\documentclass[reqno]{amsart}
\usepackage{hyperref}

\AtBeginDocument{{\noindent\small
\emph{Electronic Journal of Differential Equations},
Vol. 2014 (2014), No. 214, pp. 1--10.\newline
ISSN: 1072-6691. URL: http://ejde.math.txstate.edu or http://ejde.math.unt.edu
\newline ftp ejde.math.txstate.edu}
\thanks{\copyright 2014 Texas State University - San Marcos.}
\vspace{9mm}}

\begin{document}
\title[\hfilneg EJDE-2014/214\hfil Sturm-Picone type theorems]
{Sturm-Picone type theorems for second-order nonlinear elliptic
differential equations}

\author[A. T\.iryak\.i \hfil EJDE-2014/214\hfilneg]
{Aydin T\.iryak\.i} % in alphabetical order

\address{Aydin T\.iryak\.i \newline
Department of Mathematics and Computer Sciences,
Faculty of Arts and Sciences, Izmir University,
 35350 Uckuyular, Izmir, Turkey}
\email{aydin.tiryaki@izmir.edu.tr}

\thanks{Submitted July 16, 2014. Published October 14, 2014.}
\subjclass[2000]{35B05}
\keywords{Comparison theorem; Sturm-Picone theorem; half-linear equations,
\hfill\break\indent variational lemma; elliptic equations; oscillation}

\begin{abstract}
 The aim of this article is to give Sturm-Picone type theorems for
 the pair of second order nonlinear elliptic differential equations
 \begin{gather*}
 \operatorname{div}(p_1(x)|\nabla u|^{\alpha-1}\nabla u )
 +q_1(x)f_1(u)+r_1(x)g_1(u)=0,\\
 \operatorname{div}(p_2(x)|\nabla v|^{\alpha-1}\nabla v )
 +q_2(x)f_2(v)+r_2(x)g_2(v)=0,
 \end{gather*}
 where $|\cdot|$ denotes the Euclidean length and
 $\nabla= (\frac{\partial}{\partial x_1},\dots,
 \frac{\partial}{\partial x_{n}} )^{T}$ (the superscript $T$
 denotes the transpose). Our results include some earlier results
 and generalize to n-dimensions well-known comparison theorems
 given by Sturm, Picone and Leighton \cite{3Kreith1,Swanson}
 which play a key role in the qualitative behavior of solutions.
 By using generalization of $n$ dimensional Leigton's comparison theorem,
 an oscillation result is given as an application.
\end{abstract}

\maketitle
\numberwithin{equation}{section}
\newtheorem{theorem}{Theorem}[section]
\newtheorem{lemma}[theorem]{Lemma}
\newtheorem{remark}[theorem]{Remark}
\newtheorem{corollary}[theorem]{Corollary}
\allowdisplaybreaks


\section{Introduction}

In the qualitative theory of ordinary differential equations, the
celebrated Sturm-Picone theorem plays a crucial role. In 1836, the
first important comparison theorem was established by Sturm
\cite{1Sturm}. In 1909, Picone \cite{2Picone} modified Sturm's
theorem. For a detailed study and earlier developments of this
subject, we refer the reader to the books \cite{3Kreith1,Swanson}. 
Sturm-Picone theorem is extended in several directions,
see \cite{Ahmad} and \cite{Ahmad2} for linear systems,
\cite{Muller} for nonself adjoint differential equations,
\cite{TyagiandRaghavendra} for implicit differential equations,
\cite{JarosandKusano, Li} for half linear equations,
\cite{6Allegretto1} for degenerate elliptic equations,\cite{Zhang}
for linear equations on time scales and \cite{Tiryaki2014,
Tyagi2013} for a pair of nonlinear differential equations. On the
other hand, we emphasize that the classical proof of Sturm-Picone
theorem heavily depends on the Leighton's variational lemma
\cite{Leighton} (see \cite{Swanson} also). Since when it was
proved, it has been
extended in different contexts, see, for instance 
\cite{Dosly2003, JarosKusanoYoshida2000, Komkov1972}.

There is also a good amount of interest in the qualitative theory
of differential equations to determine whether the given equation
is oscillatory or not and Sturm-Picone theorem also plays an
important role in this direction. For earlier developments, we
refer to \cite{3Kreith1, 2Picone, 1Sturm, Swanson} and for recent
developments, we refer to Yoshida's book \cite{Yoshidakitap}.
Sturm comparison theorems for half linear elliptic equations and
Picone type identities have been studied in, for example 
\cite{7Allegretto2, 8Allegretto3, 6Allegretto1, Allegretto2,9Bognar,
Dosly2002,Fisnarova,JarosKusanoYoshida2001,JarosKusanoYoshida2,118Jaros,
11Kusano2000,Tadie,13Yoshida2,12Yoshida1,Yoshidakitap,13Yoshida,34Yoshida,35Yoshida}.

Recently, Tyagi \cite{Tyagi2013} studied a pair of second order
nonlinear elliptic partial differential equations
\begin{gather}
-\Delta u =q_1(x)f_1(u)+b_1(x)r_1(u), \label{11}\\
-\Delta v =q_2(x)f_2(v)+b_2(x)r_2(v), \label{12}
\end{gather}
under suitable conditions. By establishing a
nonlinear version of Leighton's variational lemma, he gave the
generalization of Sturm-Picone theorem for \eqref{11} and
\eqref{12}. But it is obvious that this result does not work for
the half linear elliptic case. A natural question now arises: Is
it possible to generalize the Sturm comparison results to the
nonlinear elliptic partial differential equations that contain the
half linear case by using
a nonlinear version of Leighton's variational lemma?

Motivated by the ideas in \cite{11Kusano2000, Tiryaki2014,
Tyagi2013}, extending Tyagi's results, we prove a nonlinear
analogue for n-dimensional Leighton's theorem and we give a
generalization of n-dimensional Sturm-Picone theorem by
establishing a suitable nonlinear version of Leighton's
variational lemma which contain the half linear and also linear
elliptic equations.

\section{Main results}

Let us consider a pair of second-order nonlinear elliptic type
partial differential operators:
\begin{gather}
\ell u := \nabla \cdot ( p_1(x)|\nabla u|^{\alpha-1}\nabla u )
+q_1(x)f_1(u)+r_1(x)g_1(u), \label{21} \\
L v := \nabla \cdot ( p_2(x)|\nabla v|^{\alpha-1}\nabla v )
+q_2(x)f_2(u)+r_2(x)g_2(u), \label{22}
\end{gather}
where $|\cdot|$ denotes the Euclidean length and
$\nabla= \big(\frac{\partial}{\partial x_1},\dots,
\frac{\partial}{\partial x_{n}} \big)^{T}$
(the superscript $T$ denotes the transpose). 
In this section, by establishing a nonlinear version of Leighton's 
variational Lemma, we focused on obtaining a generalization of 
n-dimensional Sturm-Picone theorem for \eqref{21} and \eqref{22}.

Let $G$ be a bounded domain in $\mathbb{R}^{n}$ with boundary $\partial G$
having a piecewise continuous unit normal. Let also 
$p_i \in C(\bar{G}, \mathbb{R})$, $q_i$, $r_i \in C^{\mu}(\bar{G}, \mathbb{R})$,
$f_1 \in C^{1}(\mathbb{R}, \mathbb{R})$, $f_2 \in C(\mathbb{R}, \mathbb{R})$,
$g_i \in C(\mathbb{R}, \mathbb{R})$,
for $i=1,2$ where $0<\mu \leq 1$, $q_i$'s are of indefinite
sign for $i=1,2$ and $p_i(x)> 0$, $r_i(x)\geq 0$ 
for all $x \in \bar{G}$ and $\alpha$ is a positive real constant. 

The domain $D_{\ell}(G)$ of $\ell$ is defined to be the set of all
functions $u$ of class $C^{1}(\bar{G},\mathbb{R})$ with the property that
$p_1(x)|\nabla u|^{\alpha-1} \nabla u \in C^{1}(G; \mathbb{R})\cap
C(\bar{G},\mathbb{R})$. The domain $D_{L}(G)$ of $L$ is defined similarly.
Note that such a function $u \in D_{\ell}(G)$ (and $v \in
D_{L}(G)$) exists for \eqref{21} (and \eqref{22}) \cite{Bergerkitap,
Renardy}. The principal part of \eqref{21} (and \eqref{22}) is reduced
to the p-Laplacian $\Delta_p u:=\nabla \cdot (|\nabla u|^{p-2}
\nabla u)$ ($p=\alpha+1$, $p_1(x)\equiv 1$). We know that a
variety of physical phenomena are modelled by equations involving the
p-Laplacian \cite{1Ahmed, 2Aris, 3Astarita, Diaz, 13Oden, 14Pelissier, 16Schoenauer}.


In what follows, we make the following hypotheses on $f_i$ and
$g_i$.

\begin{itemize}
\item[(H1)] Let $f_1 \in C^{1}(\mathbb{R},\mathbb{R})$ and there exist
    $\alpha_0$, $\alpha_1 \in (0, \infty)$ such that
    $\alpha_{0}|u|^{\alpha-1} \leq f'_1(u)$ and
    $\alpha_1|u|^{\alpha-1}u \geq f_1(u)\neq 0$ for all
    $0\neq u \in R$.
    
\item[(H1*)] Let $f_1 \in C^{1}(\mathbb{R},\mathbb{R})$ and there exists
    a $k>0$ such that
    $\frac{f'_1(u)}{|f_1(u)|^{\frac{\alpha-1}{\alpha}}}\geq
    k$ for all $0\neq u \in R$.
    
\item[(H2)] Let $g_1 \in C(\mathbb{R},\mathbb{R})$ and there exists a $\beta
    \geq 0$ such that $\frac{g_1(u)}{f_1(u)}\geq \beta$
    for all $0\neq u \in R$.
    
\item[(H3)]  Let $f_2, g_2 \in C(\mathbb{R},\mathbb{R})$ and there exists
    $\alpha_2, \alpha_3, \alpha_4 \in (0, \infty)$ such   that \\
    $\alpha_3|v|^{\alpha+1}\leq f_2(v)v \leq
    \alpha_2|v|^{\alpha+1}$ and $g_2(v)v \leq
    \alpha_4|v|^{\alpha+1}$
\end{itemize}



\begin{remark} \label{rmk2.1}\rm
Assumption (H1) motivates us to study nonlinearities of the form
$$
f_1(u)=|u|^{\alpha-1} u (1\mp \text{ a nonlinear part})
$$
 where nonlinear part is decays at $\infty$.
\end{remark}

\begin{remark} \label{rmk2.2}\rm
 Assumption (H3) simply says that
$\frac{f_2(v)}{|v|^{\alpha-1}v}$ is bounded for all $0\neq v \in
R$.
\end{remark}

\begin{remark} \label{rmk2.3}\rm
 Assumption (H1*) is a very common
condition  in the literature for half linear equations.
\end{remark}

We begin with a lemma and the definition of some concepts needed
in this article.


\begin{lemma}[\cite{11Kusano2000}] \label{lem2.4} 
 Define $\Phi (\xi)=|\xi|^{\alpha-1}\xi$,  $\xi \in \mathbb{R}^{n}$, $\alpha>0$. 
If $X, Y in \mathbb{R}^{n}$, then
\begin{equation}
X\Phi(X)+\alpha Y \Phi(Y)-(\alpha+1)X \cdot \Phi(Y)\geq 0. \label{23}
\end{equation}
where the equality holds if and only if $X=Y$.
\end{lemma}

Let $U$ be the set of all real valued continuous functions defined
on $\bar{G}$ which vanish on $\partial G$ and have uniformly
continuous firs partial derivatives on $G$. Also define the
functions $j$, $j^{*}$ and $J$: $U\to \mathbb{R}$ by
\begin{equation} \label{24}
\begin{gathered}
j(\eta)=\int_{G}\{p_1(x)|\nabla \eta|^{\alpha+1}-C_1\big(q_1(x)
+\beta r_1(x) \big) |\eta|^{\alpha+1}\}dx, \\
j^{*}(\eta)=\int_{G}\{p_1(x)|\nabla \eta|^{\alpha+1}
-C_2\big(q_1(x)+\beta r_1(x) \big) |\eta|^{\alpha+1}\}dx,\\
J(\eta)= \int_{G}\{p_2(x)|\nabla
\eta|^{\alpha+1}-(\alpha_2q_2^{+}(x)-\alpha_3q_2^{-}(x)+\alpha_4
r_2(x))|\eta|^{\alpha+1}\} dx
\end{gathered}
\end{equation}
where $C_1=(\frac{\alpha_{0}}{\alpha_1 \alpha})^{\alpha}\alpha_1$,
 $C_2=(\frac{k}{\alpha})^{\alpha}$, $q_2^{+}=max\{q_2, 0\}$ and
$q_2^{-}=max\{-q_2, 0\}$. The variation $V(\eta)$ and
$V^{*}(\eta)$ are defined as
\begin{equation} \label{26}
\begin{gathered}
V(\eta)=J(\eta)-j(\eta) ,\\
 V^{*}(\eta)=J(\eta)-j^{*}(\eta)
 \end{gathered}
\end{equation}
with domain $D:=D_{j}\cap D_{J}= D_{j^{*}}\cap D_{J}$.


To prove a nonlinear analogue of Leighton's theorem we
first establish a nonlinear version of Leighton's variational
lemma (Generalization of n-dimensional
Leighton's variational type lemma).

\begin{lemma} \label{lem2.5}  
Assume that there exists a
nontrivial function $\eta \in U$ such that $j(\eta)\leq 0$ 
(or $j^{*}(\eta)\leq 0$ ). Then under the hypotheses {\rm (H1)} 
(or {\rm (H1*)}) and {\rm (H2)}, every solution $u\in D_{j}$ of $\ell(u)=0$
vanishes at some points of $\bar{G}$.
\end{lemma}

\begin{proof}
Let us give the proof under the conditions $j(\eta)\leq 0$, (H1)
and (H2). Similarly proof holds for $j^{*}(\eta)\leq 0$,
(H1*) and (H2). Assume on the contrary that the statement is
false. Suppose that there exists a solution $u \in D_{\ell}(G)$ of
$\ell(u)=0$ satisfying $u\neq 0$ on $\bar{G}$. By (H1), we have
$f_1(u(x))\neq 0$, $\forall x \in \bar{G}$. 
Then for $\eta \in U$, the following equality is valid in $G$:
\begin{align*}
&\nabla \cdot \Big(\frac{\alpha \eta
\Phi(\eta)}{f_1(u(x))}p_1(x)|\nabla u|^{\alpha-1}\nabla u\Big)\\
&=\sum_{i=1}^{n} \{ \frac{\partial}{\partial x_i}(\alpha
\eta \Phi(\eta))\frac{p_1(x)|\nabla u|^{\alpha-1}\nabla
u}{f_1(u(x))}\\
&\quad + \alpha \eta \Phi(\eta) (\frac{\partial}{\partial x_i} 
 \frac{1}{f_1(u(x))})p_1(x)|\nabla u|^{\alpha-1}\nabla u   \\
&\quad + \frac{\alpha \eta \Phi(\eta)}{f_1(u(x))} 
  \frac{\partial}{\partial x_i} (p_1(x)|\nabla u|^{\alpha-1}\nabla u)\}   \\
&= p_1(x)\frac{|f_1(u(x))|^{\alpha-1}}{(f'_1(u(x)))^{\alpha}}|\alpha 
 \nabla \eta|^{\alpha+1}-\alpha q_1(x)|\eta|^{\alpha+1}-\alpha r_1(x) 
 \frac{g_1(u(x))}{f_1(u(x))}|u|^{\alpha+1}  \\
&\quad -p_1(x)\frac{|f_1(u(x))|^{\alpha-1}}{(f'_1(u(x)))^{\alpha}} 
 F \Big(\frac{\eta \nabla u f'_1(u(x))}{f_1(u(x))}, \alpha \nabla \eta \Big),
\end{align*}
where
\begin{align*}
& F (\frac{\eta \nabla u f'_1(u(x))}{f_1(u(x))}, \alpha \nabla \eta )\\
& =|\alpha \nabla \eta|^{\alpha+1}
 +\alpha \big|\frac{\eta \nabla u f'_1(u(x))}{f_1(u(x))} \big|^{\alpha+1}
 -(\alpha+1)\alpha \nabla\eta \cdot 
 \Phi \Big(\frac{\eta \nabla u f'_1(u(x))}{f_1(u(x))} \Big)
\end{align*}
By (H1) and (H2), we obtain
\begin{equation}
\begin{aligned}
&p_1(x)|\nabla \eta|^{\alpha+1}-C_1(q_1(x)+\beta r_1(x))|
\eta|^{\alpha+1} \\
&\geq C_1 \nabla \cdot \Big(\frac{\eta \Phi
(\eta)}{f_1(u(x))}p_1(x)|\nabla u|^{\alpha-1}\nabla u \Big)\\
&\quad +\frac{C_1}{\alpha}p_1(x)\frac{|f_1(u(x))|^{\alpha-1}}{f'_1(u(x))^{\alpha}}
F\Big(\frac{\eta \nabla u f'_1(u(x))}{f_1(u(x))}, \alpha
\nabla \eta \Big).
\end{aligned} \label{27}
\end{equation}
We integrate \eqref{27} over $G$ and then apply the divergence
theorem to obtain
$$
j(\eta)\geq \frac{C_1}{\alpha}\int_{G} p_1(x)
\frac{|f_1(u(x))|^{\alpha-1}}{f'_1(u(x))^{\alpha}}F
\Big(\frac{\alpha \nabla u f'_1(u(x))}{f_1(u(x))}, \alpha
\nabla \eta \Big)dx \geq 0 \,.
$$
Therefore,
$$
\int_{G}p_1(x)\frac{|f_1(u(x))|^{\alpha-1}}{f'_1(u(x))^{\alpha}}F
\Big(\frac{\eta \nabla u f'_1(u(x))}{f_1(u(x))}, \alpha
\nabla \eta \Big)=0.
$$
From Lemma \ref{lem2.4}, we see that
$$
\frac{\eta \nabla u f'_1(u(x))}{f_1(u(x))}=\alpha \nabla \eta \quad
\text{or}\quad
\nabla ( \frac{|\eta (x)|^{\alpha}}{|f_1(u(x))|} )=0\quad\text{in } G.
$$
Since $\eta \in U$,  there exists a nonzero constant $K$ such that
$$
|\eta(x)|^{\alpha}=|K f_1(u(x))|
$$
in $G$ and hence on $\bar{G}$ by continuity. This is not
possible because $\eta(x)=0$ on $\partial G$ but
 $f_1(u(x)) \neq 0$ on $\partial G$ ($u(x)\neq 0$ on $\partial G$). This implies
that $j(\eta)>0$, which is a contradiction and hence every
solution $u$ of $\ell u=0$ vanishes at some point of $\bar{G}$.
This completes the proof.
\end{proof}

Lemma \ref{lem2.5} plays a crucial role to establish the following
Generalization of $n$-dimensional Leighton's theorem.

\begin{theorem} \label{thm2.6}  
Let {\rm (H1)} (or {\rm (H1*)}), {\rm (H2)} and {\rm (H3)} hold. 
If there exists a nontrivial solution $v \in D$ of $L v=0$
in $\bar{G}$ such that $v=0$ on $\partial G$ and $V(v)\geq 0$ 
(or $V^{*}(v)\geq 0$), then every solution $u$ of $\ell u=0$ vanishes
at some point of $\bar{G}$.
\end{theorem}

\begin{proof}
As in the proof of Lemma \ref{lem2.5} , let us give the proof under the
conditions (H1), (H2) and $V(v)\geq 0$. Since $v$ is a
solution of $L v=0$ and $v=0$ on $\partial G$ so by an application
of Green's theorem we have
\begin{equation} \label{28}
\begin{aligned}
&\int_{G} \Big(q_2(x)f_2(v)v+r_2(x)g_2(v) v \Big)dx\\
&=-\int_{G} v \nabla \cdot (p_2(x)|\nabla v|^{\alpha-1}\nabla v)dx  \\
&=-v (p_2(x) |\nabla v|^{\alpha-1}\nabla v)\mid_{\partial G}
 + \int_{G} p_2(x)|\nabla v|^{\alpha+1}dx  \\
&=\int_{G} p_2(x)|\nabla v|^{\alpha+1}dx.
\end{aligned}
\end{equation}
In view of (H3), one can see that
\begin{equation}
\int_{G} \Big(q_2(x)f_2(v)v+r_2(x)g_2(v) v \Big)dx
\leq \int_{G} [(\alpha_2 q^{+}_2(x)-\alpha_3 q^{-}_2(x))
+ \alpha_4r_2(x)]|v|^{\alpha+1} dx. \label{2828}
\end{equation}
By \eqref{28} and \eqref{2828}, we have $J(v)\leq 0$.
Since $V(u)\geq 0$, this implies
$$
j(v)\leq J(v)\leq 0
$$
and hence by application of Lemma \ref{lem2.5} every nontrivial
solution $u$ of $\ell u=0$ vanishes at some point of $\bar{G}$.
This completes the proof.
\end{proof}


\begin{remark} \label{rmk2.7} \rm
 If the condition $V(v)\geq 0$ (or $V^{*}(v)\geq 0$) is strengthened to 
 $V(v)>0$ (or $V^{*}(v)>0$), the conclusion of Theorem \ref{thm2.6} holds
  also in the domain $G$.
\end{remark}

From Theorem \ref{thm2.6}. we immediately have the following Corollary
which is an $n$-dimensional extension of Sturm-Picone comparison
theorem for the operators \eqref{21} and \eqref{22}.

\begin{corollary} \label{coro2.8} 
Let {\rm (H1)} (or {\rm (H1*)}), {\rm (H2)} and {\rm (H3)} hold. 
Suppose there exists a nontrivial solution $v$ of $L v=0$ in
$\bar{G}$ such that $v=0$ on $\partial G$. If $p_2(x)\geq p_1(x)$ and
\begin{gather*}
C_1(q_1(x)+\beta r_1(x))\geq [\alpha_2q_2(x)
-(\alpha_3-\alpha_2)q^{-}_2(x)+\alpha_4r_2(x)],\\
\Big(\text{or }
C_2(q_1(x)+\beta r_1(x))\geq [\alpha_2q_2(x)
-(\alpha_3-\alpha_2)q^{-}_2(x)+\alpha_4r_2(x)]
\Big),
 \end{gather*}
for every $x \in \bar{G} $. Then every nontrivial
solution $u$ of $\ell u=0$ vanishes at some point of $\bar{G}$.
\end{corollary}

From Lemma \ref{lem2.5}, Theorem \ref{thm2.6} and Corollary \ref{coro2.8}
 we easily obtain the
following results which are straightforward extensions of the
variational Lemma, Leighton's theorem and the celebrated
Sturm-Picone theorem from \cite{3Kreith1, Swanson} valid for
linear second order ordinary differential equations to half linear
elliptic partial differential equations that contain linear case.


\begin{corollary} \label{coro2.9}
 Let $f_1(u)=|u|^{\alpha-1}u$ and
either $r_1(x)\equiv 0$ or $g_1(u)\equiv 0$ in \eqref{21}. If there
exists a nontrivial function $\eta \in U$ such that
\begin{eqnarray}
\int_{G} \{p_1(x)|\nabla \eta|^{\alpha+1}-q_1(x)|\eta|^{\alpha+1}  \} dx
\leq 0 \label{210}
\end{eqnarray}
 then every nontrivial solution $u$ of half linear
elliptic equation
\begin{equation}
\nabla \cdot (p_1(x)|\nabla u|^{\alpha-1}\nabla u )+q_1(x)|u|^{\alpha-1}u=0 \label{211}
\end{equation}
vanishes at some point in $\bar{G}$.
\end{corollary}

\begin{corollary} \label{coro2.10}
Suppose that there exists a nontrivial solution $v$ of
\begin{equation}
\nabla \cdot (p_2(x)|\nabla v|^{\alpha-1}\nabla v )+q_2(x)|v|^{\alpha-1}v=0 \label{212}
\end{equation}
in $\bar{G}$ such that $v=0$ on $\partial G$. If
\begin{equation}
\int_{G} \{ (p_2(x)-p_1(x))|\nabla v|^{\alpha+1} +(q_1(x)-q_2(x))|v|^{\alpha+1} \}
dx\geq 0 \label{213}
\end{equation}
then every nontrivial solution $u$ of \eqref{211}
vanishes at some point of $\bar{G}$.
\end{corollary}

\begin{corollary} \label{coro2.11}
 Let $p_2(x)\geq p_1(x)$ and
$q_1(x)\geq q_2(x)$ for every $x \in \bar{G}$. If there exists
a nontrivial solution $v$ of \eqref{212} in $\bar{G}$ such that
$v=0$ on $\partial G$, then any nontrivial solution $u$ of
\eqref{211} vanishes at some point of $\bar{G}$.
\end{corollary}

Note that the Corollaries 2.9--2.11  were also obtained in
\cite{Dosly2002, 10Dunninger, JarosKusanoYoshida2, 11Kusano2000,
Yoshidakitap}. But their proofs depend on the Picone-type and
Wirtinger type inequalities.

Recently  Bal \cite{K.Bal} gave a nonlinear version of the
Sturmian comparison principle for a special case of \eqref{21} and
\eqref{22} as the follows.

\begin{theorem}[\cite{K.Bal}] \label{thm2.12}  
Let $q_1$ and $q_2$ be the two weight functions such that
 $q_2<q_1$ and $f_1$
satisfies $f_1'(u)\geq (p-1)\big(f_1(u)^{\frac{p-2}{p-1}}\big)$. 
If there is a positive solution $v$ satisfying
\begin{equation}
-\Delta_p v=q_2(x)|v|^{p-2}v\text{ for } \Omega^{*}, \quad
v=0\text{ on }\partial \Omega^{*}, \label{*1}
\end{equation}
then any nontrivial solution $u$ of
\begin{equation}
-\Delta_p u=q_1(x)f_1(u)\quad\text{for } x \in \Omega^{*} \label{*2}
\end{equation}
must change sign, where $\Omega^{*}$ denotes any domain
in $\mathbb{R}^{n}$, $1<p<\infty$ and $f_1: (0,\infty)\to (0,\infty)$
is  a $C^{1}$ function.
\end{theorem}

From the hypothesis of $f_1$ this conclusion is not true.
Because for $u \in (0, \infty)$, $f_1(u)>0$ but for $u<0$
$f_1(u)$ is not defined.

This result can be corrected by using Corollary \ref{coro2.8}, for the
bounded domain $\bar{G}$ in $\mathbb{R}^{n}$ and we can give the following
Sturmian comparison result for the equations \eqref{*1} and
\eqref{*2} as follows:

\begin{corollary} \label{coro2.13}
 Let {\rm (H1*)} hold with $k=\alpha=p-1$. 
If there exists a nontrivial solution $v \in D$ of
\eqref{*1} in $\bar{G}$ such that $v=0$ on $\partial G$ and
$q_1(x)\geq q_2(x)$, then every solution $u$ of \eqref{*2}
vanishes at some point of $\bar{G}$.
\end{corollary}

\section{An application}

This section deals with an application of Theorem \ref{thm2.6}. This
theorem enables us to develop some oscillation criteria for the
equation $\ell u=0$.

Let $\Omega$ be an exterior domain in $\mathbb{R}^{n}$, that is, a domain
such that $\Omega \supset \{x \in \mathbb{R}^{n}: |x|\geq r_{0} \}$ for
some $r_{0}>0$, and consider the nonlinear elliptic equation
\begin{equation}
\nabla \cdot (p_1(x)|\nabla u|^{\alpha-1}\nabla u )+ q_1(x)f_1(u)=0 \label{31}
\end{equation}
in $\Omega$ where $\alpha>0$ is a constant, $p_1 \in C(\Omega, \mathbb{R}^{+})$,
$q_1 \in C(\Omega, \mathbb{R})$ and $f_1$ satisfy
the hypothesis (H1) (or (H1*)).

A nontrivial solution of \eqref{31} is said to be oscillatory if
it has a zero in $\Omega \cap \{x \in \mathbb{R}^{n}: |x|>r^{*} \}$ for any
$r^{*}>r_{0}$. For brevity,  \eqref{31} is called oscillatory
if all of its nontrivial solutions are oscillatory.

We will show that an explicit oscillation criterion for 
\eqref{31} can be obtained via the comparison principle proven in
the preceding section. Our main idea is to compare \eqref{31} with
suitably chosen equations with radial symmetry of the type
\begin{equation}
\nabla \cdot (\tilde{p_1}(|x|)|\nabla v|^{\alpha-1}\nabla v )
+ \bar{q_1}(|x|)f_1(v)=0 \label{32}
\end{equation}
in $\{x \in \mathbb{R}^{n}: |x|\geq r_{0} \}$ and employ
information about the oscillatory behavior of radially symmetric
solutions of \eqref{32}.

It is easily verified that if $v=y(|x|)$ is radially symmetric
solution of \eqref{32}, then the function $y(r)$ satisfies the
differential equation
\begin{equation}
(r^{n-1} \tilde{p_1}(r)|y'|^{\alpha-1}y' )'
+ r^{n-1}\bar{q_1}(r)f_1(y)=0,\quad r\geq r_{0} \label{33}
\end{equation}
We note that \eqref{33} is a special case of the equation
\begin{equation}
( p(r)|y'|^{\alpha-1}y' )'+ q(r)f_1(y)=0,\;\;r\geq r_{0} \label{34}
\end{equation}
the oscillatory behavior of which has been intensively
investigated in recent years by numerous authors \cite{Agarwal,
Dosly2005}.

Suppose that $p(r)$ and $q(r)$ are continuous functions defined on
$[r_{0}, \infty)$ such that $p(r)>0$ on $[r_{0}, \infty)$. A
solution of \eqref{34} is a function $y:[r_{0}, \infty)\to \mathbb{R}$ 
which is continuously differentiable on $[r_{0}, \infty)$
together with $p|y'|^{\alpha-1}y'$ and satisfies \eqref{34} at
every point of $[r_{0}, \infty)$. A nontrivial solution is said to
be oscillatory if it has a sequence of zeros clustering ar
$r=\infty$, and nonoscillatory otherwise. Now we give an
oscillation criterion for \eqref{34}. Its proof can be found, for
example, in \cite{Agarwal}.


\begin{lemma} \label{lem3.1}
 Let {\rm(H1)}  (or {\rm (H1*)}) hold. Suppose
that $p \in C([r_{0},\infty), \mathbb{R}^{+})$ and $q \in C([r_{0},\infty),
\mathbb{R})$ satisfies
$$
\int_{r_1}^{r}\Big( \int_{r_0}^{s} p(u) du \Big)^{-1/ \alpha} ds=\infty
$$
and
$$
\lim _{r \to \infty } {\frac{1}{r} \int_{r_{0}}^{r} 
\Big( \int_{r_0}^{s} q(u) du \Big)}ds=\infty.
$$
Then  \eqref{34} is oscillatory.
\end{lemma}

We first establish a principle which enables us to deduce the
oscillation of \eqref{31} from the one-dimensional oscillation of
\eqref{33}.


\begin{theorem} \label{thm3.2}
 If there exist functions $\tilde{p_1}
\in C([r_{0},\infty), \mathbb{R}^{+})$ and $\tilde{q_1} \in
C([r_{0},\infty), \mathbb{R})$ such that
$$
\tilde{p_1}(r) \geq \max_{|x|=r} {p_1(x)} 
$$
and
\begin{equation}
\alpha_2\tilde{q_1}^{+}(r)-\alpha_3\tilde{q_1}(r)
\leq C_1 \min_{|x|=r} {q_1(x)}  \label{311}
\end{equation}
\[
\Big(\text{or }\alpha_2\tilde{q_1}^{+}(r)-\alpha_3\tilde{q_1}(r)
\leq C_2 \min_{|x|=r} {q_1(x)}\Big)
\]
where $\alpha_2$, $\alpha_3$, $C_1$ and $C_2$ are
defined as before, and the ordinary differential equation
\eqref{33} is oscillatory, then \eqref{31} is oscillatory in
$\Omega$.
\end{theorem}

\begin{proof}
By hypothesis there exists an oscillatory solution $y(r)$ of
\eqref{33} on $[r_0, \infty)$. Let $\{r_i\}$ be the set of all
zeros of $y(r)$ such that $r_0\leq r_1 < r_2 <\dots< r_i<\dots$,
$\lim_{i \to \infty} {r_i}=\infty $. Then the function
$v(x)=y(|x|)$ is a radially symmetric solution of \eqref{32}
which is defined in $\{x \in \mathbb{R}^{n}: |x|\geq r_{0}  \}$ and has the
spherical nodes $|x|=r_i$, $i=1,2,\dots$. Let us compare \eqref{31}
with \eqref{32} in the annular domains $G_i= \{ x \in \mathbb{R}^{n}: r_i
<|x|< r_{i+1} \}$, $i=1,2,\dots$. For each $i$, $v$ is a solution of
\eqref{32} in $G_i$ such that $v\neq 0$ in $G_i$ and $v=0$ on
$\partial G_i$. Since \eqref{311} implies
$$
\tilde{p_1}(|x|)\geq p_1(x)
$$
and
$$
\alpha_2{\tilde{q_1}}^{+}(|x|)-\alpha_3{\tilde{q_1}}^{-}(|x|)\leq C_1 q_1(x)
$$
$$
\Big(\text{or }\alpha_2{\tilde{q_1}}^{+}(|x|)-\alpha_3{\tilde{q_1}}^{-}(|x|)\leq
C_2 q_1(x)\Big)
$$
in $\{x \in \mathbb{R}^{n}: |x|\geq r_{0}  \}$, we obtain
\begin{align*}
V(v)
&\equiv \int_{G_i} \{ (\tilde{p_1}(|x|)-p_1(x))|\nabla v|^{\alpha+1}+[C_1 q_1(x)\\
&\quad -( \alpha_2 \tilde{q_1}^{+}(|x|) -\alpha_3\tilde{q_1}^{-}(|x|)) ]
|v|^{\alpha+1}  \}dx \geq 0
\\
\Big(\text{or } V^{*}(v)
&\equiv \int_{G_i} \{ (\tilde{p_1}(|x|)
-p_1(x))|\nabla v|^{\alpha+1}+[C_2 q_1(x)\\
&\quad -( \alpha_2 \tilde{q_1}^{+}(|x|) 
-\alpha_3\tilde{q_1}^{-}(|x|)) ] |v|^{\alpha+1}  \}dx \geq 0\Big).
\end{align*}
Consequently from Theorem \ref{thm2.6}, it follows that every solution $u$
of \eqref{31} has a zero in $G_i$, $i=1,2,\dots,$ which shows that
$u$ is oscillatory in $\Omega$. This completes the proof.
\end{proof}

\begin{remark} \label{rm3.3} \rm
 An immediate consequence of Theorem \ref{thm3.2} is
that  \eqref{32} with $\tilde{p_1} \in C([r_0, \infty),
\mathbb{R}^{+})$ and 
$\tilde{a_1} \in C([r_0, \infty), \mathbb{R})$ is oscillatory
in $\{x \in \mathbb{R}^{n}: |x|\geq r_{0} \}$ if it has one radially
symmetric solution which is oscillatory there.
\end{remark}

Combining Theorem \ref{thm3.2} with Lemma \ref{lem3.1} applied to  \eqref{33}
gives the following oscillation criteria for  \eqref{31}.

\begin{theorem} \label{thm3.4}
 Let $\tilde{p_1} \in C([r_{0}, \infty),
\mathbb{R}^{+})$ and $\tilde{q_1} \in C([r_{0}, \infty), \mathbb{R})$ be functions
satisfying \eqref{311}. Let also {\rm (H1)} (or {\rm(H1*)}) hold. If the
functions $\tilde{p_1}$ and $\tilde{q_1}$ satisfy
$$
\int_{r_1}^{r}( \int_{r_0}^{s} u^{n-1}\tilde{p_1}(u) du )^{-1/ \alpha} ds=\infty
$$
and
$$
\lim _{r \to \infty } {\frac{1}{r} \int_{r_{0}}^{r} 
\Big( \int_{r_0}^{s} u^{n-1}\tilde{q_1}(u) du \Big)}ds=\infty,
$$
then  \eqref{31} is oscillatory in $\Omega$.
\end{theorem}

\begin{thebibliography}{11}

\bibitem{Agarwal} Agarwal, R. P.; Grace, S. R.; O'Regal, D.;
\emph{Oscillation Theory for Second order Linear, Half-linear,
Superlinear Dynamic Equations}, Kluwer Academia Publishers, London,
2002.

\bibitem{Ahmad} Ahmad, S.; Lazer, A., C.;
\emph{A new generalization of the Sturm comparison theorem to self adjoint systems}, 
Proc. Amer. Math. Soc., 68 (1978), 185-188.

\bibitem{Ahmad2} Ahmad, S.;
\emph{On Sturmian theory for second order
systems}, Proc. Amer. Math. Soc. 87 (1983), 661-665.


\bibitem{1Ahmed} Ahmed, N.; Sunada, D. K.;
\emph{Nonlinear flow in prous media}, J. Hydraulics Division Proc. Amer. Soc. Civil Eng., 95(1969),
1847-1857.

\bibitem{7Allegretto2} Allegretto, W.; Huang, Y. X.;
\emph{A Picone's identity for the p-Laplacian and applications}, Nonlinear Anal.,
32, (1998), 819-830.

\bibitem{8Allegretto3} Allegretto, W.; Huang, Y. X.;
\emph{Principal eigenvalues and Sturm Comprison via Picone's identity}, J.
Differential Equations 156, (1999), 427-438.

\bibitem{6Allegretto1} Allegretto, W.;
\emph{Sturm theorems for degenerate elliptic equations},
Proc. Amer. Math. Soc. 129(2001) 3031-3035. 

\bibitem{Allegretto2} Allegretto, W.;
\emph{Sturm theorems for degenerate elliptic equations}, 
Proc. Amer. Math. Soc. 129 (2001), 3031-3035.

\bibitem{2Aris} Aris, R.;
\emph{The Mathematical Theory of Diffusion and
Reaction in Permeable Catalsts, vols. I and II}, Clarendon Press,
Oxford, 1975.

\bibitem{3Astarita} Astarita, G.; Marruci, G.;
\emph{Principles of Non-Newtonian
Fluid Mechanics}, McGraw-Hill, New York, 1974.

\bibitem{K.Bal} Bal, K.;
\emph{Generalized Picone's identity and its
applications}, Electronic Journal of Differential Equations, Vol.
2013 (2013), No. 243, pp. 1-6.

\bibitem{Bergerkitap} Berger, M. S.;
\emph{Nonlinearity and Functional Analysis}, 
Lectures on Nonlinear Problems in Mathematical Analysis,
Academic Press, Inc. New York, 1997.

\bibitem{9Bognar} Bogn\'ar, G.; Do\v{s}l\'y;
\emph{The application of Picone-type identity for some nonlinear elliptic 
differential equations}, Acta
Math. Univ. Comenian. 72 (2003), 45-57.


\bibitem{Diaz} Diaz, J. I.;
\emph{Nonlinear Partial Differential
Equations and Free Boundaries, vol. I, Elliptic Equations}, Res.
Notes Math., Vol. 106, Pitman, London, 1985.

\bibitem{Dosly2002} Dosl\'y, O.;
\emph{The Picone identity for a class of partial differential equations}, 
Mathematica Bohemica, 127 (2002), 581-589.

\bibitem{Dosly2003} Dosl\'y, O.; Jaro\v(s), J., A.;
\emph{singular version of Leighton's comparison theorem for forced
quasilinear second-order differetial equations}, Arch. Math.
(BRNO), 39 (2003), 335-345.

\bibitem{Dosly2005} Dosl\'y, O.; \v{R}\'ak, P.;
\emph{Half-linear Differential Equations}, North-Holland Mathematics
Studies, 202, Elsevier Science B. V., Amsterdam, 2005.


\bibitem{10Dunninger} Dunninger, D. R.;
\emph{A sturm comparison theorem for some degenrate quasilinear elliptic operators},
 Boll. Un. Mat. Ital. A (7)9, (1995), 117-121.

\bibitem{Fisnarova} Fisnarov\'a, S.; Marik, R.;
\emph{Generalized Picone and Riccati inequalites for half-linear
differential operators with arbitrary elliptic matrices},
Electronic J. Diff. Equations, No: 111, (2010), pp. 1-13.

\bibitem{JarosandKusano} Jaro\v{s}, J.; Kusano, T.;
\emph{A Picone type identity for second order half-linear differential
equations}, Acta Math. Univ. Comenianae, Vol. LXVIII, 1(1999),
137-151.

\bibitem{JarosKusanoYoshida2000} Jaro\v{s}, J.; Kusano, T.; 
Yoshida, N.;
\emph{Forced superlinear oscillations via Picone's identity},
Acta Math. Univ. Comenianae, LXIX (2000), 107-113.


\bibitem{JarosKusanoYoshida2001} Jaro\v{s}, J.; Kusano, T.; Yoshida, N.;
\emph{Picone-type inequalities for nonlinear elliptic equations and
their applications}, J. Inequal. Appl. 6 (2001), 387-404.

\bibitem{JarosKusanoYoshida2} Jaro\v{s}, J.; Kusano, T.; Yoshida, N.;
\emph{Picone type inequalies for half linear elliptic
equations and their applications}, Adv. Math. Sci. Appl., 12(2),
(2002), 709-724.

\bibitem{118Jaros} Jaro\v{s}, J.; Kusano, T.; Yoshida, N.;
\emph{Picone-type inequalities for elliptic equations with first order
terms and their applications}, J. Inequal. Appl. (2006), 1-17.

\bibitem{Komkov1972} Komkov, V.;
\emph{A generalization of Leighton's
variational theorem}, Applicable Analysis, 2(1972), 377-383.

\bibitem{3Kreith1} Kreith, K.;
\emph{Oscillation Theory}, Lecture Notes in
Mathematics, Vol.324, (1973), (Springer-Verlag, Berlin).

\bibitem{11Kusano2000} Kusano, T., Jaro\v{s}, J.; Yoshida, N.,;
\emph{A Picone-type identity and Sturmain comparison and oscillation
theorems for a class of half-linear partial differential equations
of second order}, Nonlinear Analysis, Theory, Methods and
Applications 40 (2000), 381-395.

\bibitem{Leighton} Leighton, W.;
\emph{Comparison theorems for linear differential
equations of second order}, Proc. Amer. Math. Soc. 13 (1962),
603–610.

\bibitem{Li} Li H. J.; Yeh, C. C.;
\emph{Sturmian comparison theorem for
half-linear second order differential equations}, Proc. Roy. Soc.
Edinburgh 125A (1995), 1193–1204.

\bibitem{Muller} M\"{u}ller-Pfeiffer, E.;
\emph{Sturm Comparison theorems for
non-selfadjoint differential equations on non-compact intervals},
Math. Nachr., 159 (1992), 291-298.


\bibitem{13Oden} Oden, J. T.;
\emph{Existence theorems and approximations in
nonlinear elasticity}, in: E. Lakshmikantham (Ed.), Nonlinear
Equations in Abstract Space, North-Holland, 1978.

\bibitem{14Pelissier} Pelissier, M. C.;
\emph{Sur quelques probl\`{e}mes non
linearies en glaciologie}, Th\`{e}se, Publ. Math. d'orsay, 110,
1975.

\bibitem{2Picone} Picone, M.;
\emph{Sui valori eccezionali di un parametro
da cui dipende un'equazione differenziale lineare ordinaria del
second'ordine}, Ann. Scuola Norm. Sup. Pisa 11(1909), 1-141.

\bibitem{Renardy} Renardy, M.; Rogers, R. C.;
\emph{An introduction to Partial Differential Equations}, 
Springer Verlag, New York, Inc. 2004.

\bibitem{16Schoenauer} Schoenauer, M.;
\emph{A monodimensional model for fracturing}, 
in: A. Fasano, M. Primicerio (Eds.), Free Boundary
Problems: Theory, Applications, in: Res. Notes Math., vol. 79,
Pitman, London, 1983, pp. 701-711.

\bibitem{1Sturm} Sturm, C.;
\emph{Sur les équations différentielles
linéaires du second ordre}, J. Math. Pures. Appl. 1(1836), 106-186.

\bibitem{Swanson} Swanson, C. A.;
\emph{Comparison and Oscillation Theory of Linear Differential Equations},
 Academic Press, New York, 1968.

\bibitem{Tadie} Tadie, J.;
\emph{Oscillation criteria for semilinear elliptic
equations with a damping term in $\mathbb{R}^n$}, Electronic J. Diff.
Equations, No. 51, (2010), pp. 1-5.

\bibitem{Tiryaki2014} Tiryaki, A.;
\emph{Sturm-Picone type theorems for second-order
nonlinear differential equations}, Electronic Journal of
Differential Equations, Vol. 2014 (2014), No. 146, pp.1-11.

\bibitem{TyagiandRaghavendra} Tyagi, T.; Raghavendra, V.;
\emph{A note on generalization of Sturm's comparison theorem}, 
Nonlinear Dyn. Syst. Theory, 8(2) (2008), 213-216.

\bibitem{Tyagi2013} Tyagi, J.;
\emph{Generalization of Sturm-Picone
theorem for second-order nonlinear differential equations},
Taiwanese Journal of Mathematics, Vol. 17. No1. (2013) pp.
361-378.

\bibitem{13Yoshida2} Yoshida, N.;
\emph{Oscillation of half-linear partial differenetial equations with
first order terms}, Stud. Univ. Zilina, 17 (2003), 177-184.

\bibitem{12Yoshida1} Yoshida, N.;
\emph{Oscillation criteria for half-linear
partial differential equations via Picone's Identity}, in:
Proceedings of Equadiff-11, 2005, pp. 589-598.

\bibitem{Yoshidakitap} Yoshida, N.;
\emph{Oscillation Theory of Partial Differential Equations}, 
World Scientific Publishing Co. Pte. Ltd., 2008.

\bibitem{13Yoshida} Yoshida, N.;
\emph{Sturmian comparison and oscillation theorems for a class 
of half-linear elliptic equations}
Nonlinear Analysis, Theory, Methods and Applications,
71(2009) e1354-1359.

\bibitem{34Yoshida} Yoshida, N.;
\emph{A Picone identity for half-linear elliptic
equations and its applications to oscillatory theory}, Nonlinear
Anal. 71 (2009), 4935-4951.

\bibitem{35Yoshida} Yoshida, N.;
\emph{Sturmian comparsion and oscillation theorems for quasilinear 
elliptic equations with mixed nonlinearites via Picone-type inequality}, 
Toyama Math. J. Vol. 33 (2010), 21-41.

\bibitem{Zhang} Zhang, C.; Sun, S.;
\emph{Sturm-Picone comparison theorem
of second-order linear equationson time scales}, Adv. Diff. Eqs.,
(2009), pp. 12.

\end{thebibliography}

\end{document}
