\documentclass[reqno]{amsart} \usepackage{hyperref} \AtBeginDocument{{\noindent\small {\em Electronic Journal of Differential Equations}, Vol. 2006(2006), No. 25, pp. 1--8.\newline ISSN: 1072-6691. URL: http://ejde.math.txstate.edu or http://ejde.math.unt.edu \newline ftp ejde.math.txstate.edu (login: ftp)} \thanks{\copyright 2006 Texas State University - San Marcos.} \vspace{9mm}} \begin{document} \title[\hfilneg EJDE-2006/25\hfil On second order periodic boundary-value problems] {On second order periodic boundary-value problems with upper and lower solutions in the reversed order} \author[H. Gao, S. Weng, D. Jiang, X. Hou\hfil EJDE-2006/25\hfilneg] {Haiyin Gao, Shiyou Weng, Daqing Jiang, Xuezhang Hou} \address{Haiyin Gao \newline School of Mathematics and Statistics, Northeast Normal University, Changchun 130024, China} \email{gaohaiyinhealthy@yahoo.com.cn} \address{Shiyou Weng \newline Applied Science College, Changchun University, Jilin 130022, China} \email{wengshiyou2001@yahoo.com.cn} \address{Daqing Jiang \newline School of Mathematics and Statistics, Northeast Normal University, Changchun 130024, China} \email{daqingjiang@vip.163.com} \address{Xuezhang Hou \newline Mathematics Department, Towson University, Baltimore, MD 21252, USA} \email{xhou@towson.edu} \date{} \thanks{Submitted November 7, 2005. Published February 28, 2006.} \thanks{Supported by grant 10171010 from the National Natural Science Foundation of China} \subjclass[2000]{34B15, 34B16} \keywords{Periodic boundary value; existence; upper and lower solutions} \begin{abstract} In this paper, we study the differential equation with the periodic boundary value \begin{gather*} u''(t)=f(t, u(t), u'(t)),\quad t\in [0, 2\pi]\\ u(0)=u(2\pi), \quad u'(0)=u'(2\pi). \end{gather*} The existence of solutions to the periodic boundary problem above with appropriate conditions is proved by using an upper and lower solution method. \end{abstract} \maketitle \numberwithin{equation}{section} \newtheorem{theorem}{Theorem}[section] \newtheorem{lemma}[theorem]{Lemma} \newtheorem{remark}[theorem]{Remark} \section{Introduction and Main Results} In this paper, we study the second-order periodic boundary-value problem \begin{equation} \begin{gathered} u''(t)=f(t, u(t), u'(t)),\quad t\in [0, 2\pi]\\ u(0)=u(2\pi), \quad u'(0)=u'(2\pi), \end{gathered} \label{e1.1} \end{equation} where $f(t, u, v)$ is a Caratheodory function. A function $f: [0, 2\pi]\times \mathbb{R}^2\to \mathbb{R}$ is said to be a Carathrodary function if it possess the following three properties: \begin{itemize} \item[(i)] For all $(u, v)\in \mathbb{R}^2$, the mapping $t\to f(t, u, v)$ is measurable on $[0, 2\pi]$. \item[(ii)] For almost all $t\in [0, 2\pi]$, the mapping $(u, v)\to f(t, u, v)$ is continuous on $\mathbb{R}^2$. \item[(iii)] For any given $N>0$, there exists $g_N(t)$, a Lebesgue integrable function defined on $[0, 2\pi]$, such that $$ |f(t, u, v)|\leq g_N(t)\quad \hbox{for a. e. }t\in [0, 2\pi], $$ whenever $|u|, |v|\leq N$. \end{itemize} To develop upper and lower solutions method, we need the concepts of upper and lower solutions. We say that $\beta\in W^{2, 1}[0,2\pi]$ is an upper solution to \eqref{e1.1}, if it satisfies \begin{equation} \begin{gathered} \beta''(t)\leq f(t, \beta(t), \beta'(t)),\quad t\in [0, 2\pi]\\ \beta(0)=\beta(2\pi),\quad \beta'(0)\leq \beta'(2\pi). \end{gathered}\label{e1.2} \end{equation} Similarly, a function $\alpha\in W^{2, 1}[0, 2\pi]$ is said to be a lower solution to \eqref{e1.1}, if it satisfies \begin{equation} \begin{gathered} \alpha''(t)\geq f(t, \alpha(t), \alpha'(t)),\quad t\in [0, 2\pi]\\ \alpha(0)=\alpha(2\pi),\quad \alpha'(0)\geq \alpha'(2\pi). \end{gathered} \label{e1.3} \end{equation} We call a function $u\in W^{2, 1}[0, 2\pi]$ a solution to \eqref{e1.1}, if it is an upper and a lower solution to \eqref{e1.1}. Under the classical assumption that $\alpha(t)\leq \beta(t)$, a number of authors have studied the existence of the methods of lower and upper solutions or the monotone iterative technique \cite{c1,c3,c4,g1,l1,j1,n2,n3,s1,w1}. Only a few have study the case where $\alpha(t), \beta(t)$ satisfy the opposite ordering condition $\beta(t)\leq\alpha(t)$; see \cite{c1,c2,j2,n1,o2,r1,r2,w2} Wang \cite{w2} has investigated a special case of \eqref{e1.1} where $f(t, u,v)=-kv + F(t, u)$ and $F(t, u)$ is increasing with respect to $u$, in the presence of a lower solution $\alpha(t)$ and an upper solution $\beta(t)$ with $\beta(t)\leq \alpha(t)$. Rachunkova \cite{r2} has recently proved that \eqref{e1.1} has at least one solution $u(t)$ under the case $\beta(t)\leq\alpha(t)$. However, the proof of the result in \cite{r2} is not constructive and is not able to guarantee that $u(t)$ satisfies $\beta(t)\leq u(t)\leq \alpha(t)$. Recently, Jiang, Fan and Wan \cite{j2} have studied \eqref{e1.1} by means of a monotone iterative technique in the presence of a lower solution $\alpha(t)$ and an upper solution $\beta(t)$ with $\beta(t)\leq \alpha(t)$. To develop a monotone method, the following hypotheses are needed in \cite{j2}. \begin{itemize} \item[(A1)] For any given $\beta, \alpha\in C[0, 2\pi]$ with $\beta(t)\leq \alpha(t)$ on $[0, 2\pi]$, there exist $00$ and $B>0$ such that $B^2\geq 4A$ and \begin{equation} f(t, u_2, v_2)-f(t, u_1, v_1)\geq -A(u_2-u_1)+B(v_2-v_1) \label{e1.4} \end{equation} for a.e. $t\in [0, 2\pi]$ whenever $\beta(t)\leq u_1\leq u_2\leq\alpha(t)$, $v_1, v_2\in \mathbb{R}$, and $v_1 \leq v_2$. \item[(H1')] For any given $\beta, \alpha\in C[0, 2\pi]$ with $\beta(t)\leq \alpha(t)$ on $[0, 2\pi]$, there exist $A>0$ and $B>0$ such that $B^2\geq 4A$ and \begin{equation} f(t, u_2, v_2)-f(t, u_1, v_1)\geq -A(u_2-u_1)+B(v_1-v_2) \label{e1.4b} \end{equation} for a.e. $t\in [0, 2\pi]$ whenever $\beta(t)\leq u_1\leq u_2\leq\alpha(t)$, $v_1, v_2\in \mathbb{R}$, and $v_1 \geq v_2$. \end{itemize} %\begin{remark} \label{rmk1} \rm We remark that condition (H1') is equivalent to \begin{itemize} \item[(a1)] For any given $\beta, \alpha\in C[0, 2\pi]$ with $\beta(t)\leq \alpha(t)$ on $[0, 2\pi]$, there exists $B>0$ such that $$ f(t, u, v_1)-f(t, u, v_2)\leq -B(v_1-v_2) $$ for a.e. $t\in [0, 2\pi]$ whenever $\beta(t)\leq u\leq\alpha(t)$, $v_1, v_2\in \mathbb{R}$, and $v_1 \geq v_2$. \item[(a2)] There exists $A>0$ such that $B^2\geq 4A$ and $$ f(t, u_2, v)-f(t, u_1, v)\geq - A (u_2-u_1) $$ holds for a.e. $t\in [0, 2\pi]$, whenever $\beta(t)\leq u_1\leq u_2\leq \alpha(t)$, $v\in \mathbb{R}$. \end{itemize} Also we remark that (H1) or (H1') weaker than (A1)-(A2) in \cite{j2}. \smallskip Let $m<0$ and $M<0$ be two real roots to the equation $x^2+Bx+A=0$, then $$ m+M=-B,\quad m M=A. $$ Let $m_0>0$ and $M_0>0$ are two roots to the equation $x^2-Bx+A=0$, then $$ m_0+M_0=B,\quad m_0 M_0=A. $$ Let \begin{equation} A(t):=\alpha'(t)+m\alpha (t),\quad B(t):=\beta'(t)+m\beta(t), \label{e1.5} \end{equation} and \begin{equation} A_0(t):=\alpha'(t)+m_0\alpha (t),\quad B_0(t):=\beta'(t)+m_0\beta(t).\label{e1.5b} \end{equation} The main results of this paper are stated as follows. \begin{theorem} \label{thm1} Suppose that there exists a lower solution $\alpha(t)$ and an upper solution $\beta(t)$ of \eqref{e1.1} such that $\beta(t)\leq\alpha(t)$ on $[0,2\pi]$, and $f(t, u, v)$ is a Caratheodory function satisfying the hypothesis (H1). Then $A(t) \leq B(t)$ on $[0,2\pi]$ and \eqref{e1.1} has a solution $u\in W^{2,1}[0,2\pi]$ such that $$ \beta(t)\leq u(t)\leq \alpha(t),\quad A(t)\leq u'(t)+mu(t)\leq B(t). $$ \end{theorem} \begin{theorem} \label{thm2} Suppose that there exists a lower solution $\alpha(t)$ and an upper solution $\beta(t)$ of \eqref{e1.1} such that $\beta(t)\leq\alpha(t)$ on $[0,2\pi]$, and $f(t, u, v)$ is a Caratheodory function satisfying the hypothesis (H1'). Then $B_0(t) \leq A_0(t)$ on $[0,2\pi]$ and \eqref{e1.1} has a solution $u\in W^{2,1}[0,2\pi]$ such that $$ \beta(t)\leq u(t)\leq \alpha(t),\quad B_0(t)\leq u'(t)+m_0u(t)\leq A_0(t). $$ \end{theorem} \section{Proof of Theorems \ref{thm1} and \ref{thm2}} To prove the validity of upper and lower solutions method, we use the following maximum-minimum principle, see \cite{j2}. \begin{lemma} \label{lem1} Let $y\in W^{1, 1}[0,2\pi]$, and satisfy \begin{gather*} y'(t)+ Ly(t)\geq 0 \quad \hbox{for a. e. } t\in [0, 2\pi],\\ y(0)\geq y(2\pi), \end{gather*} where $|L|> 0$. Then $Ly(t)\geq 0$ on $[0, 2\pi]$, i.e., when $L>0$ the minimum of $y(t)$ is nonnegative; when $L<0$ the maximum of $y(t)$ is nonpositive. \end{lemma} \begin{lemma} \label{lem2} Suppose that there exists a lower solution $\alpha(t)$ and an upper solution $\beta(t)$ of \eqref{e1.1} such that $\beta(t)\leq\alpha(t)$ on $[0,2\pi]$, and $f(t, u, v)$ is a Caratheodory function satisfying the hypothesis (H1). Then $A(t)\leq B(t)$ on $[0,2\pi]$. \end{lemma} \begin{proof} It follows from \eqref{e1.2} and \eqref{e1.3} that \begin{gather*} A'(t)+MA(t)\geq f(t,\alpha (t),A(t)-m\alpha(t))+(m+M)A(t)-m^2\alpha (t),\quad t\in[0, 2\pi]\\ A(0)\geq A(2\pi), \end{gather*} and \begin{gather*} B'(t)+MB(t)\leq f(t,\beta (t),B(t)-m\beta(t))+(m+M)B(t)-m^2\beta (t),\quad t\in[0, 2\pi]\\ B(0)\leq B(2\pi). \end{gather*} Let $y(t)=A(t)-B(t)$, then $y(0)\geq y(2\pi)$. Assume that $y(t)>0$ for some $t\in[0,2\pi]$. Indeed, if $y(t)>0$ on $[0,2\pi]$, then by (H1) we have \begin{align*} y'(t)+My(t)&\geq f(t,\alpha (t),A(t)-m\alpha(t))-f(t,\beta (t),B(t)-m\beta(t))\\ &\quad +(m+M)y(t)-m^2(\alpha(t)-\beta(t))\\ &\geq -(A+Bm+m^2)(\alpha(t)-\beta(t))+(B+m+M)y(t)\\ &= 0,\quad t\in[0, 2\pi], \end{align*} then by Lemma \ref{lem1}, we have $y(t)\leq 0$ on $[0,2\pi]$, which is a contradiction. If $y(0)\leq 0$ (then $y(2\pi)\leq y(0)\leq 0$), and hence there exists $s$ $\in (0,2\pi)$ with $y(s)>0$, then there would be $0\leq a0$ in $(a,b)$ with $y(a)=y(b)=0$. By \eqref{e1.2} and \eqref{e1.3}, we have $$ y'(t)+My(t)\geq 0,\quad t\in [a,b],\quad y(a)=y(b)=0. $$ This leads to $y'(t)\geq -My(t)>0$ on $[a,b]$, which is again a contradiction. If $y(0)>0$, then there exists $a \in (0,2\pi)$ such that $y(t)>0$ on $[0,a)$ with $y(a)=0$. So we have $y'(t)+My(t)\geq 0$ on $[0,a)$, hence $y'(t)>0$ in $[0, a)$, which implies that $y(0)0$ on $[a,b]$, which is again a contradiction. If $y(2\pi)<0$, then there exists $a \in (0,2\pi)$ such that $y(t)<0$ on $(a, 2\pi]$ with $y(a)=0$. So we have $y'(t)+M_0y(t)\geq 0$ on $(a, 2\pi]$, hence $y'(t)>0$ in $(a, 2\pi]$, which implies that $y(2\pi)>y(a)=0$, this is also a contradiction. The proof of Lemma \ref{lem3} is complete. \end{proof} In the following arguments, we only give the proof of Theorem \ref{thm1}, since the proof of Theorem \ref{thm2} can be treated in a similar way. Let $$ p(t,x)=\begin{cases} A(t), & xB(t). \end{cases} $$ It is interesting to give an introduction to Lemma \ref{lem4} and a reference where it can be found. \begin{lemma} \label{lem4} If $m>0$, then for any $q(t)\in L^1[0,2\pi]$, the problem \begin{gather*} u'(t)+mu(t)=q(t),\quad \mbox{for a.e. }t\in[0,2\pi]\\ u(0)=u(2\pi), \end{gather*} has a unique solution $u\in W^{1,1}[0,2\pi]$, and $$ u(t)=L^{-1}q(t)=\int_0^{2\pi}G_m(t,s)q(s)ds, $$ where $$ G_m(t, s): =\begin{cases} \frac{e^{m(2\pi+s-t)}}{e^{2m\pi}-1}, & 0\leq s \leq t \leq 2\pi,\\ \frac{e^{m(s-t)}}{e^{2m\pi}-1}, & 0\leq t \leq s \leq 2\pi. \end{cases} $$ \end{lemma} By Lemma \ref{lem1}, we have $$ \alpha (t)=L^{-1}A(t),\quad \beta (t)=L^{-1}B(t),\quad \beta(t)\leq L^{-1}p(t,x)\leq \alpha(t). $$ Now we consider the modified problem \begin{equation} \begin{gathered} x'(t)+Mx(t)=f\big(t,L^{-1}p(t,x(t)),(I-mL^{-1})p(t,x(t))\big)\\ +(m+M)p(t,x(t))-m^2L^{-1}p(t,x(t)), x(0)=x(2\pi). \end{gathered} \label{e2.1} \end{equation} For each $x\in C[0,2\pi]$, we define the mapping $\Phi:C[0,2\pi]\to C[0,2\pi]$, \begin{equation} (\Phi x)(t)=\int_0^{2\pi}G_M(t,s)(Fx)(s)ds,\label{e2.2} \end{equation} where \begin{align*} (Fx)(t)&:=f\big(t,L^{-1}p(t,x(t)),(I-mL^{-1})p(t,x(t))\Big)\\ &\quad +(m+M)p(t,x(t))-m^2L^{-1}p(t,x(t)). \end{align*} Since $p(t,x(t))$ and $L^{-1}p(t,x(t))$ are bounded and $f(t,u,v)$ is a Caratheodary function, there exists $g(t)$, a Lebesgue integrable function defined on $[0,2\pi]$ such that $$ |(Fx)(t)|\leq g(t)\quad \hbox{for a. e. }t\in [0, 2\pi]. $$ Thus $(\Phi x)(t)$ is also bounded. We can easily prove that $\Phi:C[0,2\pi]\to C[0,2\pi]$ is completely continuous. Then Leray-Schauder fixed point Theorem assures that $\Phi$ has a fixed point $x\in C[0,2\pi]$ and \begin{equation} x(t)=\int_0^{2\pi}G_M(t,s)(Fx)(s)ds,\label{e2.3} \end{equation} thus the modified problem \eqref{e2.1} has one solution $x\in W^{1,1}[0,2\pi]$. \begin{lemma} \label{lem5} Suppose that (H1) holds. Assume that $\alpha (t)$ and $\beta (t)$ are lower and upper solutions to $\eqref{e1.1}$ and $\beta (t) \leq \alpha (t)$ on $[0,2\pi]$. Let $x\in W^{1,1}[0,2\pi]$ be a solution to \eqref{e2.1}, then $A(t)\leq x(t)\leq B(t)$ on $[0,2\pi]$. \end{lemma} \begin{remark} \label{rmk2} \rm Lemma \ref{lem4} implies $u(t)=L^{-1}x(t)=\int_0^{2\pi}G_m(t,s)x(s)ds$ is a solution to \eqref{e1.1}, since $u'(t)+mu(t)=x(t),u(0)=u(2\pi)$ and $A(t)\leq x(t)\leq B(t)$. \end{remark} \begin{proof}[Proof of Lemma \ref{lem5}] Since $\alpha (t)=L^{-1}A(t),\beta (t)=L^{-1}B(t)$, \begin{gather*} B'(t)+MB(t)\leq f(t,L^{-1}B(t),(I-mL^{-1})B(t))-m^2L^{-1}B(t)+(m+M)B(t),\\ B(0)\leq B(2\pi) \end {gather*} and \begin{gather*} A'(t)+MA(t)\geq f(t,L^{-1}A(t),(I-mL^{-1})A(t))-m^2L^{-1}A(t)+(m+M)A(t),\\ A(0)\geq A(2\pi). \end {gather*} We shall prove only that $x(t)\leq B(t)$ on $[0, 2\pi]$, because $A(t)\leq x(t)$ can be proved by a similar manner. Let $y(t)=x(t)-B(t)$, then $$y(0)\geq y(2\pi).$$ Assume that $y(t)>0$ for some $t\in[0,2\pi]$. Indeed, if $y(t)>0$ on $[0,2\pi]$, we have \begin{align*} x'(t)+Mx(t)&= f(t,L^{-1}B(t),(I-mL^{-1})B(t))-m^2L^{-1}B(t)+(m+M)B(t)\\ &\geq B'(t)+MB(t), \end{align*} i.e., $y'(t)+My(t)\geq 0$ on $[0,2\pi]$. Lemma \ref{lem1} implies $y(t)\leq 0$ on $[0,2\pi]$, which is a contradiction. Therefore, there would be a point $s\in [0,2\pi]$ with $y(s)\leq 0$. If $y(0)\leq 0$ (then $y(2\pi)\leq y(0)\leq 0$), and hence there exist $0\leq a0$ in $(a,b)$ with $y(a)=y(b)=0$. Then $p(t,x(t))=B(t)$ on $[a,b]$ and \begin{align*} &y'(t)+M y(t)\\ &\geq f(t,L^{-1}p(t,x(t)),B(t)-mL^{-1}p(t,x(t)))+(m+M)B(t)-m^2L^{-1}p(t,x(t))\\ &\quad - [f(t,L^{-1}B(t),B(t)-mL^{-1}B(t))+(m+M)B(t)-m^2L^{-1}B(t)]\\ &\geq (-A-Bm-m^2)(L^{-1}p(t,x(t))-L^{-1}B(t))\\ &= 0,\quad t\in (a,b). \end{align*} This leads to $y'(t)\geq -My(t)>0$ on $(a,b)$, which is again a contradiction. If $y(0)>0$, there exists $a\in(0,2\pi)$ such that $y(t)>0$ on $[0, a)$ with $y(a)=0$. So we have $y'(t)+My(t)\geq 0$, hence $y'(t)>0$ in $[0, a)$, which implies that $y(0)0$ or $k<0$. We say that $\beta\in W^{2, 1}[0, 2\pi]$ is an upper solution to the problem \eqref{e3.1}, if it satisfies \begin{equation} \begin{gathered} \beta''(t) + k \beta'(t) \leq F(t, \beta(t)),\quad t\in [0, 2\pi]\\ \beta(0)=\beta(2\pi), \beta'(0)\leq\beta'(2\pi). \end{gathered}\label{e3.2} \end{equation} Similarly, a function $\alpha\in W^{2, 1}[0, 2\pi]$ is said to be a lower solution to \eqref{e3.1}, if it satisfies \begin{equation} \begin{gathered} \alpha''(t) + k \alpha'(t) \geq F(t, \alpha(t)),\quad t\in [0, 2\pi]\\ \alpha(0)=\alpha(2\pi), \alpha'(0)\geq\alpha'(2\pi). \end{gathered} \label{e3.3} \end{equation} To develop the upper and lower solutions method, we also need the following hypothesis: \begin{itemize} \item[(H)] For any given $\beta, \alpha\in C[0, 2\pi]$ with $\beta(t)\leq \alpha(t)$ on $[0, 2\pi]$, the inequality $$ F(t, u_2)-F(t, u_1) \geq - \frac{k^2}{4} (u_2-u_1) $$ holds for a.e. $t\in [0, 2\pi]$, whenever $\beta(t)\leq u_1\leq u_2\leq \alpha(t)$. \end{itemize} Let $ A=k^2/4$, $B=|k|$, then (H1) holds when $k<0$, and (H1') holds when $k>0$. Hence the conclusions of Theorem \ref{thm1} hold when $k<0$, thus $\alpha'(t)+\frac{k}{2}\alpha(t)\leq \beta'(t)+\frac{k}{2}\beta(t)$ and problem \eqref{e3.1} has one solution $u\in W^{2,1}[0,2\pi]$ such that $$ \beta(t)\leq u(t)\leq \alpha (t), \alpha'(t) +\frac{k}{2}\alpha(t)\leq u'(t)+\frac{k}{2}u(t) \leq \beta'(t)+\frac{k}{2}\beta(t). $$ The conclusions of Theorem \ref{thm2} hold when $k>0$, thus $\alpha'(t)+\frac{k}{2}\alpha(t)\geq \beta'(t)+\frac{k}{2}\beta(t)$ and problem \eqref{e3.1} has one solution $u\in W^{2,1}[0,2\pi]$ such that $$ \beta(t)\leq u(t)\leq \alpha (t),~\beta'(t)+\frac{k}{2}\beta(t) \leq u'(t)+\frac{k}{2}u(t) \leq \alpha'(t)+\frac{k}{2}\alpha(t). $$ In \cite{j2,w2}, the authors obtained one solution $u\in W^{2,1}[0,2\pi]$ of \eqref{e3.1} such that $\beta(t)\leq u(t)\leq \alpha (t)$. Here we have improved the results of \cite{j2,w2}. \begin{thebibliography}{00} \bibitem{c1} A. Cabada. \emph{The method of lower and upper solutions for second,third, fourth, and higher order boundary value problems}, J. Math. Anal. Appl., {\bf 185}(1994),302-320. \bibitem{c2} A. Cabada, \emph{The Monotone Method for Boundary Value Problems, doctoral thesis}, Universidad de Santiago de Compostela, 1992. [in Spanish] \bibitem{c3} A. Cabada and J. J. Nieto, \emph{A generalization of the monotone iterative technique for linear second order periodic boundary value problems}, J. Math. Anal., {\bf 151} (1990), 181-189. \bibitem{c4} A. Cabada and J. J. Nieto, \emph{Extemal solutions of second order nonlinear periodic boundary value problems}, Appl. Math. Comput., {\bf 40}(1990), 135-145. \bibitem{g1} W. Gao and J. Wang, \emph{On a nonlinear second order periodic boundary value problem with Caratheodory functions}, Ann. Polon. Math., {\bf 62}(1995), 283-291. \bibitem{j1} D. Jiang and J. Wang, \emph{A generalized periodic boundary value problem for the one-dimensional p-Laplacian}, Ann. Polon. Math., {\bf 65}(1997), 265-270. \bibitem{j2} D. Jiang, M. Fan and A. Wan, \emph{A monotone method for constructing extremal solutions to second-order periodic boundary value problems}, J. Comput. Appl. Math., {\bf 136} (2001), 189-197. \bibitem{l1} G. S. Ladde, V. Lakshmikantham, and A. S. Vatsala, \emph{Monotone Iterative Techniques for Nonlinear Differential Equations}, Pitman Boston, 1985. \bibitem{n1} J. J. Nieto, \emph{Nonlinear second-order periodic boundary value problems}, J. Math. Anal. Appl. {\bf 30}(1988), 22-29. \bibitem{n2} J. J. Nieto, \emph{Nonlinear second-order periodic boundary value problems with Caratheodory functions}, Appl. Anal., {\bf 34}(1989), 111-128. \bibitem{n3} J. J. Nieto and A. Cabada, \emph{A generalized upper and lower solution method for nonlinear second order ordinary differential equations}, J. Appl. Math. Stochastic Anal., {\bf 5}(1992), 157-166. \bibitem{o1} P. Omari, \emph{A monotone method for constructing extremal soltions of second order scalar boundary value problems}, Appl. Math. Comput., {\bf 18}(1986), 257-275. \bibitem{o2} P. Omari, \emph{Nonordered lower and upper solutions and solvability of the periodic problem for the Lienard and the Rayleigh equations}, Rend. Istit. Mat. Univ. Trieste, {\bf 20}(1991), 54-64. \bibitem{r1} B. Rudolf and Z. Kubacek, \emph{Remarks on J. J. Niteo's paper: Nolinear second-order periodic boundary value problems}, J. Math. Anal. Appl., {\bf 146}(1990), 203-206. \bibitem{r2} I. Rachunkova, \emph{Upper and lower solutions satisfying the inverse inequality}, Ann. Polon. Math., {\bf 65}(1997), 235-244. \bibitem{s1} V. Seda, J. J. Nieto, and M. Gera, \emph{Periodic boundary value problems for nonlinear higher order ordinary differential equations}, Appl. Math. Comput., {\bf 48}(1992), 71-82. \bibitem{w1} M. X. Wang, A. Cabada, and J. J. Nieto, \emph{Monotone method for nonlinear second order periodic boundary value problemws with Caratheodory functions}, Ann. Polon. Math., {\bf 58}(1993), 221-235. \bibitem{w2} C. Wang, \emph{Generalized upper and lower solution method for the forced duffing equation}, Proc. Amer. Math. Soc., {\bf 125}(1997), 397-406. \end{thebibliography} \end{document}