Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
College of Sciences, Guangxi University for Nationalities, Nanning 530006, Guangxi Province, PR China
College of Mathematics and Information Science, Yulin Normal University, Yulin, Guangxi 537000, PR China
a r t i c l e
i n f o
a b s t r a c t
This paper is concerned with the global asymptotic stability of a stochastic delay predator
prey system with BeddingtonDeAngelis functional response. Sufcient criteria for the
global asymptotic stability of the system are established. Some simulation gures are
provided to show that our model is more realistic than existing models.
2014 Elsevier Inc. All rights reserved.
Keywords:
Predatorprey system
Global asymptotic stability
BeddingtonDeAngelis functional response
Time delays
Stochastic perturbations
1. Introduction
In a natural ecosystem, fundamental features of population interactions, such as predation and competition have been
elucidated by empirical and theoretical investigations of the dynamics between two species. One signicant component
of the predatorprey relationship is the predators functional response. Skalski and Gilliam [23] stated that the functional
response can provide better descriptions of predator feeding over a range of predatorprey abundances by comparing the
statistical evidence from 19 predatorprey systems with the three predator-dependent functional responses (HassellVarley,
BeddingtonDeAngelis and Crowley-Martin), and in some cases, the BeddingtonDeAngelis type functional response performed even better.
The original predatorprey system with BeddingtonDeAngelis type functional response is
dx
a12 y
;
x b1 a11 x
dt
1 bx cy
dy
a22 x
;
y b2 a21 y
dt
1 bx cy
where x xt and y yt stand for the preys and the predators densities at time t, respectively. For biological signicance
of each coefcient we refer the reader to [2,4].
On the other hand, population dynamics is always affected by environmental noise (see e.g. [5,6]), which is an important
component in an ecosystem. Thus many authors studied stochastic models which corresponding to deterministic models to
reveal the inuence of environmental noise on the population dynamics (see e.g. [1,3,9,18,19,1517,21,22,24,10,12]). Moreover, recently, Liu and Han [14,13] have studied integro-differential equations with impulsive integral conditions, their work
extend some existing models to a great extent of this eld.
q
Research supported by NNSF of China Grant Nos. 11271087, 61263006, Scientic research project of Guangxi Education Department, China
(No. 2013YB074), Scientic research project of State Ethnic Affairs Commission of China Nos. 12GXZ001 and 2013YJYB03.
Corresponding author.
E-mail addresses: liuqun151608@163.com (Q. Liu), yiliangliu100@126.com (Y. Liu), xuepan100@126.com (X. Pan).
http://dx.doi.org/10.1016/j.amc.2014.02.091
0096-3003/ 2014 Elsevier Inc. All rights reserved.
In 2011, Liu and Wang [11] studied the following nonlinear stochastic predatorprey system
8
h
i
a12 y
>
< dx x 1 a11 x 1bx
cy b1 dt r1 dB1 t;
h
i
>
: dy y 1 a21 x a22 y b2 dt r2 dB2 t;
1bxcy
with the initial value x0 x0 > 0; y0 y0 > 0. They showed that if the positive equilibrium point of the deterministic
system was globally stable, then the stochastic model would preserve the nice property provided the noise was small enough. However, like it or not, time delays frequently occur in almost every situation, Kuang [8] has revealed that ignoring
time delays means ignoring reality. Hence it is essential to take time delays into account. Motivated by these, in this paper,
we will propose and study a more realistic and complex stochastic delay predatorprey system with BeddingtonDeAngelis
functional response
8
h
i
R0
a14 yt
>
< dxt xt r1 a11 xt a12 xt s1 a13 1 xt hdl1 h 1bxt
cyt dt r1 xtxt x dB1 t;
h
i
R
>
: dyt yt r 2 b11 yt b12 yt s2 b13 0 yt hdl h b14 xt
dt r2 ytyt y dB2 t;
2
1
1bxtcyt
where r2i represents the intensity of the noise, Bi t is a standard Brownian motion dened on a complete probability space
X; F ; P with a ltration fF t gt2R satisfying the usual conditions, si P 0 and li h is a probability measure on
1; 0; i 1; 2; x ; y is the equilibrium state of system (3). For more biological motivation on this type of modeling we
refer the reader to Gard [5,6].
This paper is organized as follows: in Section 2, we will show that if the noise is small enough and the positive equilibrium state of the deterministic model is globally stable, the stochastic system will keep the property. In Section 3, we will
introduce some numerical simulations to support our main result. Finally we give some conclusions and discussions.
2. Global stability
As the biological signicance of xt and yt in model (3), we should rstly give some conditions under which system (3)
has a unique global positive solution.
Lemma 2.1. Consider model (3), if ri > 0; i 1; 2, then there is a unique positive local solution xt; yt on 1; se almost
surely (a.s.) for any given initial value f1 ; f2 2 BC1; 0; R2 , where se denotes the explosion time and BC1; 0; R2
represents the family of bounded and continuous functions from 1; 0 to R2 with the norm kfi k suph60 jfi hj; i 1; 2.
Because the proof of this lemma is similar to that in [11], we omit it here.
Lemma 2.2. Consider system (3), there is a unique global solution xt; yt on R for any given initial value
f1 ; f2 2 BC1; 0; R2 and the solution will remain in R2 with probability 1, where R2 fx 2 R2 jxi > 0; i 1; 2g.
Proof. The following proof is motivated by the work of Liu and Wang [11]. Let
Mx; y
a14 yt
;
1 bxt cyt
Nx; y
b14 xt
:
1 bxt cyt
Pfsk 6 T g P e;
k P k1 :
Dene
p
p
Vx; y x 1 0:5 ln x y 1 0:5 ln y:
p
r2 2 x
xt hdl1 h Fx;y dt 1
x x 2 dt
8
1
Z 0
p
p
0:5r1 x 1x x dB1 t 0:5 y 1 r 2 b11 y b12 yt s2 b13
yt hdl2 h Gx;y dt
Z
p
dVx;y 0:5 x 1 r1 a11 x a12 xt s1 a13
1
r22 2 y
y y dt 0:5r2 y 1y y dB2 t
8
Z 0
p
p
xt hdl1 h Fx;y dt 0:5 yGx;y r 2 b11 y
6 0:5 xr 1 a11 x a12 xt s1 a13
2
1
5
3
b12 yt s2 b13
yt hdl2 h dt 0:125r21 x2 2x2 2x x2 2x 2 dt
1
5
p
p
3
0:125r22 y2 2y2 2y y2 2y 2 dt 0:5r1 x 1x x dB1 t 0:5r2 y 1y y dB2 t
5
p
a14
3
a12 kxk kf1 k a13 kxk kf1 k dt 0:125r21 x2 2x2 2x x2 2x 2 dt
6 0:5 xr 1 a11 x
c
p
5
b14 y
3
r 2 b11 y b12 kyk kf2 k b13 kyk kf2 k dt 0:125r22 y2 2y2 2y y2 2y 2 dt
0:5
b
p
p
0:5r1 x 1x x dB1 t 0:5r2 y 1y y dB2 t
p
p
6 K 1 K 2 dt 0:5r1 x 1x x dB1 t 0:5r2 y 1y y dB2 t;
Z
where K 1 and K 2 are positive constants. By the above inequality, we can obtain that
EV xsk ^ T;
ysk ^ T 6 V x0 ; y0 K 1 K 2 T:
The following proof is similar to Liu and Wang [11] and hence we omit it here.
Next, we will give our main result of this paper.
Theorem 2.1. Let
C 1 a14 bx 1;
A a11
C 2 b14 cy 1;
a14 by
a12 a13 0:5r21 x ;
1 bx cy
C1
b14 cx
2
:
b11
0:5
r
y
12
13
2
C2
1 bx cy
If
A < 0;
B > 0:
6
Then the equilibrium state x ; y of model (3) is globally stochastically asymptotically stable, i.e., for any given initial value
x0 ; y0 , the solution of system (3) has the property that
lim xt x ;
t!1
lim t y
t!1
LVx V x xf t; xt 0:5 trace g T t; xt V xx xgt; xt :
Dene
V 1 x x x x ln
x
;
x
V 2 y y y y ln
y
:
y
Obviously, these two functions are nonnegative on R . We prepare the detailed computation in Appendix A in order to
derive our main result directly. If xt; yt 2 R2 , making use of Its formula yields
LV 1 x 6 a11
a14 by
2
x x 2 0:5a12 xt s1 x 2
0:5a
0:5a
0:5
r
x
12
13
1
1 bx cy
Z 0
a14 bx 1
0:5a13
xt h x 2 dl1 h
x x y y :
1
bx
LV 2 y 6 b11
b14 cx
2
y y 2 0:5b12 yt s2 y 2
0:5b
0:5b
0:5
r
y
12
13
2
1 bx cy
Z 0
b14 cy 1
0:5b13
yt h y 2 dl2 h
x x y y :
1 bx cy1 bx cy
1
Dene
V 3 t 0:5a12
xs x 2 ds;
V 4 t 0:5a13
ts1
V 5 t 0:5b12
0
1
ys y 2 ds;
V 6 t 0:5b13
ts2
1
th
th
xs x 2 dsdl1 h;
ys y 2 dsdl2 h:
Then
LV 3 t 0:5a12 xt x 2 0:5a12 xt s1 x 2 ;
2
LV 4 t 0:5a13 xt x 0:5a13
1
xt h x 2 dl1 h;
LV 5 t 0:5b12 yt y 2 0:5b12 yt s2 y 2 ;
LV 6 t 0:5b13 yt y 2 0:5b13
0
1
yt h y 2 dl2 h:
Then dene
C1
C1
C1
V 2 y V 3 t V 4 t V 5 t V 6 t
C2
C2
C2
a14 bx 1
C1
C1
V 1 x
V 2 y V 3 t V 4 t V 5 t V 6 t:
b14 cy 1
C2
C2
Vx; y; t V 1 x
We obtain that
a14 by
LVx; y; t 6 a11
a12 a13 0:5r21 x x x 2
1 bx cy
C1
b14 cx
2
y y 2
b11
0:5
r
y
12
13
2
C2
1 bx cy
Ax x 2 By y 2 :
Set jZ Z j jx x j; jy y jT , then by the above inequality, we have
LVx; y; t 6 jZ Z jT
A 0
0
jZ Z j:
Obviously, if (6) holds then by virtue of the above inequality, we can conclude that LVx; y; t < 0 along all trajectories in
R2 except x ; y . Then the conclusion follows.
3. Numerical simulations
In this section, we will apply the Milstein method mentioned in Higham [7] to illustrate our main result. Set
f1 h 0:32eh ; f2 h 0:18eh ; l1 h l2 h eh on h 2 1; 0 and r1 0:5; r 2 0:4; a11 0:5; b11 0:2; a12 0:3;
b12 0:2; a13 0:1; b13 0:1; a14 0:7; b14 0:5; b c 1:0.
xk1 xk r1 a11 xk a12 xks1 =Dt 0:16a13 ekDt a13 ekDt Rki1 xi eiDt Dt a14
yk
Dt
1 bxk cyk
p
2
r2
r1 xk xk x Dtnk 1 xk xk x nk 1Dt;
2
yk1 yk r 2 b11 yk b12 yks2 =Dt 0:09b13 ekDt b13 ekDt Rki1 yi eiDt Dt b14
xk
Dt
k
k
1 bx cy
p
r2
2
r2 yk yk y Dtgk 2 yk yk y gk 1Dt;
2
where nk and gk k 1; . . . ; n are the Gaussian random variables which follow N0; 1.
In Figs. 1 and 2, we choose r1 0:5; r 2 0:4; a11 0:5; a12 0:3; a13 0:1; a 14 0:7; b11 0:2; b12 0:2; b13
0:1; b14 0:5; r1 r2 0:008; b c 1:0 and step size Dt 0:01, then (6) holds and x 2:0392; y 0:3514. The difference in Fig. 1 is that the values of s1 and s2 are different. For populations x1 and y1 , we choose s1 1; s2 10, while for
populations x2 and y2 , we choose s1 s2 0. That is to say, the equilibrium state x ; y is asymptotically stable. Obviously,
from Fig. 1 we can see that our model is more realistic than model (2), so our model owns the better property than model (2).
The difference in Fig. 2 is that the values of r1 and r2 are different. For populations x1 and y1 , we choose r1 0:18; r2
0:06, while for populations x2 and y2 , we choose r1 r2 0, then (6) is satised. In other words, the equilibrium state
x ; y is globally stochastically asymptotically stable.
the population x1
the population y1
the population x2
the population y2
10
10
15
Time
20
25
30
s2 10, while for populations x2 and y2 , we choose s1 s2 0. Then the equilibrium state x ; y is
3
2
1
0
1
the population x1
the population y1
the population x2
the population y2
2
3
10
15
Time
20
25
30
r2 0:06, while for populations x2 and y2 , we choose r1 r2 0. Then the equilibrium state
Z
LV 1 x x x r 1 a11 x a12 xt s1 a13
1
Z
x x a11 x x a12 x xt s1 a13
1
2
a14 y
0:5r21 x x x 2
1 bx cy
x xt hdl1 h Fx ;y Fx;y 0:5r21 x x x 2
xt hdl1 h
1
x x xt h x dl1 h
a14 by
a14 bx 1
x x 2
x x y y 0:5r21 x x x 2
1 bx cy1 bx cy
1 bx cy1 bx cy
Z 0
2
6 a11 x x 2 0:5a12 x x 2 0:5a12 xt s1 x 2 0:5a13 x x 2 0:5a13
xt h x dl1 h
1
a14 by
a14 bx 1
x x 2
x x y y 0:5r21 x x x 2 :
1 bx cy1 bx cy
1 bx cy1 bx cy
As 1 bx cy P 1, one can obtain that
LV 1 x 6 a11
a14 by
2
x x 2 0:5a12 xt s1 x 2
0:5a
0:5a
0:5
r
x
12
13
1
1 bx cy
Z 0
2
a14 bx 1x x y y
0:5a13
xt h x dl1 h
:
1
bx cy1 bx cy
1
Then the desired assertion (9) follows from the Hlders inequality:
1
xt h x dl1 h
6
1
xt h x 2 dl1 h:
LV 2 y 6 b11 0:5b12 0:5b13
0:5b13
1
b14 cx
2
y y 2 0:5b12 yt s2 y 2
0:5
r
y
2
1 bx cy
yt h y 2 dl2 h
b14 cy 1x x y y
:
1 bx cy1 bx cy
References
[1] C.A. Braumann, Variable effort harvesting models in random environments: generalization to density-dependent noise intensities, Math. Biosci. 177
(2002) 229245.
[2] J.R. Beddington, Mutual interference between parasites or predators and its effect on searching efciency, J. Anim. Ecol. 44 (1975) 331341.
[3] J.R. Beddington, R.M. May, Harvesting natural populations in a randomly uctuating environment, Science 197 (1977) 463465.
[4] D.L. DeAngelis, R.A. Goldsten, R. Neill, A model for trophic interaction, Ecology 56 (1975) 881892.
[5] T.C. Gard, Persistence in stochastic food web models, Bull. Math. Biol. 46 (1984) 357370.
[6] T.C. Gard, Stability for multispecies population models in random environments, Nonlinear Anal. 10 (1986) 14111419.
[7] D.J. Higham, An algorithmic introduction to numerical simulation of stochastic differential equations, SIAM Rev. 43 (2001) 525546.
[8] Y. Kuang, Delay Differential Equations with Applications in Population Dynamics, Academic Press, New York, 1993.
[9] X. Li, X. Mao, Population dynamical behavior of non-autonomous LotkaVolterra competitive system with random perturbation, Discrete Contin. Dyn.
Syst. 24 (2009) 523545.
[10] M. Liu, K. Wang, Global asymptotic stability of a stochastic LotkaVolterra model with innite delays, Commun. Nonlinear Sci. Numer. Simul. 17 (2012)
31153123.
[11] M. Liu, K. Wang, Global stability of a nonlinear stochastic predatorprey system with BeddingtonDeAngelis functional response, Commun. Nonlinear
Sci. Numer. Simul. 16 (2011) 11141121.
[12] M. Liu, K. Wang, Global stability of stage-structured predatorprey models with BeddingtonDeAngelis functional response, Commun. Nonlinear Sci.
Numer. Simul. 16 (2011) 37923797.
[13] Z.H. Liu, J.F. Han, L.J. Fang, Integral boundary value problems for rst order integro-differential equations with impulsive integral conditions, Comput.
Math. Appl. 61 (2011) 30353043.
[14] Z.H. Liu, J.F. Han, Boundary value problems for second order impulsive functional differential equations, Dyn. Syst. Appl. 20 (2011) 369382.
[15] Z.H. Liu, BrowderTikhonov regularization of non-coercive evolution hemivariational inequalities, Inverse Prob. 21 (2005) 1320.
[16] Z.H. Liu, Existence results for quasilinear parabolic hemivariational inequalities, J. Differ. Equ. 244 (2008) 13951409.
[17] Z.H. Liu, Anti-periodic solutions to nonlinear evolution equations, J. Funct. Anal. 258 (6) (2010) 20262033.
[18] Q. Luo, X. Mao, Stochastic population dynamics under regime switching, J. Math. Anal. Appl. 334 (2007) 6984.
[19] X. Mao, Stochastic stabilisation and destabilisation, Syst. Control Lett. 23 (1994) 279290.
[20] X. Mao, Stochastic Differential Equations and Applications, Horwood Publishing, Chichester, 1997.
[21] X. Mao, C. Yuan, J. Zou, Stochastic differential delay equations of population dynamics, J. Math. Anal. Appl. 304 (2005) 296320.
[22] R. Rudnicki, K. Pichor, Inuence of stochastic perturbation on preypredator systems, Math. Biosci. 206 (2007) 108119.
[23] G.T. Skalski, J.F. Gilliam, Functional responses with predator interference: viable alternatives to the Holling type II model, Ecology 82 (2001) 3083
3092.
[24] C. Zhu, G. Yin, On hybrid competitive LotkaVolterra ecosystems, Nonlinear Anal. 71 (2009) 13701379.