\documentclass[reqno]{amsart}
\usepackage{hyperref}

\AtBeginDocument{{\noindent\small
\emph{Electronic Journal of Differential Equations},
Vol. 2014 (2014), No. 52, 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 2014 Texas State University - San Marcos.}
\vspace{9mm}}

\begin{document}
\title[\hfilneg EJDE-2014/52\hfil Periodic solutions for a food chain model]
{Periodic solutions for a food chain system with Monod-Haldane functional
 response on \\ time scales}

\author[K. Zhuang \hfil EJDE-2014/52\hfilneg]
{Kejun Zhuang}  % in alphabetical order

\address{Kejun Zhuang \newline
Institute of Applied Mathematics,
Anhui University of Finance and Economics,
Bengbu 233030, China}
\email{zhkj123@163.com}

\thanks{Submitted July 7, 2012. Published February 21, 2014.}
\subjclass[2000]{92D25, 34C25}
\keywords{Food chain system; periodic solution; time scales; coincidence degree}

\begin{abstract}
 In this article, we study a three species food chain model on time scales,
 with Monod-Haldane functional response and  time delay.
 With the help of coincidence degree theory, we establish the
 existence of periodic solutions.
\end{abstract}

\maketitle
\numberwithin{equation}{section}
\newtheorem{theorem}{Theorem}[section]
\newtheorem{lemma}[theorem]{Lemma}
\allowdisplaybreaks


\section{Introduction}\label{s1}

Research on food chain system has been a hot spot in population dynamics. 
Dynamical behavior of these models governed by differential equations 
and difference equations has been extensively studied in \cite{1,2,5,3,4}.
 Hsu, Hwang and Kuang \cite{21} considered the  ratio-dependent 
food chain model
\begin{gather*}
  \dot{x}(t)=rx(1-\frac{x}{K})-\frac{1}{\eta_1}\frac{m_1xy}{a_1y+x} ,  \\
     \dot{y}(t)=\frac{m_1xy}{a_1y+x}-d_1y-\frac{1}{\eta_2}\frac{m_2yz}{a_2z+y} ,  \\
     \dot{z}(t)=\frac{m_2yz}{a_2z+y}-d_2z,
\end{gather*}
where $x$, $y$ and $z$ sand for the population densities of prey, 
predator and top predator, respectively.  The boundness, 
extinction and periodicity were studied.

Xu, Chaplain and Davidson \cite{22} studied the delayed three-species
 Lotka-Volterra food chain system
\begin{gather*}
  \dot{x}_1(t)=x_1(t)[r_1(t)-a_{11}(t)x_1(t-\tau_{11})-a_{12}(t)x_2(t)]  ,  \\
     \dot{x}_2(t)=x_2(t)[-r_2(t)+a_{21}(t)x_1(t-\tau_{21})-a_{22}(t)x_2(t-\tau_{22}) -a_{23}(t)x_3(t)   ]  ,  \\
     \dot{x}_3(t)=x_3(t)[-r_3(t)+a_{32}(t)x_2(t-\tau_{32})-a_{33}x_3(t-\tau_{33})   ]  .
\end{gather*}
The existence, uniqueness and global stability of positive periodic solutions 
of the system were studied.
In population dynamics, the relationship between predator and prey can be 
represented as the functional response
which refers to the change in the density of prey attached per unit time 
per predator as the prey density changes.
Holling \cite{23} gave three different kinds of functional responses,
which are monotonic in the first quadrant. But some experiments and 
observations indicate that nonmonotonic response is more realistic \cite{6}. 
To model such an inhibitory effect, Andrews
 suggested the so-called Monod-Haldane function proposed in \cite{6}. 
From above all, we can get the following non-autonomous food chain model 
with time delays,
\begin{equation}
\begin{gathered}
\dot{u}_1(t)=u_1(t)[ r_1(t)-d_1(t)u_1(t)-\frac{m_{12}(t)u_2(t)}{a_1(t)+b_1(t)u_1(t)
 +u_1^2(t)}  ],  \\
 \begin{aligned}
\dot{u}_2(t)
&=u_2(t)[ -r_2(t)+\frac{m_{21}u_1(t-\tau)}{a_1(t)+b_1(t)u_1(t-\tau)
 +u_1^2(t-\tau)}-d_2(t)u_2(t)\\ 
&\quad -\frac{m_{23}(t)u_3(t)}{a_2(t)+b_2(t)u_2(t)+u_2^2(t)}  ] ,
\end{aligned} \\
 \dot{u}_3(t)=u_3(t)[  -r_3(t)+\frac{m_{32}(t)u_2(t-\sigma)}{a_2(t)+b_2(t)
 u_2(t-\sigma)+u_2^2(t-\sigma)}  -d_3(t)u_3(t)  ],
\end{gathered} \label{1.1}
\end{equation}
where $u_1(t)$, $u_2(t)$ and $u_3(t)$ stand for the population density of prey, 
predator and top--predator at time $t$, respectively. All coefficients 
are positive continuous functions. $m_{i,i+1}(t)$ is the capture rate of 
the predator, $m_{i+1,i}(t)$  is a measure of the food quality that the prey 
provided for conversion into predator birth, where $i=1,2$.

On the other hand, if the populations have non-overlapping generations,
the discrete model governed difference equations is more appropriate
\begin{equation}
\begin{gathered}
u_1(n+1)=u_1(n)\exp[ r_1(n)-d_1(n)u_1(n)  
 -\frac{m_{12}(n)u_2(n)}{a_1(n)+b_1(n)u_1(n)+u_1^2(n)} ] ,  \\
\begin{aligned}
     u_2(n+1)&=u_2(n)\exp[  \frac{m_{21}u_1(n-\tau)}{a_1(n)+b_1(n)u_1(n-\tau)
 +u_1^2(n-\tau)}-r_2(n)   -d_2(n)u_2(n)\\ 
&\quad -\frac{m_{23}(n)u_3(n)}{a_2(n)+b_2(n)u_2(n)+u_2^2(n)}  ], 
\end{aligned}  \\
     u_3(n+1)=u_3(n)\exp[  \frac{m_{32}(n)u_2(n-\sigma)}{a_2(n)+b_2(n)u_2(n-\sigma)
+u_2^2(n-\sigma)}  -r_3(n) -d_3(n)u_3(n)   ],
\end{gathered} \label{1.3}
\end{equation}
where all the coefficients are positive periodic sequences.

To explore the periodic solutions of differential equation and difference 
equation models, coincidence degree theory is a common method. 
However, for these two types of systems, the methods and results are 
significantly similar.
Enlightened by  the idea of Stefan Hilger \cite{7}, to unify the continuous 
and discrete dynamic systems,  we consider the following dynamic system 
on time scales,
\begin{equation}
\begin{gathered}
x^\Delta(t)=r_1(t)-d_1(t)e^{x(t)}-\frac{m_{12}(t)e^{y(t)}}{a_1(t)
+b_1(t)e^{x(t)}+e^{2x(t)}},  \\
\begin{aligned}
 y^\Delta(t)&= -r_2(t)+\frac{m_{21}(t)e^{x(t-\tau)}}{a_1(t)+b_1e^{x(t-\tau)}
+e^{2x(t-\tau)}}-d_2(t)e^{y(t)} \\
&\quad   -\frac{m_{23}(t)e^{z(t)}}{a_2(t)+b_2(t)e^{y(t)}+e^{2y(t)}} ,
\end{aligned}   \\
     z^\Delta(t)=-r_3(t)+\frac{m_{32}(t)e^{y(t-\sigma)}}{a_2(t)+b_2(t)
e^{y(t-\sigma)}+e^{2y(t-\sigma)}}-d_3(t)e^{z(t)},
\end{gathered} \label{1.2}
\end{equation}
where $t\in\mathbb{T}$ and $\mathbb{T}$ is a time scale that is unbounded above. 
 $x^\Delta(t)$ is the delta derivative of $x$ at $t$, which is defined in \cite{12}.
 All the coefficients are positive $\omega-$periodic functions.
Set $u_1(t)=e^{x(t)}$,  $u_2(t)=e^{y(t)}$,  $u_3(t)=e^{z(t)}$,  
then \eqref{1.2} can be reduced to \eqref{1.1} and \eqref{1.3} when
 $\mathbb{T}=\mathbb{R}$ and $\mathbb{T}=\mathbb{Z}$, respectively.

The  purpose of this paper is to study the periodicity of  three-species 
food chain system on time scales and this model has not been investigated 
before. We would like to mention that there are several papers on 
periodicity in dynamic systems on time scales by using the coincidence 
degree theory, see \cite{11,8,9,10}.
 The remainder of the paper is organized as follows. In the following section, 
some preliminary results about calculus on time scales and the
continuation theorem are stated. Next, the sufficient conditions 
for the existence of periodic solutions are explored.

\section{Preliminaries}\label{s2}

For  convenience,  we first present the useful lemma about time scales 
and the continuation theorem of the coincidence degree theory;
 more details can be found in \cite{13,14}.

A time scale $\mathbb{T}$ is an arbitrary nonempty closed subset of
 real numbers $\mathbb{R}$. Throughout this paper, we assume that
the time scale $\mathbb{T}$ is unbounded above and below, such as
$\mathbb{R}$, $\mathbb{Z}$ and $\cup_{k\in\mathbb{Z}}[2k,2k+1]$.
The following definitions and lemmas about time scales are from \cite{13,14}.

\begin{lemma}[\cite{13}] \label{lem2.1}
 Let $t_1$, $t_2\in I_\omega$ and $t\in\mathbb{T}$. 
If $g:\mathbb{T}\to\mathbb{R}\in C_{rd}(\mathbb{T})$ is $\omega$-periodic,
 then
\begin{gather*}
g(t)\leq g(t_1)+\frac{1}{2}\int_k^{k+\omega}|g^\Delta(s)|\Delta s,\\
g(t)\geq g(t_2)-\frac{1}{2}\int_k^{k+\omega}|g^\Delta(s)|\Delta s,
\end{gather*}
where the constant factor $1/2$ is the best possible.
\end{lemma}

For simplicity, we use the following notation throughout this
paper. Let $\mathbb{T}$ be $\omega$-periodic; that is,
$t\in\mathbb{T}$ implies $t+\omega\in\mathbb{T}$,
\begin{gather*}
k=\min \{\mathbb{R}^+\cap\mathbb{T}\}, \quad
I_\omega=[k,k+\omega]\cap\mathbb{T},\quad
g^L=\inf_{t\in\mathbb{T}}g(t),\\
g^M=\sup_{t\in\mathbb{T}}g(t), \quad
\bar{g}=\frac{1}{\omega}\int_{I_\omega}g(s)\Delta
s=\frac{1}{\omega}\int_k^{k+\omega}g(s)\Delta s,
\end{gather*}
where $g\in C_{rd}(\mathbb{T})$ is an $\omega$-periodic real
function; i.e., $g(t+\omega)=g(t)$ for all $t\in \mathbb{T}$.

Next, we state the Mawhin's continuation theorem, which is a main
tool in the proof of our theorem.

\begin{lemma}[\cite{14}] \label{lem2.2} 
 Let $L$ be a Fredholm mapping of index zero and $N$ be $L$-compact on
$\bar{\Omega}$. Suppose
\begin{itemize}
\item[(a)] for each $\lambda\in(0,1)$,
every solution $u$ of $Lu=\lambda Nu$ is such that
$u\notin\partial\Omega$;

\item[(b)]  $QNu\neq 0$ for each
$u\in\partial\Omega\cap\ker L$  and the Brouwer degree 
$\deg \{JQN,\Omega\cap\ker L,0\}\neq 0$.
\end{itemize}
Then the operator equation $Lu=Nu$ has at least one solution lying
in $\operatorname{Dom}L\cap\bar{\Omega}$.
\end{lemma}

\section{Main Results}\label{s3}


\begin{theorem} \label{thm3.1}
If $m_{32}^Le^{L_2}>r_3^M(a_2^M+b_2^Me^{M_2}+e^{2M_2})$
 holds, where
 $$
L_2=\ln\frac{a_2^Lr_3^L}{m_{32}^M} - \frac{m_{21}^M\omega}{b_1^L}
$$
  and
$$
M_2=\ln\frac{r_1^M(a_1^M+\frac{b_1^Mr_1^M}{d_1^L}
+\frac{2r_1^M}{d_1^L})}{m_{12}^L} +\frac{m_{21}^M\omega}{b_1^L},
$$
then \eqref{1.2} has at least one $\omega$-periodic solution.
\end{theorem}

\begin{proof}
Let 
\begin{gather*}
\begin{aligned}
X=Z=\Big\{ &(x,y,z)^T\in C(\mathbb{T},\mathbb{R}^3):
x(t+\omega)=x(t),  y(t+\omega)=y(t),\\
& z(t+\omega)=z(t),  \forall t\in \mathbb{T} \Big\},
\end{aligned} \\
\| (x,y,z)^T\| =\max_{t\in I_\omega}|x(t)|+\max_{t\in I_\omega}|y(t)|
+\max_{t\in I_\omega}|z(t)|,\quad (x,y,z)^T\in X \quad(\text{or in } Z).
\end{gather*}
Then $X$ and $Z$ are both Banach spaces when they are endowed with
the above norm $\| \cdot \|$. Let
\begin{gather*}
N \begin{bmatrix}
     x  \\
     y  \\
     z  \end{bmatrix}
= \begin{bmatrix}
     N_1  \\
     N_2  \\
     N_3    \end{bmatrix},
\end{gather*}
where
\begin{gather*}
N_1=r_1(t)-d_1(t)e^{x(t)}-\frac{m_{12}(t)e^{y(t)}}{a_1(t)+b_1(t)e^{x(t)}+e^{2x(t)}},
\\
\begin{aligned}
N_2&=-r_2(t)+\frac{m_{21}(t)e^{x(t-\tau)}}{a_1(t)+b_1e^{x(t-\tau)}+e^{2x(t-\tau)}}
-d_2(t)e^{y(t)} \\
&\quad -\frac{m_{23}(t)e^{z(t)}}{a_2(t)+b_2(t)e^{y(t)}+e^{2y(t)}},
\end{aligned} \\
N_3=-r_3(t)+\frac{m_{32}(t)e^{y(t-\sigma)}}{a_2(t)+b_2(t)e^{y(t-\sigma)}
+e^{2y(t-\sigma)}}-d_3(t)e^{z(t)}.
\\
L \begin{bmatrix}
     x  \\
     y  \\
     z    \end{bmatrix}
=\begin{bmatrix}
     x^\Delta  \\
     y^\Delta  \\
     z^\Delta   \end{bmatrix},\quad
P \begin{bmatrix}
     x  \\
     y  \\
     z  \end{bmatrix}
=Q \begin{bmatrix}
     x  \\
     y  \\
     z \end{bmatrix}
= \begin{bmatrix}
     \frac{1}{\omega}\int_k^{k+\omega}x(t)\Delta t  \\
     \frac{1}{\omega}\int_k^{k+\omega}t(t)\Delta t  \\
     \frac{1}{\omega}\int_k^{k+\omega}z(t)\Delta t \end{bmatrix}.
\end{gather*}
Obviously,  $\ker L=\mathbb{R}^3$,  
$\operatorname{Im}L=\big\{ (x,y,z)^T\in Z :
\bar{x}=\bar{y}=\bar{z}=0,t\in\mathbb{T} \big\}$,
$\dim \ker L=3=\operatorname{codim} \operatorname{Im}L$.
Since $\operatorname{Im}L$ is closed in $Z$, then $L$ is a Fredholm
mapping of index zero. It is easy to show that $P$ and $Q$ are
continuous projections such that $\operatorname{Im}P=\ker L$ and
$\operatorname{Im}L=\ker Q=\operatorname{Im}(I-Q)$. Furthermore, the
generalized inverse (of $L$) $K_P:\operatorname{Im}L\to \ker P\cap
\operatorname{Dom}L$ exists and is given by
$$
K_P \begin{bmatrix}
     x  \\
     y  \\
     z    \end{bmatrix}
= \begin{bmatrix}
     \int_k^tx(s)\Delta s-\frac{1}{\omega}\int_k^{k+\omega}
\int_k^t x(s)\Delta s\Delta t  \\
     \int_k^ty(s)\Delta s-\frac{1}{\omega}\int_k^{k+\omega}
\int_k^t y(s)\Delta s\Delta t  \\
     \int_k^tz(s)\Delta s-\frac{1}{\omega}\int_k^{k+\omega}
\int_k^t z(s)\Delta s\Delta t
   \end{bmatrix} .
$$
Thus
$$
QN\begin{bmatrix}
     x  \\
     y  \\
     z    \end{bmatrix}
     = \begin{bmatrix}
     \frac{1}{\omega}\int_k^{k+\omega} (   r_1(t)-d_1(t)e^{x(t)}  -\frac{m_{12}(t)e^{y(t)}}{a_1(t)+b_1(t)e^{x(t)}+e^{2x(t)}}
          )\Delta t  \\
     \frac{1}{\omega}\int_k^{k+\omega} (   -r_2(t)+\frac{m_{21}(t)e^{x(t-\tau)}}{a_1(t)+b_1e^{x(t-\tau)}+e^{2x(t-\tau)}}\\
     \quad\quad\quad\quad\quad\quad -d_2(t)e^{y(t)}-\frac{m_{23}(t)e^{z(t)}}{a_2(t)+b_2(t)e^{y(t)}+e^{2y(t)}}           )\Delta t  \\
     \frac{1}{\omega}\int_k^{k+\omega}  (     -r_3(t)+\frac{m_{32}(t)e^{y(t-\sigma)}}{a_2(t)+b_2(t)e^{y(t-\sigma)}+e^{2y(t-\sigma)}}\\
     \quad\quad\quad\quad\quad\quad  \quad\quad\quad\quad\quad\quad     -d_3(t)e^{z(t)}
           )\Delta t
   \end{bmatrix},
$$
and
$$
 K_P(I-Q)N \begin{bmatrix}
     x  \\
     y  \\
     z    \end{bmatrix} \\
= \begin{bmatrix}
     \int_k^tx(s)\Delta s-\frac{1}{\omega}\int_k^{k+\omega}
\int_k^tx(s)\Delta s\Delta t\\
\quad\quad\quad\quad\quad\quad\quad  -\left( t-k-\frac{1}{\omega}
\int_k^{k+\omega}(t-k)\Delta t \right)\bar{x}  \\
     \int_k^ty(s)\Delta s-\frac{1}{\omega}\int_k^{k+\omega}
\int_k^ty(s)\Delta s\Delta t\\
\quad\quad\quad\quad\quad\quad\quad   -\left( t-k-\frac{1}{\omega}
\int_k^{k+\omega}(t-k)\Delta t \right)\bar{y}  \\
     \int_k^tz(s)\Delta s-\frac{1}{\omega}\int_k^{k+\omega}
\int_k^tz(s)\Delta s\Delta t-\\
\quad\quad\quad\quad\quad\quad\quad   \left( t-k-\frac{1}{\omega}
\int_k^{k+\omega}(t-k)\Delta t \right)\bar{z}
   \end{bmatrix}.
$$
Clearly, $QN$ and $K_P(I-Q)N$ are continuous. According to the
Arzela-Ascoli theorem, it is not difficulty to show that
$K_P(I-Q)N(\bar{\Omega})$ is compact for any open bounded set
$\Omega\subset X$ and $QN(\bar{\Omega})$ is bounded. Thus, $N$ is
$L$-compact on $\bar{\Omega}$.


Now, we shall search an appropriate open bounded subset $\Omega$ for
the application of the continuation theorem, Lemma \ref{lem2.2}. For the
operator equation $Lu=\lambda Nu$, where $\lambda\in (0,1)$, we have
\begin{equation}
\begin{gathered}
     u^\Delta_1(t)=\lambda(   r_1(t)-d_1(t)e^{x(t)}
-\frac{m_{12}(t)e^{y(t)}}{a_1(t)+b_1(t)e^{x(t)}+e^{2x(t)}}) ,  \\
\begin{aligned}
     u^\Delta_2(t)&=\lambda( -r_2(t)+\frac{m_{21}(t)e^{x(t-\tau)}}{a_1(t)
 +b_1e^{x(t-\tau)}+e^{2x(t-\tau)}}   -d_2(t)e^{y(t)}\\
&\quad -\frac{m_{23}(t)e^{z(t)}}{a_2(t)+b_2(t)e^{y(t)}+e^{2y(t)}}), 
\end{aligned}  \\
     u^\Delta_3(t)=\lambda(  -r_3(t)+\frac{m_{32}(t)e^{y(t-\sigma)}}{a_2(t)
+b_2(t)e^{y(t-\sigma)}+e^{2y(t-\sigma)}}   -d_3(t)e^{z(t)}).
   \end{gathered}  \label{3.1}
\end{equation}
Assume that $(u_1,u_2,u_3)^T\in X$ is a solution of system
\eqref{3.1} for a certain $\lambda\in (0,1)$. Integrating
\eqref{3.1} on both sides from $k$ to $k+\omega$, we obtain
\begin{equation}
\begin{gathered}
     \int_k^{k+\omega}[ d_1(t)e^{x(t)}+\frac{m_{12}(t)e^{y(t)}}{a_1(t)+b_1(t)
 e^{x(t)}+e^{2x(t)}}      ] \Delta t=\bar{r}_1\omega ,  \\
\begin{aligned}
 & \int_k^{k+\omega}[  r_2(t)+d_2(t)e^{y(t)}+\frac{m_{23}(t)e^{z(t)}}{a_2(t)
+b_2(t)e^{y(t)}+e^{2y(t)}}   ]  \Delta t  \\
&\quad =\int_k^{k+\omega} 
\frac{m_{21}(t)e^{x(t-\tau)}}{a_1(t)+b_1e^{x(t-\tau)}
+e^{2x(t-\tau)}}   \Delta t,  
\end{aligned} \\
       \int_k^{k+\omega}[  r_3(t)+d_3(t)e^{z(t)}  ]   \Delta t 
 =\int_k^{k+\omega}  \frac{m_{32}(t)e^{y(t-\sigma)}}{a_2(t)+b_2(t)
e^{y(t-\sigma)}+e^{2y(t-\sigma)}}   \Delta t.
   \end{gathered}  \label{3.2}
\end{equation}
Since $(x,y,z)^T\in X$, there  exist
$\xi_i,\eta_i\in I_\omega$, $i=1,2,3$, such that
\begin{equation}
\begin{gathered}
x(\xi_1)=\min_{t\in I_\omega}\{x(t)\},\quad  x(\eta_1)
=\max_{t\in I_\omega}\{x(t)\},\\
y(\xi_2)=\min_{t\in I_\omega}\{y(t)\},\quad  y(\eta_2)
=\max_{t\in I_\omega}\{y(t)\},\\
z(\xi_3)=\min_{t\in I_\omega}\{z(t)\},\quad  z(\eta_3)
=\max_{t\in I_\omega}\{z(t)\}.
\end{gathered}
\label{3.3}
\end{equation}
From \eqref{3.1} and \eqref{3.2}, we have
\begin{gather*}
\int_k^{k+\omega} | x^\Delta (t) |\Delta t\leq 2\bar{r}_1\omega,  \\
\int_k^{k+\omega} | y^\Delta (t) |\Delta t\leq 2\frac{m_{21}^M\omega}{b_1^L}, \\
\int_k^{k+\omega} | z^\Delta (t) |\Delta t \leq 2\frac{m_{32}^M\omega}{b_2^L}.
\end{gather*}
By the first equation of \eqref{3.2} and \eqref{3.3},
$$
d_1(\xi_1)e^{x(\xi_1)}<r_1(\xi_1);
$$
 that is,
 $$
x(\xi_1)<\ln\frac{r_1^M}{d_1^L}.
$$
From the second equation of \eqref{3.2}, we have
$$
r_2(\eta_2)<\frac{m_{21}(\eta_2)e^{x(\eta_2-\tau)}}{a_1(\eta_2)}
$$
and
$$
x(\eta_1)\geq x(\eta_2-\tau)>\ln\frac{r_2^La_1^L}{m_{21}^M}. 
$$
According to Lemma \ref{lem2.1}, we have
\begin{gather*}
x(t) \leq  x(\xi_1)+\frac{1}{2}\int_k^{k+\omega}| x^\Delta (t)|\Delta t
 \leq  \ln\frac{r_1^M}{d_1^L}+\bar{r}_1\omega:= M_1,
\\
x(t) \geq  x(\eta_1)-\frac{1}{2}\int_k^{k+\omega}| x^\Delta (t)|\Delta t
\geq    \ln\frac{r_2^La_1^L}{m_{21}^M} -\bar{r}_1\omega:= L_1.
\end{gather*}
From the first equation of \eqref{3.2} and \eqref{3.3}, we  obtain
\begin{gather*}
\frac{m_{12}(\xi_1)e^{y(\xi_1)}}{a(\xi_1)+b_1(\xi_1)e^{x(\xi_1)}+e^{2x(\xi_1)}}
<r_1(\xi_1),
\\
y(\xi_2)<y(\xi_1)<\ln\frac{r_1^M(a_1^M+\frac{b_1^Mr_1^M}{d_1^L}
+\frac{2r_1^M}{d_1^L})}{m_{12}^L}.
\end{gather*}
Then
$$
y(t) \leq  y(\xi_2)+\frac{1}{2}\int_k^{k+\omega}| y^\Delta (t)|\Delta t 
<  \ln\frac{r_1^M(a_1^M+\frac{b_1^Mr_1^M}{d_1^L}+\frac{2r_1^M}{d_1^L})}{m_{12}^L} 
+\frac{m_{21}^M\omega}{b_1^L}
 :=  M_2.
$$
From the third equation of \eqref{3.2}, we have
$$
\frac{m_{32}(\xi_3)}{a_2(\xi_3)}e^{y(\xi_3-\sigma)}>r_3(t),
$$
this reduces to
$$
y(\eta_2)\geq y(\xi_3-\sigma)>\ln\frac{a_2^Lr_3^L}{m_{32}^M}.
$$
Then
$$
y(t) \geq  y(\eta_2)-\frac{1}{2}\int_k^{k+\omega}\left| y^\Delta (t)
\right|\Delta t
 \geq   \ln\frac{a_2^Lr_3^L}{m_{32}^M} - \frac{m_{21}^M\omega}{b_1^L}:= L_2.
$$

According to the first equation of \eqref{3.2}, we have
$$
\frac{m_{32}(\xi_3)}{b_2(\xi_3)}>d_3(\xi_3)e^{z(\xi_3)}.
$$
Then
$$
z(t) \leq  z(\xi_3)+\frac{1}{2}\int_k^{k+\omega}| z^\Delta (t)|\Delta  t\\
 \leq   \ln\frac{m_{32}^M}{b_2^Ld_3^L}   +\frac{m_{32}^M\omega}{b_2^L}:= M_3.
$$
Also we have
$$
d_3^Me^{z(\eta_3)}>\frac{m_{32}^Le^{L_2}}{a_2^M+b_2^Me^{M_2}+e^{2M_2}}-r_3^M.
$$
Then
$$
z(t) \geq  z(\eta_3)-\frac{1}{2}\int_k^{k+\omega}| z^\Delta (t) |\Delta t\\
 \geq    \ln\frac{\frac{m_{32}^Le^{L_2}}{a_2^M+b_2^Me^{M_2}+e^{2M_2}}-r_3^M}{d_3^M} 
-\frac{m_{32}^M\omega}{b_2^L}  := L_3.
$$
Therefore, we have
\begin{gather*}
\max_{t\in[k,k+\omega]}|x(t)|\leq\max\{ | M_1|, |L_1| \}:= R_1, \\
\max_{t\in[k,k+\omega]}|y(t)|\leq\max\{ | M_2|, |L_2| \}:= R_2, \\
\max_{t\in[k,k+\omega]}|z(t)|\leq\max\{ | M_3|, |L_3| \}:= R_3.
\end{gather*}
Clearly, $R_1, R_2$ and $R_3$ are independent of $\lambda$. Let
$R=R_1+R_2+R_3+R_0$, where $R_0$ is taken sufficiently large such
that for  the algebraic equations
\begin{equation}
   \begin{gathered}
      \bar{r}_1-\bar{d}_1e^x- \frac{1}{\omega}\int_\kappa^{\kappa+\omega} 
 \frac{m_{12}(t)e^y}{a_1(t)+b_1(t)e^x+e^{2x}}\Delta t=0 ,  \\
 \begin{aligned}
& -\bar{r}_2+ \frac{1}{\omega}\int_\kappa^{\kappa+\omega}  
 \frac{m_{21}(t)e^x}{a_1(t)+b_1(t)e^x+e^{2x}}  \Delta t-\bar{d}_2e^y  \\
& -\frac{1}{\omega}\int_\kappa^{\kappa+\omega}  
  \frac{\bar{m}_{23}e^z}{a_2(t)+b_2(t)e^y+e^{2y}}  \Delta t=0, 
\end{aligned}  \\
      -\bar{r}_3+\frac{1}{\omega}\int_\kappa^{\kappa+\omega}   \frac{m_{32}(t)e^y}{a_2(t)+b_2(t)e^y+e^{2y}} \Delta t   -\bar{d}_3e^z=0,
   \end{gathered} \label{3.4}
\end{equation}
every solution $(x^\ast,y^\ast,z^\ast)^T$ of \eqref{3.4} satisfies
$\| (x^\ast,y^\ast,z^\ast)^T \|<R$. 
Now, we define $\Omega=\{ (x,y,z)^T\in X:  \| (x,y,z)^T \|<R \}$. Then it is clear
that $\Omega$ satisfies the requirement (a) of Lemma \ref{lem2.2}. If
$(u_1,u_2,u_3)^T\in \partial\Omega\cap\ker L=\partial\Omega\cap\mathbb{R}^3$, then
$(x,y,z)^T$ is a constant vector in $\mathbb{R}^3$ with
$\|(x,y,z)^T\|=|x|+|y|+|z|=R$, so we have
$$
QN \begin{bmatrix}
     x  \\
     y  \\
     z   \end{bmatrix}
      \neq \begin{bmatrix}
     0  \\
     0  \\
     0   \end{bmatrix}.
$$

By the assumption in Theorem \ref{thm3.1} and the definition of topological 
degree, a direct calculation yields
$\deg (JQN,\Omega\cap\ker L,0)\neq 0$.
We have verified that $\Omega$ satisfies all requirements of
Lemma \ref{lem2.2}; therefore, system \eqref{1.2}  has at least one
$\omega$-periodic solution in $\operatorname{Dom}L\cap\bar{\Omega}$. 
This completes the proof.
\end{proof}


\subsection*{Conclusion}

In this article, a three species  food chain model on time scales is proposed. 
This model not only unifies the food chain system with Monod-Haldane functional 
response and time delay governed by differential equations and their 
discrete analogues in form of difference equations, but also extends 
the results to more general time scales. By using the Mawhin's 
continuation theorem of coincidence degree theory, the existence of 
periodic solutions is established, which means that we do not have 
to investigate the same problem in systems \eqref{1.1}  
and \eqref{1.3} repeatedly. Moreover, based on the sharp inequalities 
in \cite{13},  the priori  estimates of periodic solutions are better 
than previous work.


\subsection*{Acknowledgments}
The author would like to acknowledge the support from National Natural
 Science Foundation of China (11301001), Anhui Provincial Natural 
Science Funds (1208085QA11), Excellent Youth Scholars Foundation 
and the Natural Science Foundation of Anhui Province  (2013SQRL030ZD) 
and  Humanities and Social Science Research Projects of Anhui Universities 
(SK2013B018).

\begin{thebibliography}{99}

\bibitem{6} Andrews, J. F.;  
 A mathematical model for the continuous culture of microorganisms 
utilizing inhabitory substrates, \emph{ Biotechnol Bioengrg}, 
Vol. 10(1968), 707--723.

\bibitem{11}    Bohner, M.;  Fan, M.;  Zhang, J. M.; 
Existence of periodic solutions in predator--prey and competition dynamic systems, 
 \emph{ Nonl. Anal: RWA},  Vol. 7(2006),  1193--1204.

\bibitem{8}    Bohner, M.;  Fan, M.;  Zhang, J. M.;  
  Periodicity of scalar dynamic equations and applications to population models, 
\emph{J. Math. Anal. Appl.} Vol. 330(2007), 1--9.

\bibitem{12}    Bohner, M.;  Peterson, A.; 
  \emph{Dynamic Equations on Time Scales: An Introduction with Applications},  
Birkh\"{a}user,  Boston, 2001.

\bibitem{1}  Francoise, J. P.;  Llibre, J.;  
Analytical study of a triple Hopf bifurcation in a tritrophic food chain model,  
\emph{Applied Mathematics and Computation},  Vol. 217(2011), 7146--7154.

\bibitem{14}   Gaines,  R. E.; Mawhin, J. L.; 
\emph{ Coincidence Degree and Nonlinear Differential Equations},
 Lecture Notes in Mathematics, Springer-Verlag,  Berlin,  1977.

\bibitem{7}   Hilger, S.;   Analysis on measure chains-a unified
approach to continuous and discrete calculus, 
\emph{Results Math.}, Vol. 18(1990),  18--56.

\bibitem{23}   Holling C. S.;  
The functional response of predator to prey density and its role in mimicry 
and population regulations,
 \emph{Mem. Ent. Sec. Can.  Vol.},  45(1965), 1--60.

\bibitem{21}  Hsu,  S. B.; Hwang,  T.W.;  Kuang, Y.; 
  A ratio-dependent food chain model and its applications to biological control, 
\emph{ Math. Biosci.},  Vol. 181(2003),  55--83.

\bibitem{2} Mukhopadhyay B., Bhattacharyya R.; 
 A stage-structured food chain model with stage dependent predation: 
Existence of codimension one and codimension two bifurcations, 
 \emph{Nonlinear Analysis: RWA},  Vol. 12(2011), 3056--3072.

\bibitem{9}  Tong, Y.;  Liu, Z. J.;  Gao, Z. Y.;   Wang, Y. H.; 
 Existence of periodic solutions for a predator-prey system with sparse 
effect and functional response on time scales,
  \emph{ Commun. Nonlinear Sci. Numer. Simulat.},   Vol. 17(2012), 3360--3366.

\bibitem{5}  Wang, W. M.;  Wang, H. L.;   Li, Z. Q.; 
 The dynamic complexity of a three--species Beddington-type food 
chain with impulsive control strategy,  
\emph{ Chaos, Solitons and Fractals},  Vol. 32(2007), 1772--1785.

\bibitem{3}  Xu, C. Y.;  Wang,  M. J.; 
 Permanence for a delayed discrete three-level food-chain model with 
Beddington-DeAngelis functional response, 
 \emph{ Applied Mathematics and Computation},  Vol. 187(2007),   1109--1119.

\bibitem{22}  Xu, R.; Chaplain, M. A. J.;  Davidson, F. A.; 
 Periodic solution for a three-species Lotka-Volterra food-chain model 
with time delays,  \emph{Mathematical and Computer Modelling}, 
 Vol. 40(2004), 823--837.

\bibitem{4}  Xu, R.;  Chen, L. S.;   Hao, F. L.; 
Periodic solutions of a discrete time Lotka-Volterra type food-chain model 
with delays, \emph{Applied Mathematics and Computation},  Vol. 171(2005), 91--103.

\bibitem{13}   Zhang, B. B.,  Fan, M.; 
 A remark on the application of coincidence degree to periodicity of dynamic 
equtions on time scales,
 \emph{J. Northeast Normal University(Natural Science Edition)}, Vol. 39(2007), 
1--3. (in Chinese)

\bibitem{10}  Zhuang, K. J.;  
Periodic solutions for a three species predator--prey system on time scales, 
\emph{ Int. J. Dynamical Systems and Differential Equations},    
Vol. 3(2011),  3-11.

\end{thebibliography}

\end{document}
