You are on page 1of 6

Applied Mathematics Letters 19 (2006) 13611366

www.elsevier.com/locate/aml
Periodic solutions for discrete predatorprey systems with the
BeddingtonDeAngelis functional response
Jimin Zhang

, Jing Wang
School of Mathematics and Statistics, KLAS and KLVE, Northeast Normal University, 5268 Renmin Street, Changchun, Jilin, 130024, PR China
Received 18 April 2005; received in revised form 22 January 2006; accepted 15 February 2006
Abstract
Sufcient criteria are established for the existence of positive periodic solutions of discrete nonautonomous predatorprey
systems with the BeddingtonDeAngelis functional response using a continuation theorem.
c 2006 Elsevier Ltd. All rights reserved.
Keywords: Periodic solutions; Difference equations; Predatorprey; BeddingtonDeAngelis functional response; Coincidence degree
1. Introduction
Predatorprey interaction is one of the major forces shaping food webs. Since the great work of Lotka (in 1925)
and Volterra (in 1926), modelling these interactions has been one of the central themes in mathematical ecology.
One signicant component of the predatorprey relationship is the predators rate of feeding upon prey, i.e., the so-
called predators functional response. In general, the functional responses can be either prey dependent or predator
dependent. Functional response equations that are strictly prey dependent, such as the Holling family ones, are
predominant in the literature.
However, the prey-dependent ones fail to model the interference among predators, and have been facing challenges
from the biology and physiology communities. The predator-dependent functional responses can provide better
descriptions of predator feeding over a range of predatorprey abundances, as is supported by much signicant
laboratory and eld evidence (see [5] and references therein). The BeddingtonDeAngelis functional response,
rst proposed by Beddington [3] and DeAngelis [2], performed even better. So, predatorprey systems with the
BeddingtonDeAngelis response have been studied extensively in the literature [1,6,7,9,10].
It is well known that the discrete time models governed by difference equations are more appropriate than the
continuous ones when the populations have nonoverlapping generations. In addition, discrete time models can also
provide efcient computational models of continuous for numerical simulations. However, no such work has been
done for predatorprey systems with the BeddingtonDeAngelis functional response.

Corresponding author.
E-mail address: Zhangjm588@nenu.edu.cn (J. Zhang).
0893-9659/$ - see front matter c 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.aml.2006.02.004
1362 J. Zhang, J. Wang / Applied Mathematics Letters 19 (2006) 13611366
In this work, we will study the existence of positive periodic solutions of the discrete analogue of the predatorprey
system with BeddingtonDeAngelis functional response explored by Fan and Kuang [5]. Following the clues in [4],
with the help of differential equations with piecewise constant arguments, one can reach its discrete analogue
x(k +1) = x(k) exp
_
a(k) b(k)x(k)
c(k)y(k)
(k) +(k)x(k) + (k)y(k)
_
y(k +1) = y(k) exp
_
d(k) +
f (k)y(k)
(k) +(k)x(k) + (k)y(k)
_ , k Z (1.1)
where Z denotes the set of integers. In the following, we will focus our attention on system (1.1). Considering the
biological signicance, we consider (1.1) with positive initial values and assume that the parameters in system (1.1)
are nonnegative and -periodic with Z and > 1.
2. Existence of positive periodic solutions
Let Z
+
, R
+
and R
2
denote the set of nonnegative integers, nonnegative real numbers, and two-dimensional
Euclidean vector space, respectively, and let
I
w
= {0, 1, 2, . . . , w 1} g =
1
w
w1

k=0
g(k) g
u
= max
kI
w
g(k) g
l
= min
kI
w
g(k)
where g(k) is a w-periodic sequence of nonnegative real numbers dened for k Z.
Let X, Y be normed vector spaces, L : Dom L X Y be a linear mapping, N : X Y be a continuous
mapping. The mapping L will be called a Fredholm mapping of index zero if dimKer L = codimIm L < + and
Im L is closed in Y. If L is a Fredholm mapping of index zero and there exist continuous projections P : X X and
Q : Y Y such that Im P = Ker L, Im L = Ker Q = Im(I Q), it follows that L|Dom L Ker P : (I P)X
Im L is invertible. We denote the inverse of that map by K
P
. If is an open bounded subset of X, the mapping N will
be called L-compact on

if QN(

) is bounded and K
P
(I Q)N :

X is compact. Since Im Q is isomorphic
to Ker L, there exists an isomorphism J : Im Q Ker L.
Lemma 2.1 (Continuation Theorem [8]). Let L be a Fredholm mapping of index zero and N be L-compact on

.
Suppose:
(a) for each (0, 1), every solution x of Lx = Nx is such that x ;
(b) QNx = 0 for each x Ker L and the Brouwer degree deg{JQN, Ker L, 0} = 0.
Then the operator equation Lx = Nx has at least one solution lying in Dom L

.
Lemma 2.2 ([4, Lemma 3.2]). Let g : Z R be w periodic, i.e., g(k +w) = g(k). Then for any xed k
1
, k
2
I
w
,
and any k Z, one has
g(k) g(k
1
) +
w1

s=0
|g(s +1) g(s)| g(k) g(k
2
)
w1

s=0
|g(s +1) g(s)|.
Dene l
2
= {u = u(k) : u(k) R
2
, k Z
+
}. Let l
w
l
2
denote the subspace of all w periodic sequences with
the usual supremum norm , i.e.,
u = max
kI
w
|u
1
(k)| +max
kI
w
|u
2
(k)|, for any u = {u(k) : k Z
+
} l
w
.
It is not difcult to show l
w
is a nite-dimensional Banach space. Let
l
w
0
=
_
u = {u(k)} l
w
:
w1

k=0
u(k) = 0, k Z
+
_
l
w
c
= {u = {u(k)} l
w
: u(k) = h R
2
, k Z
+
};
then it follows that l
w
0
and l
w
c
are both closed linear subspaces of l
w
and l
w
= l
w
0

l
w
c
, dim l
w
c
= 2, so we reach our
main result as follows:
J. Zhang, J. Wang / Applied Mathematics Letters 19 (2006) 13611366 1363
Theorem 2.1. If a > (c/ ) and (

f

d
u
)( a (c/ ))

b
1
exp{2 a}

d
u
> 0, then system (1.1) has at least one
positive periodic solution.
Proof. First, let x(k) = exp{u
1
(k)}, y(k) = exp{u
2
(k)}, so that (1.1) becomes
u
1
(k +1) u
1
(k) = a(k) b(k) exp{u
1
(k)}
c(k) exp{u
2
(k)}
(k) +(k) exp{u
1
(k)} + (k) exp{u
2
(k)}
u
2
(k +1) u
2
(k) = d(k) +
f (k) exp{u
1
(k)}
(k) +(k) exp{u
1
(k)} + (k) exp{u
2
(k)}
.
(2.1)
Dene X = Y = l
w
, (Lu)(k) = u(k +1) u(k), and
Nu = N
_
u
1
u
2
_
(k) =
_
_
_
_
a(k) b(k) exp{u
1
(k)}
c(k) exp{u
2
(k)}
(k) +(k) exp{u
1
(k)} + (k) exp{u
2
(k)}
d(k) +
f (k) exp{u
1
(k)}
(k) +(k) exp{u
1
(k)} + (k) exp{u
2
(k)}
_
_
_
_
for any u X and k Z
+
. It is trivially easy to see that L is a bounded linear operator and Ker L = l
w
c
, Im L = l
w
0
,
as well as that dimKer L = 2 = codimIm L. Since Im L is closed in Y, it follows that L is a Fredholm mapping of
index zero. Dene
Pu =
1
w
w1

s=0
u(s), u X, Qz =
1
w
w1

s=0
z(s), z Y.
It is easy to show that P and Q are continuous projections such that Im P = Ker L and Im L = Ker Q = Im(I Q).
Then, the generalized inverse (to L) K
P
: Im L Ker P
_
Dom L exists and is given by
K
P
(z) =
w1

s=0
z(s)
1
w
w1

s=0
(w s)z(s).
Obviously, QN and K
P
(I Q)N are continuous. Since X is a nite-dimensional Banach space, on the basis of the
ArzelaAscoli theorem, it is not difcult to show that K
P
(I Q)N(

) is compact for any open bounded set X.
Furthermore, QN(

) is bounded, so N is L-compact on

with any open bounded set X.
For the application of the continuation theorem, we must search for an appropriate open, bounded subset .
Corresponding to the operator equation Lu = Nu, (0, 1), we have
u
1
(k +1) u
1
(k) =
_
a(k) b(k) exp{u
1
(k)}
c(k) exp{u
2
(k)}
(k) +(k) exp{u
1
(k)} + (k) exp{u
2
(k)}
_
u
2
(k +1) u
2
(k) =
_
d(k) +
f (k) exp{u
1
(k)}
(k) +(k) exp{u
1
(k)} + (k) exp{u
2
(k)}
_
.
(2.2)
Assume that u(k) = {(u
1
(k), u
2
(k))
T
} X is an arbitrary solution of system (2.2) from 0 to w 1 with respect to k;
we reach
aw =
w1

k=0
_
b(k) exp{u
1
(k)} +
c(k) exp{u
2
(k)}
(k) +(k) exp{u
1
(k)} + (k) exp{u
2
(k)}
_

dw =
w1

k=0
_
f (k) exp{u
1
(k)}
(k) +(k) exp{u
1
(k)} + (k) exp{u
2
(k)}
_
(2.3)
From (2.2) and (2.3), we obtain
w1

k=0
|u
1
(k +1) u
1
(k)|
w1

k=0
_
a(k) +b(k) exp{u
1
(k)} +
c(k) exp{u
2
(k)}
(k) +(k) exp{u
1
(k)} + (k) exp{u
2
(k)}
_
= 2 aw
1364 J. Zhang, J. Wang / Applied Mathematics Letters 19 (2006) 13611366
w1

k=0
|u
2
(k +1) u
2
(k)|
w1

k=0
_
d(k) +
f (k) exp{u
1
(k)}
(k) +(k) exp{u
1
(k)} + (k) exp{u
2
(k)}
_
= 2

dw.
Since (u
1
(k), u
2
(k)) X, there exist
i
,
i
I
w
, such that
u
i
(
i
) = min
kI
w
{u
i
(k)} u
i
(
i
) = max
kI
w
{u
i
(k)}, i = 1, 2. (2.4)
It follows from (2.3) and (2.4) that
aw
w1

k=0
b(k) exp{u
1
(k)}
w1

k=0
b(k) exp{u
1
(
1
)} = exp{u
1
(
1
)}

bw
which reduces to u
1
(
1
) ln( a/

b) := L
1
, and hence
u
1
(k) u
1
(
1
) +
w1

s=0
|u
1
(s +1) u
1
(s)| ln{ a/

b} +2 aw := H
1
. (2.5)
On the other hand, from (2.3) and (2.4), we also have
aw
w1

k=0
_
b(k) exp{u
1
(
1
)} +
c(k)
(k)
_
= (c/ )w +

bwexp{u
1
(
1
)}
which reduces to u
1
(
1
) ln{
a(c/ )

b
} := l
1
, and hence
u
1
(k) u
1
(
1
)
w1

s=0
|u
1
(s +1) u
1
(s)| ln
_
a (c/ )

b
_
2 aw := H
2
which, together with (2.5), leads to max
kI
w
|u(k)| max{|H
1
|, |H
2
|} := B
1
.
From (2.3) and (2.4), we obtain

dw
w1

k=0
f (k) exp{u
1
(k)}

l
exp{u
1
(k)} +
l
exp{u
2
(k)}

w1

k=0
f (k)( a/

b) exp{2 aw}

l
( a/

b) exp{2 aw} +
l
exp{u
2
(
2
)}
.
Then u
2
(
2
) ln{
(

f

d
l
) a exp{2 aw}

b

d
l
} := L
2
; we have
u
2
(k) u
2
(
2
) +
w1

s=0
|u
2
(s +1) u
2
(s)| ln
_
(

f

d
l
) a exp{2 aw}

b

d
l
_
+2

dw := H
3
. (2.6)
We also can obtain from (2.3) that

dw
w1

k=0
f (k) exp{u
1
(k)}

u
+
u
exp{u
1
(k)} +
u
exp{u
2
(
2
)}

w1

k=0
f (k)( a (c/ ))

b
1
exp{2 aw}

u
+
u
a(c/)

b
exp{2 aw} +
u
exp{u
2
(
2
)}
.
Then u
2
(
2
) ln{
(

f

d
u
)( ac/ )

b
1
exp{2 aw}

d
u

d
u
} := l
2
; it follows that
u
2
(k) u
2
(
2
)
w1

s=0
|u
2
(s +1) u
2
(s)| ln
_
(

f

d
u
)( a c/ )

b
1
exp{2 aw}

d
u

d
u
_
2

dw:= H
4
so max
kI
w
|u
2
(k)| max{|H
3
|, |H
4
|} = B
2
. Obviously, B
1
and B
2
are independent of . Take B = B
1
+ B
2
+ B
0
where B
0
is taken sufciently large that B
0
> |l
1
| +|L
1
| +|l
2
| +|L
2
|.
J. Zhang, J. Wang / Applied Mathematics Letters 19 (2006) 13611366 1365
Fig. 1. A solution (1.1) with a(k) = 1 +0.2 sin(
k
2
), b(k) = 0.002(2 + sin(
k
2
)), c(k) = 0.5 +0.2 cos(
k
2
), d(k) = 0.6 +0.1 sin(
k
2
), f (k) =
0.8(0.5 +0.2 cos(
k
2
)), (k) = 0.2 +0.1 cos(
k
2
), (k) = 0.2 +0.1 sin(
k
2
), (k) = 1 and initial conditions x(0) = 20, y(0) = 5. The solution
tends to the 4-periodic solution (x

(k), y

(k)).
Considering the algebraic equations
_
_
_
_
_
_
a

b exp{u
1
}
1

w1

k=0
vc(k) exp{u
2
}
(k) +(k) exp{u
1
} + (k) exp{u
2
}

d +
1

w1

k=0
f (k) exp{u
1
}
(k) +(k) exp{u
1
} + (k) exp{u
2
}
_

_
=
_
0
0
_
(2.7)
where v [0, 1] is a parameter and (u
1
, u
2
) R
2
. Similarly, one can easily show that any solution (u

1
, u

2
) of (2.7)
with v [0, 1] satises
l
1
u

1
L
1
l
2
u

2
L
2
. (2.8)
Let = {(u
1
, u
2
)
T
X|(u
1
, u
2
) < B}; then is an open, bounded set in X and veries requirement (a)
of Lemma 2.1. When (u
1
, u
2
)
_
Ker L =
_
R
2
, (u
1
, u
2
) is a constant vector in R
2
with (u
1
, u
2
) =
|u
1
| +|u
2
| = B. Then we have QN[(u
1
, u
2
)
T
] = 0; that is, the rst part of (b) of Lemma (2.1) is valid.
Consider the homotopy for computing the Brouwer degree
A

((u
1
, u
2
)
T
) = QN((u
1
, u
2
)
T
) +(1 )F((u
1
, u
2
)
T
), [0, 1],
where
F((u
1
, u
2
)
T
) =
_
_
_
a

b exp{u
1
}

d
1

w1

k=0
f (k) exp{u
1
}
(k) +(k) exp{u
1
} + (k) exp{u
2
}
_
_
_
.
By the invariance property of homotopy, direct calculation produces
deg
_
JQN,
_
Ker L, 0
_
= deg
_
QN,
_
Ker L, 0
_
= deg
_
F,
_
Ker L, 0
_
= 0,
where deg(, , ) is the Brouwer degree and J is the identity mapping since Im Q = Ker L. We have proved that
veries all requirements of Lemma 2.1; then Lx = Nx has at least one solution in Dom L
_

, so system (2.1) has at
least one periodic solution in Dom L
_

, say (u

1
(k), u

2
(k))
T
. Let x

(k) = exp{u

1
(k)}, y

(k) = exp{u

2
(k)}, so
(x

(k), y

(k))
T
is an periodic solution of system(1.1) with strictly positive components. The proof is complete.
Remark 2.1. In (1.1), we have assumed that > 1. In fact, if = 1, then the parameters in (1.1) are all constants
and (1.1) is autonomous. The 1-periodic solutions of (1.1) must be constant solutions, i.e., the equilibria of (1.1).
Finally, in order to illustrate some features of our main results, let
a(k) = 1 +0.2 sin
_
k
2
_
, b(k) = 0.002
_
2 +sin
_
k
2
__
, c(k) = 0.5 +0.2 cos
_
k
2
_
,
1366 J. Zhang, J. Wang / Applied Mathematics Letters 19 (2006) 13611366
d(k) = 0.6 +0.1 sin
_
k
2
_
, f (k) = 0.8
_
0.5 +0.2 cos
_
k
2
__
,
(k) = 0.2 +0.1 cos
_
k
2
_
, (k) = 0.2 +0.1 sin
_
k
2
_
, (k) = 1.
It is trivial to show that the conditions in Theorem 2.1 are veried. Therefore, (1.1) admits at least one -periodic
solution. Our numerical simulation supports our theoretical ndings (see Fig. 1).
Acknowledgements
The authors would like express their gratitude to Prof. Meng Fan for helpful discussion and suggestions, and to
the referees for their excellent comments, which greatly improved the presentation of the work. The rst author was
supported by the NSFC.
References
[1] C. Cosner, D.L. DeAngelis, J.S. Ault, D.B. Olson, Effects of spatial grouping on the functional response of predators, Theoret. Population
Biol. 56 (1999) 6575.
[2] D.L. DeAngelis, R.A. Goldstein, R.V. ONeill, A model for trophic interaction, Ecology 56 (1975) 881892.
[3] J.R. Beddington, Mutual interference between parasites or predators and its effect on searching efciency, J. Animal Ecol. 44 (1975) 331340.
[4] M. Fan, K. Wang, Periodic solutions of a discrete time nonautonomous ratio-dependent predatorprey system, Math. Comput. Modelling 35
(910) (2002) 951961.
[5] M. Fan, Y. Kuang, Dynamics of a nonautonomous predatorprey system with the BeddingtonDeAngelis functional response, J. Math. Anal.
Appl. 295 (2004) 1539.
[6] R.S. Cantrell, C. Cosner, On the dynamics of predatorprey models with the BeddingtonDeAngelis functional response, J. Math. Anal. Appl.
257 (2001) 206222.
[7] R.S. Cantrell, C. Cosner, Effects of domain size on the persistence of populations in a diffusive food chain model with DeAngelisBeddginton
functional response, Nat. Resour. Modelling 14 (2001) 335367.
[8] R.E. Gaines, R.M. Mawhin, Coincidence Degree and Nonlinear Differential Equations, Springer-Verlag, Berlin, 1977.
[9] T.W. Hwang, Global analysis of the predatorprey system with BeddingtonDeAngelis functional response, J. Math. Anal. Appl. 281 (2003)
395401.
[10] T.W. Hwang, Uniqueness of limit cycles of the predatorprey system with BeddingtonDeAngelis functional response, J. Math. Anal. Appl.
290 (2004) 113122.