\documentclass[reqno]{amsart} \usepackage{hyperref} \AtBeginDocument{{\noindent\small \emph{Electronic Journal of Differential Equations}, Vol. 2009(2009), No. 99, pp. 1--9.\newline ISSN: 1072-6691. URL: http://ejde.math.txstate.edu or http://ejde.math.unt.edu \newline ftp ejde.math.txstate.edu} \thanks{\copyright 2009 Texas State University - San Marcos.} \vspace{9mm}} \begin{document} \title[\hfilneg EJDE-2009/99\hfil Existence and exponential stability] {Existence and exponential stability for anti-periodic solutions for shunting inhibitory cellular neural networks with continuously distributed delays} \author[Y. Wu, Z. Zhou\hfil EJDE-2009/99\hfilneg] {Yuanheng Wu, Zhan Zhou} % in alphabetical order \address{Yuanheng Wu \newline College of Mathematics and Information Sciences, Guangzhou University\\ Guangzhou 510006, China} \email{wyhcd2006@yahoo.com.cn} \address{Zhan Zhou \newline College of Mathematics and Information Sciences, Guangzhou University\\ Guangzhou 510006, China} \email{zzhou@gzhu.edu.cn} \thanks{Submitted July 25, 2009. Published August 19, 2009.} \subjclass[2000]{34C25, 34K13, 34K25} \keywords{Shunting inhibitory cellular neural networks; anti-periodic solution; \hfill\break\indent exponential stability; continuously distributed delays} \begin{abstract} This article concerns anti-periodic solutions for shunting inhibitory cellular neural networks (SICNNs), with continuously distributed delays, arising from the description of the neurons state in delayed neural networks. Without assuming global Lipschitz conditions of activation functions, we obtain existence and local exponential stability of anti-periodic solutions. \end{abstract} \maketitle \numberwithin{equation}{section} \newtheorem{theorem}{Theorem}[section] \newtheorem{lemma}[theorem]{Lemma} \newtheorem{remark}[theorem]{Remark} \section{Introduction} We consider shunting inhibitory cellular neural networks (SICNNs) with continuously distributed delays described by \begin{equation} \begin{aligned} x'_{ij}(t)&=-a_{ij}(t)x_{ij}(t)\\ &\quad -\sum_{C_{kl}\in N_{r}(i,j)}C_{ij}^{kl}(t) \int_{0}^{\infty}K_{ij}(u) f(x_{kl}(t-u))dux_{ij}(t)+L_{ij}(t), \end{aligned} \label{e1.1} \end{equation} where $i=1, 2,\dots, m$, $j=1,2,\dots,n$, $C_{ij}$ denotes the cell at the $(i,j)$ position of the lattice, the $r$-neighborhood $N_{r}(i,j)$ of $C_{ij}$ is given by $$ N_{r}(i,j)=\{C_{kl}:\max(|k-i|,|l-j|)\leq r,1\leq k\leq m,1\leq l\leq n\}. $$ Here $x_{ij}$ represents the activity of the cell $C_{ij}$, $L_{ij}(t)$ is the external input to $C_{ij}$, the constant $a_{ij}(t)>0$ represents the passive decay rate of the cell activity, $C_{ij}^{kl}(t)\geq0$ is the connection or coupling strength of postsynaptic activity of the cell transmitted to the cell $C_{ij}$, and the activity function $f(\cdot)$ is a continuous function representing the output or firing rate of the cell $C_{kl}$. It is well known that studies on SICNNs not only involve a discussion of stability properties, but also involve many dynamic behaviors such as periodic oscillatory behavior and almost periodic oscillatory properties. Therefore, considerable effort has been devoted to study dynamic behaviors, in particular, the existence and stability of periodic and almost periodic solutions of SICNNs in the literature (see, e.g., \cite{c1,c2,c3,h2,l1,l2} and the references therein). Recently, in \cite{p1,s1} the authors studied the existence and stability of anti-periodic solutions for SICNNs with the following assumption: \begin{itemize} \item[(T0)] The function $f$ is global Lipschitz continuous; that is, there exists a constant $\mu>0$ such that for all $x,y \in \mathbb{R}$, $$ |f(x)-f(y)|\leq \mu |x-y|. $$ \end{itemize} To the best of our knowledge, very few authors have considered problems of anti-periodic solutions of SICNNs \eqref{e1.1} without Assumption (T0). Moreover, it is well known that the existence and stability of anti-periodic solutions play a key role in characterizing the behavior of nonlinear differential equations; see \cite{a1,a2,c4,w1}. Since SICNNs can be analog voltage transmission, and voltage transmission process often a anti-periodic process. Thus, it is worth while to continue to investigate the existence and stability of anti-periodic solutions of SICNNs \eqref{e1.1} without Assumption (T0). The purpose of this article is to present sufficient conditions for the existence and local exponential stability of anti-periodic solutions of \eqref{e1.1}. Moreover, we do not assume (T0). An example is provided to illustrate our results. Let $u(t):\mathbb{R}\to \mathbb{R}$ be continuous in $t$. $u(t)$ is said to be $T$-anti-periodic on $\mathbb{R}$ if, $$ u(t+T)=-u(t) \quad \text{for all } t\in \mathbb{R}. $$ In this article, for $i=1, 2, \dots, m$, $j=1, 2, \dots, n$, it is assumed that $a_{ij}, C^{kl}_{ij}:\mathbb{R}\to [0,+\infty) $, $K_{ij} :[0, +\infty)\to \mathbb{R} $ and $L_{ij}:\mathbb{R}\to \mathbb{R} $ are continuous functions, and \begin{equation} \begin{gathered} a_{ij}(t+T)=a_{ij}(t), \quad C^{kl}_{ij}(t+T)=C^{kl}_{ij}(t), \\ f(-u)=f( u), \quad L_{ij}(t+T) =-L_{ij}(t ) , \end{gathered} \label{e1.2} \end{equation} for all $t,u \in \mathbb{R}$. Let $$ \{x_{ij}(t)\}=(x_{11}(t), \dots,x_{1n}(t),\dots,x_{i1}(t),\dots,x_{in}(t),\dots, x_{m1}(t),\dots,x_{mn}(t)) $$ be an element in $\mathbb{R}^{m\times n}$. For $x(t)=\{x_{ij}(t) \}$ in $\mathbb{R}^{m\times n}$, define the norm $\|x(t)\|=\max_{(i,j)}\{ |x_{ij}(t)|\}$. We shall use the following conditions \begin{itemize} \item[(H1)] For $i\in\{1, 2, \dots, m \}$, $j\in\{1, 2, \dots, n \}$, the delay kernels $K_{ij}:[0, \infty)\to \mathbb{R}$ are continuous and integrable; \item[(H2)] there exists a function $L:\mathbb{R}^{+}\to\mathbb{R}^{+}$ such that for each $r>0$, \begin{equation} |f(u )-f(v )| \leq L(r)|u -v |,\quad |u|,|v|\leq r. \label{e1.3} \end{equation} \item[(H3)] there exists a constant $r_{0}>0$ such that $$ D[|f(0)|r_{0}+L(r_{0})r^{2}_{0}]+L^{+}\leq r_{0} , $$ where \begin{gather*} D=\max_{(i,j)}\big\{\frac{\sum_{C_{kl}\in{N_{r}(i,j)}} \bar{C}_{ij}^{kl} \int_{0}^{\infty}|K_{ij}(u)|du} {\underline{a}_{ij}}\big\}>0,\quad L^{+}=\max_{(i,j)}\frac{\overline{L}_{ij}}{\underline{a}_{ij}}, \\ \overline{L}_{ij}=\sup_{t\in{\mathbb{R}}}|L_{ij}(t)|, \quad \bar{C}_{ij}^{kl}=\sup_{t\in \mathbb{R}}C_{ij}^{kl}(t), \quad \underline{a}_{ij}=\inf_{t\in \mathbb{R}}a_{ij}(t)>0. \end{gather*} \end{itemize} The initial conditions associated with system \eqref{e1.1} are \begin{equation} x_{ij}(s)=\varphi_{ij}(s),\quad s\in (-\infty, 0], \; i=1,2,\dots,m,\; j=1,2,\dots,n, \label{e1.4} \end{equation} where $ \varphi_{ij}(\cdot)$ denotes real-valued bounded continuous function defined on $(-\infty, 0]$. The remaining parts of this paper are organized as follows. In Section 2, sufficient conditions are derived for the boundedness of solution of \eqref{e1.1}. In Section 3, we present sufficient conditions for the existence and local exponential stability of anti-periodic solution of \eqref{e1.1}. In Section 4, an illustrative example is given to show the proposed theory and method. \section{Preliminary Results} The following lemmas will be used to prove our main results in Section 3. \begin{lemma} \label{lem2.1} Assume {\rm (H1)--(H3)}. Suppose that $\widetilde{x}(t)= \{\widetilde{x}_{ij}(t)\} $ is a solution of \eqref{e1.1} with initial conditions \begin{equation} \widetilde{x}_{ij}(s)=\widetilde{\varphi}_{ij}(s), \quad |\widetilde{\varphi}_{ij}(s)|< r_{0} , \quad s\in (-\infty, 0],\; ij=11, 12, \dots, mn \label{e2.1} \end{equation} Then \begin{equation} |\widetilde{x}_{ij}(t)| 0$ such that \begin{equation} \widetilde{x}_{ij}(\rho) =r_{0}, \quad \widetilde{x}_{\bar{ij}}(t) 0$ such that $$ \int_{0}^{\infty}|K_{ij}(s)|e^{\lambda_{0}s}ds<+\infty . $$ \end{itemize} Then there is a positive constant $\lambda$ such that for every solution $x(t)=\{x_{ij}(t)\}$ of \eqref{e1.1} with any initial value $\sup_{t\in (-\infty, 0]}\|\varphi (t)\|\leq r_{1}$, $$ \|x(t)-x^{*}(t)\|\leq M e^{-\lambda t}, \quad \forall t>0, $$ where $ M =\sup_{ -\infty\leq t\leq 0}\|\varphi (t)-x^{*}(t)\|$. \end{lemma} \begin{proof} In view \eqref{e2.4} and Lemma \ref{lem2.1}, we have \begin{equation} |x^{*}_{ij}(t)| < r_{0}, \quad\text{for all } t\in \mathbb{R},\; ij=11, 12, \dots, mn. \label{e2.5} \end{equation} Set \begin{align*} \Gamma_{ij}(\alpha) &=\alpha-\underline{a}_{ij} +\sum_{C_{kl}\in{N_{r}(i,j)}}\bar{C}_{ij}^{kl}(|f(0)| +L(r_{0})r_{0}\int_{0}^{\infty}|K_{ij}(s)| ds\\ &\quad +L(r_{1})r_{1}\int_{0}^{\infty}|K_{ij}(s)|e^{\alpha s}ds) \end{align*} where $i=1,2,\dots,m$ and $ j=1,2,\dots,n$. It is not difficult to prove that $\Gamma_{ij}(ij=11, 12, \dots, mn)$ are continuous functions on $[0,\lambda_{0}]$. Moreover, by (H4) and (H5), we have $$ \Gamma_{ij}(0)=-\underline{a}_{ij} +\sum_{C_{kl}\in{N_{r}(i,j)}}\bar{C}_{ij}^{kl}(|f(0)| +L(r_{0})r_{0}+L(r_{1})r_{1})\int_{0}^{\infty}|K_{ij}(s)| ds<0 $$ Thus, there exists a sufficiently small constant $\lambda>0$ such that \begin{equation} \Gamma_{ij}(\lambda)<0,\quad ij=11, 12, \dots, mn.\label{e2.6} \end{equation} Take $\varepsilon>0$, and set $$ z_{ij}(t)=|x_{ij}(t)-x^{*}_{ij}(t)|e^{\lambda t},\quad ij=11, 12, \dots, mn. $$ It follows that $$ z_{ij}(t)\leq M0,\; ij=11, 12, \dots, mn.\label{e2.7} \end{equation} If this is not true, then there exists $i_{0}\in\{1,2,\dots,m\}$ and $j_{0}\in\{1,2,\dots,n\}$ such that \begin{equation} \{t>0|z_{i_{0}j_{0}}(t)> M+\varepsilon\}\neq \phi\label{e2.8} \end{equation} Let $$ t_{ij}= \begin{cases} \inf\{t>0|z_{ij}(t)> M+\varepsilon\}, &\text{if }\{ t>0: z_{ij}(t)> M+\varepsilon\}\neq \emptyset\\ +\infty,& \text{otherwise}. \end{cases} $$ Then $t_{ij}>0$ and \begin{equation} z_{ij}(t)\leq M+\varepsilon,\quad \forall t\in(-\infty,t_{ij}],\; ij=11, 12, \dots, mn.\label{e2.9} \end{equation} We denote $t_{pq}=\min_{(i,j)}t_{ij}$, where $p\in\{1,2,\dots,m\}$ and $q\in\{1,2,\dots,n\}$. In view of \eqref{e2.8}, we have $00:z_{pq}(t)> M+\varepsilon\}$, we obtain \begin{equation} z_{pq}(t_{pq})=M+\varepsilon,\quad D^{+}z_{pq}(t_{pq})\geq 0.\label{e2.11} \end{equation} Since $x(t)$ and $x^{*}(t)$ are solutions of \eqref{e1.1}, we have \begin{align*} 0 &\leq D^{+}z_{pq}(t_{pq})\\ &= D^{+}[|x_{pq}(t)-x^{*}_{pq}(t)|e^{\lambda t}]|_{t=t_{pq}}\\ &\leq |x_{pq}(t_{pq})-x^{*}_{pq}(t_{pq})|\lambda e^{\lambda t_{pq}}-\underline{a}_{pq}|x_{pq}(t_{pq})-x^{*}_{pq}(t_{pq})|e^{\lambda t_{pq}}\\ &\quad +\sum_{C_{kl}\in{N_{r}(p,q)}}\bar{C}_{pq}^{kl}|\int_{0}^{\infty}K_{ij}(u) f( x_{kl}(t_{pq}-u))dux_{pq}(t_{pq})\\ &\quad -\int_{0}^{\infty}K_{ij}(u) f(x^{*}_{kl}(t_{pq}-u))du x^{*}_{pq}(t_{pq})|e^{\lambda t_{pq}} \\ &= |x_{pq}(t_{pq})-x^{*}_{pq}(t_{pq})|\lambda e^{\lambda t_{pq}}-\underline{a}_{pq}|x_{pq}(t_{pq})-x^{*}_{pq}(t_{pq})|e^{\lambda t_{pq}}\\ &\quad +\sum_{C_{kl}\in{N_{r}(p,q)}}\bar{C}_{pq}^{kl}|\int_{0}^{\infty}K_{ij}(u) f(x^{*}_{kl}(t_{pq}-u) du (x_{pq}(t_{pq})-x^{*}_{pq}(t_{pq})) \\ &\quad +\int_{0}^{\infty}K_{ij}(u)(f(x _{kl}(t_{pq}-u))- f(x^{*}_{kl}(t_{pq}-u)))dux _{pq}(t_{pq}) |e^{\lambda t_{pq}}\\ &\leq (\lambda-\underline{a}_{pq})z_{pq}(t_{pq}) +\sum_{C_{kl}\in{N_{r}(p,q)}}\bar{C}_{pq}^{kl} \int_{0}^{\infty}|K_{ij}(u)|| f(x^{*}_{kl}(t_{pq}-u)|du\\ &\quad\times |x_{pq}(t_{pq})-x^{*}_{pq}(t_{pq})|e^{\lambda t_{pq}} \\ &\quad +\sum_{C_{kl}\in{N_{r}(p,q)}}\bar{C}_{pq}^{kl} \int_{0}^{\infty}|K_{ij}(u)||f(x _{kl}(t_{pq}-u))- f(x^{*}_{kl}(t_{pq}-u))|du\\ &\quad \times |x_{pq}(t_{pq})|e^{\lambda t_{pq}}. % \eqref{e2.12} \end{align*} Now, combining the above inequality, \eqref{e2.10}, \eqref{e2.11}, (H2) and (H3), we deduce \begin{align*} 0 &\leq D^+z_{pq }(t_{pq })\\ &\leq (\lambda-\underline{a}_{pq})(M+\varepsilon) +\sum_{C_{kl}\in{N_{r}(p,q)}}\bar{C}_{pq}^{kl} (|f(0)|+L(r_{0})r_{0})\int_{0}^{\infty}|K_{ij}(u)|du\cdot z_{pq}(t_{pq}) \\ & \quad + \sum_{C_{kl}\in{N_{r}(p,q)}}\bar{C}_{pq}^{kl} L(r_{1})\int_{0}^{\infty}|K_{ij}(u)|| x _{kl}(t_{pq}-u) - x^{*}_{kl}(t_{pq}-u)|\\ &\quad\times e^{\lambda (t_{pq}-u)}e^{\lambda u}du\cdot r_{1}\\ &\leq (\lambda-\underline{a}_{pq})(M+\varepsilon) +\sum_{C_{kl}\in{N_{r}(p,q)}}\bar{C}_{pq}^{kl} (|f(0)|+L(r_{0})r_{0})\cdot (M+\varepsilon) \\ & \quad + \sum_{C_{kl}\in{N_{r}(p,q)}}\bar{C}_{pq}^{kl} L(r_{1})r_{1}\int_{0}^{\infty}|K_{ij}(u)| e^{\lambda u}du (M+\varepsilon)\\ &\leq (\lambda-\underline{a}_{pq})(M+\varepsilon)+\sum_{C_{kl}\in{N_{r}(p,q)}}\bar{C}_{pq}^{kl} (|f(0)|\\ &\quad +L(r_{0})r_{0})\int_{0}^{\infty}|K_{ij}(u)| du\cdot (M+\varepsilon) \\ & \quad + \sum_{C_{kl}\in{N_{r}(p,q)}}\bar{C}_{pq}^{kl} L(r_{1})r_{1}\int_{0}^{\infty}|K_{ij}(u)| e^{\lambda u}du\cdot (M+\varepsilon)\,. \end{align*} It follow that \begin{align*} &\lambda-\underline{a}_{pq}+\sum_{C_{kl}\in{N_{r}(p,q)}} \bar{C}_{pq}^{kl}\cdot (|f(0)|\\ &+L(r_{0})r_{0}\int_{0}^{\infty}|K_{ij}(u)| du +L(r_{1})r_{1}\int_{0}^{\infty}|K_{ij}(u)| e^{\lambda u}du) \geq 0 \end{align*} which contradicts with \eqref{e2.6}. Hence, \eqref{e2.7} holds; i.e. $$ |x_{ij}(t)-x^{*}_{ij}(t)|e^{\lambda t}=z_{ij}(t)\leq M+\varepsilon, \quad \forall t>0,\; i=1,2,\dots,m, \; j=1,2,\dots,n. $$ Therefore, $$ \|x(t)-x^{*}(t)\|=\max_{(i,j)}|x_{ij}(t)-x^{*}_{ij}(t)|\leq (M+\varepsilon)e^{-\lambda t},\quad \forall t>0. $$ Letting $\varepsilon\to0^+$, we obtain $$ \|x(t)-x^{*}(t)\|\leq M e^{-\lambda t}, \quad \forall t>0. $$ The proof is complete. \end{proof} \begin{remark} \label{rmk2.2} \rm Let $x^{*}(t)=(x^{*}_{11}(t), x^{*}_{12}(t),\dots,x^{*}_{mn}(t))^{T} $ is the $T$-anti-periodic solution of \eqref{e1.1}. It follows from Lemma \ref{lem2.2} that $x^{*}(t)$ is globally exponentially stable. \end{remark} \section{ Main Results} \begin{theorem} \label{thm3.1} Assume {\rm (H1)-(H5)}. Then \eqref{e1.1} has at least one $T$-anti-periodic solution $x^{*}(t)$. Moreover, $x^{*}(t)$ is locally exponentially stable. \end{theorem} \begin{proof}. Let $v(t)= (v_{11}(t), v_{12}(t),\dots, v_{mn}(t))^{T} $ be a solution of \eqref{e1.1} with initial conditions \begin{equation} v_{ij}(s)=\varphi^{v}_{i}(s), \quad |\varphi^{v}_{ij}(s)|0, \; ij=11, 12, \dots, mn. \end{aligned} \label{e3.4} \end{equation} Thus, for any natural number $p $, we obtain \begin{equation} \begin{aligned} &(-1)^{p+1} v_{ij} (t + (p+1)T)\\ &= v_{ij} (t ) +\sum_{k=0}^{p}[(-1)^{k+1} v _{ij}(t + (k+1)T)-(-1)^{k} v_{ij} (t + kT)]. \end{aligned} \label{e3.5} \end{equation} Then \begin{equation} \begin{aligned} & |(-1)^{p+1} v_{ij} (t + (p+1)T)|\\ & \leq | v_{ij} (t )| +\sum_{k=0}^{p}| (-1)^{k+1} v _{ij}(t + (k+1)T)-(-1)^{k} v_{ij} (t + kT)|, \end{aligned}\label{e3.6} \end{equation} where $ij =11,12,\dots,mn$. In view of \eqref{e3.4}, we can choose a sufficiently large constant $N>0$ and a positive constant $\alpha $ such that \begin{equation} |(-1)^{k+1} v_{ij} (t + (k+1)T)-(-1)^{k } v_{ij} (t + kT)| \leq \alpha (e^{-\lambda T } )^{k} , \label{e3.7} \end{equation} for all $k >N$, $i=1, 2, \dots, n$, on any compact set of $\mathbb{R}$. It follows from \eqref{e3.6} and \eqref{e3.7} that $ \{(-1)^{p} v (t + pT)\}$ uniformly converges to a continuous function $x^{*}(t)=(x^{*}_{11}(t), x^{*}_{12}(t),\dots,x^{*}_{mn}(t))^{T}$ on any compact set of $\mathbb{R}$. Now we show that $x^{*}(t)$ is $T$-anti-periodic solution of \eqref{e1.1}. First, $x^{*}(t)$ is $T$-anti-periodic, since $$ x^{*}(t+T)=\lim_{p\to \infty}(-1)^{p } v (t +T+ p T)=-\lim_{(p+1)\to \infty}(-1)^{p+1 } v (t + (p +1)T)=-x^{*}(t ). $$ Next, we prove that $x^{*}(t)$ is a solution of \eqref{e1.1}. In fact, together with the continuity of the right side of \eqref{e1.1}, \eqref{e3.3} implies that $ \{((-1)^{p+1} v (t + (p+1)T))'\}$ uniformly converges to a continuous function on any compact set of $\mathbb{R}$. Thus, letting $p \to\infty$, we obtain \begin{equation} \begin{aligned} \frac{d}{dt}\{x^{*}_{ij}(t)\} &= -a_{ij}(t)x^{*}_{ij}(t )\\ &\quad -\sum_{C_{kl}\in N_{r}(i,j)}C_{ij}^{kl}(t) \int_{0}^{\infty}K_{ij}(u) f(x^{*}_{kl}(t-u))du x^{*}_{ij}(t ) + L_{ij }(t ). \end{aligned} \label{e3.8} \end{equation} Therefore, $x^{*}(t)$ is a solution of \eqref{e1.1}. Finally, by Lemma \ref{lem2.2}, we can prove that $x^{*}(t)$ is locally exponentially stable. This completes the proof. \end{proof} \section{An Example} In this section, we give an example to illustrate the results obtained in previous sections. Consider the following SICNNs with continuously distributed delays: \begin{equation} x_{ij}'(t) =-a_{ij}(t)x_{ij}(t)-\sum_{C_{kl}\in{N_{r}(i,j)}}C_{ij}^{kl}(t) \int_{0}^{\infty}K_{ij}(u) f(x_{kl}(t-u))dux_{ij}(t)+L_{ij}(t), \label{e4.1} \end{equation} where $ i=1, 2, 3$, $j=1, 2, 3$, $K_{ij}(u)=(\sin u)e^{-u}$, $f(x)=\frac{x^{4}+1}{6}$, \begin{gather*} \begin{pmatrix} a_{11}(t)&a_{12}(t)&a_{13}(t)\\ a_{21}(t)&a_{22}(t)&a_{23}(t)\\ a_{31}(t)&a_{32}(t)&a_{33}(t)\end{pmatrix} =\begin{pmatrix} 5+|\sin t| & 5 +|\sin t|& 9+|\sin t|\\ 6 +|\sin t|& 6 +|\sin t|& 7+|\sin t|\\ 8 +|\sin t|& 8+|\sin t| & 5+|\sin t|\end{pmatrix}, \\ \begin{pmatrix} C_{11}(t)&C_{12}(t)&C_{13}(t)\\ C_{21}(t)&C_{22}(t)&C_{23}(t)\\ C_{31}(t)&C_{32}(t)&C_{33}(t)\end{pmatrix} =\begin{pmatrix} 0.1 |\sin t|& 0.3|\sin \sqrt{3}t| & 0.5|\sin t|\\ 0.2 |\sin t|& 0.1 |\sin t|& 0.2|\sin t|\\ 0.1|\sin t| & 0.2 |\sin t|& 0.1|\sin \sqrt{3}t|\end{pmatrix},\\ \begin{pmatrix} L_{11}(t)&L_{12}(t)&L_{13}(t)\\ L_{21}(t)&L_{22}(t)&L_{23}(t)\\ L_{31}(t)&L_{32}(t)&L_{33}(t)\end{pmatrix} =\begin{pmatrix} \sin t & \sin t & \cos t\\ \sin t & \cos t & \cos t\\ \cos t & \sin t & \sin t\end{pmatrix}\,. \end{gather*} Obviously, (H1) holds. Set $L(r)=\frac{2}{3}r^{3}$. Since $f'(x)=\frac{2}{3}x^{3}$, (H2) holds. Next, let us check (H3). Clearly, we have $f(0)=\frac{1}{6}$, \begin{gather*} \sum_{C_{kl}\in{N_{1}(1,1)}}\bar{C}_{11}^{kl}=0.7,\quad \sum_{C_{kl}\in{N_{1}(1,2)}}\bar{C}_{12}^{kl}=1.4, \quad \sum_{C_{kl}\in{N_{1}(1,3)}}\bar{C}_{13}^{kl}=1.1,\\ \sum_{C_{kl}\in{N_{1}(2,1)}}\bar{C}_{21}^{kl}=1,\quad \sum_{C_{kl}\in{N_{1}(2,2)}}\bar{C}_{22}^{kl}=1.8, \quad \sum_{C_{kl}\in{N_{1}(2,3)}}\bar{C}_{23}^{kl}=1.4, \\ \sum_{C_{kl}\in{N_{1}(3,1)}}\bar{C}_{31}^{kl}=0.6,\quad \sum_{C_{kl}\in{N_{1}(3,2)}}\bar{C}_{32}^{kl}=0.9,\quad \sum_{C_{kl}\in{N_{1}(3,3)}}\bar{C}_{33}^{kl}=0.6,\\ D=\max_{(i,j)}\big\{\frac{\sum_{C_{kl}\in{N_{r}(i,j)}} \bar{C}_{ij}^{kl}\int_{0}^{\infty}|K_{ij}(u)|du}{\underline{a}_{ij}}\big\} \leq \max_{(i,j)}\big\{\frac{\sum_{C_{kl}\in{N_{r}(i,j)}} \bar{C}_{ij}^{kl}}{\underline{a}_{ij}}\big\}\leq 0.3,\\ L^{+}=\max_{(i,j)}\frac{\overline{L_{ij}}}{\underline{a}_{ij}}=0.2 \end{gather*} Take $r_{0}=1$. Then $$ D[|f(0)|r_{0}+L(r_{0})r^{2}_{0}]+L=0.3\cdot (\frac{1}{6}+\frac{2}{3})+0.2=0.45<1=r_{0} $$ and $$ D|f(0)|+2DL(r_{0})r_{0}=0.005+0.4=0.45<1. $$ Thus, (H3) holds for $r_{0}=1$. Take $r_{1}=(7/2)^{1/4}$. Then $$ D|f(0)|+DL(r_{0})r_{0}+ \max_{(i,j)}\big\{\frac{\sum_{C_{kl}\in{N_{r}(i,j)}} \bar{C}_{ij}^{kl}}{\underline{a}_{ij}}\big\}L(r_{1})r_{1} =0.25+0.7=0.95<1 $$ So (H4) holds. By Theorem \ref{thm3.1}, system \eqref{e4.1} has a $\pi$-anti-periodic solution $x^{*}(t)$ with initial value $\sup_{t\in [-1, 0]}\|\varphi^{*}\|< 1$. Moreover, all solutions of \eqref{e4.1} with initial value $\sup_{t\in (-\infty, 0]}\|\varphi (t)\|\leq (7/2)^{1/4}$ converge exponentially to $x^{*}(t)$ as $t\to+\infty$. \begin{remark} \label{rmk4.1} \rm SICNNs \eqref{e4.1} is a very simple form of shunting inhibitory cellular neural networks with continuously distributed delays. Since $f(x)=\frac{x^{4}+1}{6} $. One can observe that the condition (T0) is not satisfied. Therefore, the results in \cite{p1,s1} and the references therein can not be applicable to \eqref{e4.1}. This implies that our results are essentially new. \end{remark} \begin{thebibliography}{00} \bibitem{a1} A. R. Aftabizadeh, S. Aizicovici, N. H. Pavel; \emph{On a class of second-order anti-periodic boundary value problems}, J. Math. Anal. Appl. 171 (1992), 301--320. \bibitem{a2} S. Aizicovici, M. McKibben, S. Reich; \emph{Anti-periodic solutions to nonmonotone evolution equations with discontinuous nonlinearities}, Nonlinear Anal. 43 (2001), 233--251. \bibitem{c1} M. Cai, et al.??; \emph{Positive almost periodic solutions for shunting inhibitory cellular neural networks with time-varying delays}, Math. Comput. Simul. 78(4) (2008), 548-558. \bibitem{c2} A. Chen, J. Cao and L. Huang; \emph{Almost periodic solution of shunting inhibitory CNNs with delays}, Phys. Lett. A, 298 (2002), 161-170. \bibitem{c3} A. Chen and X. Huang; \emph{Almost periodic attractor of delayed neural networks with variable coefficients}, Phys. Lett. A, 340:1-4 (2005), 104-120. \bibitem{c4} Y. Chen, J. J. Nieto and D. O¡¯Regan; \emph{Anti-periodic solutions for fully nonlinear first-order differential equations}, Mathematical and Computer Modelling, 46 (2007), 1183--1190. \bibitem{h1} J. K. Hale; \emph{Theory of Functional Differential Equations}, Springer-Verlag, New York, 1977. \bibitem{h2} X. Huang and J. Cao; \emph{Almost periodic solutions of inhibitory cellular neural networks with time-vary delays}, Phys. Lett. A, 314 (2003), 222-231. \bibitem{l1} Y. Li, C. Liu and L. Zhu; \emph{Global exponential stability of periodic solution of shunting inhibitory CNNs with delays}, Phys. Lett. A, 337 (2005), 46-54. \bibitem{l2} B. Liu and L. Huang; \emph{Existence and stability of almost periodic solutions for shunting inhibitory cellular neural networks with continuously distributed delays}, Phys. Lett. A, 349 (2006), 177-186. \bibitem{p1} Guoqiang Peng, Lihong Huang; \emph{Anti-periodic solutions for shunting inhibitory cellular neural networks with continuously distributed delays}, Nonlinear Analysis: Real World Applications, 10(4) (2009) 2434-2440. \bibitem{s1} Jianying Shao; \emph{Anti-periodic solutions for shunting inhibitory cellular neural networks with time-varying delays}, Physics Letters A, 372(30) (2008) 5011-5016. \bibitem{w1} R. Wu; \emph{An anti-periodic LaSalle oscillation theorem}, Applied Mathematics Letters, 21(9) (2008), 928-933. \end{thebibliography} \end{document}