Download as pdf or txt
Download as pdf or txt
You are on page 1of 11

Stability Analysis

of
Two prey and One predator model

Sumi Dey
Ausmita Barman
Mohammad Arifur Rahman

University of Texas at El Paso

Friday 26th June, 2015

The model
In this section, we propose a system consists of two teams of preys with densities x(t), y(t), respectively,
interacting with one team of predator with densities z(t). The assumptions of this model are as follows:
In the absence of any predation, each team of preys grows logistically; this is ax(t)(1-x(t)) and
ay(t)(1-y(t)).
The effect of the predation is to reduce the prey growth rate by a term proportional to the prey and
predator populations; this is the -x(t)z(t) and -y(t)z(t) terms.
The teams of preys help each other against the predator, that is a x(t)y(t)z(t) term exist.
In the absence of any prey for sustenance, the predators death rate results in inverse decay, that is
the term cz 2 (t).
The preys contribution to the predator growth rate are dx(t)z(t), ey(t)z(t); that is proportional to the
available prey as well as the size of the predator population.
Using the above assumptions, the following model is proposed:
dx(t)
dt = ax(t)(1 x(t)) x(t)z(t) + x(t)y(t)z(t)
dy(t)
dt = by(t)(1 y(t)) y(t)z(t) + x(t)y(t)z(t)
dz(t)
2
dt = cz (t) + dx(t)z(t) + ey(t)z(t)

(1)

where the coefficients a, b, c, d and e are positive constants and x(0),y(0), z(0) > 0. It is clear that the two
teams of preys help each other e.g. in foraging and in early warning against predation. Note that this help
occurs only in the presence of predator. This is presented by the term x(t)y(t)z(t) in the preys equations.

The analysis of the model


By Solving the equations,
dx(t)
= ax(t)(1 x(t)) x(t)z(t) + x(t)y(t)z(t) = 0
dt

(2)

dy(t)
= by(t)(1 y(t)) y(t)z(t) + x(t)y(t)z(t) = 0
dt

(3)

dz(t)
= cz 2 (t) + dx(t)z(t) + ey(t)z(t) = 0
dt

(4)

From (7) we get,


x = 0 or a(1 x) z + yz = 0
For x = 0 (8) becomes,
y = 0 or b(1) z = 0
For x = 0, y = 0 (9) becomes,
z=0
E0 (0, 0, 0)
For x = 0 (8) becomes,
b(1 y) z = 0 y =
Now (9) becomes,
z = 0 or cz +

bz
b

eb ez
=0
b

E1 (0, 1, 0)
For x = 0, y =

bz
b

(9) becomes,
cz +

Now y =

eb ez
eb
=0z=
b
e + bc

cb
e+bc

E2 (0,

cb
eb
,
)
e + bc e + bc

From (7) we get,


a(1 x) z + yz = 0 x =
Now (8)becomes,
y = 0 or (1 y) z +

a z + yz
a

a z + yz
z=0
a
5

From (9) we get,


z = 0 or cz +

da dz
=0
a

E3 (1, 0, 0)
az
a ,y

For x =

= 0 Now (9) becomes,


cz +

da
da dz
=0z=
a
ca + d

E4 (

ca
da
, 0,
)
ca + d
ca + d

Now (8)becomes,
b(1 y) z +

a z + yz
z2
z2
z = 0 y( b) = ( b)
a
a
a

If ( za b) 6= 0 then y = 1
q
z2
If ( a b) = 0 then z = ab
For y = 1, x = 1 From (9) we get,
z = 0 or cz + d + e = 0
E5 (1, 1, 0)
For x = 1, y = 1 (9) becomes,
z=

d+e
c

E6 (1, 1,
For z =

b
a

d+e
)
c

we get infinitely many values of y.

Let y = where is a positive parameter.


From (7)

x=1+

E7 (1 +

If we assume,y = =
under the conditions

c abd(1 ab )
q
e+d ab

ab( 1)
a

ab( 1)
, , ab)
a

then x =

q
cb+e(1 ab )
q
e+d ab

r
b
b
cb + e(1
) > 0 cb + e > e
a
a
r
r

b
b
c ab d(1
) > 0 d c ab < d
a
a
Definitions: First, we define some types of stability for solutions of the equation:
dx(t)
= f (x, t)...................()
dt

Definition 1: A solution x
(t) of (*) is (Lyapunov) stable if for each  > 0 and t0 R there exists
= (, t0 ) > 0 such that if x(t) is a solution of (*) and |x(t0 ) x
(t0 )| <  then |x(t) x
(t)| < for
all t t0 .
Definition 2: A solution x
(t) of (*) is asymptotically stable if it is (Lyapunov) and if for every t0 R there
exists = (t0 ) > 0 such that if x(t) is a solution of (*) and |x(t0 ) x
(t0 )| < then |x(t) x
(t)| 0
for all t .
Definition 3: A solution x
(t) of (*) is uniformly (Lyapunov) stable if for each  > 0 and t0 R there
exists = () > 0 such that if x(t) is a solution of (*) and |x(t0 ) x
(t0 )| <  then |x(t) x
(t)| <
for all t t0 .
Proposition 1: Local stability analysis [18] shows that the system (6) has four unstable equilibrium
solutions E0 , E1 , E2 and E3 .
Proof: The Jaconian matrix of the above system (6) is given by:

a(1 2x) z(1 y)


xz
x(1 y)

J =
yz
b(1 2y) z(1 x)
y(1 x)

dz
ez
2cz + dx + ey
Substituting by the point E0 (0, 0, 0) in the above matrix,

a 0 0

JE0 = 0 b 0
0 0 0
det(JE0 I) = 0 = 0, b, a
which has two positive eigenvalues. So, it is unstable equilibrium point.
Similarly,

a 0 1

JE3 = 0 b 0
0
0
d
det(JE3 I) = 0 = d, b, a
which has two positive eigenvalues. So, it is unstable equilibrium point.
Similarly,

a 0
0

JE1 = 0 b 1
det(JE1

0 0
e
I) = 0 = a, b, e

E1 (0, 1, 0) has two positive eigenvalues. So, it is unstable equilibrium point.


By the same way,

a 0

JE5 = 0 b
det(JE5

0
0
0 d+e
I) = 0 = a, b, d + e

E5 (0, 1, 0) has two positive eigenvalues. So, it is unstable equilibrium point.


So, all of them are unstable equilibrium points.
Now,
a

JE2 =

cb
eb
bc+e (1 cb+e )
2
ceb
(cb+e)2
ebd
cb+e

0
b(1

2cb
cb+e )
be2
cb+e

eb
bc+e

cb
cb+e

2ceb
cb+e
+

ceb
cb+e

det(JE2 I) = 0
=a

be2
(cb+e)2

or 2 + ( bce+cb
bc+e ) +

ceb2
(bc+e)

=0

The equilibrium solution E4 is locally asymptotically stable under the condition,

a<

be2
(cb + e)2

Numerical simulations agreed with these results. Let a = 1.0, b = 2.0, c = 0.1, d = 1.0, e = 1.4, we get the
stable solution (0, 0.125, 1.75) as shown in Fig. 1.

Fig. 1. In this figure we used the values a = 1.0, b = 2.0, c = 0.1, d = 1.0, and e = 1.4. So, we get the stable
solution (x = 0, y = 0.125, z = 1.75).
Similarly

a(1

JE4 =

2ca
ca+d )

ad
ca+d

0
ad2
ca+d

cda2
(ca+d)2
ad
ca
ca+d (1 ca+d )
ead
ca+d

ca
ca+d

2cad
ca+d
+

cad
ca+d

det(JE4 I) = 0
=b

ad2
(ca+d)2

or 2 + ( cad+ca
ca+d ) +

dca2
(ca+d)

=0

The equilibrium solution E4 is locally asymptotically stable under the condition,

b<

ad2
(ca + d)2

Numerical simulations agreed with these results. Let a = 1.0, b = 0.4, c = 0.3, d = 1.0, e = 1.4, we get the
stable solution (0.231, 0, 0.769) as shown in Fig. 2.

Fig. 2. In this figure we used the values a = 1.0, b = 0.4, c = 0.3, d = 1.0, and e = 1.4. So, we get the stable
solution (x = 0.231, y = 0, z = 0.769).

d+e
0
a
c

Now JE6 = d+e


b
0
c
d2 +ed
c

ed+e2
c

d e

det(JE6 I) = 0
2
The first internal equilibrium solution E6 is locally asymptotically stable if ab > ( d+e
c ) . Numerical

simulations agreed with these results. Let a = 1.0, b = 2.0, c = 2.0, d = 1.0, e = 1.6, we get the stable
solution (1, 1, 1.3) as shown in Fig. 3.

Fig. 3. In this figure we used the values a = 1.0, b = 2.0, c = 2.0, d = 1.0, and e = 1.6. So, we get the stable

solution (x = 1, y = 1, z = 1.3).
The second internal equilibrium solution

ab( 1)
, , ab)
a

a + 3 ab( 1)
ab + b b
a + ab + ab2 2 ab

ab ab2
JE7 =

ab
b
a

d ab
e ab
2ac ab + ad + d ab d ab
det(JE7 I) = 0
E7 (1 +

We get the characteristic equation,


3 + a1 2 + a2 + a3 = 0

where

a1 = a 3 ab + 3 ab + b + 2ac ab ad b ab d ab

a2 = ab 3 abb2 + 3b ab + 2ca2 ab a2 d ab ab abad 6a2 cb 3ab ab + 32 b2 a +

3dab + 6a2 cb 3ad ab 3b2 a 3abd ab 2 b ab + b ab + ad ab dab 2 abd +

2abd + 2abc ab abd bd ab bd ab 2b + 2b + 2b2

a3 = b(a + 3 ab 3 ab)(2ac ab ad d ab d ab) + 2b a + 3 ab( 1) ( ab +

b b)db(b b2 ) (a + ab + ab2 2 ab)eab (a + ab + ab2 2 ab)bd ab +


( ab + b b)(2ac ab + ad + d ab d ab) ab
E7 is locally asymptotically stable under the Routh-Hurwitz conditions: a1 > 0, a3 > 0anda1 a2 >
a3 .Numerical simulations agreed with these results. Let a = 1.2, b = 1.4, c = 1.0, d = 1.0, e = 2.0, we
get the stable solution (0.40, 0.45, 1.29). Also, let a = 1.0, b = 1.44, c = 1.0, d = 1.0, e = 1.2, we get the
internal solution (0.51, 0.59, 1.23) as shown in Figs. 4 and 5.

Fig. 4. In this figure we used the values a = 1.2, b = 1.4, c = 1.0, d = 1.0, and e = 2.0. So, we get the

internal stable solution (x = 0.40, y = 0.45, z = 1.29).

Fig. 5. In this figure we used the values a = 1.0, b = 1.44, c = 1.0, d = 1.0, and e = 1.2. So, we get the
internal stable solution (x = 0.51, y = 0.59, z = 1.23).

You might also like