L. Yao / Applied Mathematical Modelling 64 (2018) 185–195 187
2. Preliminary results
In this section, a few Lemmas, notations and definitions are introduced which will be used in Section 3 . Designate
R
+
= [0 , + ∞ ) , C = C([ −r, 0] , R ) , C
+
= C([ −r, 0] , R
+
) ,
C
0
(R
+
, J ) = { ν : ν ∈ C(R
+
, J ) , lim
t→ + ∞
ν(t) = 0 } ,
and BC(J
1
, J
2
) be the collection of all continuous and bounded functions from J
1
to J
2
, where J , J
1
, J
2
⊆ R . In addition, for
ψ ∈ BC(R , R ) ,
ψ
+
= sup
t∈ R
ψ (t) , ψ
−
= inf
t∈ R
ψ (t) .
Throughout this paper, it will be assumed that
r = max
j∈ J
σ
+
j
> 0 , lim inf
t→ + ∞
γ
j
(t) ≥ 1 , j ∈ J, (2.1)
which is a weaker condition than (1.4) in [32] and [33] . If N ( t ) is continuous and defined on [ −r + t
0
, σ ) with t
0
, σ ∈ R ,
then we define N
t
∈ C where N
t
(θ ) = N(t + θ ) for all θ ∈ [ −r, 0] .
Definition 2.1. (See [21,22] ). Q ⊆ R is called to be relatively dense in R provided that there is a number l > 0 such that
[ t, t + l]
Q = ∅ ( t ∈ R ). u ∈ BC (R , J ) is called to be almost periodic on R provided that, for any > 0, the collection
T (u, ) = { τ : | u (t + τ ) − u (t) | < , ∀ t ∈ R } is relatively dense.
Definition 2.2. (See [21,22] ). u ∈ C(R
+
, J ) is called to be asymptotically almost periodic provided that u = h + g, where g ∈
C
0
(R
+
, J ) and h is an almost periodic function.
For J ⊆ R , we designate the collections of the almost periodic functions and the asymptotically almost periodic functions
from R to J by AP (R , J ) and AAP (R , J ) , respectively. In particular, the decomposition involved in Definition 2.2 is single
[21,22] .
Hereafter, we presume that a, b, β
j
, γ
j
∈ AAP (R , (0 , + ∞ )) , σ
j
∈ AAP (R , R
+
) and
a = a
h
+ a
g
, b = b
h
+ b
g
, β
j
= β
h
j
+ β
g
j
, γ
j
= γ
h
j
+ γ
g
j
, σ
j
= σ
h
j
+ σ
g
j
,
where a
h
, b
h
, β
h
j
, γ
h
j
∈ AP (R , (0 , + ∞ )) , σ
h
j
∈ AP (R , R
+
) , a
g
, b
g
, β
g
j
, γ
g
j
, σ
g
j
∈ C
0
(R
+
, R
+
) , and j ∈ J .
For further analysis, we set up the following nonlinear auxiliary equation:
N
(t) = −
a
h
(t ) N(t )
b
h
(t) + N(t )
+
m
j=1
β
h
j
(t ) N(t −σ
h
j
(t )) e
−γ
h
j
(t) N(t−σ
h
j
(t))
, t ≥ t
0
. (1.2)
0
The initial condition associated with system (1.2) and (1.2)
0
can be described as follows:
N
t
0
= ϕ, ϕ ∈ C
+
and ϕ(0) > 0 . (2.2)
Lemma 2.1. Designate N ( t ; t
0
, ϕ)( N
t
( t
0
, ϕ)) for a solution of the initial value problem (1.2)
0
and (2.2) . Presume that (2.1) is
satisfied, and there is a constant M > 0 satisfying (1.7) , (1.9) and
sup
t∈ [ t
0
, + ∞ )
−
a
h
(t) M
b
h
(t) + M
+
1
e
m
j=1
β
h
j
(t)
γ
h
j
(t)
< 0 . (2.3)
Then, N ( t ; t
0
, ϕ)) is bounded on [ t
0
, + ∞ ) , and there is t
ϕ
∈ [ t
0
, + ∞ ) such that
κ < N(t;t
0
, ϕ) < M for all t ∈ [ t
ϕ
, + ∞ ) . (2.4)
Proof. For ϕ ∈ C
+
, from Theorem 5.2.1 in [ 34 , p 81], we have N
t
(t
0
, ϕ) ∈ C
+
for all t ∈ [ t
0
, η( ϕ)), where [ t
0
, η( ϕ)) is the
maximal right existence interval of N
t
( t
0
, ϕ). Let N(t ) = N(t;t
0
, ϕ) . In view of the fact that
N(t
0
) = ϕ(0) > 0 , N
(t) ≥−
a
h
(t)
b
h
(t)
N(t) +
m
j=1
β
h
j
(t) N(t − δ
h
j
(t)) e
−γ
h
j
(t) N(t−δ
h
j
(t))
, (2.5)
we obtain
N(t) ≥ e
−
t
t
0
a
h
(u )
b
h
(u )
du
N(t
0
) + e
−
t
t
0
a
h
(u )
b
h
(u )
du
t
t
0
e
s
t
0
a
h
(v )
b
h
(v )
dv
m
j=1
β
h
j
(s ) N(s − σ
h
j
(s )) e
−γ
h
j
(s ) N(s −σ
h
j
(s ))
ds
> 0 , for all t ∈ [ t
0
, η(ϕ)) . (2.6)
For each t ∈ [ t
0
− r, η(ϕ)) , define
(t ) = max { ξ : ξ ≤ t, N(ξ ) = max
t
0
−r≤s ≤t
N(s ) } .