\documentclass[reqno]{amsart} \usepackage{hyperref} \AtBeginDocument{{\noindent\small \emph{Electronic Journal of Differential Equations}, Vol. 2017 (2017), No. 291, pp. 1--9.\newline ISSN: 1072-6691. URL: http://ejde.math.txstate.edu or http://ejde.math.unt.edu} \thanks{\copyright 2017 Texas State University.} \vspace{8mm}} \begin{document} \title[\hfilneg EJDE-2017/291\hfil A singular problem with variable potentials] {Existence and non-existence of solutions for a singular problem with variable potentials} \author[K. Saoudi \hfil EJDE-2017/291\hfilneg] {Kamel Saoudi} \address{Kamel Saoudi \newline College of sciences at Dammam, University of Imam Abdulrahman Bin Faisal\\ 31441 Dammam, Saudi Arabia} \email{kasaoudi@gmail.com} \dedicatory{Communicated by Claudianor O. Alves} \thanks{Submitted March 14, 2017. Published November 21, 2017.} \subjclass[2010]{35J35, 35R05} \keywords{Singular elliptic equation; variational methods; existence; \hfill\break\indent non-existence} \begin{abstract} The purpose of this article is to prove some existence and nonexistence theorems for the inhomogeneous singular Dirichlet problem $$ - \Delta_p u = \frac{\lambda k(x)}{u^\delta}\pm h(x) u^q. $$ For proving our results we use the sub and super solution method, and monotonicity arguments. \end{abstract} \maketitle \numberwithin{equation}{section} \newtheorem{theorem}{Theorem}[section] \newtheorem{definition}[theorem]{Definition} \allowdisplaybreaks \section{Introduction} \label{S1} In this paper we are interested in the following quasilinear and singular problem with variable potentials: \begin{equation} \label{103} \begin{gathered} - \Delta_p u = \lambda k(x) u^{-\delta} \pm h(x) u^q \quad \text{in } \Omega,\\ u|_{\partial\Omega} = 0 ,\quad u > 0\quad\text{in }\Omega, \end{gathered} \end{equation} where $\Omega\subset\mathbb{R}^N$, $(N\geq 2)$ is a bounded domain with smooth boundary, $\lambda$ is a positive parameter, $1 < p < \infty$, $p-1 < q\leq p^{*} - 1$, and $0 < \delta < 1$. As usual, $p^{*} = \frac{Np}{N-p}$ if $1 < p < N$, $p^{*}\in (p,\infty)$ is arbitrarily large if $p = N$, and $p^{*} = \infty$ if $p > N$, and the variable weight functions $h,\, k \in L^\infty(\Omega)$ satisfy \begin{equation} \label{104} \operatorname{ess\,inf}_{ x\in\Omega} k(x) >0 \quad \text{and}\quad \operatorname{ess\,inf}_{ x\in\Omega} h(x) >0. \end{equation} Associated with problem \eqref{103} we have the singular functional $E_\lambda:$ $W^{1,p}_0(\Omega)\to\mathbb{R}$ defined by \begin{equation}\label{energy} E_{\lambda}(u) = \frac{1}{p}\int_{\Omega} |\nabla u|^p \,\mathrm{d}x - \frac{\lambda}{1-\delta}\int_{\Omega} k(x)u^{1-\delta} \,\mathrm{d}x \pm \frac{1}{q+1}\int_{\Omega}h(x) u^{q+1} \,\mathrm{d}x \end{equation} in the Sobolev space $W^{1,p}_0(\Omega)$. \begin{definition} \rm $u\in W^{1,p}_0 (\Omega)$ is called a \emph{weak} solution (or solution, for short) of problem~\eqref{103}, that is, for functions $u\in W_0^{1,p}(\Omega)$ satisfying $\operatorname{ess\,inf}_K u > 0$ over every compact set $K\subset\Omega$ and \begin{equation} \label{e1} \int_{\Omega} |\nabla u|^{p-2}\nabla u\cdot \nabla\phi \,{\rm d}x = \lambda\int_{\Omega}k(x) u^{-\delta} \phi \,{\rm d}x \pm \int_{\Omega} h(x)u^q\phi \,{\rm d}x \end{equation} for all $\phi\in C^{\infty}_\mathrm{c}(\Omega)$. As usual, $C^{\infty}_\mathrm{c}(\Omega)$ denotes the space of all $C^{\infty}$ functions $\phi\colon \Omega\to \mathbb{R}$ with compact support. \end{definition} Obviously, every critical point of $E_\lambda$ is a weak solution of the problem \eqref{103}. $\Delta_{p}u := \operatorname{div}(|\nabla u|^{p-2}\nabla u)$, where $p > 1$ is a real constant is called the $p-$Laplacian or the $p$-Laplace operator. The $p$-Laplacian is an elliptic partial differential equation, which is degenerate if $p > 2$ and singular if $p < 2$. If $p = 2$, then the $p$-Laplacian reduces to the simpler classical linear Laplace equation $\Delta u := \nabla.\nabla u$ and in the case of one spatial dimension, we have $\Delta_{p}u =(| u'|^{p-2} u')'$. The class of problems \eqref{103} appears in many nonlinear phenomena, for instance, in the theory of quasi-regular and quasi-conformal mappings (for this see \cite{Mi,Uh}), in the generalized reaction-diffusion theory \cite{HeVa}, in the turbulent flow of a gas in a porous medium and in the non-Newtonian fluid theory \cite{EsVa}. In the non-Newtonian fluid theory, the quantity $p$ is the characteristic of the medium. If $p < 2$, the fluids are called pseudo-plastics, if $p = 2$, the fluids are called Newtonian, and if $p > 2$, the fluids are called dilatants. This kind of problems with convex and concave nonlinearities have been extensively studied by many authors. We refer the reader to the celebrate paper of Ambrosetti-Brezis-Cerami \cite{AmBrCe}, Saoudi \cite{Sa2}, Santos \cite{Sa} with their references therein. For $p = 2$, we refer the reader to \cite{RaRe,BrKa} and references therein. The basic work in our direction is the paper \cite{CP} where Coclite-Palmieri have been considered the nonlinear elliptic equation containing singular term \begin{equation}\label{e1.2} \begin{gathered} -\Delta u=u^{p}+\lambda u^{-\gamma}, \quad\text{in }\Omega,\\ u>0, \quad\text{in }\Omega,\\ u=0, \quad\text{on }\partial \Omega, \end{gathered} \end{equation} where $\Omega\subset R^{N}$ $(N\geq 3)$ is a bounded domain with smooth boundary $\partial \Omega$ and $\lambda$ is a positive parameter. The exponent $p$ of the sublinear satisfies $0
0$ small enough, and has no solution when
$\lambda$ is large. We mention that in the
work \cite{CP} the authors have been extended the results of
Crandall-Rabinowitz-Tartar \cite{CrRaTa}.
Problem \eqref{e1.2} have been also studied with different elliptic operators.
We refer the reader to \cite{CP,CrRaTa,GaSa, GaSa1, GiSa1,GiScTa,LM,AA,Sa5, SaKr}
and references therein.
The aim of this work is to extend the results obtained in \cite{CP}
to the more general problems \eqref{103}. Precisely, the main goal of this
paper is to prove some existence and non-existence theorems for the non-linear
singular elliptic problem \eqref{103}. Firstly, we state the following definitions.
\begin{definition} \rm
A function $ \underline{u}\in W^{1,p}_0(\Omega)$ is called a weak sub-solution
to \eqref{103}$_+$ if $\underline{u}\in C^2(\Omega)\cap C(\overline{\Omega})$ and
\begin{gather*}
- \Delta_p \underline{u} \leq \lambda k(x) \underline{u}^{-\delta}
+ h(x) \underline{u}^q \quad \text{in } \Omega,\\
\underline{u}|_{\partial\Omega}
= 0 ,\quad \underline{u} > 0\quad\text{ in }\Omega,
\end{gather*}
A function $ \overline{u}\in W^{1,p}_0(\Omega)$ is called a weak super-solution
to \eqref{103}$_+$ if $\overline{u}\in C^2(\Omega)\cap C(\overline{\Omega})$ and
\begin{gather*}
- \Delta_p \overline{u} \leq \lambda k(x) \overline{u}^{-\delta}
+ h(x) \overline{u}^q \quad \text{in } \Omega,\\
\overline{u}|_{\partial\Omega}
= 0 ,\quad \overline{u} > 0\quad\text{ in }\Omega,
\end{gather*}
\end{definition}
\begin{definition} \rm
A solution $u_\lambda$ of problem \eqref{103}$_+$ is called minimal if
$u_\lambda\leq v$ almost everywhere in $\Omega$ for any further solution
$v$ of problem \eqref{103}$_+$.
\end{definition}
We state below the results that we will prove.
\begin{theorem} \label{T101}
Assume $0< \delta<1$, $p-1 \Lambda^*$.
\end{enumerate}
\end{theorem}
\begin{theorem} \label{T102}
Assume $0< \delta<1$, $p-1
\Lambda_*$, then problem \eqref{103}$_-$ has at least one solution;
\item If $\lambda < \Lambda_*$, then problem \eqref{103}$_-$ does not have
any solution.
\end{enumerate}
\end{theorem}
A comparison between our main result (Theorems \ref{T101} and \ref{T102})
and some of those the previously cited ones, is now in order:
in the present paper, we extended the main result of
Giacomoni-Schindler-Tak\'{a}\v{c} \cite[Theorem 2.1]{GiScTa}
to a class of perturbed singular functionals, this feature gains a remarkable
importance in the applications. Moreover, it is worth noticing that,
since parameter $k(x)$ and $h(x)$ in problem \eqref{103}$_\pm$, is variable,
causes that the quasilinear singular problem is investigate in a complete form.
On the other hand, the main difference between Theorems \ref{T101}
and \ref{T102} above and the main result of
R\v{a}dulescu-Repov\v{s} \cite[Theorems 1.1 and 1.2]{RaRe}
in applications consists in different from two directions:
one is the operator considered in this work is more general than in \cite{RaRe},
the other is with considering singular term instead of R\v{a}dulescu and
Repov\v{s} in \cite{RaRe}.
\section{Proof of Theorem \ref{T101}}
The proof is organized in several steps.
\smallskip
\noindent\textbf{Step 1:}
Existence of minimal solution for $0<\lambda<\Lambda^*$.
Let us define
\begin{equation}\label{def:Lambda1}
\Lambda^*= \sup\{ \lambda > 0\colon \eqref{103}_+\text{ has a weak solution} \}
\end{equation}
and let $\lambda_1(\Omega, m)\equiv \lambda_1$ be the first (principal) eigenvalue
of $-\Delta_p$ and let $\Phi_m$ denote an eigenfunction of $-\Delta_p$ associated
to $\lambda_1$ i.e., $\Phi_m$ solves
\begin{gather*}
-\Delta_p \Phi_m = \lambda_1 m(x) |\Phi_m|^{p-2}\Phi_m \quad \text{in } \Omega\\
\Phi_m > 0 \quad \text{in } \Omega\\
\Phi_m = 0 \quad \text{in } \partial\Omega.
\end{gather*}
It is well-known that $\Phi_m$ belongs to $C^1(\overline{\Omega})$,
that $\Phi_m$ may be chosen positive in $\Omega$ and that $|\nabla\Phi|$ is
positive on a neighborhood of $\partial\Omega$.
To show the existence of a solution to the problem \eqref{103}$_+$, we
construct a well ordered pair of sub-solution $\underline{u}_\lambda$,
and a super-solution $\overline{u}_\lambda$,
such that $\underline{u}_\lambda \leq \overline{u}_\lambda$.
To find a sub-solution, we assume that $m(x)=\min\{ k(x),h(x)\}$ and
$\lambda_1\leq \lambda$. Define $\psi_c=c\Phi_m^{\frac{p}{p-1+\delta}}$.
By a straightforward calculation, we have
$$
\nabla \psi_c= c\Big(\frac{p}{p-1+\delta}\Big)
\Phi_m ^{\frac{1-\delta}{p-1+\delta}}\nabla\Phi_m
$$
and
\begin{align*}
&-\Delta_p(\psi_c)\\
&= -\operatorname{div}(|\nabla \psi_c|^{p-2}\nabla\psi_c)\\
& =\frac{(pc)^{p-1}(\delta-1)(p-1)}{(p-1+\delta)^p}|\nabla \Phi_m|^p
\Phi_m^{\frac{-\delta p}{p-1+\delta}}
+\lambda_1 \Big(\frac{pc}{p-1+\delta}\Big)^{p-1} m(x)\phi_m^p
\Phi_m^{\frac{-\delta p}{p-1+\delta}}
\end{align*}
Thus,
\begin{align*}
&-\Delta_p(\psi_c) \\
&=\frac{(pc)^{p-1}(\delta-1)(p-1)}{(p-1+\delta)^p}|\nabla
\Phi_m|^p\Phi_m^{\frac{-\delta p}{p-1+\delta}}
+\lambda_1 m(x) \Big(\frac{pc}{p-1+\delta}\Big)^{p-1} \phi_m^p
\Phi_m^{\frac{-\delta p}{p-1+\delta}} \\
& \leq m(x)\Big(
\Big(\frac{p}{(p-1+\delta)^p}\Big)^p
\frac{c^{p-1+\delta}(\delta-1)(p-1)}{p}|\nabla \Phi_m|^p\psi_c^{-\delta} \\
&\quad + \lambda_1 \Big(\frac{p}{p-1+\delta}\Big)^{p-1}c^{p-1-q}
c^q \Phi_m^{\frac{p(p-1)}{p-1+\delta}}\Big)\\
& \leq m(x)\Big(
\Big(\frac{p}{(p-1+\delta)^p}\Big)^p \frac{c^{p-1+\delta}
(\delta-1)(p-1)}{p}|\nabla \Phi_m|^p\psi_c^{-\delta} \\
&\quad +\lambda_1 \Big(\frac{p}{p-1+\delta}\Big)^{p-1}c^{p-1-q}
c^q \Phi_m^{\frac{pq}{p-1+\delta}}\Big)\\
& \leq m(x)\Big(
\Big(\frac{p}{(p-1+\delta)^p}\Big)^p \frac{c^{p-1+\delta}
(\delta-1)(p-1)}{p}|\nabla \Phi_m|^p\psi_c^{-\delta} \\
&\quad +\lambda_1 \Big(\frac{p}{p-1+\delta}\Big)^{p-1}c^{p-1-q} \psi_c^{q}\Big)
\end{align*}
Therefore, for $c>0$ small enough, we have
\begin{align*}
-\Delta_p(\psi_c)\leq m(x)\left(\lambda\psi_c^{-\delta}
+ \psi_c^{q}\right)\leq \lambda k(x)\psi_c^{-\delta} + h(x) \psi_c^{q}
\end{align*}
This shows that $\psi_c$ is a sub-solution of the problem \eqref{103}$_+$.
Let us now show that problem \eqref{103}$_+$ has a super-solution.
Now, we put $m(x)=\max\{ k(x),h(x)\}$ and $\lambda_1\geq \lambda$.
Define $\psi_M= M \Phi_m^{\frac{p}{p-1+\delta}}$ for $M > c$ large enough.
Using similar arguments as above we have
$$
\nabla \psi_M= M\Big(\frac{p}{p-1+\delta}\Big)
\Phi_m ^{\frac{1-\delta}{p-1+\delta}}\nabla\Phi_m
$$
and
\begin{align*}
&-\Delta_p(\psi_M) \\
&= -\operatorname{div}(|\nabla \psi_M|^{p-2}\nabla\psi_M)\\
&=\frac{(pM)^{p-1}(\delta-1)(p-1)}{(p-1+\delta)^p}|
\nabla \Phi_m|^p\Phi_m^{\frac{-\delta p}{p-1+\delta}}
+ \lambda_1 m(x) \Big(\frac{pM}{p-1+\delta}\Big)^{p-1}\phi_m^p
\Phi_m^{\frac{-\delta p}{p-1+\delta}}
\end{align*}
Thus,
\begin{align*}
& -\Delta_p(\psi_M) \\
&=\Big(\frac{pM}{p-1+\delta}\Big)^{p-1}\Phi_m^{\frac{-\delta p}{p-1+\delta}}
\Big[ \frac{(\delta-1)(p-1)}{p-1+\delta}|\nabla \Phi_m|^p+\lambda_1 m(x)\phi_m^p\Big] \\
& =\Big(\frac{pM}{p-1+\delta}\Big)^{p-1}\Phi_m^{\frac{-\delta p}{p-1+\delta}}
\Big[ \frac{(\delta-1)(p-1)}{p-1+\delta}|\nabla \Phi_m|^p
+\frac{\lambda_1 m(x)}{2}\phi_m^p\Big]\\
&\quad + \frac{\lambda_1 m(x)}{2}\Big(\frac{pM}{p-1+\delta}\Big)^{p-1}
\Phi_m^{\frac{-\delta p}{p-1+\delta}}\phi_m^p\\
&=\Big(\frac{p}{p-1+\delta}\Big)^{p-1}M^{p-1+\delta}\left[
\frac{(\delta-1)(p-1)}{p-1+\delta}|\nabla \Phi_m|^p
+\frac{\lambda_1(m) m(x)}{2}\phi_m^p\right]\psi_M^{-\delta}\\
&\quad + \frac{\lambda_1 m(x)}{2}\Big(\frac{p}{p-1+\delta}\Big)^{p-1}
M^{p-1-q}\Phi_m^{\frac{p(p-1-q)}{p-1+\delta}}\psi_M^q
\end{align*}
Therefore, for $M>0 $ may be chosen arbitrarily large, we have
\begin{align*}
-\Delta_p(\psi_M)\geq m(x)\left(\lambda\psi_M^{-\delta} + \psi_M^{q}\right)
\geq \lambda k(x)\psi_M^{-\delta} + h(x) \psi_M^{q}
\end{align*}
This shows that $\psi_M$ is a super-solution of the problem \eqref{103}$_+$.
It remains to show that $\psi_c=\underline{u}_\lambda\leq \psi_M= \overline{u}_\lambda$.
Therefore, for $c>0$ small enough and $M>0$ large enough, we obtain
\begin{align*}
&-\Delta_p(\underline{u}_\lambda) \\
&= \Big(\frac{pc}{p-1+\delta}\Big)^{p-1}\Phi_m^{\frac{-\delta p}{p-1+\delta}}
\Big[ \frac{(\delta-1)(p-1)}{p-1+\delta}|\nabla \Phi_m|^p+\lambda_1 m(x)\phi_m^p\Big]\\
&\leq \Big(\frac{pM}{p-1+\delta}\Big)^{p-1}\Phi_m^{\frac{-\delta p}{p-1+\delta}}
\Big[ \frac{(\delta-1)(p-1)}{p-1+\delta}|\nabla \Phi_m|^p+\lambda_1 m(x)\phi_m^p\Big]
= -\Delta_p(\overline{u}_\lambda).
\end{align*}
Consequently, we may apply the weak comparison principle
(see in \cite[Theorem 2.3]{GiScTa}) in order to conclude that
$\underline{u}_\lambda\leq \overline{u}_\lambda$. Thus,
By the classical iteration method \eqref{103}+ has a solution between
the sub-solution and the super-solution.
Let us now prove that $u_\lambda$ is a minimal weak solution of \eqref{103}$_+$.
We use here the weak comparison principle (see Proposition 2.3 in
Cuesta and Tak\'a\v{c} \cite{CuTa}) and the following monotone iterative
scheme:
\begin{equation}\label{scheme}
\begin{gathered}
-\Delta_pu_n-\lambda k(x) u_{n}^{-\delta}= h(x)u_{n-1}^q\quad\text{in }\Omega;\\
u_n|_{\partial\Omega}=0,
\end{gathered}
\end{equation}
where $u_0=\underline{u}_\lambda$, according to Giacomoni, Schindler and
Tak\'{a}\v{c} \cite{GiScTa}, is the unique solution to the following purely
singular problem
\begin{gather*}
-\Delta_p u=\lambda k(x)u^{-\delta}\quad \text{ in }\Omega ,\\
u|_{\partial\Omega}=0, \quad u>0 \quad \text{ in }\Omega.
\end{gather*}
Note that $u_0$ is a weak subsolution to \eqref{103}$_+$ and
$u_0\leq U$ where $U$ is any weak solution to \eqref{103}$_+$.
Then, from the weak comparison principle, we obtain easily
that $u_0 \leq u_1$ and $\{u_n\}_{n=1}^\infty$ is a
nondecreasing sequence. Furthermore, $u_n \leq U$ and $\{u_n\}_{n=1}^\infty$
is uniformly bounded in $W^{1,p}_0(\Omega)$.
Hence, it is easy to prove that
$\{u_n\}$ converges weakly in $W^{1,p}_0(\Omega)$ and pointwise to $u_\lambda$,
a weak solution to the problem \eqref{103}$_+$. Let us show that $u_\lambda$
is the minimal solution to \eqref{103}$_+$ for
any $0<\lambda<\Lambda^*$. Let $v_\lambda$ a weak solution to \eqref{103}$_+$
for any $0<\lambda<\Lambda^*$. Then, $u_0 = \underline{u}_\lambda \leq v_\lambda$.
From the weak comparison principle, $u_n \leq v_\lambda$ for any $n \geq 0$.
Letting $n\to \infty$, we obtain
$u_\lambda \leq v_\lambda$. This completes the proof of the Step 1.
\smallskip
\noindent\textbf{Step 2:} \eqref{103}$_+$ has
no positive solution for $\lambda > \Lambda^*$.
Firstly, from Step 1 we have that $\Lambda^*>0$. Now, let us show that
$\Lambda^*<\infty$.
We argue by contradiction:
suppose there exists a sequence $\lambda_n \to \infty$ such that \eqref{103}$_+$
admits a solution $u_n$. Denote
$$
m := \min \{\operatorname{ess\,inf}_{x\in\Omega} k(x),
\operatorname{ess\,inf}_{x\in\Omega}\inf h(x)\} > 0.
$$
There exists $\lambda^{*}>0$ such that
\[
m\left(\lambda t^{-\delta} + t^q\right)
\ge (\lambda_1+\epsilon) t^{p-1}
\quad \text{for all } t>0,\; \epsilon\in(0,1),\; \lambda > \lambda^{*}
\]
where $\lambda_1$ is the first Dirichlet eigenvalue of $- \Delta_p$ is
positive and is given by
\begin{equation} \label{201}
\lambda_1 = \min_{u \ne 0} \frac{ \int_\Omega |\nabla u|^p}{
\int_\Omega |u|^p}
\end{equation}
(see Lindqvist \cite{Lin1}).
Choose $\lambda_n > \lambda^{*}$. Clearly $u_n$ is a supersolution of the
problem
\begin{equation}\label{problem-max-Lambda}
\begin{gathered}
-\Delta_p u = (\lambda_1 + \epsilon ) u^{p-1}\quad \text{in }\Omega;\\
u>0,\quad u|_{\partial\Omega}=0.
\end{gathered}
\end{equation}
for all $\epsilon\in (0,1)$. We now use the \cite[Lemma 3.1]{GiScTa} to
choose $\mu< \lambda_1 + \epsilon$ small enough so that $\mu
\phi_1(x) < u_n (x)$ and $\mu \phi_1$ is a subsolution to problem
\eqref{problem-max-Lambda}. By a monotone interation procedure we obtain a
solution to \eqref{problem-max-Lambda} for any $\epsilon \in (0,1)$,
contradicting the fact that $\lambda_1$ is an isolated point in the spectrum
of $- \Delta_p$ in $W^{1,p}_0(\Omega)$ (see Anane \cite{An}).
This proves the claim and completes
the proof of the step 2.
\smallskip
\noindent\textbf{Step 3:} Existence of at least one positive weak solution
for $\lambda=\Lambda^*$ to \eqref{103}$_+$.
Let $\{\lambda_k\}_{k\in\mathbb{N}}$ such that
$\lambda_k\uparrow \Lambda^*$ as $k\to \infty$. Then, from Step 1, there
exists
$u_k=u_{\lambda_k}\geq \underline{u}_{\lambda_k}$
to a weak positive solution to \eqref{103}$_+$ for
$\lambda=\lambda_k$. Therefore, for any $\phi\in C^{\infty}_c(\Omega)$,
we have:
\begin{equation}\label{eq30}
\int_\Omega|\nabla u_k|^{p-2}\nabla u_k\nabla\phi\,\mathrm{d}x
=\lambda_k \int_\Omega k(x)u_k^{-\delta}\phi\,\mathrm{d}x
+\int_\Omega h(x) u_k^q\phi\,\mathrm{d}x.
\end{equation}
Since $u_k\in W^{1,p}_0(\Omega)$ and $u_k\geq\underline{u}_{\lambda_k}$ it is easy
to see that \eqref{eq30} holds also for $\phi\in W^{1,p}_0(\Omega)$.
Moreover, from above
\begin{equation}\label{eq31}
E_{\lambda_k}(u_k)\leq
E_{\lambda_k}(\underline{u}_{\lambda_k})
<\frac{1}{p}\int_\Omega|\nabla\underline{u}_{\lambda_k}|^p\,\mathrm{d}x-
\frac{\lambda_k } {1-\delta}
\int_\Omega k(x){\underline{u}_{\lambda_k}}^{1-\delta}\,\mathrm{d}x<0,
\end{equation}
Thus, by Sobolev imbedding and using the fact that $k,h\in L^\infty (\Omega )$
it follows that
\begin{equation}\label{eq31b}
\sup_k\|u_k\|_p<\infty.
\end{equation}
Hence, there exists $u_{\Lambda^*}\geq\underline{u}_{\lambda_k}$ such that
$u_k\rightharpoonup u_{\Lambda^*}$ in $W^{1,p}_0(\Omega)$ as $k\to \infty$
and
\begin{equation}\label{eq32}
u_k\rightharpoonup u_{\Lambda^*}\text{ in $L^q(\Omega)$ since
$p-1
0$
\begin{equation}\label{coerci}
J_\lambda(u) \geq \frac{1}{p}\|u\|^p - C_1\|u\|^{1-\delta} + C_2\|u\|_{q+1}^{q+1}
\end{equation}
where $C_1 =\lambda |\Omega|^{D+E(1-\delta)}S^{\frac{\delta-1}{p}}
\frac{ ||k||_{L^\infty}}{( 1-\delta)}$
with $D=\frac{q+\delta}{q+1}$, $E=\frac{p^*-q-1}{p^*(q+1)}$ and
$S > 0$ is the best Sobolev constant and
$C_2 = (q + 1)^{-1} \operatorname{ess\,inf}_{x\in\Omega} h(x)$ are positive
constants. It follows from \eqref{coerci} that
\begin{equation}\label{coerci1}
J_\lambda(u) \geq \frac{1}{p}\|u\|^p - C_1\|u\|^{1-\delta} .
\end{equation}
and hence
$ J_\lambda(u) \to +\infty$ as $\|u\| \to \infty$.
This completes the proof of our Claim.
Now, let $n \mapsto u_n$ be a minimizing sequence of $J_\lambda$ in
$W^{1,p}_0(\Omega)$. The coercivity of $J_\lambda$ implies the boundedness
of $u_n$ in $W^{1,p}_0(\Omega)$.
Since $J_\lambda(u)=J_\lambda(|u|)$, without loss of
generality, we may assume that $(u_n)_n$ is non-negative, converges weakly
to some $u$ in $W^{1,p}_0(\Omega)$ and converges
also pointwise. Moreover, by the weak lower semicontinuity
of the norm $\|\cdot\|$ and the boundedness of $(u_n)_n$ in $W^{1,p}_0(\Omega)$
we obtain
$$
J_\lambda(u)\leq\underset{n\to \infty}{\lim}\inf J_\lambda(u_n).
$$
Hence $u$ is a global minimizer of $J_\lambda$ in $W^{1,p}_0(\Omega)$.
Which completes the proof of the
Step 1.
\smallskip
\noindent\textbf{Step 2:} The weak limit $u$ is a non-negative weak solution
of problem \eqref{103}$_-$ if
$\lambda > 0$ is sufficiently large.
Firstly, observe that $J_\lambda(0)=0$. So, to prove that the
non-negative solution is non-trivial, it suffices to prove that there exists
$\lambda_\ast > 0$ such that
\begin{equation}\label{nega}
\inf_{u\in W^{1,p}_0(\Omega)} J_\lambda(u)<0\quad \text{for all } \lambda>0.
\end{equation}
For this purpose, take any positive
$u$ and consider $\epsilon u$. Then, for a fixed $\lambda > 0$,
$J_\lambda(\epsilon u) < 0$ if $\epsilon > 0$ is small enough.
Therefore the minimum is negative for all $\lambda > 0$.
Now, we consider the variational problem with constraints,
\begin{equation}\label{mini}
\begin{aligned}
\lambda_\ast & = \inf\Big\{\frac{1}{p}\int_{\Omega} |\nabla w|^p \,\mathrm{d}x
+ \frac{1}{q+1}\int_{\Omega} h(x)|w|^{q+1} \,\mathrm{d}x: w\in W^{1,p}_0(\Omega)
\text{ and }\\
& \quad \frac{1}{1-\delta}\int_{\Omega}k(x) |w|^{1-\delta} \,\mathrm{d}x =1\Big\}.
\end{aligned}
\end{equation}
and define
\begin{equation}
\label{def:Lambda2}
\Lambda_*=
\inf\{ \lambda > 0\colon \eqref{103}_-\text{ admits a nontrivial weak solution} \}.
\end{equation}
From above, we have
$$
J_\lambda(u) = \lambda_\ast - \lambda < 0\quad \text{for any }\lambda > \lambda_\ast.
$$
Therefore, the above remarks show that $\lambda_\ast \geq \Lambda_*$ and that
problem \eqref{103}$_-$ has a solution for all $\lambda> \lambda_\ast$.
We now argue that problem \eqref{103}$_-$ has a solution for all
$\lambda > \Lambda_*$.
Fixed $\lambda>\Lambda_*$, by the definition of $\Lambda_*$, we can take
$\mu \in (\Lambda_* , \lambda)$ such that
that $J_\mu$ has a non-trivial critical point $u_\mu \in W^{1,p}_0(\Omega)$.
Since $\mu < \lambda$, $u_\mu$ is a sub-solution of the problem \eqref{103}$_-$.
In order to find a super-solution of the problem \eqref{103}$_-$ which
dominates $u_\mu$. For this purpose we consider the
constrained minimization problem
\begin{equation}\label{mini1}
\inf \{J_\lambda(w):\; w \in W^{1,p}_0(\Omega) \text{ and } w \geq u_\mu.\}
\end{equation}
Arguments similar to those used to treat \eqref{mini} show that the above
minimization problem has a solution
$u_\lambda > u_\mu$. Moreover, $u_\lambda$ is also a weak solution of problem
\eqref{103}$_-$ for all
$\lambda > \Lambda_*$. With the arguments developed in \cite{GiScTa}
we deduce that problem \eqref{103}$_-$ has a solution if $\lambda=\Lambda_*$.
Thus, one applies \cite[Theorem A.1]{An}, based on the Moser iteration,
shows that $u \in L^\infty_{\rm loc}$.
Next, again by a bootstrap regularity
due to Giacomoni-Schindler-Tak\'{a}\v{c} \cite[Theorem B.1]{GiScTa} shows
that the weak solution $u\in C^{1,\alpha}(\Omega)$ where $\alpha\in(0,1)$.
Finally, the non-negative follows immediately
by the strong maximum principle (see \cite[Theorem 2.3]{GiScTa})
since $u$ is a $C^1$ non-negative weak solution of the differential inequality
$$
-\nabla(|\nabla u|^{p-2} \nabla u) + h(x) u^q \geq 0\;\text{ in} \; \Omega.
$$
We deduce that $u$ is positive everywhere in $\Omega$.
The proof of the step 2 is now complete.
\smallskip
\noindent\textbf{Step 3:} Non-existence for $\lambda > 0$ small.
The same monotonicity arguments as in Step 2 show that \eqref{103}$_-$
does not have any solution if
$\lambda < \Lambda_*$. Which completes the proof of the Theorem \ref{T102}.
\subsection*{Acknowledgments}
I would like to thank the referees for their suggestions and helpful comments
which improved the presentation of the original manuscript.
\begin{thebibliography}{99}
\bibitem{AmBrCe} A. Ambrosetti, H. Brezis, G. Cerami;
\emph{Combined effects of concave and convexe nonlinearities
in some elliptic problems},
Journal of Functional Analysis, \textbf{122} (1994), 519--543.
\bibitem{An} A. Anane;
\emph{Simplicit\'e et isolation de la premi\`ere valeur propre
du $p$-laplacien avec poids},
Comptes Rendus Acad\'emie Sciences Paris, S\'erie I, Math.,
\textbf{305}, (1987), 725-728.
\bibitem{BrKa} H. Brezis, S. Kamin;
\emph{Sublinear elliptic equations in $R^N$ }, Manuscripta Mathematica,
\textbf{74} (1) (1992) 87--106.
\bibitem{CP} M. M. Coclite, G. Palmieri;
\emph{On a singular nonlinear Dirichlet problem},
Communication in Partial Differential Equations, \textbf{14} (1989) 1315--1327.
\bibitem{CrRaTa} M. G. Crandall, P.H. Rabinowitz, L. Tartar;
\emph{On a Dirichlet problem with a singular nonlinearity}, Comm. Partial
Differential Equations, \textbf{2} (1977) p. 193-222.
\bibitem{CuTa} M. Cuesta and P. Tak\'a\v{c};
\emph{A strong comparison principle for positive solutions of
degenerate elliptic equations},
Differential and Integral Equations, \textbf{13}(4--6) (2000), 721--746.
\bibitem{EsVa} R. Esteban, J. L. V\'asquez;
\emph{On the equation of turbulent filtration in one-dimensional porous media,}
Nonlinear Analysis \textbf{10}, (11) (1986) 1303--1325.
\bibitem{GaSa} A. Ghanmi, K. Saoudi;
\emph{The Nehari manifold for a singular elliptic equation involving the
fractional Laplace operator,} Fractional Differential Calculus, \textbf{6},
Number 2 (2016), 201--217.
\bibitem{GaSa1} A. Ghanmi, K. Saoudi;
\emph{A multiplicity results for a singular problem involving the fractional
$p$-Laplacian operator,} Complex variables and elliptic equations \textbf{61},
9 (2016) 1199--1216.
\bibitem{GiSa1} J. Giacomoni, K. Saoudi;
\emph{Multiplicity of positive solutions for a singular and critical problem,}
Nonlinear Analysis \textbf{71} (2009), no. 9, 4060--4077.
\bibitem{GiScTa} J. Giacomoni, I. Schindler, P. Tak\'{a}\v{c};
\emph{Sobolev versus H\"{o}lder local minimizers and
global multiplicity for a singular and quasilinear equation}, annali della scuola
normale superiore di pisa, classe di scienze,
s\'erie V \textbf{6} No.1 (2007) 117--158.
\bibitem{GuWe} Z. Guo and J. R. L. Webb;
\emph{Uniqueness of positive solutions for quasilinear elliptic equations
when a parameter is large,} Proceedings of the Royal Society of Edinburgh,
\textbf{124}, (1994) 189--198.
\bibitem{HeVa} M. A. Herrero, J. L. v\'asquez;
\emph{On the propagation properties of a nonlinear degenerate parabolic equation,}
Communication in Partial Differential
Equations, \textbf{7} (12) (1982) 1381--1402.
\bibitem{LM} A. C. Lazer, P. J. Mckenna;
\emph{On a singular nonlinear elliptic boundary value problem},
Proceedings of the American Mathematical Society \textbf{111} (1991) 721--730.
\bibitem{AA} A. V. Lair, A. W. Shaker;
\emph{Classical and weak solutions of a singular semilinear elliptic problem},
Journal of Mathematical Analysis and Applications, \textbf{211} (1997) 371--385.
\bibitem{Lin1} P. Lindqvist;
\emph{On the equation ${\rm div}(| \nabla u| \sp {p-2}\nabla
u)+\lambda| u| \sp {p-2}u=0$,}
Proceedings of the American Mathematical Society, \textbf{109}(1)(1990) 157--164.
\bibitem{Mi} V. Mikljukov;
\emph{On the asymptotic properties of subsolutions of quasilinear equations
of elliptic type and mappings with bounded distortion,} Sbornik
Mathematics (N.S.) \textbf{111} (1980) (in Russian).
\bibitem{RaRe} V. R\v{a}dulescu, D. Repov\v{s};
\emph{Combined effects in nonlinear problems arising in the study of anisotropic
continuous media}, Nonlinear Analysis, \textbf{75} (2012), 1524--1530.
\bibitem{Sa2} K. Saoudi;
\emph{Existence and nonexistence of positive solutions for quasilinear elliptic
problem,} Abstract and Applied Analysis 2012, Art. ID 275748, 9 pp.
\bibitem{Sa5} K. Saoudi;
\emph{Existence and non-existence of solution for a singular nonlinear
Dirichlet problem involving the $p(x)-$Laplace operator,} J. Adv. Math. Stud.
Vol. 9(2016), No. 2, 292-303.
\bibitem{SaKr} K. Saoudi, M. Kratou;
\emph{Existence of multiple solutions for a singular and quasilinear equation,}
Complex variables and elliptic equations, \textbf{60} (2015), 893--925.
\bibitem{Sa} C. A. Santos;
\emph{Non-existence and existence of entire solutions for a quasi-linear
problem with singular and super-linear terms}, Nonlinear Analysis \textbf{72}
(2010), 3813--3819.
\bibitem{Uh} K. Uhlenbeck;
\emph{Regularity for a class of non-linear elliptic systems,}
Acta Mathematica, \textbf{138} (1977) 219--240.
\end{thebibliography}
\end{document}