}}P_\mathrm{{max}}Q_\mathrm{{max}}}\). For each \(\xi \in {\mathcal {V}}\cap {\mathcal {P}}\cap {\mathcal {Q}}\), let
Then, there exists a unique \(t_\varepsilon >0\) such that \(t_\varepsilon \Psi _{\varepsilon ,y}\in {\mathcal {N}}_\varepsilon \). Define \(\Phi _\varepsilon : {\mathcal {V}}\cap {\mathcal {P}}\cap {\mathcal {Q}}\rightarrow {\mathcal {N}}_\varepsilon \) by setting \(\Phi _\varepsilon (\xi )=t_\varepsilon \Psi _{\varepsilon ,\xi }\).
Lemma 5.1
\(\lim \limits _{\varepsilon \rightarrow 0}I_\varepsilon (\Phi _\varepsilon (\xi ))=m_{V_\mathrm{{min}}P_\mathrm{{max}}Q_\mathrm{{max}}}~~\hbox {uniformly~in}~\xi \in {\mathcal {V}}\cap {\mathcal {P}}\cap {\mathcal {Q}}.\)
Proof
Suppose that the result is false. Then, there exists some \(\alpha >0\), \(\{\xi _n\}\subset {\mathcal {V}}\cap {\mathcal {P}}\cap {\mathcal {Q}}\) and \(\varepsilon _n\rightarrow 0\) such that
The compactness of \({\mathcal {V}}\cap {\mathcal {P}}\cap {\mathcal {Q}}\) implies that there exists \(\xi \in {\mathcal {V}}\cap {\mathcal {P}}\cap {\mathcal {Q}}\) such that \(\xi _n\rightarrow \xi \), up to a subsequence if necessary. Now we claim that \(\displaystyle \lim _{n\rightarrow \infty }t_{\varepsilon _n}=1\). Indeed, from \(t_{\varepsilon _n}\Psi _{\varepsilon _n,\xi _n}\in {\mathcal {N}}_{\varepsilon _n}\), we have
By using a change of variables and Lebesgue Dominated Convergence Theorem, we can obtain
and
Then \(t_n\) is bounded from above. Thus we can obtain
It follows from Lemma 2.5 that \(t_n\) is has a positive lower bound. Without loss of generality, we assume that \(t_{\varepsilon _n}\rightarrow T>0\). Letting \(n\rightarrow \infty \) in the above expression, we can get
It follows from \(w\in {\mathcal {N}}_{V_\mathrm{{min}}P_\mathrm{{max}}Q_\mathrm{{max}}}\) that \(T=1\). Then, we have
which is a contradiction. \(\square \)
For any \(\delta >0,\) let \(\rho =\rho (\delta )>0\) such that \(({\mathcal {V}}\cap {\mathcal {P}}\cap {\mathcal {Q}})_\delta \subset B_\rho (0)\). Consider \(\chi :{\mathbb {R}}^3\rightarrow {\mathbb {R}}^3\) defined as \(\chi (x)=x\) for \(|x|\le \rho \) and \(\chi (x)=\frac{\rho x}{|x|}\) for \(|x|\ge \rho \). Define \(\beta _\varepsilon : {\mathcal {N}}_\varepsilon \rightarrow {\mathbb {R}}^3\) given by
Lemma 5.2
\( \lim \limits _{\varepsilon \rightarrow 0}\beta _{\varepsilon }(\Phi _\varepsilon (\xi ))=\xi ~\text {uniformly}~\text {in}~\xi \in {\mathbb {V}}\cap {\mathbb {P}}\cap {\mathbb {Q}}. \)
Proof
Suppose by contradiction that there exist \(\delta _0>0,~\{\xi _n\}\subset {\mathcal {V}}\cap {\mathcal {P}}\cap {\mathcal {Q}}\) and \(\varepsilon _n\rightarrow 0\) such that
By the definition of \(\beta _\varepsilon \), we have
Since \(\{\xi _n\}\subset {\mathcal {V}}\cap {\mathcal {P}}\cap {\mathcal {Q}}\subset B_\rho (0)\) and \(\chi \big |_{B_\rho }\equiv id\), we conclude that
which contradicts (5.1) and the desired conclusion holds. \(\square \)
Define the set
where \(h(\varepsilon )=\sup \limits _{\xi \in {\mathcal {V}} \cap {\mathcal {P}} \cap {\mathcal {Q}}}|I_\varepsilon (\Phi _\varepsilon (\xi ))-m_{V_\mathrm{{min}}P_\mathrm{{max}}Q_\mathrm{{max}}}|\). We conclude from Lemma 5.1 that \(h(\varepsilon )\rightarrow 0\) as \(\varepsilon \rightarrow 0\). By the definition of \(h(\varepsilon )\), for any \(\xi \in {\mathcal {V}} \cap {\mathcal {P}} \cap {\mathcal {Q}}\) and \(\varepsilon >0\), \(\Phi _\varepsilon (\xi )\in \tilde{{\mathcal {N}}}_\varepsilon \) and \(\tilde{ {\mathcal {N}}}_\varepsilon \ne \emptyset \).
Lemma 5.3
Let \(\varepsilon _{n} \rightarrow 0\) and \(u_{n} \in {\mathcal {N}}_{\varepsilon _{n}}\) such that \(I_{\varepsilon _{n}}\left( u_{n}\right) \rightarrow m_{V_\mathrm{{min}}P_\mathrm{{max}}Q_\mathrm{{max}}}.\) Then there exists \(\left\{ y_{n}\right\} \subset {\mathbb {R}}^{N}\) such that the sequence \(u_{n}\left( x+y_{n}\right) \) has a convergent subsequence in E. Moreover, up to a subsequence, \(\varepsilon _{n} y_{n} \rightarrow \xi \in {\mathcal {V}} \cap {\mathcal {P}} \cap {\mathcal {Q}}\).
Proof
Since
then \(\left\{ u_{n}\right\} \) is bounded in E. We can have a sequence \(\left\{ y_{n}\right\} \subset {\mathbb {R}}^{3}\) and positive constants \(R, \beta \) such that
If not, for any \(R>0\), one has
Then, we have
Noting \(0=\langle I'_\varepsilon (u_n),u_n\rangle \), we can obtain
Up to a subsequence, assume that
It follows from Lemma 2.5 that \(l>0\). From condition \((H_2)\),
By the definition of \(S_{H,L}\), we can get
It follows that
Since \(I_{\varepsilon _n}(u_n)=m_{V_\mathrm{{min}}P_\mathrm{{max}}Q_\mathrm{{max}}}+o_n(1)\), we can deduce that
which is a contradiction with Lemma 2.9. Therefore, the conclusion follows. Denote \({\tilde{u}}_{n}(x)=u_{n}\left( x+y_{n}\right) \), going if necessary to a subsequence, we can assume that
Let \(t_{n}>0\) be such that \(t_{n} {\tilde{u}}_{n} \in {\mathcal {N}}_{V_\mathrm{{min}}P_\mathrm{{max}}Q_\mathrm{{max}}}\). By the definition of \(I_{V_\mathrm{{min}}P_\mathrm{{max}}Q_\mathrm{{max}}}\) and \( m_{V_\mathrm{{min}}P_\mathrm{{max}}Q_\mathrm{{max}}}\), we obtain
so \(I_{V_\mathrm{{min}}P_\mathrm{{max}}Q_\mathrm{{max}}}\left( t_{n} {\tilde{u}}_{n}\right) \rightarrow m_{V_\mathrm{{min}}P_\mathrm{{max}}Q_\mathrm{{max}}}\). Then \(\left\{ t_{n} {\tilde{u}}_{n}\right\} \) is bounded in E. Since \(t_{n} {\tilde{u}}_{n} \in {\mathcal {N}}_{V_\mathrm{{min}}P_\mathrm{{max}}Q_\mathrm{{max}}}\), it follows from Lemma 2.5 that \(\Vert t_{n} {\tilde{u}}_{n}\Vert \ge C_0\). Noting \(u_n\) is bounded in E, then there exists \(C>0\) such that \(t_nC\ge \Vert t_{n} {\tilde{u}}_{n}\Vert \ge C_0\). Thus \(t_n\) has a positive lower bound. On the other hand, \({\tilde{u}}_{n}\) does not converge to 0 in E, so there exists a \(\delta ^{\prime }>0\) such that \(\left\| {\tilde{u}}_{n}\right\| \ge \delta ^{\prime }\). Therefore, \(t_{n} \delta ^{\prime } \le \) \(\left\| t_{n} {\tilde{u}}_{n}\right\| \le C .\) Thus \(\left\{ t_{n}\right\} \) is bounded from above. Then, up to a subsequence, \(t_{n} \rightarrow t_{0}>0\).
Denote \({\hat{u}}_{n}:=t_{n} {\tilde{u}}_{n}, {\hat{u}}:=t_{0} {\tilde{u}}\), we have
By the Ekeland’s Variational Principle, there exists a sequence \(\left\{ {\hat{w}}_{n}\right\} \subset \) \({\mathcal {N}}_{V_\mathrm{{min}}P_\mathrm{{max}}Q_\mathrm{{max}}}\) satisfying
Therefore,
and \({\hat{u}}=t_{0} {\tilde{u}}\) is a nontrivial critical point of \(I_{V_\mathrm{{min}}P_\mathrm{{max}}Q_\mathrm{{max}}}\). Then
Thus
Now, we are going to prove that \(\varepsilon _{n} y_{n} \rightarrow \xi \in {\mathcal {V}} \cap {\mathcal {P}}\cap {\mathcal {Q}}\). We first claim that \(\{\varepsilon _ny_n\}\) must be bounded. Otherwise, \(|\varepsilon _ny_n|\rightarrow \infty \) as \(n\rightarrow \infty \). For any small \(\delta >0\), there exists \(\rho =\rho (\delta )>0\), such that, for \(|x|\ge \rho \),
For \(u\in E\), define
Then, we can introduce
where \({\mathcal {N}}_\delta =\{u\in E: \langle I'_\delta (u), u\rangle =0\}\).
By Lemma 3.2 and condition \((H_1)\), we have \(m_{V_\mathrm{{min}}P_\mathrm{{max}}Q_\mathrm{{max}}}<m_\infty \). Noting the continuity of \(m_\delta \) about \(\delta \), we can obtain \(m_{V_\mathrm{{min}}P_\mathrm{{max}}Q_\mathrm{{max}}}<m_\delta \) for \(\delta \) small. For \(u_n\), there exists \({\tilde{t}}_n>0\) such that \({\tilde{t}}_nu_n\in {\mathcal {N}}_\delta \). It is easy to see that \(\{{\tilde{t}}_n\}\) is bounded. For any small \(\sigma >0\), from (5.2), there exists \(R>0\) and N big enough, such that
Thus,
From \(|\varepsilon _ny_n|\rightarrow \infty \) as \(n\rightarrow \infty \), we can get \(B_R(y_n)\cap B_{\frac{\rho }{\varepsilon _n}}(0)=\emptyset \). Then, by using (5.4), we have
and
Thus, noting (5.3), we can get
Therefore, \(m_{V_\mathrm{{min}}P_\mathrm{{max}}Q_\mathrm{{max}}}\ge m_\delta \), which is a contradiction.
Up to a subsequence, assume that \(\varepsilon _{n} y_{n} \rightarrow \xi \). Hence, it suffices to show that \(V(\xi )=V_\mathrm{{min}}\), \(P(\xi )=P_\mathrm{{max}}\) and \(Q(\xi )=Q_\mathrm{{max}}\). Arguing by contradiction again, we assume that \(V(\xi )>V_\mathrm{{min}}\), \(P(\xi )<P_\mathrm{{max}}\) or \(Q(\xi )<Q_\mathrm{{max}}\). Since
and
we can obtain
which is a contradiction. Therefore, \(V(\xi )=V_\mathrm{{min}}\), \(P(\xi )=P_\mathrm{{max}}\), and \(Q(\xi )=Q_\mathrm{{max}}\), and the proof is completed. \(\square \)
Lemma 5.4
For any \(\delta >0\), there holds that
Proof
Let \(\{\varepsilon _n\}\subset (0,+\infty )\) be such that \(\varepsilon _n\rightarrow 0\). By definition, there exists \(\{u_n\}\subset \tilde{{\mathcal {N}}}_{\varepsilon _n}\) such that
So, it suffices to find a sequence \(\{\xi _n\}\subset ({\mathcal {V}}\cap {\mathcal {P}}\cap {\mathcal {Q}})_\delta \) satisfying
By Lemma 5.3, we can obtain \({\tilde{u}}\in E\) such that \(u_n(x+y_n)\rightarrow {\tilde{u}}\) in E, and, up to a subsequence, \(\varepsilon _ny_n\rightarrow \xi \in {\mathcal {V}}\cap {\mathcal {P}}\cap {\mathcal {Q}}.\) Thus, \(\varepsilon _ny_n\in ({\mathcal {V}}\cap {\mathcal {P}}\cap {\mathcal {Q}})_\delta \) for n large enough. It is easy to see that
Set \(\xi _n=\varepsilon _ny_n\). We have that the sequence \(\{\xi _n\}\) satisfies (5.5). This completes the proof. \(\square \)
Lemma 5.5
[4] Let I be a \(C^1\) functional defined on a \(C^1\) Finsler manifold M. If I is bounded from below and satisfies the (PS) condition, then I has at least \(cat_MM\) distinct critical points.
Lemma 5.6
[2] Let \(\Gamma ,\Omega ^+,\Omega ^-\) be closed sets with \(\Omega ^-\subset \Omega ^+.\) Let \(\Phi :\Omega ^-\rightarrow \Gamma ,\) \(\beta :\Gamma \rightarrow \Omega ^+\) be two continuous maps such that \(\beta \circ \Phi \) is homotopically equivalent to the embedding \(Id:\Omega ^-\rightarrow \Omega ^+.\) Then \(cat_\Gamma \Gamma \ge cat_{\Omega ^+}\Omega ^-.\)
The proof of Theorem 1.1:
(i) For a fixed \(\delta >0\), by Lemmas 5.1 and 5.4, we know that there exists a \(\varepsilon _\delta >0\) such that for any \(\varepsilon \in (0,\varepsilon _\delta )\), the diagram
is well defined. From Lemma 5.2, for \(\varepsilon \) small enough, there is a function \(\lambda (\xi )\) with \(|\lambda (\xi )|<\frac{\delta }{2}\) uniformly in \(\xi \in {\mathcal {V}} \cap {\mathcal {P}}\cap {\mathcal {Q}}\), such that \(\beta _\varepsilon (\Phi _\varepsilon (\xi ))=\xi +\lambda (\xi )\) for all \(\xi \in {\mathcal {V}} \cap {\mathcal {P}}\cap {\mathcal {Q}}\). Define \(H(t,\xi )=\xi +(1-t)\lambda (\xi )\). Then, \(H:[0,1]\times {\mathcal {V}} \cap {\mathcal {P}}\cap {\mathcal {Q}}\rightarrow ({\mathcal {V}} \cap {\mathcal {P}}\cap {\mathcal {Q}})_\delta \) is continuous. Obviously, \(H(0,\xi )=\beta _\varepsilon (\Phi _\varepsilon (\xi )),H(1,\xi )=\xi \) for all \(\xi \in {\mathcal {V}} \cap {\mathcal {P}}\cap {\mathcal {Q}}\). That is, \(H(t,\xi )\) is homotopy between \(\beta _\varepsilon \circ \Phi _\varepsilon \) and the inclusion map \(id:{\mathcal {V}} \cap {\mathcal {P}}\cap {\mathcal {Q}}\rightarrow ({\mathcal {V}} \cap {\mathcal {P}}\cap {\mathcal {Q}})_\delta \). By Lemma 5.6, we obtain
On the other hand, using the definition of \({\tilde{\mathcal {N}}}_\varepsilon \) and choosing \(\varepsilon _\delta \) small if necessary, we see that \(I_\varepsilon \) satisfies the (PS) condition in \({\tilde{N}}_\varepsilon \) recalling \((H_1)\) and Lemma 4.4. By Lemma 5.5, we know that \(I_\varepsilon \) has at least \(cat_{\tilde{\mathcal {N}}_\varepsilon }(\tilde{\mathcal {N}}_\varepsilon )\) critical points on \({\mathcal {N}}_\varepsilon \). By Lemma 2.11, these points are critical points of \(I_\varepsilon \) in E. Consequently, we see that the problem (1.1) has at least \(cat_{({\mathcal {V}} \cap {\mathcal {P}}\cap {\mathcal {Q}})_\delta }({\mathcal {V}} \cap {\mathcal {P}}\cap {\mathcal {Q}})\) solutions.
(ii) By the definition of \(c_\varepsilon \) and \(m_{V_\mathrm{{min}}P_\mathrm{{max}}Q_\mathrm{{max}}}\), we have \(m_{V_\mathrm{{min}}P_\mathrm{{max}}Q_\mathrm{{max}}}\le c_\varepsilon \). Then,
It follows from Lemma 3.3 that
By Lemma 5.3, there exist \(\{y_n\}\subset {\mathbb {R}}^3\) and \(v\in E\) such that
Now we prove that v is a ground state solution of equation 1.2. For any \(\psi \in \mathcal {C}_0^\infty ({\mathbb {R}}^3)\), since \(I'_{\varepsilon _n}(u_n)=0\), we have \(\langle I'_{\varepsilon _n}(u_n(x)), \psi (x-y_n)\rangle =0\). By direct computation, it is easy to get
On the other hand,
Thus, v is a ground solution of Eq. (1.2). \(\square \)
References
Alves, C.O., Gao, F., Squassina, M., Yang, M.: Singularly perturbed critical Choquard equations. J. Differ. Equ. 263(7), 3943–3988 (2017)
Benci, V., Cerami, G.: The effect of the domain topology on the number of positive solutions of nonlinear elliptic problems. Arch. Rational Mech. Anal. 114(1), 79–93 (1991)
Brézis, H., Nirenberg, L.: Positive solutions of nonlinear elliptic equations involving critical Sobolev exponents. Commun. Pure Appl. Math. 36(4), 437–477 (1983)
Chang, K.: Infinite-dimensional Morse theory and multiple solution problems. Progress in Nonlinear Differential Equations and their Applications, vol. 6. Birkhäuser Boston Inc, Boston, MA (1993)
Chen, S., Li, Y., Yang, Z.: Multiplicity and concentration of nontrivial nonnegative solutions for a fractional Choquard equation with critical exponent. Rev. R. Acad. Cienc. Exactas Fís. Nat. Ser. A Mat. RACSAM, 114(1): Paper No. 33, 35 (2020)
Gao, F., Yang, M.: The Brezis–Nirenberg type critical problem for the nonlinear Choquard equation. Sci. China Math. 61(7), 1219–1242 (2018)
Gao, F., Yang, M.: Infinitely many non-radial solutions for a Choquard equation. Adv. Nonlinear Anal. 11(1), 1085–1096 (2022)
Guo, L., Hu, T., Peng, S., Shuai, W.: Existence and uniqueness of solutions for Choquard equation involving Hardy-Littlewood-Sobolev critical exponent. Calc. Var. Part. Differ. Equ. 58(4), 1–34 (2019)
Ji, C., Rădulescu, V.D.: Multi-bump solutions for the nonlinear magnetic Choquard equation with deepening potential well. J. Differ. Equ. 306, 251–279 (2022)
Lieb, E.H.: Existence and uniqueness of the minimizing solution of Choquard’s nonlinear equation. Stud. Appl. Math., 57(2):93–105, (1976/77)
Lions, P.-L.: The Choquard equation and related questions. Nonlinear Anal. 4(6), 1063–1072 (1980)
Liu, M., Tang, Z.: Multiplicity and concentration of solutions for Choquard equation via Nehari method and pseudo-index theory. Discrete Contin. Dyn. Syst. 39(6), 3365–3398 (2019)
Luo, P., Peng, S., Wang, C.: Uniqueness of positive solutions with concentration for the Schrödinger-Newton problem. Calc. Var. Part. Differ. Equ. 59(2), 41 (2020)
Ma, Li., Zhao, Lin: Classification of positive solitary solutions of the nonlinear Choquard equation. Arch. Ration. Mech. Anal. 195(2), 455–467 (2010)
Moroz, V., Van Schaftingen, J.: Groundstates of nonlinear Choquard equations: existence, qualitative properties and decay asymptotics. J. Funct. Anal. 265(2), 153–184 (2013)
Moroz, V., Van Schaftingen, J.: A guide to the Choquard equation. J. Fixed Point Theory Appl. 19(1), 773–813 (2017)
Pekar, S.: Untersuchung über die Elektronentheorie der Kristalle. Akademie Verlag, Berlin (1954)
Penrose, R.: Quantum computation, entanglement and state reduction. R. Soc. Lond. Philos. Trans. Ser. A 356(1743), 1927–1939 (1998)
Shen, Z., Gao, F., Yang, M.: On critical Choquard equation with potential well. Discrete Contin. Dyn. Syst. 38(7), 3567–3593 (2018)
Wang, X.: On concentration of positive bound states of nonlinear Schrödinger equations. Commun. Math. Phys. 153(2), 229–244 (1993)
Wang, X., Zeng, B.: On concentration of positive bound states of nonlinear Schrödinger equations with competing potential functions. SIAM J. Math. Anal. 28(3), 633–655 (1997)
Wei, J., Winter, M.: Strongly interacting bumps for the Schrödinger–Newton equations. J. Math. Phys. 50(1), 22 (2009)
Yang, M., Zhao, F., Zhao, S.: Classification of solutions to a nonlocal equation with doubly Hardy–Littlewood–Sobolev critical exponents. Discrete Contin. Dyn. Syst. 41(11), 5209–5241 (2021)
Yang, Z., Yu, Y.: Existence and concentration of solution for Schrödinger-Poisson system with local potential. Partial Differ. Equ. Appl. 2(4), 22 (2021)
Yang, Z., Yu, Y., Zhao, F.: The concentration behavior of ground state solutions for a critical fractional Schrödinger-Poisson system. Math. Nachr. 292(8), 1837–1868 (2019)
Yang, Z., Zhao, F.: Multiplicity and concentration behaviour of solutions for a fractional Choquard equation with critical growth. Adv. Nonlinear Anal. 10(1), 732–774 (2021)
Yu, Y., Zhao, F., Zhao, L.: The concentration behavior of ground state solutions for a fractional Schrödinger–Poisson system. Calc. Var. Part. Differ. Equ. 56(4), 25 (2017)
Zhang, F., Zhang, H.: Existence and concentration of ground states for a Choquard equation with competing potentials. J. Math. Anal. Appl. 465(1), 159–174 (2018)
Zhang, H., Zhang, F.: Multiplicity and concentration of solutions for Choquard equations with critical growth. J. Math. Anal. Appl. 481(1), 21 (2020)
Zhang, J., Zhang, W.: Semiclassical states for coupled nonlinear Schrödinger system with competing potentials. J. Geom. Anal. 32(4), 36 (2022)
Zhang, Y., Tang, X., Rădulescu, V.D.: High and low perturbations of Choquard equations with critical reaction and variable growth. Discrete Contin. Dyn. Syst. 42(4), 1971–2003 (2022)
Acknowledgements
Y. Chen is supported by National Natural Science Foundation of China (No. 12161007), Guangxi science and technology base and talent project (AD21238019), and Doctoral Foundation of Guangxi University of Science and Technology (20Z30). Z. Yang is supported by National Natural Science Foundation of China (Nos. 11771385, 11961081).
Author information
Authors and Affiliations
Corresponding author
Ethics declarations
Conflict of interest
The authors have no competing interests to declare that are relevant to the content of this article.
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
About this article
Cite this article
Chen, Y., Yang, Z. Existence and Asymptotical Behavior of Multiple Solutions for the Critical Choquard Equation. J Geom Anal 32, 238 (2022). https://doi.org/10.1007/s12220-022-00976-2
Received:
Accepted:
Published:
DOI: https://doi.org/10.1007/s12220-022-00976-2