J. Math. Anal. Appl.

325 (2007) 542553

www.elsevier.com/locate/jmaa

Equilibrium points, stability and numerical solutions

of fractional-order predatorprey and rabies models
E. Ahmed a , A.M.A. El-Sayed b , H.A.A. El-Saka c,
a Faculty of Science, Mansoura University, Egypt
b Faculty of Science, Alexandria University, Alexandria, Egypt
c Damietta Faculty of Science, Mansoura University, New Damietta, Egypt

Received 16 October 2005

Available online 2 March 2006
Submitted by William F. Ames

Abstract
In this paper we are concerned with the fractional-order predatorprey model and the fractional-order
rabies model. Existence and uniqueness of solutions are proved. The stability of equilibrium points are
studied. Numerical solutions of these models are given. An example is given where the equilibrium point is
a centre for the integer order system but locally asymptotically stable for its fractional-order counterpart.
2006 Elsevier Inc. All rights reserved.

Keywords: Existence and uniqueness; Predatorprey system; Rabies system; Fractional-order differential equations;
Equilibrium points; Stability; Numerical solution; Predictorcorrector method

1. Introduction

Biology is a rich source for mathematical ideas. We argue that fractional-order differential
equations [18] are, at least, as stable as their integer order counterpart. The relation between
memory and fractional mathematics is pointed out. Then in Sections 25 sufficient conditions
for the local asymptotic stability of some biologically inspired, fractional non-autonomous equa-
tions are derived. An example is given whose internal solution is a centre, in the case of integer
order, and numerical solutions indicate that it is stable in the case of fractional order.

* Corresponding author.
E-mail addresses: magd45@yahoo.com (E. Ahmed), amasayed@maktoob.com (A.M.A. El-Sayed),
halaelsaka@yahoo.com (H.A.A. El-Saka).

0022-247X/$ see front matter 2006 Elsevier Inc. All rights reserved.
doi:10.1016/j.jmaa.2006.01.087
 E. Ahmed et al. / J. Math. Anal. Appl. 325 (2007) 542553 543

Definition 1. A complex adaptive system consists of inhomogeneous, interacting adaptive agents.

Definition 2. An emergent property of a CAS is a property of the system as a whole which does
not exist at the individual elements (agents) level. Typical examples are the brain, the immune
system, the economy, social systems, ecology, insects swarm, etc.

Therefore to understand a complex system one has to study the system as a whole and not
to decompose it into its constituents. This totalistic approach is against the standard reductionist
one, which tries to decompose any system to its constituents and hopes that by understanding the
elements one can understand the whole system.
Recently [19] it has became apparent that fractional equations solve some of the above men-
tioned problems for the PDE approach. To see this consider the following evolution equation:
t
df/dt = 2
k(t t  )f (t  ) dt  .
0

If the system has no memory then k(t t  ) = (t t  ). If the system has an ideal memory
then

1 if t > t  ,
k(t t  ) =
0 if t  > t.
Using Laplace transform, L[f ] = 1 if there is no memory and L[f ] = 1/s for perfect memory
hence the case of non-ideal memory is expected to be given by L[f ] = 1/s , 0 < < 1. In this
case the above system becomes
t
df/dt = 2 (t t  )1 f (t  ) dt  ,
0
 
f (t) = f0 E+1 2 t +1 ,

where E (z) = k=0 z / (k + 1) is Mittag-Leffler function.
k

It is also argued that there is a relevance between fractals and fractional differentiation [19].
Moreover since it is known that systems with memory are typically more stable than their mem-
oryless counterpart we expect the following conclusion:

Fractional order differential equations are, at least, as stable as their integer order counter-
part.

In the following sections both analytical and numerical results will be given which supports
this conclusion.
Now we give the definition of fractional-order integration and fractional-order differentiation:

Definition 3. The fractional integral of order R + of the function f (t), t > 0, is defined by
t
(t s)1
I f (t) =

f (s) ds (1)
()
0
544 E. Ahmed et al. / J. Math. Anal. Appl. 325 (2007) 542553

and the fractional derivative of order (n 1, n) of f (t), t > 0, is defined by

d
D f (t) = I n D n f (t), D = . (2)
dt
The following properties are some of the main ones of the fractional derivatives and integrals
(see [915,17]).
Let , R + and (0, 1). Then

+
(i) Ia : L1 L1 , and if f (x) L1 , then Ia Ia f (x) = Ia f (x).
x
(ii) limn Ia f (x) = Ian f (x) uniformly on [a, b], n = 1, 2, 3, . . . , where Ia1 f (x) = a f (s) ds.

(iii) lim0 Ia f (x) = f (x) weakly.
(iv) If f (x) is absolutely continuous on [a, b], then lim1 D f (x) = dfdx(x) .
(v) If f (x) = k = 0, k is a constant, then D k = 0.

The following lemma can be easily proved (see [14]).

Lemma 1. Let (0, 1) if f C[0, T ], then I f (t)|t=0 = 0.

2. Existence and uniqueness

Consider the fractional-order LotkaVolterra predatorprey system

 
D x1 (t) = x1 (t) r ax1 (t) bx2 (t) , t (0, T ], (3)
 
D x2 (t) = x2 (t) d + cx1 (t) , t (0, T ], (4)
with the initial values
x1 (t)|t=0 = x1 (0) and x2 (t)|t=0 = x2 (0), (5)
where 0 <  1, x1  0, x2  0 are prey and predator densities, respectively, and all constants
r, a, b, c and d are positive.

Lemma 2. The initial value problem (3)(5) can be written in the form
D X(t) = A1 X(t) x1 (t)A2 X(t), t (0, T ] and X(0) = X0 , (6)
where
   
x1 (t) r 0 a b x1 (0)
X(t) = , A1 = , A2 = , and X0 = .
x2 (t) 0 d 0 c x2 (0)

Definition 4. Let C [0, T ] be the class of continuous column vector X(t) whose compo-
nents x1 , x2 C[0, T ], the class of continuous functions on the interval [0, T ]. The norm of
X C [0, T ] is given by

2
 
X = supxi (t).
t
i=1

Definition 5. By a solution of the initial value problem (6) we mean a column vector
X C [0, T ]. This vector satisfies the system (6).
 E. Ahmed et al. / J. Math. Anal. Appl. 325 (2007) 542553 545

Now we have the following existence theorem.

Theorem 1. The initial value problem (6) has a unique solution.

Proof. The proof follows from Theorems 2.1 and 2.2 of [6]. 2

3. Equilibrium points and their asymptotic stability

Let (0, 1] and consider the system

D x1 (t) = f1 (x1 , x2 ),
D x2 (t) = f2 (x1 , x2 ), (7)
with the initial values
x1 (0) = x01 and x2 (0) = x02 . (8)
To evaluate the equilibrium points, let
 eq eq 
D xi (t) = 0 fi x1 , x2 = 0, i = 1, 2,
eq eq
from which we can get the equilibrium points x1 , x2 .
To evaluate the asymptotic stability, let
eq
xi (t) = xi + i (t),
then
 eq   eq eq 
D xi + i = fi x1 + 1 , x2 + 2
which implies that
 eq eq 
D i (t) = fi x1 + 1 , x2 + 2
but
 
 eq   eq eq  fi  
eq
fi x1 + 1 , x2 + 2
fi x1 , x2 +  1 + fi  2 +

x1 eq x2 eq
 
 eq  fi  
eq
fi x1 + 1 , x2 + 2
 1 + fi  2 ,
x  x 
1 eq 2 eq
eq eq
where fi (x1 , x2 ) = 0, then
 
fi  fi 
D i (t)
+ 2
x1 eq x2 eq
1

and we obtain the system

D = A (9)
with the initial values
eq eq
1 (0) = x1 (0) x1 and 2 (0) = x2 (0) x2 , (10)
where
 
1 a11 a12
= , A=
2 a21 a22
546 E. Ahmed et al. / J. Math. Anal. Appl. 325 (2007) 542553

and

fi 
aij = , i, j = 1, 2.
xj eq
We have
B 1 AB = C,
where C is a diagonal matrix of A given by

1 0
C= ,
0 2
where 1 and 2 are the eigenvalues of A and B is the eigenvectors of A, then

AB = BC, A = BCB 1 ,
which implies that
     
D = BCB 1 , D B 1 = C B 1 ,
then
D = C, = B 1 , (11)
where

1
= ,
2
i.e.

D 1 = 1 1 , (12)
D 2 = 2 2 , (13)
the solutions of Eqs. (12)(13) are given by Mittag-Leffler functions (see [10])


(1 )n t n  
1 (t) = 1 (0) = E 1 t 1 (0), (14)
(n + 1)
n=0

(2 )n t n  
2 (t) = 2 (0) = E 2 t 2 (0). (15)
(n + 1)
n=0

Using the result of Matignon [16] then if

   
arg(1 ) > /2 and arg(2 ) > /2

then 1 (t), 2 (t) are decreasing and then 1 (t), 2 (t) are decreasing.
eq eq
So the equilibrium point (x1 , x2 ) is locally asymptotically stable if both the eigenvalues of
the matrix A are negative (|arg(1 )| > /2, |arg(2 )| > /2). This confirms our statement in
Section 1 that fractional-order differential equations are, at least, as stable as their integer order
counterpart.
 E. Ahmed et al. / J. Math. Anal. Appl. 325 (2007) 542553 547

4. Fractional-order LotkaVolterra predatorprey model

Consider the fractional-order LotkaVolterra predatorprey system

 
D x1 (t) = x1 (t) r ax1 (t) bx2 (t) , (16)
 
D x2 (t) = x2 (t) d + cx1 (t) . (17)
To evaluate the equilibrium points, let
D xi (t) = 0, i = 1, 2,
eq eq
then (x1 , x2 ) = (0, 0), ( ar , 0), ( dc , crad
cb ), are the equilibrium points.
eq eq
For (x1 , x2 ) = (0, 0) we find that

r 0
A= ,
0 d
its eigenvalues are

1 = r > 0, 2 = d < 0.
eq eq
Hence the equilibrium point (x1 , x2 ) = (0, 0) is unstable.
eq eq
For (x1 , x2 ) = ( ar , 0) we find that

r br
A= a ,
0 cra d
its eigenvalues are
cr
1 = r < 0, 2 = d <0 if cr < ad.
a
eq eq
Hence the equilibrium point (x1 , x2 ) = ( ar , 0) is locally asymptotically stable if cr < ad.
eq eq
For (x1 , x2 ) = ( dc , crad
cb ) we find that
 ad
c bd c
A = crad ,
b 0
its eigenvalues are

ad + a 2 d 2 4cd(cr ad)
1 = ,
2c

ad a 2 d 2 4cd(cr ad)
2 = .
2c
eq eq
A sufficient condition for the local asymptotic stability of the equilibrium point (x1 , x2 ) =
cb ) is |arg(1 )| > /2, |arg(2 )| > /2.
( dc , crad
In the special case a = 0 it is known that the internal equilibrium point is a centre (arg(1 ) =
/2, arg(1 ) = /2) for the integer order system ( = 1). In the fractional case 0 < < 1 the
internal equilibrium point is locally asymptotically stable. The numerical simulations in the next
section will support this result.
548 E. Ahmed et al. / J. Math. Anal. Appl. 325 (2007) 542553

Fig. 1.

Fig. 2.

5. Fractional-order rabies model

In Eqs. (16) and (17) when we take r = 0, a = 0, c = b, we obtain the fractional-order rabies
model

D x1 (t) = bx1 x2 , (18)

D x2 (t) = bx1 x2 dx2 , (19)

 E. Ahmed et al. / J. Math. Anal. Appl. 325 (2007) 542553 549

Fig. 3.

Fig. 4.

where 0 <  1, x1  0, x2  0 are healthy and infected foxes, respectively, and all constants
eq eq eq eq
b, d are positive and (x1 , x2 ) = (0, 0), ( db , 0) are the equilibrium points. For (x1 , x2 ) = (0, 0)
we find that

0 0
A= ,
0 d
its eigenvalues are

1 = 0, 2 = d.
550 E. Ahmed et al. / J. Math. Anal. Appl. 325 (2007) 542553

Fig. 5.

Fig. 6.

eq eq
For (x1 , x2 ) = ( db , 0) we find that

0 d
A= ,
0 0

its eigenvalues are

1 = 0, 2 = 0.

Hence the equilibrium is a centre.

 E. Ahmed et al. / J. Math. Anal. Appl. 325 (2007) 542553 551

Fig. 7.

Fig. 8.

6. Numerical methods and results

An Adams-type predictorcorrector method has been introduced in [2,3] and investigated

further in [1,49,13]. In this paper we use an Adams-type predictorcorrector method for the
numerical solution of fractional integral equation.
The key to the derivation of the method is to replace the original problem (3)(5) by an
equivalent fractional integral equation
 
X(t) = X0 + I A1 X(t) x1 (t)A2 X(t) (20)
and then apply the PECE (Predict, Evaluate, Correct, Evaluate) method.
552 E. Ahmed et al. / J. Math. Anal. Appl. 325 (2007) 542553

The approximate solutions displayed in Figs. 18 for the step size 0.05 and different
0 <  1. In Fig. 1 we take b = 1, c = 1, r = 2, d = 3, a = 1, x1 (0) = 0.15 and x2 (0) = 0.33.
In Fig. 2 we take r = 0, a = 0, b = 1, d = 1, x1 (0) = 1 and x2 (0) = 2.5. In Figs. 38 we take
a = 0, b = 1, c = 1, r = 2, d = 3, x1 (0) = 1 and x2 (0) = 2. In Figs. 3 and 6 we take = 1. In
Figs. 4 and 7 we take = 0.9. In Figs. 5 and 8 we take = 0.8.

7. Conclusion

Existence and uniqueness of solutions of fractional order systems have been studied. We have
argued that fractional-order differential equations are, at least, as stable as their integer order
counterpart. We studied equilibrium points, existence, uniqueness, stability, numerical solution,
of LotkaVolterra predatorprey system and used numerical solutions to show that although the
internal solution for integer order case is only a centre, it is stable for its fractional-order coun-
terpart.

References

[1] E. Ahmed, A.M.A. El-Sayed, E.M. El-Mesiry, H.A.A. El-Saka, Numerical solution for the fractional replicator
equation, IJMPC 16 (7) (2005) 19.
[2] K. Diethelm, A. Freed, On the solution of nonlinear fractional order differential equations used in the modelling of
viscoplasticity, in: F. Keil, W. Mackens, H. Vo, J. Werther (Eds.), Scientific Computing in Chemical Engineering
IIComputational Fluid Dynamics, Reaction Engineering, and Molecular Properties, Springer, Heidelberg, 1999,
pp. 217224.
[3] K. Diethelm, A. Freed, The FracPECE subroutine for the numerical solution of differential equations of fractional
order, in: S. Heinzel, T. Plesser (Eds.), Forschung und wissenschaftliches Rechnen 1998, Gesellschaft fr Wiss-
eschaftliche Datenverarbeitung, Gttingen, 1999, pp. 5771.
[4] K. Diethelm, Predictorcorrector strategies for single- and multi-term fractional differential equations, in: E.A. Lip-
itakis (Ed.), Proceedings of the 5th HellenicEuropean Conference on Computer Mathematics and Its Applications,
LEA Press, Athens, 2002, pp. 117122 [Zbl. Math. 1028.65081].
[5] K. Diethelm, N.J. Ford, A.D. Freed, A predictorcorrector approach for the numerical solution of fractional differ-
ential equations, Nonlinear Dynam. 29 (2002) 322.
[6] K. Diethelm, N.J. Ford, Analysis of fractional differential equations, J. Math. Anal. Appl. 265 (2002) 229248.
[7] K. Diethelm, N.J. Ford, A.D. Freed, Detailed error analysis for a fractional Adams method, Numer. Algorithms 36
(2004) 3152.
[8] K. Diethelm, N.J. Ford, Multi-order fractional differential equations and their numerical solution, Appl. Math.
Comput. 154 (2004) 621640.
[9] E.M. El-Mesiry, A.M.A. El-Sayed, H.A.A. El-Saka, Numerical methods for multi-term fractional (arbitrary) orders
differential equations, Appl. Math. Comput. 160 (3) (2005) 683699.
[10] A.M.A. El-Sayed, Fractional differential-difference equations, J. Fract. Calc. 10 (1996) 101106.
[11] A.M.A. El-Sayed, Nonlinear functional differential equations of arbitrary orders, Nonlinear Anal. 33 (2) (1998)
181186.
[12] A.M.A. El-Sayed, F.M. Gaafar, Fractional order differential equations with memory and fractional-order relaxation-
oscillation model, (PU.M.A) Pure Math. Appl. 12 (2001).
[13] A.M.A. El-Sayed, E.M. El-Mesiry, H.A.A. El-Saka, Numerical solution for multi-term fractional (arbitrary) orders
differential equations, Comput. Appl. Math. 23 (1) (2004) 3354.
[14] A.M.A. El-Sayed, F.M. Gaafar, H.H. Hashem, On the maximal and minimal solutions of arbitrary orders nonlinear
functional integral and differential equations, Math. Sci. Res. J. 8 (11) (2004) 336348.
[15] R. Gorenflo, F. Mainardi, Fractional calculus: Integral and differential equations of fractional order, in: A. Carpinteri,
F. Mainardi (Eds.), Fractals and Fractional Calculus in Continuum Mechanics, Springer, Wien, 1997, pp. 223
276.
 E. Ahmed et al. / J. Math. Anal. Appl. 325 (2007) 542553 553

[16] D. Matignon, Stability results for fractional differential equations with applications to control processing,
Computational Eng. in Sys. Appl., Vol. 2, Lille, France 963, 1996.
[17] I. Podlubny, A.M.A. El-Sayed, On Two Definitions of Fractional Calculus, Slovak Academy of Science, Institute
of Experimental Phys., ISBN 80-7099-252-2, 1996, UEF-03-96.
[18] I. Podlubny, Fractional Differential Equations, Academic Press, 1999.
[19] A.A. Stanislavsky, Memory effects and macroscopic manifestation of randomness, Phys. Rev. E 61 (2000) 4752.