Show Summary Details
More options …

Editor-in-Chief: Radulescu, Vicentiu / Squassina, Marco

IMPACT FACTOR 2018: 6.636

CiteScore 2018: 5.03

SCImago Journal Rank (SJR) 2018: 3.215
Source Normalized Impact per Paper (SNIP) 2018: 3.225

Mathematical Citation Quotient (MCQ) 2018: 3.18

Open Access
Online
ISSN
2191-950X
See all formats and pricing
More options …
Volume 8, Issue 1

# A fractional Kirchhoff problem involving a singular term and a critical nonlinearity

Alessio Fiscella
Published Online: 2017-08-03 | DOI: https://doi.org/10.1515/anona-2017-0075

## Abstract

In this paper, we consider the following critical nonlocal problem:

where Ω is an open bounded subset of ${ℝ}^{N}$ with continuous boundary, dimension $N>2s$ with parameter $s\in \left(0,1\right)$, ${2}_{s}^{*}=2N/\left(N-2s\right)$ is the fractional critical Sobolev exponent, $\lambda >0$ is a real parameter, $\gamma \in \left(0,1\right)$ and M models a Kirchhoff-type coefficient, while ${\left(-\mathrm{\Delta }\right)}^{s}$ is the fractional Laplace operator. In particular, we cover the delicate degenerate case, that is, when the Kirchhoff function M is zero at zero. By combining variational methods with an appropriate truncation argument, we provide the existence of two solutions.

MSC 2010: 35J75; 35R11; 49J35; 35A15; 45G05; 35S15

## 1 Introduction

This paper is devoted to the study of a class of Kirchhoff-type problems driven by a nonlocal fractional operator and involving a singular term and a critical nonlinearity. More precisely, we consider

(1.1)

where Ω is an open bounded subset of ${ℝ}^{N}$ with continuous boundary, dimension $N>2s$ with parameter $s\in \left(0,1\right)$, ${2}_{s}^{*}=2N/\left(N-2s\right)$ is the fractional critical Sobolev exponent, $\lambda >0$ is a real parameter, $\theta \in \left(1,{2}_{s}^{*}/2\right)$, while $\gamma \in \left(0,1\right)$. Here ${\left(-\mathrm{\Delta }\right)}^{s}$ is the fractional Laplace operator defined, up to normalization factors, by the Riesz potential as

${\left(-\mathrm{\Delta }\right)}^{s}\phi \left(x\right)={\int }_{{ℝ}^{N}}\frac{2\phi \left(x\right)-\phi \left(x+y\right)-\phi \left(x-y\right)}{{|y|}^{N+2s}}𝑑y,x\in {ℝ}^{N},$

along any $\phi \in {C}_{0}^{\mathrm{\infty }}\left(\mathrm{\Omega }\right)$; we refer to [11] and the recent monograph [22] for further details on the fractional Laplacian and the fractional Sobolev spaces ${H}^{s}\left({ℝ}^{N}\right)$ and ${H}_{0}^{s}\left(\mathrm{\Omega }\right)$.

As is well explained in [11, 22], problem (1.1) is the fractional version of the following nonlinear problem:

(1.2)

where Δ denotes the classical Laplace operator while, just for a general discussion, $M\left(t\right)={t}^{\theta -1}$ for any $t\in {ℝ}_{0}^{+}$. In literature, problems like (1.1) and (1.2) are called of Kirchhoff type whenever the function $M:{ℝ}_{0}^{+}\to {ℝ}_{0}^{+}$ models the Kirchhoff prototype, given by

$M\left(t\right)=a+b{t}^{\theta -1},a,b\ge 0,a+b>0,\theta \ge 1.$(1.3)

In particular, when $M\left(t\right)\ge \mathrm{constant}>0$ for any $t\in {ℝ}_{0}^{+}$, Kirchhoff problems are said to be non-degenerate and this happens for example if $a>0$ in the model case (1.3). While if $M\left(0\right)=0$ but $M\left(t\right)>0$ for any $t\in {ℝ}^{+}$, Kirchhoff problems are called degenerate. Of course, for (1.3) this occurs when $a=0$.

This kind of nonlocal problems has been widely studied in recent years. We refer to [17, 18, 19, 21, 20] for different Kirchhoff problems with M like in (1.3), driven by the Laplace operator and involving a singular term of type ${u}^{-\gamma }$. In [21], Liu and Sun study a Kirchhoff problem with a singular term and a Hardy potential by using the Nehari method. The same approach is used in [19] for a singular Kirchhoff problem with also a subcritical term. In [17], strongly assuming $a>0$ in (1.3), Lei, Liao and Tang prove the existence of two solutions for a Kirchhoff problem like (1.2) by combining perturbation and variational methods. While in [18], Liao, Ke, Lei and Tang provide a uniqueness result for a singular Kirchhoff problem involving a negative critical nonlinearity by a minimization argument. By arguing similarly to [17], Liu, Tang, Liao and Wu [20] give the existence of two solutions for a critical Kirchhoff problem with a singular term of type ${|x|}^{-\beta }{u}^{-\gamma }$.

Problem (1.1) has been studied by Barrios, De Bonis, Medina and Peral [4] when $\theta =1$, namely without a Kirchhoff coefficient. They prove the existence of two solutions by applying the sub/supersolutions and Sattinger methods. In [8], Canino, Montoro, Sciunzi and Squassina generalize the results of [4, Section 3] to the delicate case of the p-fractional Laplace operator ${\left(-\mathrm{\Delta }\right)}_{p}^{s}$. While in the last section of [1], Abdellaoui, Medina, Peral and Primo provide the existence of a solution for nonlinear fractional problems with a singularity like ${u}^{-\gamma }$ and a fractional Hardy term by perturbation methods. Concerning fractional Kirchhoff problems involving critical nonlinearities, we refer to [2, 9, 13, 14, 16, 23] for existence results and to [5, 12, 24, 25, 29] for multiplicity results. In particular, in [9, 13, 14, 23] different singular terms appear, but are given by the fractional Hardy potential.

Inspired by the above works, we study a multiplicity result for problem (1.1). As far as we know, a fractional Kirchhoff problem involving a singular term of type ${u}^{-\gamma }$ has not been studied yet. We can state our result as follows.

#### Theorem 1.1.

Let $s\mathrm{\in }\mathrm{\left(}\mathrm{0}\mathrm{,}\mathrm{1}\mathrm{\right)}$, $N\mathrm{>}\mathrm{2}\mathit{}s$, $\theta \mathrm{\in }\mathrm{\left(}\mathrm{1}\mathrm{,}{\mathrm{2}}_{s}^{\mathrm{*}}\mathrm{/}\mathrm{2}\mathrm{\right)}$, $\gamma \mathrm{\in }\mathrm{\left(}\mathrm{0}\mathrm{,}\mathrm{1}\mathrm{\right)}$ and let Ω be an open bounded subset of ${\mathrm{R}}^{N}$ with $\mathrm{\partial }\mathit{}\mathrm{\Omega }$ continuous. Then there exists $\overline{\lambda }\mathrm{>}\mathrm{0}$ such that for any $\lambda \mathrm{\in }\mathrm{\left(}\mathrm{0}\mathrm{,}\overline{\lambda }\mathrm{\right)}$ problem (1.1) has at least two different solutions.

The first solution of problem (1.1) is obtained by a suitable minimization argument, where we must pay attention to the nonlocal nature of the fractional Laplacian. Concerning the second solution, because of the presence of ${u}^{-\gamma }$, we can not apply the usual critical point theory to problem (1.1). For this, we first study a perturbed problem obtained by truncating the singular term ${u}^{-\gamma }$. Then by approximation we get our second solution of (1.1).

Finally, we observe that Theorem 1.1 generalizes in several directions the first part of [4, Theorem 4.1] and [17, Theorem 1.1].

The paper is organized as follows: In Section 2, we discuss the variational formulation of problem (1.1), and we introduce the perturbed problem. In Section 3, we prove the existence of the first solution of (1.1), and we give a possible generalization of this existence result at the end of the section. In Section 4, we prove the existence of a mountain pass solution for the perturbed problem. In Section 5, we prove Theorem 1.1.

## 2 Variational setting

Throughout this paper, we assume without further mentioning that $s\in \left(0,1\right)$, $N>2s$, $\theta \in \left(1,{2}_{s}^{*}/2\right)$, $\gamma \in \left(0,1\right)$ and Ω is an open bounded subset of ${ℝ}^{N}$ with $\partial \mathrm{\Omega }$ continuous. As a matter of notations, we denote with ${\phi }^{+}=\mathrm{max}\left\{\phi ,0\right\}$ and ${\phi }^{-}=\mathrm{max}\left\{-\phi ,0\right\}$ respectively the positive and negative part of a function φ.

Problem (1.1) has a variational structure, and the natural space where to find solutions is the homogeneous fractional Sobolev space ${H}_{0}^{s}\left(\mathrm{\Omega }\right)$. In order to study (1.1) it is important to encode the “boundary condition” $u=0$ in ${ℝ}^{N}\setminus \mathrm{\Omega }$ in the weak formulation, by considering also that the interaction between Ω and its complementary in ${ℝ}^{N}$ gives a positive contribution in the so-called Gagliardo norm, given as

${\parallel u\parallel }_{{H}^{s}\left({ℝ}^{N}\right)}={\parallel u\parallel }_{{L}^{2}\left({ℝ}^{N}\right)}+{\left({\iint }_{{ℝ}^{2N}}\frac{{|u\left(x\right)-u\left(y\right)|}^{2}}{{|x-y|}^{N+2s}}𝑑x𝑑y\right)}^{1/2}.$(2.1)

The functional space that takes into account this boundary condition will be denoted by ${X}_{0}$ and it is defined as

We refer to [26] for a general definition of ${X}_{0}$ and its properties. We also would like to point out that when $\partial \mathrm{\Omega }$ is continuous, by [15, Theorem 6] the space ${C}_{0}^{\mathrm{\infty }}\left(\mathrm{\Omega }\right)$ is dense in ${X}_{0}$, with respect to the norm (2.1). This last point will be used to overcome the singularity in problem (1.1).

In ${X}_{0}$ we can consider the norm

${\parallel u\parallel }_{{X}_{0}}={\left({\iint }_{{ℝ}^{2N}}\frac{{|u\left(x\right)-u\left(y\right)|}^{2}}{{|x-y|}^{N+2s}}𝑑x𝑑y\right)}^{1/2},$

which is equivalent to the usual one defined in (2.1) (see [26, Lemma 6]). We also recall that $\left({X}_{0},{\parallel \cdot \parallel }_{{X}_{0}}\right)$ is a Hilbert space, with the scalar product defined as

${〈u,v〉}_{{X}_{0}}={\iint }_{{ℝ}^{2N}}\frac{\left(u\left(x\right)-u\left(y\right)\right)\left(v\left(x\right)-v\left(y\right)\right)}{{|x-y|}^{N+2s}}𝑑x𝑑y.$

From now on, in order to simplify the notation, we will denote ${\parallel \cdot \parallel }_{{X}_{0}}$ and ${〈\cdot ,\cdot 〉}_{{X}_{0}}$ by $\parallel \cdot \parallel$ and $〈\cdot ,\cdot 〉$, respectively, and ${\parallel \cdot \parallel }_{{L}^{q}\left(\mathrm{\Omega }\right)}$ by ${\parallel \cdot \parallel }_{q}$ for any $q\in \left[1,\mathrm{\infty }\right]$.

In order to present the weak formulation of (1.1) and taking into account that we are looking for positive solutions, we will consider the following Kirchhoff problem:

(2.2)

We say that $u\in {X}_{0}$ is a (weak) solution of problem (2.2) if u satisfies

${\parallel u\parallel }^{2\left(\theta -1\right)}〈u,\phi 〉=\lambda {\int }_{\mathrm{\Omega }}\frac{\phi }{{\left({u}^{+}\right)}^{\gamma }}𝑑x+{\int }_{\mathrm{\Omega }}{\left({u}^{+}\right)}^{{2}_{s}^{*}-1}\phi$(2.3)

for any $\phi \in {X}_{0}$. Problem (2.2) has a variational structure and ${J}_{\lambda }:{X}_{0}\to ℝ$, defined by

${J}_{\lambda }\left(u\right)=\frac{1}{2\theta }{\parallel u\parallel }^{2\theta }-\frac{\lambda }{1-\gamma }{\int }_{\mathrm{\Omega }}{\left({u}^{+}\right)}^{1-\gamma }𝑑x-\frac{1}{{2}_{s}^{*}}{\parallel {u}^{+}\parallel }_{{2}_{s}^{*}}^{{2}_{s}^{*}},$

is the underlying functional associated to (2.2). Because of the presence of a singular term in (2.2), the functional ${J}_{\lambda }$ is not differentiable on ${X}_{0}$. Therefore, we can not apply directly the usual critical point theory to ${J}_{\lambda }$ in order to solve problem (2.2). However, it is possible to find a first solution of (2.2) by using a local minimization argument. In order to get the second solution of (2.2) we have to study an associated approximating problem. That is, for any $n\in ℕ$, we consider the following perturbed problem:

(2.4)

For this, we say that $u\in {X}_{0}$ is a (weak) solution of problem (2.4) if u satisfies

${\parallel u\parallel }^{2\left(\theta -1\right)}〈u,\phi 〉=\lambda {\int }_{\mathrm{\Omega }}\frac{\phi }{{\left({u}^{+}+\frac{1}{n}\right)}^{\gamma }}𝑑x+{\int }_{\mathrm{\Omega }}{\left({u}^{+}\right)}^{{2}_{s}^{*}-1}\phi$(2.5)

for any $\phi \in {X}_{0}$. In this case, solutions of (2.4) correspond to the critical points of the functional ${J}_{n,\lambda }:{X}_{0}\to ℝ$, set as

${J}_{n,\lambda }\left(u\right)=\frac{1}{2\theta }{\parallel u\parallel }^{2\theta }-\frac{\lambda }{1-\gamma }{\int }_{\mathrm{\Omega }}\left[{\left({u}^{+}+\frac{1}{n}\right)}^{1-\gamma }-{\left(\frac{1}{n}\right)}^{1-\gamma }\right]𝑑x-\frac{1}{{2}_{s}^{*}}{\parallel {u}^{+}\parallel }_{{2}_{s}^{*}}^{{2}_{s}^{*}}.$(2.6)

It is immediate to see that ${J}_{n,\lambda }$ is of class ${C}^{1}\left({X}_{0}\right)$.

We conclude this section by recalling the best constant of the fractional Sobolev embedding, which will be very useful to study the compactness property of the functional ${J}_{n,\lambda }$. That is, we consider

$S=\underset{\begin{array}{c}v\in {H}^{s}\left({ℝ}^{N}\right)\\ v\not\equiv 0\end{array}}{inf}\frac{{\iint }_{{ℝ}^{2N}}\frac{{|v\left(x\right)-v\left(y\right)|}^{2}}{{|x-y|}^{N+2s}}𝑑x𝑑y}{{\left({\int }_{{ℝ}^{N}}{|v\left(x\right)|}^{{2}_{s}^{*}}𝑑x\right)}^{2/{2}_{s}^{*}}},$(2.7)

which is well defined and strictly positive, as shown in [10, Theorem 1.1].

## 3 A first solution for problem (1.1)

In this section, we prove the existence of a solution for problem (1.1) by a local minimization argument. For this, we first study the geometry of the functional ${J}_{\lambda }$.

#### Lemma 3.1.

There exist numbers $\rho \mathrm{\in }\mathrm{\left(}\mathrm{0}\mathrm{,}\mathrm{1}\mathrm{\right]}$, ${\lambda }_{\mathrm{0}}\mathrm{=}{\lambda }_{\mathrm{0}}\mathit{}\mathrm{\left(}\rho \mathrm{\right)}\mathrm{>}\mathrm{0}$ and $\alpha \mathrm{=}\alpha \mathit{}\mathrm{\left(}\rho \mathrm{\right)}\mathrm{>}\mathrm{0}$ such that ${J}_{\lambda }\mathit{}\mathrm{\left(}u\mathrm{\right)}\mathrm{\ge }\alpha$ for any $u\mathrm{\in }{X}_{\mathrm{0}}$, with $\mathrm{\parallel }u\mathrm{\parallel }\mathrm{=}\rho$, and for any $\lambda \mathrm{\in }\mathrm{\left(}\mathrm{0}\mathrm{,}{\lambda }_{\mathrm{0}}\mathrm{\right]}$. Furthermore, set

${m}_{\lambda }=inf\left\{{J}_{\lambda }\left(u\right):u\in {\overline{B}}_{\rho }\right\},$

where ${\overline{B}}_{\rho }\mathrm{=}\mathrm{\left\{}u\mathrm{\in }{X}_{\mathrm{0}}\mathrm{:}\mathrm{\parallel }u\mathrm{\parallel }\mathrm{\le }\rho \mathrm{\right\}}$. Then ${m}_{\lambda }\mathrm{<}\mathrm{0}$ for any $\lambda \mathrm{\in }\mathrm{\left(}\mathrm{0}\mathrm{,}{\lambda }_{\mathrm{0}}\mathrm{\right]}$.

#### Proof.

Let $\lambda >0$. From the Hölder inequality and (2.7) for any $u\in {X}_{0}$ we have

${\int }_{\mathrm{\Omega }}{\left({u}^{+}\right)}^{1-\gamma }𝑑x\le {|\mathrm{\Omega }|}^{\left({2}_{s}^{*}-1+\gamma \right)/{2}_{s}^{*}}{\parallel u\parallel }_{{2}_{s}^{*}}^{1-\gamma }\le {|\mathrm{\Omega }|}^{\left({2}_{s}^{*}-1+\gamma \right)/{2}_{s}^{*}}{S}^{-\left(1-\gamma \right)/2}{\parallel u\parallel }^{1-\gamma }.$(3.1)

Hence, using again (2.7) and (3.1), we get

${J}_{\lambda }\left(u\right)\ge \frac{1}{2\theta }{\parallel u\parallel }^{2\theta }-\frac{{S}^{-{2}_{s}^{*}/2}}{{2}_{s}^{*}}{\parallel u\parallel }^{{2}_{s}^{*}}-\frac{\lambda }{1-\gamma }{|\mathrm{\Omega }|}^{\left({2}_{s}^{*}-1+\gamma \right)/{2}_{s}^{*}}{S}^{-\left(1-\gamma \right)/2}{\parallel u\parallel }^{1-\gamma }.$

Since $1-\gamma <1<2\theta <{2}_{s}^{*}$, the function

$\eta \left(t\right)=\frac{1}{2\theta }{t}^{2\theta -1+\gamma }-\frac{{S}^{-{2}_{s}^{*}/2}}{{2}_{s}^{*}}{t}^{{2}_{s}^{*}-1+\gamma },t\in \left[0,1\right],$

admits a maximum at some $\rho \in \left(0,1\right]$ small enough, that is, ${\mathrm{max}}_{t\in \left[0,1\right]}\eta \left(t\right)=\eta \left(\rho \right)>0$. Thus, let

${\lambda }_{0}=\frac{\left(1-\gamma \right){S}^{\left(1-\gamma \right)/2}}{2{|\mathrm{\Omega }|}^{\left({2}_{s}^{*}-1+\gamma \right)/{2}_{s}^{*}}}\eta \left(\rho \right).$

Then for any $u\in {X}_{0}$ with $\parallel u\parallel =\rho \le 1$ and for any $\lambda \le {\lambda }_{0}$, we get ${J}_{\lambda }\left(u\right)\ge {\rho }^{1-\gamma }\eta \left(\rho \right)/2=\alpha >0$.

Furthermore, fixed $v\in {X}_{0}$ with ${v}^{+}\not\equiv 0$, for $t\in \left(0,1\right)$ sufficiently small we have

${J}_{\lambda }\left(tv\right)=\frac{{t}^{2\theta }}{2\theta }{\parallel v\parallel }^{2\theta }-{t}^{1-\gamma }\frac{\lambda }{1-\gamma }{\int }_{\mathrm{\Omega }}{\left({v}^{+}\right)}^{1-\gamma }𝑑x-\frac{{t}^{{2}_{s}^{*}}}{{2}_{s}^{*}}{\parallel {v}^{+}\parallel }_{{2}_{s}^{*}}^{{2}_{s}^{*}}<0$

since $1-\gamma <1<2\theta <{2}_{s}^{*}$. ∎

We are now ready to prove the existence of the first solution of (1.1).

#### Theorem 3.2.

Let ${\lambda }_{\mathrm{0}}$ be given as in Lemma 3.1. Then for any $\lambda \mathrm{\in }\mathrm{\left(}\mathrm{0}\mathrm{,}{\lambda }_{\mathrm{0}}\mathrm{\right]}$ problem (1.1) has a solution ${u}_{\mathrm{0}}\mathrm{\in }{X}_{\mathrm{0}}$ with ${J}_{\lambda }\mathit{}\mathrm{\left(}{u}_{\mathrm{0}}\mathrm{\right)}\mathrm{<}\mathrm{0}$.

#### Proof.

Fix $\lambda \in \left(0,{\lambda }_{0}\right]$ and let ρ be as given in Lemma 3.1. We first prove that there exists ${u}_{0}\in {\overline{B}}_{\rho }$ such that ${J}_{\lambda }\left({u}_{0}\right)={m}_{\lambda }<0$. Let ${\left\{{u}_{k}\right\}}_{k}\subset {\overline{B}}_{\rho }$ be a minimizing sequence for ${m}_{\lambda }$, that is, such that

$\underset{k\to \mathrm{\infty }}{lim}{J}_{\lambda }\left({u}_{k}\right)={m}_{\lambda }.$(3.2)

Since ${\left\{{u}_{k}\right\}}_{k}$ is bounded in ${X}_{0}$, by applying [26, Lemma 8] and [6, Theorem 4.9], there exist a subsequence, still denoted by ${\left\{{u}_{k}\right\}}_{k}$, and a function ${u}_{0}\in {\overline{B}}_{\rho }$ such that, as $k\to \mathrm{\infty }$, we have

(3.3)

Since $\gamma \in \left(0,1\right)$, by the Hölder inequality, for any $k\in ℕ$ we have

$|{\int }_{\mathrm{\Omega }}{\left({u}_{k}^{+}\right)}^{1-\gamma }𝑑x-{\int }_{\mathrm{\Omega }}{\left({u}_{0}^{+}\right)}^{1-\gamma }𝑑x|\le {\int }_{\mathrm{\Omega }}{|{u}_{k}^{+}-{u}_{0}^{+}|}^{1-\gamma }𝑑x\le {\parallel {u}_{k}^{+}-{u}_{0}^{+}\parallel }_{2}^{1-\gamma }{|\mathrm{\Omega }|}^{\left(1+\gamma \right)/2},$

which yields, by (3.3),

$\underset{k\to \mathrm{\infty }}{lim}{\int }_{\mathrm{\Omega }}{\left({u}_{k}^{+}\right)}^{1-\gamma }𝑑x={\int }_{\mathrm{\Omega }}{\left({u}_{0}^{+}\right)}^{1-\gamma }𝑑x.$(3.4)

Let ${w}_{k}={u}_{k}-{u}_{0}$; by [7, Theorem 2] it holds true that

${\parallel {u}_{k}\parallel }^{2}={\parallel {w}_{k}\parallel }^{2}+{\parallel {u}_{0}\parallel }^{2}+o\left(1\right),{\parallel {u}_{k}\parallel }_{{2}_{s}^{*}}^{{2}_{s}^{*}}={\parallel {w}_{k}\parallel }_{{2}_{s}^{*}}^{{2}_{s}^{*}}+{\parallel {u}_{0}\parallel }_{{2}_{s}^{*}}^{{2}_{s}^{*}}+o\left(1\right)$(3.5)

as $k\to \mathrm{\infty }$. Since ${\left\{{u}_{k}\right\}}_{k}\subset {\overline{B}}_{\rho }$, by (3.5) for k sufficiently large, we have ${w}_{k}\in {\overline{B}}_{\rho }$. Lemma 3.1 implies that for any $u\in {X}_{0}$, with $\parallel u\parallel =\rho$, we get

$\frac{1}{2\theta }{\parallel u\parallel }^{2\theta }-\frac{1}{{2}_{s}^{*}}{\parallel {u}^{+}\parallel }_{{2}_{s}^{*}}^{{2}_{s}^{*}}\ge \alpha >0,$

and from this, since $\rho \le 1$, for k sufficiently large we have

$\frac{1}{2\theta }{\parallel {w}_{k}\parallel }^{2\theta }-\frac{1}{{2}_{s}^{*}}{\parallel {w}_{k}^{+}\parallel }_{{2}_{s}^{*}}^{{2}_{s}^{*}}>0.$(3.6)

Thus, by (3.2), (3.4)–(3.6) and considering $\theta \ge 1$, it follows that, as $k\to \mathrm{\infty }$,

${m}_{\lambda }={J}_{\lambda }\left({u}_{k}\right)+o\left(1\right)$$=\frac{1}{2\theta }{\left({\parallel {w}_{k}\parallel }^{2}+{\parallel {u}_{0}\parallel }^{2}\right)}^{\theta }-\frac{\lambda }{1-\gamma }{\int }_{\mathrm{\Omega }}{\left({u}_{0}^{+}\right)}^{1-\gamma }𝑑x-\frac{1}{{2}_{s}^{*}}\left({\parallel {w}_{k}^{+}\parallel }_{{2}_{s}^{*}}^{{2}_{s}^{*}}+{\parallel {u}_{0}^{+}\parallel }_{{2}_{s}^{*}}^{{2}_{s}^{*}}\right)+o\left(1\right)$$\ge {J}_{\lambda }\left({u}_{0}\right)+\frac{1}{2\theta }{\parallel {w}_{k}\parallel }^{2\theta }-\frac{1}{{2}_{s}^{*}}{\parallel {w}_{k}^{+}\parallel }_{{2}_{s}^{*}}^{{2}_{s}^{*}}+o\left(1\right)\ge {J}_{\lambda }\left({u}_{0}\right)+o\left(1\right)\ge {m}_{\lambda }$

since ${u}_{0}\in {\overline{B}}_{\rho }$. Hence, ${u}_{0}$ is a local minimizer for ${J}_{\lambda }$, with ${J}_{\lambda }\left({u}_{0}\right)={m}_{\lambda }<0$, which implies that ${u}_{0}$ is nontrivial.

Now, we prove that ${u}_{0}$ is a positive solution of (2.2). For any $\psi \in {X}_{0}$, with $\psi \ge 0$ a.e. in ${ℝ}^{N}$, let us consider a $t>0$ sufficiently small so that ${u}_{0}+t\psi \in {\overline{B}}_{\rho }$. Since ${u}_{0}$ is a local minimizer for ${J}_{\lambda }$, we have

$0\le {J}_{\lambda }\left({u}_{0}+t\psi \right)-{J}_{\lambda }\left({u}_{0}\right)$$=\frac{1}{2\theta }\left({\parallel {u}_{0}+t\psi \parallel }^{2\theta }-{\parallel {u}_{0}\parallel }^{2\theta }\right)-\frac{\lambda }{1-\gamma }{\int }_{\mathrm{\Omega }}\left[{\left({\left({u}_{0}+t\psi \right)}^{+}\right)}^{1-\gamma }-{\left({u}_{0}^{+}\right)}^{1-\gamma }\right]𝑑x-\frac{1}{{2}_{s}^{*}}\left({\parallel {u}_{0}+t\psi \parallel }_{{2}_{s}^{*}}^{{2}_{s}^{*}}-{\parallel {u}_{0}^{+}\parallel }_{{2}_{s}^{*}}^{{2}_{s}^{*}}\right).$

From this, by dividing by $t>0$ and passing to the limit as $t\to {0}^{+}$, it follows that

$\underset{t\to {0}^{+}}{lim inf}\frac{\lambda }{1-\gamma }{\int }_{\mathrm{\Omega }}\frac{{\left({\left({u}_{0}+t\psi \right)}^{+}\right)}^{1-\gamma }-{\left({u}_{0}^{+}\right)}^{1-\gamma }}{t}𝑑x\le {\parallel {u}_{0}\parallel }^{2\left(\theta -1\right)}〈{u}_{0},\psi 〉-{\int }_{\mathrm{\Omega }}{\left({u}_{0}^{+}\right)}^{{2}_{s}^{*}-1}\psi 𝑑x.$(3.7)

We observe that

with $\xi \in \left(0,1\right)$ and ${\left({\left({u}_{0}+\xi t\psi \right)}^{+}\right)}^{-\gamma }\to {\left({u}_{0}^{+}\right)}^{-\gamma }\psi$ a.e. in Ω as $t\to {0}^{+}$. Thus, by the Fatou lemma, we obtain

$\lambda {\int }_{\mathrm{\Omega }}{\left({u}_{0}^{+}\right)}^{-\gamma }\psi 𝑑x\le \underset{t\to {0}^{+}}{lim inf}\frac{\lambda }{1-\gamma }{\int }_{\mathrm{\Omega }}\frac{{\left({\left({u}_{0}+t\psi \right)}^{+}\right)}^{1-\gamma }-{\left({u}_{0}^{+}\right)}^{1-\gamma }}{t}𝑑x.$(3.8)

Therefore, combining (3.7) and (3.8), we get

${\parallel {u}_{0}\parallel }^{2\left(\theta -1\right)}〈{u}_{0},\psi 〉-\lambda {\int }_{\mathrm{\Omega }}{\left({u}_{0}^{+}\right)}^{-\gamma }\psi 𝑑x-{\int }_{\mathrm{\Omega }}{\left({u}_{0}^{+}\right)}^{{2}_{s}^{*}-1}\psi 𝑑x\ge 0$(3.9)

for any $\psi \in {X}_{0}$ with $\psi \ge 0$ a.e. in ${ℝ}^{N}$.

Since ${J}_{\lambda }\left({u}_{0}\right)<0$ and by Lemma 3.1, we have ${u}_{0}\in {B}_{\rho }$. Hence, there exists $\delta \in \left(0,1\right)$ such that $\left(1+t\right){u}_{0}\in {\overline{B}}_{\rho }$ for any $t\in \left[-\delta ,\delta \right]$. Let us define ${I}_{\lambda }\left(t\right)={J}_{\lambda }\left(\left(1+t\right){u}_{0}\right)$. Since ${u}_{0}$ is a local minimizer for ${J}_{\lambda }$ in ${\overline{B}}_{\rho }$, the functional ${I}_{\lambda }$ has a minimum at $t=0$, that is,

${I}_{\lambda }^{\prime }\left(0\right)={\parallel {u}_{0}\parallel }^{2\theta }-\lambda {\int }_{\mathrm{\Omega }}{\left({u}_{0}^{+}\right)}^{1-\gamma }𝑑x-{\int }_{\mathrm{\Omega }}{\left({u}_{0}^{+}\right)}^{{2}_{s}^{*}}𝑑x=0.$(3.10)

For any $\phi \in {X}_{0}$ and any $\epsilon >0$, let us define ${\psi }_{\epsilon }={u}_{0}^{+}+\epsilon \phi$. Then by (3.9) we have

$0\le {\parallel {u}_{0}\parallel }^{2\left(\theta -1\right)}〈{u}_{0},{\psi }_{\epsilon }^{+}〉-\lambda {\int }_{\mathrm{\Omega }}{\left({u}_{0}^{+}\right)}^{-\gamma }{\psi }_{\epsilon }^{+}𝑑x-{\int }_{\mathrm{\Omega }}{\left({u}_{0}^{+}\right)}^{{2}_{s}^{*}-1}{\psi }_{\epsilon }^{+}𝑑x$$={\parallel {u}_{0}\parallel }^{2\left(\theta -1\right)}〈{u}_{0},{\psi }_{\epsilon }+{\psi }_{\epsilon }^{-}〉-\lambda {\int }_{\mathrm{\Omega }}{\left({u}_{0}^{+}\right)}^{-\gamma }\left({\psi }_{\epsilon }+{\psi }_{\epsilon }^{-}\right)𝑑x-{\int }_{\mathrm{\Omega }}{\left({u}_{0}^{+}\right)}^{{2}_{s}^{*}-1}\left({\psi }_{\epsilon }+{\psi }_{\epsilon }^{-}\right)𝑑x.$(3.11)

We observe that, for a.e. $x,y\in {ℝ}^{N}$, we obtain

$\left({u}_{0}\left(x\right)-{u}_{0}\left(y\right)\right)\left({u}_{0}^{-}\left(x\right)-{u}_{0}^{-}\left(y\right)\right)=-{u}_{0}^{+}\left(x\right){u}_{0}^{-}\left(y\right)-{u}_{0}^{-}\left(x\right){u}_{0}^{+}\left(y\right)-{\left[{u}_{0}^{-}\left(x\right)-{u}_{0}^{-}\left(y\right)\right]}^{2}$$\le -{|{u}_{0}^{-}\left(x\right)-{u}_{0}^{-}\left(y\right)|}^{2},$(3.12)

from which we immediately get

$\left({u}_{0}\left(x\right)-{u}_{0}\left(y\right)\right)\left({u}_{0}^{+}\left(x\right)-{u}_{0}^{+}\left(y\right)\right)\le {|{u}_{0}\left(x\right)-{u}_{0}\left(y\right)|}^{2}.$

From the last inequality it follows that

$〈{u}_{0},{\psi }_{\epsilon }+{\psi }_{\epsilon }^{-}〉={\iint }_{{ℝ}^{2N}}\frac{\left({u}_{0}\left(x\right)-{u}_{0}\left(y\right)\right)\left({\psi }_{\epsilon }\left(x\right)+{\psi }_{\epsilon }^{-}\left(x\right)-{\psi }_{\epsilon }\left(y\right)-{\psi }_{\epsilon }^{-}\left(y\right)\right)}{{|x-y|}^{N+2s}}𝑑x𝑑y$$\le {\iint }_{{ℝ}^{2N}}\frac{{|{u}_{0}\left(x\right)-{u}_{0}\left(y\right)|}^{2}}{{|x-y|}^{N+2s}}𝑑x𝑑y+\epsilon {\iint }_{{ℝ}^{2N}}\frac{\left({u}_{0}\left(x\right)-{u}_{0}\left(y\right)\right)\left(\phi \left(x\right)-\phi \left(y\right)\right)}{{|x-y|}^{N+2s}}𝑑x𝑑y$$+{\iint }_{{ℝ}^{2N}}\frac{\left({u}_{0}\left(x\right)-{u}_{0}\left(y\right)\right)\left({\psi }_{\epsilon }^{-}\left(x\right)-{\psi }_{\epsilon }^{-}\left(y\right)\right)}{{|x-y|}^{N+2s}}𝑑x𝑑y.$(3.13)

Hence, denoting ${\mathrm{\Omega }}_{\epsilon }=\left\{x\in {ℝ}^{N}:{u}_{0}^{+}\left(x\right)+\epsilon \phi \left(x\right)\le 0\right\}$ and combining (3.11) with (3.13), we get

$0\le {\parallel {u}_{0}\parallel }^{2\theta }+\epsilon {\parallel {u}_{0}\parallel }^{2\left(\theta -1\right)}〈{u}_{0},\phi 〉+{\parallel {u}_{0}\parallel }^{2\left(\theta -1\right)}〈{u}_{0},{\psi }_{\epsilon }^{-}〉-\lambda {\int }_{\mathrm{\Omega }}{\left({u}_{0}^{+}\right)}^{-\gamma }\left({u}_{0}^{+}+\epsilon \phi \right)𝑑x-{\int }_{\mathrm{\Omega }}{\left({u}_{0}^{+}\right)}^{{2}_{s}^{*}-1}\left({u}_{0}^{+}+\epsilon \phi \right)𝑑x$$+\lambda {\int }_{{\mathrm{\Omega }}_{\epsilon }}{\left({u}_{0}^{+}\right)}^{-\gamma }\left({u}_{0}^{+}+\epsilon \phi \right)𝑑x+{\int }_{{\mathrm{\Omega }}_{\epsilon }}{\left({u}_{0}^{+}\right)}^{{2}_{s}^{*}-1}\left({u}_{0}^{+}+\epsilon \phi \right)𝑑x$$\le {\parallel {u}_{0}\parallel }^{2\theta }-\lambda {\int }_{\mathrm{\Omega }}{\left({u}_{0}^{+}\right)}^{1-\gamma }𝑑x-{\int }_{\mathrm{\Omega }}{\left({u}_{0}^{+}\right)}^{{2}_{s}^{*}}𝑑x+{\parallel {u}_{0}\parallel }^{2\left(\theta -1\right)}〈{u}_{0},{\psi }_{\epsilon }^{-}〉$$+\epsilon \left[{\parallel {u}_{0}\parallel }^{2\left(\theta -1\right)}〈{u}_{0},\phi 〉-\lambda {\int }_{\mathrm{\Omega }}{\left({u}_{0}^{+}\right)}^{-\gamma }\phi 𝑑x-{\int }_{\mathrm{\Omega }}{\left({u}_{0}^{+}\right)}^{{2}_{s}^{*}-1}\phi 𝑑x\right]$$={\parallel {u}_{0}\parallel }^{2\left(\theta -1\right)}〈{u}_{0},{\psi }_{\epsilon }^{-}〉+\epsilon \left[{\parallel {u}_{0}\parallel }^{2\left(\theta -1\right)}〈{u}_{0},\phi 〉-\lambda {\int }_{\mathrm{\Omega }}{\left({u}_{0}^{+}\right)}^{-\gamma }\phi 𝑑x-{\int }_{\mathrm{\Omega }}{\left({u}_{0}^{+}\right)}^{{2}_{s}^{*}-1}\phi 𝑑x\right],$(3.14)

where the last equality follows from (3.10). Arguing similarly to (3.12), for a.e. $x,y\in {ℝ}^{N}$ we have

$\left({u}_{0}\left(x\right)-{u}_{0}\left(y\right)\right)\left({u}_{0}^{+}\left(x\right)-{u}_{0}^{+}\left(y\right)\right)\ge {|{u}_{0}^{+}\left(x\right)-{u}_{0}^{+}\left(y\right)|}^{2}.$(3.15)

Thus, denoting

${\mathcal{𝒰}}_{\epsilon }\left(x,y\right)=\frac{\left({u}_{0}\left(x\right)-{u}_{0}\left(y\right)\right)\left({\psi }_{\epsilon }^{-}\left(x\right)-{\psi }_{\epsilon }^{-}\left(y\right)\right)}{{|x-y|}^{N+2s}},$

by the symmetry of the fractional kernel and (3.15), we get

$〈{u}_{0},{\psi }_{\epsilon }^{-}〉={\iint }_{{\mathrm{\Omega }}_{\epsilon }×{\mathrm{\Omega }}_{\epsilon }}{\mathcal{𝒰}}_{\epsilon }\left(x,y\right)𝑑x𝑑y+2{\iint }_{{\mathrm{\Omega }}_{\epsilon }×\left({ℝ}^{N}\setminus {\mathrm{\Omega }}_{\epsilon }\right)}{\mathcal{𝒰}}_{\epsilon }\left(x,y\right)𝑑x𝑑y$$\le -\epsilon \left({\iint }_{{\mathrm{\Omega }}_{\epsilon }×{\mathrm{\Omega }}_{\epsilon }}\mathcal{𝒰}\left(x,y\right)𝑑x𝑑y+2{\iint }_{{\mathrm{\Omega }}_{\epsilon }×\left({ℝ}^{N}\setminus {\mathrm{\Omega }}_{\epsilon }\right)}\mathcal{𝒰}\left(x,y\right)𝑑x𝑑y\right)$$\le 2\epsilon {\iint }_{{\mathrm{\Omega }}_{\epsilon }×{ℝ}^{N}}|\mathcal{𝒰}\left(x,y\right)|𝑑x𝑑y,$(3.16)

where we set

$\mathcal{𝒰}\left(x,y\right)=\frac{\left({u}_{0}\left(x\right)-{u}_{0}\left(y\right)\right)\left(\phi \left(x\right)-\phi \left(y\right)\right)}{{|x-y|}^{N+2s}}.$

Clearly $\mathcal{𝒰}\in {L}^{1}\left({ℝ}^{2N}\right)$, so that for any $\sigma >0$ there exists ${R}_{\sigma }$ sufficiently large such that

${\iint }_{\left(\mathrm{supp}\phi \right)×\left({ℝ}^{N}\setminus {B}_{{R}_{\sigma }}\right)}|\mathcal{𝒰}\left(x,y\right)|𝑑x𝑑y<\frac{\sigma }{2}.$

Also, by the definition of ${\mathrm{\Omega }}_{\epsilon }$, we have ${\mathrm{\Omega }}_{\epsilon }\subset \mathrm{supp}\phi$ and $|{\mathrm{\Omega }}_{\epsilon }×{B}_{{R}_{\sigma }}|\to 0$ as $\epsilon \to {0}^{+}$. Thus, since $\mathcal{𝒰}\in {L}^{1}\left({ℝ}^{2N}\right)$, there exist ${\delta }_{\sigma }>0$ and ${\epsilon }_{\sigma }>0$ such that for any $\epsilon \in \left(0,{\epsilon }_{\sigma }\right]$,

$|{\mathrm{\Omega }}_{\epsilon }×{B}_{{R}_{\sigma }}|<{\delta }_{\sigma }\mathit{ }\text{and}\mathit{ }{\iint }_{{\mathrm{\Omega }}_{\epsilon }×{B}_{{R}_{\sigma }}}|\mathcal{𝒰}\left(x,y\right)|𝑑x𝑑y<\frac{\sigma }{2}.$

Therefore, for any $\epsilon \in \left(0,{\epsilon }_{\sigma }\right]$,

${\iint }_{{\mathrm{\Omega }}_{\epsilon }×{ℝ}^{N}}|\mathcal{𝒰}\left(x,y\right)|𝑑x𝑑y<\sigma ,$

from which we get

$\underset{\epsilon \to {0}^{+}}{lim}{\iint }_{{\mathrm{\Omega }}_{\epsilon }×{ℝ}^{N}}|\mathcal{𝒰}\left(x,y\right)|𝑑x𝑑y=0.$(3.17)

Combining (3.14) with (3.16), dividing by ε, letting $\epsilon \to {0}^{+}$ and using (3.17), we obtain

${\parallel {u}_{0}\parallel }^{2\left(\theta -1\right)}〈{u}_{0},\phi 〉-\lambda {\int }_{\mathrm{\Omega }}{\left({u}_{0}^{+}\right)}^{-\gamma }\phi 𝑑x-{\int }_{\mathrm{\Omega }}{\left({u}_{0}^{+}\right)}^{{2}_{s}^{*}-1}\phi 𝑑x\ge 0$

for any $\phi \in {X}_{0}$. By the arbitrariness of φ, we prove that ${u}_{0}$ verifies (2.3), that is, ${u}_{0}$ is a nontrivial solution of (2.2).

Finally, considering $\phi ={u}_{0}^{-}$ in (2.3) and using (3.12), we see that $\parallel {u}_{0}^{-}\parallel =0$, which implies that ${u}_{0}$ is nonnegative. Moreover, by the maximum principle in [28, Proposition 2.17], we can deduce that ${u}_{0}$ is a positive solution of (2.2), and so also solves problem (1.1). This concludes the proof. ∎

We end this section by observing that the result in Theorem 3.2 can be extended to more general Kirchhoff problems. That is, we can consider the problem

(3.18)

where ${p}_{s}^{*}=pN/\left(N-ps\right)$, with $N>ps$ and $p>1$, while the Kirchhoff coefficient M satisfies the following condition:

• ${(\mathcal{M})}$

$M:{ℝ}_{0}^{+}\to {ℝ}_{0}^{+}$ is continuous and nondecreasing. There exist numbers $a>0$ and ϑ such that for any $t\in {ℝ}_{0}^{+}$,

The main operator ${\left(-\mathrm{\Delta }\right)}_{p}^{s}$ is the fractional p-Laplacian which may be defined, for any function $\phi \in {C}_{0}^{\mathrm{\infty }}\left(\mathrm{\Omega }\right)$, as

${\left(-\mathrm{\Delta }\right)}_{p}^{s}\phi \left(x\right)=2\underset{\epsilon \to {0}^{+}}{lim}{\int }_{{ℝ}^{N}\setminus {B}_{\epsilon }\left(x\right)}\frac{{|\phi \left(x\right)-\phi \left(y\right)|}^{p-2}\left(\phi \left(x\right)-\phi \left(y\right)\right)}{{|x-y|}^{N+ps}}𝑑y,x\in {ℝ}^{N},$

where ${B}_{\epsilon }\left(x\right)=\left\{y\in {ℝ}^{N}:|x-y|<\epsilon \right\}$. Then, arguing as in the proof of Theorem 3.2 and observing that we have not used yet the assumption that $\partial \mathrm{\Omega }$ is continuous, we can prove the following result.

#### Theorem 3.3.

Let $s\mathrm{\in }\mathrm{\left(}\mathrm{0}\mathrm{,}\mathrm{1}\mathrm{\right)}$, $p\mathrm{>}\mathrm{1}$, $N\mathrm{>}p\mathit{}s$, $\gamma \mathrm{\in }\mathrm{\left(}\mathrm{0}\mathrm{,}\mathrm{1}\mathrm{\right)}$ and let Ω be an open bounded subset of ${\mathrm{R}}^{N}$. Let M satisfy $\mathrm{\left(}\mathcal{M}\mathrm{\right)}$. Then there exists ${\lambda }_{\mathrm{0}}\mathrm{>}\mathrm{0}$ such that for any $\lambda \mathrm{\in }\mathrm{\left(}\mathrm{0}\mathrm{,}{\lambda }_{\mathrm{0}}\mathrm{\right]}$ problem (3.18) admits a solution.

## 4 A mountain pass solution for problem (2.4)

In this section, we prove the existence of a positive solution for the perturbed problem (2.4) by the mountain pass theorem. For this, throughout this section we assume $n\in ℕ$ without further mentioning. Now, we first prove that the related functional ${J}_{n,\lambda }$ satisfies all the geometric features required by the mountain pass theorem.

#### Lemma 4.1.

Let $\rho \mathrm{\in }\mathrm{\left(}\mathrm{0}\mathrm{,}\mathrm{1}\mathrm{\right]}$, ${\lambda }_{\mathrm{0}}\mathrm{=}{\lambda }_{\mathrm{0}}\mathit{}\mathrm{\left(}\rho \mathrm{\right)}\mathrm{>}\mathrm{0}$ and $\alpha \mathrm{=}\alpha \mathit{}\mathrm{\left(}\rho \mathrm{\right)}\mathrm{>}\mathrm{0}$ be given as in Lemma 3.1. Then, for any $\lambda \mathrm{\in }\mathrm{\left(}\mathrm{0}\mathrm{,}{\lambda }_{\mathrm{0}}\mathrm{\right]}$ and any $u\mathrm{\in }{X}_{\mathrm{0}}$ with $\mathrm{\parallel }u\mathrm{\parallel }\mathrm{\le }\rho$, one has ${J}_{n\mathrm{,}\lambda }\mathit{}\mathrm{\left(}u\mathrm{\right)}\mathrm{\ge }\alpha$. Furthermore, there exists $e\mathrm{\in }{X}_{\mathrm{0}}$, with $\mathrm{\parallel }e\mathrm{\parallel }\mathrm{>}\rho$, such that ${J}_{n\mathrm{,}\lambda }\mathit{}\mathrm{\left(}e\mathrm{\right)}\mathrm{<}\mathrm{0}$.

#### Proof.

Since $\gamma \in \left(0,1\right)$, by the subadditivity of $t↦{t}^{1-\gamma }$, we have

(4.1)

for any $u\in {X}_{0}$ and any $n\in ℕ$. Thus, we have ${J}_{n,\lambda }\left(u\right)\ge {J}_{\lambda }\left(u\right)$ for any $u\in {X}_{0}$ and the first part of the lemma directly follows by Lemma 3.1.

For any $v\in {X}_{0}$, with ${v}^{+}\not\equiv 0$, and $t>0$, we have

since $1-\gamma <1<2\theta <{2}_{s}^{*}$. Hence, we can find $e\in {X}_{0}$, with $\parallel e\parallel >\rho$ sufficiently large, such that ${J}_{n,\lambda }\left(e\right)<0$. ∎

We discuss now the compactness property for the functional ${J}_{n,\lambda }$, given by the Palais–Smale condition. We recall that ${\left\{{u}_{k}\right\}}_{k}\subset {X}_{0}$ is a Palais–Smale sequence for ${J}_{n,\lambda }$ at level $c\in ℝ$ if

(4.2)

We say that ${J}_{n,\lambda }$ satisfies the Palais–Smale condition at level c if any Palais–Smale sequence ${\left\{{u}_{k}\right\}}_{k}$ at level c admits a convergent subsequence in ${X}_{0}$.

Before proving this compactness condition, we introduce the following positive constants, which will help us for a better explanation:

${D}_{1}=\left(\frac{1}{2\theta }-\frac{1}{{2}_{s}^{*}}\right){S}^{{2}_{s}^{*}\theta /\left({2}_{s}^{*}-2\theta \right)},{D}_{2}=\frac{{\left[\left(\frac{1}{1-\gamma }+\frac{1}{{2}_{s}^{*}}\right){|\mathrm{\Omega }|}^{\left({2}_{s}^{*}-1+\gamma \right)/{2}_{s}^{*}}{S}^{-\left(1-\gamma \right)/2}\right]}^{2\theta /\left(2\theta -1+\gamma \right)}}{{\left(\frac{1}{2\theta }-\frac{1}{{2}_{s}^{*}}\right)}^{\left(1-\gamma \right)/\left(2\theta -1+\gamma \right)}}.$(4.3)

#### Lemma 4.2.

Let $\lambda \mathrm{>}\mathrm{0}$. Then the functional ${J}_{n\mathrm{,}\lambda }$ satisfies the Palais–Smale condition at any level $c\mathrm{\in }\mathrm{R}$ verifying

$c<{D}_{1}-{D}_{2}{\lambda }^{2\theta /\left(2\theta -1+\gamma \right)},$(4.4)

with ${D}_{\mathrm{1}}$, ${D}_{\mathrm{2}}\mathrm{>}\mathrm{0}$ given as in (4.3).

#### Proof.

Let $\lambda >0$ and let ${\left\{{u}_{k}\right\}}_{k}$ be a Palais–Smale sequence in ${X}_{0}$ at level $c\in ℝ$, with c satisfying (4.4). We first prove the boundedness of ${\left\{{u}_{k}\right\}}_{k}$. By (4.2), there exists $\sigma >0$ such that, as $k\to \mathrm{\infty }$,

$c+\sigma \parallel {u}_{k}\parallel +o\left(1\right)\ge {J}_{n,\lambda }\left({u}_{k}\right)-\frac{1}{{2}_{s}^{*}}〈{J}_{n,\lambda }^{\prime }\left({u}_{k}\right),{u}_{k}〉$$=\left(\frac{1}{2\theta }-\frac{1}{{2}_{s}^{*}}\right){\parallel {u}_{k}\parallel }^{2\theta }-\frac{\lambda }{1-\gamma }{\int }_{\mathrm{\Omega }}\left[{\left({u}_{k}^{+}+\frac{1}{n}\right)}^{1-\gamma }-{\left(\frac{1}{n}\right)}^{1-\gamma }\right]𝑑x+\frac{\lambda }{{2}_{s}^{*}}{\int }_{\mathrm{\Omega }}{\left({u}_{k}^{+}+\frac{1}{n}\right)}^{-\gamma }{u}_{k}𝑑x$$\ge \left(\frac{1}{2\theta }-\frac{1}{{2}_{s}^{*}}\right){\parallel {u}_{k}\parallel }^{2\theta }-\lambda \left(\frac{1}{1-\gamma }+\frac{1}{{2}_{s}^{*}}\right){\int }_{\mathrm{\Omega }}{|{u}_{k}|}^{1-\gamma }𝑑x$$\ge \left(\frac{1}{2\theta }-\frac{1}{{2}_{s}^{*}}\right){\parallel {u}_{k}\parallel }^{2\theta }-\lambda \left(\frac{1}{1-\gamma }+\frac{1}{{2}_{s}^{*}}\right){|\mathrm{\Omega }|}^{\left({2}_{s}^{*}-1+\gamma \right)/{2}_{s}^{*}}{S}^{-\left(1-\gamma \right)/2}{\parallel {u}_{k}\parallel }^{1-\gamma },$

where the last two inequalities follow by (2.7), (4.1) and the Hölder inequality. Therefore, ${\left\{{u}_{k}\right\}}_{k}$ is bounded in ${X}_{0}$ since $1-\gamma <1<2\theta$. Also, ${\left\{{u}_{k}^{-}\right\}}_{k}$ is bounded in ${X}_{0}$, and by (4.2) we have

$\underset{k\to \mathrm{\infty }}{lim}〈{J}_{n,\lambda }^{\prime }\left({u}_{k}\right),-{u}_{k}^{-}〉=\underset{k\to \mathrm{\infty }}{lim}{\parallel {u}_{k}\parallel }^{2\left(\theta -1\right)}〈{u}_{k},-{u}_{k}^{-}〉=0.$

Thus, by inequality (3.12) we deduce that $\parallel {u}_{k}^{-}\parallel \to 0$ as $k\to \mathrm{\infty }$, which yields

Hence, we can suppose that ${\left\{{u}_{k}\right\}}_{k}$ is a sequence of nonnegative functions.

By the boundedness of ${\left\{{u}_{k}\right\}}_{k}$ and by using [26, Lemma 8] and [6, Theorem 4.9], there exist a subsequence, still denoted by ${\left\{{u}_{k}\right\}}_{k}$, and a function $u\in {X}_{0}$ such that

(4.5)

as $k\to \mathrm{\infty }$, with $h\in {L}^{p}\left(\mathrm{\Omega }\right)$ for a fixed $p\in \left[1,{2}_{s}^{*}\right)$. If $\mu =0$, then immediately ${u}_{k}\to 0$ in ${X}_{0}$ as $k\to \mathrm{\infty }$. Hence, let us assume that $\mu >0$.

Since $n\in ℕ$, by (4.5) it follows that

so by the dominated convergence theorem and (4.5) we have

$\underset{k\to \mathrm{\infty }}{lim}{\int }_{\mathrm{\Omega }}\frac{{u}_{k}-u}{{\left({u}_{k}+\frac{1}{n}\right)}^{\gamma }}𝑑x=0.$(4.6)

By (4.5) and [7, Theorem 2], we have

${\parallel {u}_{k}\parallel }^{2}={\parallel {u}_{k}-u\parallel }^{2}+{\parallel u\parallel }^{2}+o\left(1\right),{\parallel {u}_{k}\parallel }_{{2}_{s}^{*}}^{{2}_{s}^{*}}={\parallel {u}_{k}-u\parallel }_{{2}_{s}^{*}}^{{2}_{s}^{*}}+{\parallel u\parallel }_{{2}_{s}^{*}}^{{2}_{s}^{*}}+o\left(1\right)$(4.7)

as $k\to \mathrm{\infty }$. Consequently, we deduce from (4.2), (4.5), (4.6) and (4.7) that, as $k\to \mathrm{\infty }$,

$o\left(1\right)=〈{J}_{n,\lambda }^{\prime }\left({u}_{k}\right),{u}_{k}-u〉$$={\parallel {u}_{k}\parallel }^{2\left(\theta -1\right)}〈{u}_{k},{u}_{k}-u〉-\lambda {\int }_{\mathrm{\Omega }}\frac{{u}_{k}-u}{{\left({u}_{k}+\frac{1}{n}\right)}^{\gamma }}𝑑x-{\int }_{\mathrm{\Omega }}{u}_{k}^{{2}_{s}^{*}-1}\left({u}_{k}-u\right)𝑑x$$={\mu }^{2\left(\theta -1\right)}\left({\mu }^{2}-{\parallel u\parallel }^{2}\right)-{\parallel {u}_{k}\parallel }_{{2}_{s}^{*}}^{{2}_{s}^{*}}+{\parallel u\parallel }_{{2}_{s}^{*}}^{{2}_{s}^{*}}+o\left(1\right)$$={\mu }^{2\left(\theta -1\right)}{\parallel {u}_{k}-u\parallel }^{2}-{\parallel {u}_{k}-u\parallel }_{{2}_{s}^{*}}^{{2}_{s}^{*}}+o\left(1\right).$

Therefore, we have proved the crucial formula

${\mu }^{2\left(\theta -1\right)}\underset{k\to \mathrm{\infty }}{lim}{\parallel {u}_{k}-u\parallel }^{2}=\underset{k\to \mathrm{\infty }}{lim}{\parallel {u}_{k}-u\parallel }_{{2}_{s}^{*}}^{{2}_{s}^{*}}.$(4.8)

If $\mathrm{\ell }=0$, since $\mu >0$, by (4.5) and (4.8) we have ${u}_{k}\to u$ in ${X}_{0}$ as $k\to \mathrm{\infty }$, concluding the proof.

Thus, let us assume by contradiction that $\mathrm{\ell }>0$. By (2.7), the notation in (4.5) and (4.8), we get

${\mathrm{\ell }}^{{2}_{s}^{*}}\ge S{\mu }^{2\left(\theta -1\right)}{\mathrm{\ell }}^{2}.$(4.9)

Noting that (4.8) implies in particular that

${\mu }^{2\left(\theta -1\right)}\left({\mu }^{2}-{\parallel u\parallel }^{2}\right)={\mathrm{\ell }}^{{2}_{s}^{*}},$

by using (4.9), it follows that

${\left({\mathrm{\ell }}^{{2}_{s}^{*}}\right)}^{2s/N}={\left({\mu }^{2\left(\theta -1\right)}\right)}^{2s/N}{\left({\mu }^{2}-{\parallel u\parallel }^{2}\right)}^{2s/N}\ge S{\mu }^{2\left(\theta -1\right)}.$

From this we obtain

${\mu }^{4s/N}\ge {\left({\mu }^{2}-{\parallel u\parallel }^{2}\right)}^{2s/N}\ge S{\left({\mu }^{2\left(\theta -1\right)}\right)}^{\left(N-2s\right)/N}.$

Considering $N<2s\theta /\left(\theta -1\right)=2s{\theta }^{\prime }$, we have

${\mu }^{2}\ge {S}^{N/\left(2s\theta -N\left(\theta -1\right)\right)}.$(4.10)

Indeed, the restriction $N/\left(2{\theta }^{\prime }\right) follows directly from the fact that $1<\theta <{2}_{s}^{*}/2=N/\left(N-2s\right)$. By (4.1), considering that $n\in ℕ$, for any $k\in ℕ$ we have

${J}_{n,\lambda }\left({u}_{k}\right)-\frac{1}{{2}_{s}^{*}}〈{J}_{n,\lambda }^{\prime }\left({u}_{k}\right),{u}_{k}〉\ge \left(\frac{1}{2\theta }-\frac{1}{{2}_{s}^{*}}\right){\parallel {u}_{k}\parallel }^{2\theta }-\lambda \left(\frac{1}{1-\gamma }+\frac{1}{{2}_{s}^{*}}\right){\int }_{\mathrm{\Omega }}{u}_{k}^{1-\gamma }𝑑x.$

From this, as $k\to \mathrm{\infty }$, since $\theta \ge 1$, by (4.2), (4.5), (4.7), (4.10), the Hölder inequality and the Young inequality, we obtain

$c\ge \left(\frac{1}{2\theta }-\frac{1}{{2}_{s}^{*}}\right)\left({\mu }^{2\theta }+{\parallel u\parallel }^{2\theta }\right)-\lambda \left(\frac{1}{1-\gamma }+\frac{1}{{2}_{s}^{*}}\right){|\mathrm{\Omega }|}^{\left({2}_{s}^{*}-1+\gamma \right)/{2}_{s}^{*}}{S}^{-\left(1-\gamma \right)/2}{\parallel u\parallel }^{1-\gamma }$$\ge \left(\frac{1}{2\theta }-\frac{1}{{2}_{s}^{*}}\right)\left({\mu }^{2\theta }+{\parallel u\parallel }^{2\theta }\right)-\left(\frac{1}{2\theta }-\frac{1}{{2}_{s}^{*}}\right){\parallel u\parallel }^{2\theta }$$-{\left(\frac{1}{2\theta }-\frac{1}{{2}_{s}^{*}}\right)}^{-\left(1-\gamma \right)/\left(2\theta -1+\gamma \right)}{\left[\lambda \left(\frac{1}{1-\gamma }+\frac{1}{{2}_{s}^{*}}\right){|\mathrm{\Omega }|}^{\left({2}_{s}^{*}-1+\gamma \right)/{2}_{s}^{*}}{S}^{-\left(1-\gamma \right)/2}\right]}^{2\theta /\left(2\theta -1+\gamma \right)}$$\ge \left(\frac{1}{2\theta }-\frac{1}{{2}_{s}^{*}}\right){S}^{{2}_{s}^{*}\theta /\left({2}_{s}^{*}-2\theta \right)}-{\left(\frac{1}{2\theta }-\frac{1}{{2}_{s}^{*}}\right)}^{-\left(1-\gamma \right)/\left(2\theta -1+\gamma \right)}{\left[\lambda \left(\frac{1}{1-\gamma }+\frac{1}{{2}_{s}^{*}}\right){|\mathrm{\Omega }|}^{\left({2}_{s}^{*}-1+\gamma \right)/{2}_{s}^{*}}{S}^{-\left(1-\gamma \right)/2}\right]}^{2\theta /\left(2\theta -1+\gamma \right)},$

which contradicts (4.4) since (4.3). This concludes the proof. ∎

We now give a control from above for the functional ${J}_{n,\lambda }$ under a suitable situation. For this, we assume, without loss of generality, that $0\in \mathrm{\Omega }$. By [10], we know that the infimum in (2.7) is attained at the function

(4.11)

that is, it holds true that

${\iint }_{{ℝ}^{2N}}\frac{{|{u}_{\epsilon }\left(x\right)-{u}_{\epsilon }\left(y\right)|}^{2}}{{|x-y|}^{N+2s}}𝑑x𝑑y=S{\parallel {u}_{\epsilon }\parallel }_{{L}^{{2}_{s}^{*}}\left({ℝ}^{N}\right)}^{2}.$

Let us fix $r>0$ such that ${B}_{4r}\subset \mathrm{\Omega }$, where ${B}_{4r}=\left\{x\in {ℝ}^{N}:|x|<4r\right\}$, and let us introduce a cut-off function $\varphi \in {C}^{\mathrm{\infty }}\left({ℝ}^{N},\left[0,1\right]\right)$ such that

(4.12)

For any $\epsilon >0$, we set

${\psi }_{\epsilon }=\frac{\varphi {u}_{\epsilon }}{{\parallel \varphi {u}_{\epsilon }\parallel }_{{2}_{s}^{*}}^{2}}\in {X}_{0}.$(4.13)

Then we can prove the following result.

#### Lemma 4.3.

There exist $\psi \mathrm{\in }{X}_{\mathrm{0}}$ and ${\lambda }_{\mathrm{1}}\mathrm{>}\mathrm{0}$ such that for any $\lambda \mathrm{\in }\mathrm{\left(}\mathrm{0}\mathrm{,}{\lambda }_{\mathrm{1}}\mathrm{\right)}$,

$\underset{t\ge 0}{sup}{J}_{n,\lambda }\left(t\psi \right)<{D}_{1}-{D}_{2}{\lambda }^{2\theta /\left(2\theta -1+\gamma \right)},$

with ${D}_{\mathrm{1}}$, ${D}_{\mathrm{2}}\mathrm{>}\mathrm{0}$ given as in (4.3).

#### Proof.

Let $\lambda ,\epsilon >0$. Let ${u}_{\epsilon }$ and ${\psi }_{\epsilon }$ be as in (4.11) and (4.13), respectively. By (2.6), we have ${J}_{n,\lambda }\left(t{\psi }_{\epsilon }\right)\to -\mathrm{\infty }$ as $t\to \mathrm{\infty }$, so that there exists ${t}_{\epsilon }>0$ such that ${J}_{n,\lambda }\left({t}_{\epsilon }{\psi }_{\epsilon }\right)={\mathrm{max}}_{t\ge 0}{J}_{n,\lambda }\left(t{\psi }_{\epsilon }\right)$. By Lemma 4.1, we know that ${J}_{n,\lambda }\left({t}_{\epsilon }{\psi }_{\epsilon }\right)\ge \alpha >0$. Hence, by the continuity of ${J}_{n,\lambda }$ there exist two numbers ${t}_{0}$, ${t}^{*}>0$ such that ${t}_{0}\le {t}_{\epsilon }\le {t}^{*}$.

Now, since ${\parallel {u}_{\epsilon }\parallel }_{{L}^{{2}_{s}^{*}}\left({ℝ}^{N}\right)}$ is independent of ε, by [27, Proposition 21] we have

${\parallel {\psi }_{\epsilon }\parallel }^{2}\le \frac{{\iint }_{{ℝ}^{2N}}\frac{{|{u}_{\epsilon }\left(x\right)-{u}_{\epsilon }\left(y\right)|}^{2}}{{|x-y|}^{N+2s}}𝑑x𝑑y}{{\parallel \varphi {u}_{\epsilon }\parallel }_{{2}_{s}^{*}}^{2}}=S+O\left({\epsilon }^{N-2s}\right),$

from which, by the elementary inequality

with $p=2\theta$, it follows that, as $\epsilon \to {0}^{+}$,

${\parallel {\psi }_{\epsilon }\parallel }^{2\theta }\le {S}^{\theta }+O\left({\epsilon }^{N-2s}\right).$

Hence, by the last inequality, (2.6) and since ${t}_{0}\le {t}_{\epsilon }\le {t}^{*}$, for any $\epsilon >0$ sufficiently small, we have

${J}_{n,\lambda }\left({t}_{\epsilon }{\psi }_{\epsilon }\right)\le \frac{{t}_{\epsilon }^{2\theta }}{2\theta }{S}^{\theta }+{C}_{1}{\epsilon }^{N-2s}-\frac{\lambda }{1-\gamma }{\int }_{\mathrm{\Omega }}\left[{\left({t}_{0}{\psi }_{\epsilon }+\frac{1}{n}\right)}^{1-\gamma }-{\left(\frac{1}{n}\right)}^{1-\gamma }\right]𝑑x-\frac{{t}_{\epsilon }^{{2}_{s}^{*}}}{{2}_{s}^{*}},$(4.14)

with a suitable positive constant ${C}_{1}$. We observe that

$\underset{t\ge 0}{\mathrm{max}}\left(\frac{{t}^{2\theta }}{2\theta }{S}^{\theta }-\frac{{t}^{{2}_{s}^{*}}}{{2}_{s}^{*}}\right)=\left(\frac{1}{2\theta }-\frac{1}{{2}_{s}^{*}}\right){S}^{{2}_{s}^{*}\theta /\left({2}_{s}^{*}-2\theta \right)}.$

Thus, by (4.14) it follows that

${J}_{n,\lambda }\left({t}_{\epsilon }{\psi }_{\epsilon }\right)\le \left(\frac{1}{2\theta }-\frac{1}{{2}_{s}^{*}}\right){S}^{{2}_{s}^{*}\theta /\left({2}_{s}^{*}-2\theta \right)}+{C}_{1}{\epsilon }^{N-2s}-\frac{\lambda }{1-\gamma }{\int }_{\mathrm{\Omega }}\left[{\left({t}_{0}{\psi }_{\epsilon }+\frac{1}{n}\right)}^{1-\gamma }-{\left(\frac{1}{n}\right)}^{1-\gamma }\right]𝑑x.$(4.15)

Now, let us consider a positive number q, less than 1, satisfying

$\frac{\left(N-2s\right)\left(1-\gamma \right)-2q\left(N-2s\right)\left(1-\gamma \right)+{2}_{s}^{*}qN}{{2}_{s}^{*}}\cdot \frac{2\theta }{2\theta -1+\gamma }\cdot \frac{1}{N-2s}+1-\frac{2\theta }{2\theta -1+\gamma }<0,$(4.16)

that is, since $2<2\theta <{2}_{s}^{*}$, $N>2s$ and $\gamma \in \left(0,1\right)$, such that

$0

By the elementary inequality

with $p={2}_{s}^{*}/2$, considering $\epsilon <{r}^{1/q}$ sufficiently small, with r given by (4.12), and since $q<1$, we have

$-\frac{1}{1-\gamma }{\int }_{\left\{x\in \mathrm{\Omega }:|x|\le {\epsilon }^{q}\right\}}\left[{\left({t}_{0}{\psi }_{\epsilon }+\frac{1}{n}\right)}^{1-\gamma }-{\left(\frac{1}{n}\right)}^{1-\gamma }\right]𝑑x\le -\stackrel{~}{C}{\epsilon }^{\left(N-2s\right)\left(1-\gamma \right)/{2}_{s}^{*}}{\int }_{\left\{x\in \mathrm{\Omega }:|x|\le {\epsilon }^{q}\right\}}{\left[\frac{1}{{\left({|x|}^{2}+{\epsilon }^{2}\right)}^{\left(N-2s\right)/2}}\right]}^{2\left(1-\gamma \right)/{2}_{s}^{*}}𝑑x$$\le -{C}_{2}{\epsilon }^{\left(\left(N-2s\right)\left(1-\gamma \right)-2q\left(N-2s\right)\left(1-\gamma \right)+{2}_{s}^{*}qN\right)/{2}_{s}^{*}},$(4.17)

with two positive constants $\stackrel{~}{C}$ and ${C}_{2}$ independent of ε. By combining (4.15) with (4.17), we get

${J}_{n,\lambda }\left({t}_{\epsilon }{\psi }_{\epsilon }\right)\le \left(\frac{1}{2\theta }-\frac{1}{{2}_{s}^{*}}\right){S}^{{2}_{s}^{*}\theta /\left({2}_{s}^{*}-2\theta \right)}+{C}_{1}{\epsilon }^{N-2s}-{C}_{2}\lambda {\epsilon }^{\left(\left(N-2s\right)\left(1-\gamma \right)-2q\left(N-2s\right)\left(1-\gamma \right)+{2}_{s}^{*}qN\right)/{2}_{s}^{*}}.$(4.18)

Thus, let us consider ${\lambda }^{*}>0$ such that

and let us set

${\nu }_{1}=\frac{2q\theta }{\left(2\theta -1+\gamma \right)\left(N-2s\right)},$${\nu }_{2}=\frac{2\theta \left[\left(N-2s\right)\left(1-\gamma \right)-2q\left(N-2s\right)\left(1-\gamma \right)+{2}_{s}^{*}qN\right]}{{2}_{s}^{*}\left(2\theta -1+\gamma \right)\left(N-2s\right)}+1,$${\nu }_{3}={\nu }_{2}-\frac{2\theta }{2\theta -1+\gamma },$${\lambda }_{1}=\mathrm{min}\left\{{\lambda }^{*},{r}^{1/{\nu }_{1}},{\left(\frac{{C}_{2}}{{C}_{1}+{D}_{2}}\right)}^{-1/{\nu }_{3}}\right\},$

where r and q are given in (4.12) and (4.16), respectively, while we still consider ${D}_{1}$ and ${D}_{2}$ as defined in (4.3). Then, by considering $\epsilon ={\lambda }^{{\nu }_{1}/q}$ in (4.18), since (4.16) implies that ${\nu }_{3}<0$, for any $\lambda \in \left(0,{\lambda }_{1}\right)$ we have

${J}_{n,\lambda }\left({t}_{\epsilon }{\psi }_{\epsilon }\right)\le {D}_{1}+{C}_{1}{\lambda }^{2\theta /\left(2\theta -1+\gamma \right)}-{C}_{2}{\lambda }^{{\nu }_{2}}={D}_{1}+{\lambda }^{2\theta /\left(2\theta -1+\gamma \right)}\left({C}_{1}-{C}_{2}{\lambda }^{{\nu }_{3}}\right)<{D}_{1}-{D}_{2}{\lambda }^{2\theta /\left(2\theta -1+\gamma \right)},$

which concludes the proof. ∎

We can now prove the existence result for (2.4) by applying the mountain pass theorem.

#### Theorem 4.4.

There exists $\overline{\lambda }\mathrm{>}\mathrm{0}$ such that, for any $\lambda \mathrm{\in }\mathrm{\left(}\mathrm{0}\mathrm{,}\overline{\lambda }\mathrm{\right)}$, problem (2.4) has a positive solution ${v}_{n}\mathrm{\in }{X}_{\mathrm{0}}$ with

$\alpha <{J}_{n,\lambda }\left({v}_{n}\right)<{D}_{1}-{D}_{2}{\lambda }^{2\theta /\left(2\theta -1+\gamma \right)},$(4.19)

where α, ${D}_{\mathrm{1}}$ and ${D}_{\mathrm{2}}$ are given in Lemma 3.1 and (4.3), respectively.

#### Proof.

Let $\overline{\lambda }=\mathrm{min}\left\{{\lambda }_{0},{\lambda }_{1}\right\}$, with ${\lambda }_{0}$ and ${\lambda }_{1}$ given in Lemmas 3.1 and 4.3, respectively. Let us consider $\lambda \in \left(0,\overline{\lambda }\right)$. By Lemma 4.1, the functional ${J}_{n,\lambda }$ verifies the mountain pass geometry. For this, we can set the critical mountain pass level as

${c}_{n,\lambda }=\underset{g\in \mathrm{\Gamma }}{inf}\underset{t\in \left[0,1\right]}{\mathrm{max}}{J}_{n,\lambda }\left(g\left(t\right)\right),$

where

$\mathrm{\Gamma }=\left\{g\in C\left(\left[0,1\right],{X}_{0}\right):g\left(0\right)=0,{J}_{n,\lambda }\left(g\left(1\right)\right)<0\right\}.$

By Lemmas 4.1 and 4.3, we get

$0<\alpha <{c}_{n,\lambda }\le \underset{t\ge 0}{sup}{J}_{n,\lambda }\left(t\psi \right)<{D}_{1}-{D}_{2}{\lambda }^{2\theta /\left(2\theta -1+\gamma \right)}.$

Hence, by Lemma 4.2 the functional ${J}_{n,\lambda }$ satisfies the Palais–Smale condition at level ${c}_{n,\lambda }$. Thus, the mountain pass theorem gives the existence of a critical point ${v}_{n}\in {X}_{0}$ for ${J}_{n,\lambda }$ at level ${c}_{n,\lambda }$. Since

${J}_{n,\lambda }\left({v}_{n}\right)={c}_{n,\lambda }>\alpha >0={J}_{n,\lambda }\left(0\right),$

we obtain that ${v}_{n}$ is a nontrivial solution of (2.4). Furthermore, by (2.5) with test function $\phi ={v}_{n}^{-}$ and inequality (3.12), we can see that $\parallel {v}_{n}^{-}\parallel =0$, which implies that ${v}_{n}$ is nonnegative. By the maximum principle in [28, Proposition 2.17], we have that ${v}_{n}$ is a positive solution of (2.4), concluding the proof. ∎

## 5 A second solution for problem (1.1)

In this last section, we prove the existence of a second solution for problem (1.1), as a limit of solutions of the perturbed problem (2.4). For this, here we need the assumption that $\partial \mathrm{\Omega }$ is continuous in order to apply a density argument for the space ${X}_{0}$.

#### Proof of Theorem 1.1.

Let us consider $\overline{\lambda }$ as given in Theorem 4.4, and let $\lambda \in \left(0,\overline{\lambda }\right)$. Since $\overline{\lambda }\le {\lambda }_{0}$, by Theorem 3.2 we know that problem (1.1) admits a solution ${u}_{0}$ with ${J}_{\lambda }\left({u}_{0}\right)<0$.

In order to find a second solution for (1.1) let ${\left\{{v}_{n}\right\}}_{n}$ be a family of positive solutions of (2.4). By (2.7), (4.1), (4.19) and the Hölder inequality, we have

${D}_{1}-{D}_{2}{\lambda }^{2\theta /\left(2\theta -1+\gamma \right)}>{J}_{n,\lambda }\left({v}_{n}\right)-\frac{1}{{2}_{s}^{*}}〈{J}_{n,\lambda }^{\prime }\left({v}_{n}\right),{v}_{n}〉$$=\left(\frac{1}{2\theta }-\frac{1}{{2}_{s}^{*}}\right){\parallel {v}_{n}\parallel }^{2\theta }-\frac{\lambda }{1-\gamma }{\int }_{\mathrm{\Omega }}\left[{\left({v}_{n}+\frac{1}{n}\right)}^{1-\gamma }-{\left(\frac{1}{n}\right)}^{1-\gamma }\right]𝑑x+\frac{\lambda }{{2}_{s}^{*}}{\int }_{\mathrm{\Omega }}{\left({v}_{n}+\frac{1}{n}\right)}^{-\gamma }{v}_{n}𝑑x$$\ge \left(\frac{1}{2\theta }-\frac{1}{{2}_{s}^{*}}\right){\parallel {v}_{n}\parallel }^{2\theta }-\frac{\lambda }{1-\gamma }{\int }_{\mathrm{\Omega }}{v}_{n}^{1-\gamma }𝑑x$$\ge \left(\frac{1}{2\theta }-\frac{1}{{2}_{s}^{*}}\right){\parallel {v}_{n}\parallel }^{2\theta }-\frac{\lambda }{1-\gamma }{|\mathrm{\Omega }|}^{\left({2}_{s}^{*}-1+\gamma \right)/{2}_{s}^{*}}{S}^{-\left(1-\gamma \right)/2}{\parallel {v}_{n}\parallel }^{1-\gamma },$

which yields that ${\left\{{v}_{n}\right\}}_{n}$ is bounded in ${X}_{0}$ since $1-\gamma <1<2\theta$. Hence, by using [26, Lemma 8] and [6, Theorem 4.9], there exist a subsequence, still denoted by ${\left\{{v}_{n}\right\}}_{n}$, and a function ${v}_{0}\in {X}_{0}$ such that

(5.1)

We want to prove that ${v}_{n}\to {v}_{0}$ in ${X}_{0}$ as $n\to \mathrm{\infty }$. When $\mu =0$, by (5.1) we have ${v}_{n}\to 0$ in ${X}_{0}$ as $n\to \mathrm{\infty }$. For this, we suppose $\mu >0$. We observe that

so by the Vitali convergence theorem and (5.1) it follows that

$\underset{n\to \mathrm{\infty }}{lim}{\int }_{\mathrm{\Omega }}\frac{{v}_{n}}{{\left({v}_{n}+\frac{1}{n}\right)}^{\gamma }}𝑑x={\int }_{\mathrm{\Omega }}{v}_{0}^{1-\gamma }𝑑x.$(5.2)

Using (2.5) for ${v}_{n}$ and test function $\phi ={v}_{n}$, by (5.1) and (5.2), as $n\to \mathrm{\infty }$, we have

${\mu }^{2\theta }-\lambda {\int }_{\mathrm{\Omega }}{v}_{0}^{1-\gamma }𝑑x+{\parallel {v}_{n}\parallel }_{{2}_{s}^{*}}^{{2}_{s}^{*}}=o\left(1\right).$(5.3)

For any $n\in ℕ$, by an immediate calculation in (2.4) we see that

Thus, since ${\left\{{v}_{n}\right\}}_{n}$ is bounded in ${X}_{0}$ and by using a standard comparison argument (see [3, Lemma 2.1]) and the maximum principle in [28, Proposition 2.17], for any $\stackrel{~}{\mathrm{\Omega }}⋐\mathrm{\Omega }$ there exists a constant ${c}_{\stackrel{~}{\mathrm{\Omega }}}>0$ such that

(5.4)

Now, let $\phi \in {C}_{0}^{\mathrm{\infty }}\left(\mathrm{\Omega }\right)$ with $\mathrm{supp}\phi =\stackrel{~}{\mathrm{\Omega }}⋐\mathrm{\Omega }$. By (5.4), we have

so that by (5.1) and the dominated convergence theorem we obtain

$\underset{n\to \mathrm{\infty }}{lim}{\int }_{\mathrm{\Omega }}\frac{\phi }{{\left({v}_{n}+\frac{1}{n}\right)}^{\gamma }}𝑑x={\int }_{\mathrm{\Omega }}{v}_{0}^{-\gamma }\phi 𝑑x.$(5.5)

Thus, by considering (2.5) for ${v}_{n}$, sending $n\to \mathrm{\infty }$ and using (5.1) and (5.5), for any $\phi \in {C}_{0}^{\mathrm{\infty }}\left(\mathrm{\Omega }\right)$ it follows that

${\mu }^{2\left(\theta -1\right)}〈{v}_{0},\phi 〉-\lambda {\int }_{\mathrm{\Omega }}{v}_{0}^{-\gamma }\phi 𝑑x+{\int }_{\mathrm{\Omega }}{v}_{0}^{{2}_{s}^{*}-1}\phi 𝑑x=0.$(5.6)

However, since $\partial \mathrm{\Omega }$ is continuous, by [15, Theorem 6] the space ${C}_{0}^{\mathrm{\infty }}\left(\mathrm{\Omega }\right)$ is dense in ${X}_{0}$. Thus, by a standard density argument, (5.6) holds true for any $\phi \in {X}_{0}$. By combining (5.3) and (5.6) with test function $\phi ={v}_{0}$, as $n\to \mathrm{\infty }$ we get

${\mu }^{2\left(\theta -1\right)}\left({\mu }^{2}-{\parallel {v}_{0}\parallel }^{2}\right)={\parallel {v}_{n}\parallel }_{{2}_{s}^{*}}^{{2}_{s}^{*}}-{\parallel {v}_{0}\parallel }_{{2}_{s}^{*}}^{{2}_{s}^{*}}+o\left(1\right),$

and by (5.1) and [7, Theorem 2] we have

${\mu }^{2\left(\theta -1\right)}\underset{n\to \mathrm{\infty }}{lim}{\parallel {v}_{n}-{v}_{0}\parallel }^{2}={\mathrm{\ell }}^{{2}_{s}^{*}}.$(5.7)

If $\mathrm{\ell }=0$, then ${v}_{n}\to {v}_{0}$ in ${X}_{0}$ as $n\to \mathrm{\infty }$ since $\mu >0$.

Let us suppose $\mathrm{\ell }>0$ by contradiction. Arguing as in Lemma 4.2, by (5.1) and (5.7) we get (4.10). Therefore, since $\theta \ge 1$, by (4.1), (4.10), (4.19), (5.1), the Hölder inequality and the Young inequality we have

${D}_{1}-{D}_{2}{\lambda }^{2\theta /\left(2\theta -1+\gamma \right)}$$\mathrm{ }>{J}_{n,\lambda }\left({v}_{n}\right)-\frac{1}{{2}_{s}^{*}}〈{J}_{n,\lambda }^{\prime }\left({v}_{n}\right),{v}_{n}〉$$\mathrm{ }\ge \left(\frac{1}{2\theta }-\frac{1}{{2}_{s}^{*}}\right)\left({\mu }^{2\theta }+{\parallel {v}_{0}\parallel }^{2\theta }\right)-\lambda \left(\frac{1}{1-\gamma }+\frac{1}{{2}_{s}^{*}}\right){|\mathrm{\Omega }|}^{\left({2}_{s}^{*}-1+\gamma \right)/{2}_{s}^{*}}{S}^{-\left(1-\gamma \right)/2}{\parallel {v}_{0}\parallel }^{1-\gamma }$$\mathrm{ }\ge \left(\frac{1}{2\theta }-\frac{1}{{2}_{s}^{*}}\right){S}^{{2}_{s}^{*}\theta /\left({2}_{s}^{*}-2\theta \right)}-{\left(\frac{1}{2\theta }-\frac{1}{{2}_{s}^{*}}\right)}^{-\left(1-\gamma \right)/\left(2\theta -1+\gamma \right)}{\left[\lambda \left(\frac{1}{1-\gamma }+\frac{1}{{2}_{s}^{*}}\right){|\mathrm{\Omega }|}^{\left({2}_{s}^{*}-1+\gamma \right)/{2}_{s}^{*}}{S}^{-\left(1-\gamma \right)/2}\right]}^{2\theta /\left(2\theta -1+\gamma \right)},$

which is the desired contradiction, thanks to (4.3).

Therefore, ${v}_{n}\to {v}_{0}$ in ${X}_{0}$ as $n\to \mathrm{\infty }$, and by (2.3) and (2.5) we immediately see that ${v}_{0}$ is a solution of problem (2.2). Furthermore, by (4.19) we have ${J}_{\lambda }\left({v}_{0}\right)\ge \alpha >0$, which also implies that ${v}_{0}$ is nontrivial. Reasoning as at the end of the proof of Theorem 4.4, we conclude that ${v}_{0}$ is a positive solution of (2.2), and so ${v}_{0}$ also solves problem (1.1). Finally, ${v}_{0}$ is different from ${u}_{0}$ since ${J}_{\lambda }\left({v}_{0}\right)>0>{J}_{\lambda }\left({u}_{0}\right)$. ∎

## References

• [1]

B. Abdellaoui, M. Medina, I. Peral and A. Primo, The effect of the Hardy potential in some Calderón–Zygmund properties for the fractional Laplacian, J. Differential Equations 260 (2016), no. 11, 8160–8206.

• [2]

G. Autuori, A. Fiscella and P. Pucci, Stationary Kirchhoff problems involving a fractional elliptic operator and a critical nonlinearity, Nonlinear Anal. 125 (2015), 699–714.

• [3]

B. Barrios, M. Medina and I. Peral, Some remarks on the solvability of non-local elliptic problems with the Hardy potential, Commun. Contemp. Math. 16 (2014), no. 4, Article ID 1350046.

• [4]

B. Barrios, I. De Bonis, M. Medina and I. Peral, Semilinear problems for the fractional Laplacian with a singular nonlinearity, Open Math. 13 (2015), 390–407.

• [5]

Z. Binlin, A. Fiscella and S. Liang, Infinitely many solutions for critical degenerate Kirchhoff type equations involving the fractional p-Laplacian, submitted.  Google Scholar

• [6]

H. Brézis, Functional Analysis, Sobolev Spaces and Partial Differential Equations, Universitext, Springer, New York, 2011.  Google Scholar

• [7]

H. Brézis and E. A. Lieb, A relation between pointwise convergence of functions and convergence of functionals, Proc. Amer. Math. Soc. 88 (1983), no. 3, 486–490.

• [8]

A. Canino, L. Montoro, B. Sciunzi and M. Squassina, Nonlocal problems with singular nonlinearity, Bull. Sci. Math. 141 (2017), no. 3, 223–250.

• [9]

M. Caponi and P. Pucci, Existence theorems for entire solutions of stationary Kirchhoff fractional p-Laplacian equations, Ann. Mat. Pura Appl. 195 (2016), no. 6, 2099–2129.

• [10]

A. Cotsiolis and N. K. Tavoularis, Best constants for Sobolev inequalities for higher order fractional derivatives, J. Math. Anal. Appl. 295 (2004), no. 1, 225–236.

• [11]

E. Di Nezza, G. Palatucci and E. Valdinoci, Hitchhiker’s guide to the fractional Sobolev spaces, Bull. Sci. Math. 136 (2012), no. 5, 521–573.

• [12]

A. Fiscella, Infinitely many solutions for a critical Kirchhoff type problem involving a fractional operator, Differential Integral Equations 29 (2016), no. 5–6, 513–530.  Google Scholar

• [13]

A. Fiscella and P. Pucci, Kirchhoff–Hardy fractional problems with lack of compactness, Adv. Nonlinear Stud. 17 (2017), no. 3, 429–456.

• [14]

A. Fiscella and P. Pucci, p-fractional Kirchhoff equations involving critical nonlinearities, Nonlinear Anal. Real World Appl. 35 (2017), 350–378.

• [15]

A. Fiscella, R. Servadei and E. Valdinoci, Density properties for fractional Sobolev spaces, Ann. Acad. Sci. Fenn. Math. 40 (2015), no. 1, 235–253.

• [16]

A. Fiscella and E. Valdinoci, A critical Kirchhoff type problem involving a nonlocal operator, Nonlinear Anal. 94 (2014), 156–170.

• [17]

C.-Y. Lei, J.-F. Liao and C.-L. Tang, Multiple positive solutions for Kirchhoff type of problems with singularity and critical exponents, J. Math. Anal. Appl. 421 (2015), no. 1, 521–538.

• [18]

J.-F. Liao, X.-F. Ke, C.-Y. Lei and C.-L. Tang, A uniqueness result for Kirchhoff type problems with singularity, Appl. Math. Lett. 59 (2016), 24–30.

• [19]

J.-F. Liao, P. Zhang, J. Liu and C.-L. Tang, Existence and multiplicity of positive solutions for a class of Kirchhoff type problems with singularity, J. Math. Anal. Appl. 430 (2015), no. 2, 1124–1148.

• [20]

R.-Q. Liu, C.-L. Tang, J.-F. Liao and X.-P. Wu, Positive solutions of Kirchhoff type problem with singular and critical nonlinearities in dimension four, Commun. Pure Appl. Anal. 15 (2016), no. 5, 1841–1856.

• [21]

X. Liu and Y. Sun, Multiple positive solutions for Kirchhoff type problems with singularity, Commun. Pure Appl. Anal. 12 (2013), no. 2, 721–733.

• [22]

G. Molica Bisci, V. D. Radulescu and R. Servadei, Variational Methods for Nonlocal Fractional Problems, Encyclopedia Math. Appl. 162, Cambridge University Press, Cambridge, 2016. Google Scholar

• [23]

P. Piersanti and P. Pucci, Entire solutions for critical p-fractional Hardy Schrödinger Kirchhoff equations, Publ. Mat., to appear.  Google Scholar

• [24]

P. Pucci and S. Saldi, Critical stationary Kirchhoff equations in ${ℝ}^{N}$ involving nonlocal operators, Rev. Mat. Iberoam. 32 (2016), no. 1, 1–22.  Google Scholar

• [25]

P. Pucci, M. Xiang and B. Zhang, Existence and multiplicity of entire solutions for fractional p-Kirchhoff equations, Adv. Nonlinear Anal. 5 (2016), no. 1, 27–55.

• [26]

R. Servadei and E. Valdinoci, Mountain pass solutions for non-local elliptic operators, J. Math. Anal. Appl. 389 (2012), no. 2, 887–898.

• [27]

R. Servadei and E. Valdinoci, The Brezis–Nirenberg result for the fractional Laplacian, Trans. Amer. Math. Soc. 367 (2015), no. 1, 67–102.  Google Scholar

• [28]

L. Silvestre, Regularity of the obstacle problem for a fractional power of the Laplace operator, Comm. Pure Appl. Math. 60 (2007), no. 1, 67–112.

• [29]

M. Xiang, B. Zhang and H. Qiu, Existence of solutions for a critical fractional Kirchhoff type problem in ${ℝ}^{N}$, Sci. China Math., to appear.  Google Scholar

Revised: 2017-05-18

Accepted: 2017-06-05

Published Online: 2017-08-03

Funding Source: Coordenação de Aperfeiçoamento de Pessoal de Nível Superior

Award identifier / Grant number: PNPD–CAPES 33003017003P5

The author is supported by Coordenação de Aperfeiçoamento de Pessoal de Nível Superior through the fellowship PNPD–CAPES 33003017003P5. The author is member of the Gruppo Nazionale per l’Analisi Matematica, la Probabilità e le loro Applicazioni (GNAMPA) of the Istituto Nazionale di Alta Matematica “G. Severi” (INdAM).

Citation Information: Advances in Nonlinear Analysis, Volume 8, Issue 1, Pages 645–660, ISSN (Online) 2191-950X, ISSN (Print) 2191-9496,

Export Citation

## Citing Articles

[1]
Li Wang, Kun Cheng, and Binlin Zhang
Applied Mathematics & Optimization, 2019
[2]
M. Khiddi and S. M. Sbai
Complex Variables and Elliptic Equations, 2019, Page 1
[3]
Alessio Fiscella and Pawan Kumar Mishra
Nonlinear Analysis, 2018
[4]
Li Wang, Kun Xie, and Binlin Zhang
Journal of Mathematical Analysis and Applications, 2017