Professional Documents
Culture Documents
Abstract. First, we prove the uniform persistence for discrete model An+1 =
An Fn (An , An1 , , Anm ) of population growth, where Fn : (0, )m+1
(0, ) are continuous all. Second, we investigation the effect of delay m on
the global attractivity of the unique positive equilibrium.
1. Introduction
Consider the model
(1.1)
n = 0, 1, ,
where Fn : (0, )m+1 (0, ) are continuous all. This model is potentially
appeared in medicine (for example, the population of blood cells) and was investigated by several authors [Graef, Liz, Tkachenko et al.] with more restrictions
on Fn . If Fn (x, y) = exp( x y) with , > 0 we get back a model investigated by Tkachenko et al. (But they found no explicit conditions for the global
attractivity of the positive equilibrium.) A positive solution {An }
n=m is called
persistent if
0 < lim inf An 6 lim sup An < .
n
(1.2)
lim inf
min
Fn (x0 , x1 , , xm ) > 0
lim sup
n,x0 ,x1 , ,xm
Fn (x0 , x1 , , xm ) < 1,
300
DANG VU GIANG
(1.5)
lim inf
Fn (x0 , x1 , , xm ) > 1.
(1.6)
(1.7)
(1.8)
Now take lim sup on both sides we have lim supn Akn < which contradicts
(1.6). Thus, {An }
n=m is bounded from above. Let K be an upper bound of
{An }
.
n=m
Next, we prove that lim inf n An > 0. Assume, for the sake of a contradiction, that lim inf n An = 0. For each integer n m, we define
sn := max{ : m 6 6 n, A =
min Ai }.
m6i6n
lim Asn = 0.
inf
min
FN (x0 , , xm ) > 0,
so
(1.10)
lim Asn 1 = 0.
301
Let n0 > 0 such that sn0 > 0. We have for any n > n0 ,
Asn = Asn 1 Fsn 1 (Asn 1 , , Asn 1m ) Asn Fsn 1 (Asn 1 , , Asn 1m )
and therefore,
Fsn 1 (Asn 1 , , Asn 1m ) 6 1.
By (1.5) and (1.10), this implies that
lim inf max{Asn 2 , , Asn 1m } > 0.
n
Now take lim inf as n on both sides we have lim inf n Asn > 0 which
contradicts (1.9) The proof is complete.
2. The global attractivity
In this section we assume that there is a unique positive equilibrium x
of (1.1)
and
1 = Fn (
x, , x
),
(2.1)
all. Then Fn0 (An0 , An0 m+1 , , An0 ) 6 1, so An0 +1 6 An0 . Similarly, An+1 6 An
for all n n0 , . Therefore, there is a limit of {An }
n=m . This limit is exactly
x
.
To investigate the effect of delay, we suppose further that
n x x
xm o
0
1
(2.2) lim sup | ln Fn (x0 , x1 , , xm )| 6 L max ln , ln , , ln
x
n
for all x0 , x1 , , xm > 0.
302
DANG VU GIANG
Suppose further
for every
t tn 2m.
Then
A
n A
Atm o
t+1
t
6 L1
ln
= | ln Ft (At , , Atm )| 6 L max ln , , ln
At
x
A
A
t+1
t
t=s
t=s
for all s [t1 m, t ]. This is right because the last sum is of (t + 1 s) terms
and each of them is 6 L1 . Furthermore,
A
n A
Atm o
t
t+1
= | ln Ft (At , , Atm )| 6 L max ln , , ln
ln
At
x
6 L2 1 (t + m + 1 t)
2
2
for all
t > t1 + m.
At > x
,
o
At
ln > 1 L(m + 3 ) 1 .
x
2
2
303
At
At
t=t
t=t
1
t +m[ L
]
L1 +
t=t
tX
+m
L2 1 (t + m + 1 t)
1
t=t +m[ L
]+1
1
1
1 1
) + 1 L2 [ ]([ ] + 1)
L
2
L L
1
3
6 1 L(m + ) .
2
2
= L1 m + 1 [
and consequently,
(2.3)
At
At
t=t
t=t
1
1
]
t1 +m[ L
L1 +
t=t1
tX
1 +m
L2 1 (t1 + m + 1 t)
1
t=t1 +m[ L
]+1
1
1 1
1
) + 1 L2 [ ]([ ] + 1)
L
2
L L
3
1
6 1 L(m + ) .
2
2
= L1 m + 1 [
2
2
This result permits us to choose
for all
t > t1 + m.
3
1
2 = 1 L(m + ) .
2
2
Repeat the above argument (with t1 and 1 replaced by t2 and 2 ) we have
At
ln 6 2 L(m + 3 ) 1
for all t > t2 + m.
x
2
2
Using the assumption L(m + 32 ) 12 < 1, we complete the proof.
304
DANG VU GIANG
3. Application
A tipical example is the equation
An+1 = An exp( An An1 ).
Here m = 1 and we easily compute
x
=
,
+
L = e2 .
and
m
X
j,n =
is fixed.
j=1
We easily compute
1
1
, L=
.
Hence, if (m + 32 ) 1
<
3
2
Acknowledgement
The author wish to express his sincere thanks to the referee for various advices
and suggestions.
References
[1] D. V. Giang, Persistence and global attractivity in the model An+1 = An F (Anm ), (submitted).
[2] J. R. Graef, C. Qian and P. W. Spikes, Stability in a population model, Appl. Math.
Comput. 89 (1998), 119132.
[3] J. R. Graef and C. Qian, Global attractivity in a nonlinear difference equation and application, Dynam. Systems Appl. 15 (2006), 8996.
[4] E. Liz, A sharp global stability result for a discrete population model, J. Math. Anal.
Appl. 330 (2007), 740743.
[5] E. Liz, V. Tkachenko and S. Trofimchuk, Global stability in discrete population models
with delayed-sensity dependence, Math. Biosci. 199 (2006), 2637.
[6] V. Tkachenko and S. Trofimchuk, A global attractivity criterion for nonlinear nonautonomous difference equation, J. Math. Anal. Appl. 322 (2006), 901912.
Hanoi Institute of Mathematics
18 Hoang Quoc Viet, 10307 Hanoi, Vietnam
E-mail address: dangvugiang@yahoo.com