Zhang et al. Advances in Difference Equations (2015) 2015:357 DOI 10.1186/s13662-015-0689-y
RESEARCH
Open Access
Asymptotics and oscillation of nth-order nonlinear differential equations with p-Laplacian like operators Shao-Yan Zhang1 , Qi-Ru Wang2* and Ravi P Agarwal3,4 *
Correspondence:
[email protected] 2 School of Mathematics and Computational Science, Sun Yat-sen University, West Xingang Road 135, Guangzhou, 510275, P.R. China Full list of author information is available at the end of the article
Abstract This paper is concerned with nth-order nonlinear differential equations of the form (a(t)|x(n–1) (t)|p–2 x(n–1) (t)) + r(t)|x(n–1) (t)|p–2 x(n–1) (t) + q(t)|x(g(t))|p–2 x(g(t)) = 0 with n ≥ 2. By discussing the signs of ith-order derivatives of eventually positive solutions, for i = 1, . . . , n – 1, and using the generalized Riccati technique and integral averaging technique, we derive new criteria for oscillation and asymptotic behavior of the equation. Our results generalize and improve many existing results in the literature. Keywords: nth-order nonlinear differential equations; asymptotic behavior; oscillation; p-Laplacian
1 Introduction In this paper, we study the nth-order nonlinear differential equation with Laplacian and deviating argument
p– p– p– a(t)x(n–) (t) xn– (t) + r(t)xn– (t) x(n–) (t) + q(t)x g(t) x g(t) = ,
(.)
where t ∈ [t , ∞). Throughout this paper, we assume the following: (H) a(t) ∈ C ([t , ∞), (, ∞)), r(t), q(t) ∈ C([t , ∞), R), q(t) ≥ , and a (t) + r(t) ≥ ; (H) p > is a real number, g(t) ∈ C([t , ∞), R) such that limt→∞ g(t) = ∞. Asymptotics and oscillation of (.) and related equations have been discussed by many authors; see [–] and the references therein. In particular in , Zhang et al. [] established oscillation criteria for (.) when n ≥ is even via the integral averaging technique and two kinds of functions H(t, s) and H∗ (t, s), employed the comparison technique to discuss the oscillation of (.) when n ≥ is even and the oscillation and asymptotic behavior of (.) when n ≥ is odd. By imposing some additional assumptions, in the present paper we shall discuss the signs of ith-order derivatives of eventually positive solutions for i = , . . . , n – , and we establish concrete criteria for the asymptotics and oscillation of (.) for both the evenorder and the odd-order cases, where the deviating arguments may be retarded, advanced, or mixed. Our results will generalize and improve those in [] and many other papers for the even-order case and develop new results for the odd-order case. © 2015 Zhang et al. This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.
Zhang et al. Advances in Difference Equations (2015) 2015:357
Page 2 of 16
By a solution of (.) we mean a nontrivial real-valued function x(t) ∈ C n– ([t , ∞), R) such that a(t)|x(n–) (t)|p– xn– (t) ∈ C ([t , ∞), R), which satisfies (.). Our attention is restricted to those solutions of (.) that satisfy sup{|x(t)| : t ≥ tx } > for any tx ≥ t . A solution x(t) of (.) is said to be oscillatory if it is neither eventually positive nor eventually negative. Otherwise, it is called nonoscillatory. The equation itself is called oscillatory if all its solutions are oscillatory. This paper is organized as follows: After this introduction, we present the main results in Section , followed by illustrative examples in Section . We then introduce some preliminary lemmas in Section , which are used to prove the main results in Section . The conclusion is drawn in Section .
2 Main results In this section, we present our main results which provide conditions for every solution of (.) to be oscillatory on [t , ∞) or convergent to as t → ∞. In order to state the main theorems, we need the following notation. For t, T ∈ R such that t ≥ T, we define β (t, T) = a(t) exp
t t
r(τ ) dτ a(τ )
– p–
,
t
βi (t, T) =
βi– (s, T) ds, T
β ∗ (t, T) = βn– (t, T)βn– (t, T), β∗ (t, T) = min βn– (t, T), βn– g(t), T ,
n– ( g tn–(t) )p– , g(t) ≤ t, Q(t) = q(t)G(t), G(t) = , g(t) > t. p–
For D = {(t, s) ∈ R : t ≥ s ≥ }, we define H = H(t, s) ∈ C D, [, ∞) : H(t, t) = , H(t, s) > and Hs (t, s) ≤ for t > s ≥ ,
z (s) r(s) z (s) r(s) , , Hs (t, s) + H(t, s) + – = max Hs (t, s) + H(t, s) + – z(s) a(s) + z(s) a(s) where z ∈ C (R, (, ∞)) is to be given in Theorem ., z+ (t) = max{z (t), }, and [ r(s) ] a(s) +
z+ (s) = max{[ z(s)
z+ (s) z(s)
–
r(s) ], }. a(s)
– The results in the first theorem are valid for all p > .
Theorem . Let p > . Assume that (H)-(H) and
∞ exp( s r(τ ) t a(τ ) a(s)
t
dτ )
p–
ds = ∞
(.)
hold, and either
∞
t
exp
s
t
r(τ ) dτ q(s) ds = ∞, a(τ )
(.)
Zhang et al. Advances in Difference Equations (2015) 2015:357
Page 3 of 16
or
a– (s) exp –
∞ ∞
t
v
s
r(τ ) dτ a(τ )
t
∞
exp
s
u t
p– r(τ ) dτ q(u) du ds dv a(τ )
= ∞.
(.)
Furthermore, for sufficiently large T ∈ R, one of the following conditions is satisfied: (a) there exists a z ∈ C (R, (, ∞)) such that t z(s) exp lim sup t→∞
T
s
t
z+ (s) r(τ ) dτ Q(s) – p– ds = ∞, a(τ ) βn– (s, T)
(b) there exists a z ∈ C (R, (, ∞)) such that t z+ (s) r(s) p z(s) lim sup – z(s)Q(s) – p s p– p z(s) a(s) t→∞ T + βn– (s, T) exp(
r(τ ) t a(τ )
dτ )
ds = ∞,
(c) there exist a z ∈ C (R, (, ∞)) and an H ∈ H such that lim sup t→∞ H(t, T)
t H(t, s)z(s)Q(s) – T
r(s) p )] z(s) a(s) + ds s r(τ ) p– pp βn– (s, T)H p– (t, s) exp( t a(τ dτ ) ) [Hs (t, s) + H(t, s)(
z+ (s) z(s)
–
= ∞. Then: (i) every solution x(t) of (.) is either oscillatory or tends to zero as t → ∞ when n is odd, and (ii) every solution x(t) of (.) is oscillatory when n is even. The results in the second theorem hold only for g(t) ≥ t or g(t) ≤ t with g (t) > . Theorem . Let p > . Assume that (H)-(H) and (.) hold. If either (.) is satisfied, or (.) is satisfied and for sufficiently large T ∈ R, β∗p– (t, T)
t
∞
exp
s t
r(τ ) dτ q(s) ds > , a(τ )
(.)
then: (i) every solution x(t) of (.) is either oscillatory or tends to zero as t → ∞ when n is odd, and (ii) every solution x(t) of (.) is oscillatory when n is even. The results in the third theorem hold only for p ≥ . Theorem . Let p ≥ . Assume that (H)-(H) and (.) hold and either (.) or (.) is satisfied. Furthermore, for sufficiently large T ∈ R, there exists a z ∈ C (R, (, ∞)) such that one of the following conditions is satisfied:
Zhang et al. Advances in Difference Equations (2015) 2015:357
Page 4 of 16
(a) t z(s)Q(s) – lim sup t→∞
T
(s) r(s) [ zz(s) – a(s) ] z(s) ds = ∞, s r(τ ) (p – ) exp( t a(τ ) dτ )β ∗ (s, T)
(b) there exists an H ∈ H such that lim sup t→∞ H(t, T)
t H(t, s)z(s)Q(s) – T
z+ (s) r(s) – a(s) )] z(s) z(s) ds s r(τ ) (p – )β ∗ (s, T) exp( t a(τ ) dτ )H(t, s)
[Hs (t, s) + H(t, s)(
= ∞. Then: (i) every solution x(t) of (.) is either oscillatory or tends to zero as t → ∞ when n is odd, and (ii) every solution x(t) of (.) is oscillatory when n is even.
3 Examples In this section, we give two examples to illustrate our main results. Example Consider the equation
(n–) – (n–) x (t) x (t) t
– – + t x(n–) (t) x(n–) (t) + t – x(t – ) x(t – ) = ,
(.)
where n ≥ , t ∈ [, ∞). Here we have: (i) p = , a(t) = t , r(t) = t , q(t) = t – , g(t) = t – ; (ii) a (t) + r(t) = – t + t > ,
∞ exp( s r(τ ) a(τ ) a(s)
dτ )
p–
ds =
∞
– ·
≥e
s
se( –
)
ds = e–
·
∞
s
se
ds
∞
s ds = ∞;
s r(τ ) ∞ s ∞ ∞ dτ )q(s) ds = e( – ) s– ds ≥ e– s– ds = ∞. (iii) t exp( a(τ ) Hence (H)-(H), (.), and (.) hold. For sufficiently large T ∈ [, ∞), we set z(t) = and have s t z (s) r(τ ) ds dτ Q(s) – p–+ z(s) exp lim sup t→∞ βn– (s, T) T t a(τ ) t (s – )n– s ds e( – ) s– = lim sup sn– t→∞ T t s n– ( ) – s– n– ds ≥ lim sup e s t→∞ T t = lim sup e– n– ds = ∞. t→∞ T s Hence Condition (a) of Theorem . is satisfied.
Zhang et al. Advances in Difference Equations (2015) 2015:357
Page 5 of 16
By Theorem ., every solution x(t) of (.) is oscillatory or tends to as t → ∞ when n is odd, and (.) is oscillatory when n is even. Remark . In Example above, we see that a(t) = t is strictly decreasing on [, ∞); but in [], we see that its condition (H) requires a (t) > , so the results in [] are not applicable to this example. Example Consider the equation (n–) (n–) (n–) (n–) t x (t) x (t) – x (t) x (t) + t x(t + ) x(t + ) = ,
(.)
where n ≥ , t ∈ [, ∞). Here we have: (i) p = , a(t) = t, r(t) = –, q(t) = t , and g(t) = t + ; (ii) a (t) + r(t) = > ,
∞ exp( s r(τ ) a(τ )
dτ )
a(s)
p–
∞
ds =
(– ln s+ln ) e s
ds =
∞
s
ds = ∞;
s r(τ ) ∞ s ∞ ∞ dτ )q(s) ds = exp( – dτ )s ds = s ds = ∞. (iii) t exp( a(τ ) τ Hence (H)-(H), (.), and (.) hold. With z(t) = t we see that for sufficiently large T ∈ [, ∞), t
(s) r(s) [ zz(s) – a(s) ] z(s) ds s r(τ ) t→∞ (p – ) exp( t a(τ dτ )β ∗ (s, T) T ) t t t z(s)Q(s) ds = lim sup s · ds = lim sup s ds = ∞. = lim sup s t→∞ t→∞ t→∞ T T T
lim sup
z(s)Q(s) –
Hence Condition (a) of Theorem . is satisfied. By Theorem ., every solution x(t) of (.) is oscillatory or tends to zero as t → ∞ when n is odd, and (.) is oscillatory when n is even. Remark . In Example above, we see that r(t) = – < and g(t) = t + ≥ t on [, ∞). But in [], we see that its condition (H) requires r(t) > , and in [], its condition (H) requires g(t) ≤ t, so the results in [, ] are not applicable to this example.
4 Preliminary lemmas In this section, we present several technical lemmas which will be used in the proofs of the main results. The first one is on the signs of derivatives of certain classes of functions. Lemma . ([]) Let f (t) ∈ C m ([t , ∞), (, ∞)). Assume f (m) (t) is of one sign and not identically zero on [t , ∞) for any t ≥ t . Then there exist a tx ≥ t and an integer ≤ l ≤ m, with m + l even for f m (t) ≥ or m + l odd for f m (t) ≤ such that l > implies f (i) (t) > l≤m–
for t ≥ tx , i = , , . . . , l – and
implies (–)l+i f (i) (t) > for t ≥ tx , i = l, l + , . . . , m – .
(.) (.)
Zhang et al. Advances in Difference Equations (2015) 2015:357
Page 6 of 16
The next lemma concerns the signs of derivatives of eventually positive solutions of (.). In particular, we derive conditions for the following inequalities to hold eventually: x(j) (t) > ,
j = , , . . . , n – .
(.)
Lemma . Assume that conditions (H)-(H), (.) and either (.) or (.) hold. Let x(t) be an eventually positive solution of (.). Then there exists a T ∈ R sufficiently large such that a(t) exp
t
t
p– r(τ ) dτ x(n–) (t) xn– (t) ≤ a(τ )
and x(n) (t) < for t ≥ T. Moreover: (i) (.) holds when n is even, and (ii) either (.) holds or limt→∞ x(t) = when n is odd. Proof If x(t) is an eventually positive solution of (.), then by (H), there exists a t ∈ [t , ∞) such that x(t) > and
x g(t) > ,
t ≥ t .
From (.) and (H), we have
p– p– a(t)x(n–) (t) x(n–) (t) + r(t)x(n–) (t) x(n–) (t) p– = –q(t)x g(t) x g(t) ≤ , t ≥ t ,
(.)
which implies that a(t) exp
t
t
(n–) p– n– r(τ ) dτ x (t) x (t) ≤ , a(τ )
t ≥ t .
(.)
t r(τ ) dτ )|x(n–) (t)|p– xn– (t) is strictly decreasing on [t , ∞) and eventually Then a(t) exp( t a(τ ) of one sign. So, x(n–) (t) is either eventually positive or eventually negative. By (.), from the proof of Lemma in [], we have x(n–) (t) > . Then we can write (.) in the form a(t) exp
t
t
p– r(τ ) dτ x(n–) (t) ≤ , a(τ )
t ≥ t ,
which implies that for t ≥ t , p– r(τ ) dτ a (t) + r(t) x(n–) (t) a(τ ) t t p– (n) r(τ ) dτ (p – )a(t) x(n–) (t) + exp x (t) ≤ . t a(τ )
exp
t
Thus, x(n) (t) < eventually.
Zhang et al. Advances in Difference Equations (2015) 2015:357
Page 7 of 16
From Lemma . with m = n – , there exist a t ∈ [t , ∞)T and an integer l, ≤ l ≤ m such that (.) and (.) are satisfied. That is, x(j) (t) > for t ≥ t , j = , . . . , l – , (–)l+j x(j) (t) >
(.)
for t ≥ t , j = l, l + , . . . , m – ,
(.)
and x(t) is eventually monotone. When n is even, by Lemma ., l must be an odd number. By (.) and (.), we can get x (t) > . Hence lim x(t) exists and is positive
t→∞
or
lim x(t) = ∞.
(.)
t→∞
In this case, we claim that l = m = n – . Otherwise, we obtain the odd integer l ≤ m – = n – . By (.), we get (–)l+m– x(m–) (t) >
and (–)l+m– x(m–) (t) > .
(.)
This means x(m–) (t) = x(n–) (t) < ,
x(m–) (t) = x(n–) (t) > ,
t ∈ [t , ∞).
(.)
It follows from (.) that there exist a T ≥ t and b > such that x(g(t)) ≥ b for t ≥ T. From (.) and (.), and x(n–) (t) > we have a(t) exp
t
t
t p– r(τ ) r(τ ) dτ x(n–) (t) dτ q(t). ≤ –bp– exp a(τ ) t a(τ )
(.)
If (.) holds, by integrating (.) from T to t with t ≥ T we obtain a(t) exp
p– r(τ ) dτ x(n–) (t) a(τ ) t T s t (n–) p– r(τ ) r(τ ) p– dτ x dτ q(s) ds ≤ a(T) exp (T) –b exp t a(τ ) T t a(τ ) t
→ –∞ as t → ∞, which contradicts the fact that x(n–) (t) > for t ∈ [t , ∞). Hence, l = m = n – and (.) holds. If (.) holds, by integrating (.) from t to u with T ≤ t ≤ u we obtain p– r(τ ) dτ x(n–) (t) a(τ ) t u s u (n–) p– r(τ ) r(τ ) p– dτ x dτ q(s) ds ≥ a(u) exp (u) +b exp t a(τ ) t t a(τ ) s u r(τ ) dτ q(s) ds. exp ≥ bp– t t a(τ )
a(t) exp
t
Zhang et al. Advances in Difference Equations (2015) 2015:357
Page 8 of 16
Taking the limit as u → ∞, we have x(n–) (t) ≥ b a– (t) exp –
t t
r(τ ) dτ a(τ )
∞
exp
t
s
t
p– r(τ ) . dτ q(s) ds a(τ )
Since x(n–) (t) < , integrating the above inequality from t to v with T ≤ t ≤ v we get –x(n–) (t) ≥ x(n–) (v) – x(n–) (t) v – a (s) exp – ≥b t
s
r(τ ) dτ a(τ )
t
∞
exp
s
p– r(τ ) dτ q(u) du ds. a(τ )
u t
Taking the limit as v → ∞, we obtain –x
(n–)
(t) ≥ b
a (s) exp –
∞
s
–
t
t
r(τ ) dτ a(τ )
∞
exp
s
u t
p– r(τ ) ds. dτ q(u) du a(τ )
Since x(n–) (t) > , integrating the above inequality from T to t with t ≥ T, we get x(n–) (T) ≥ –x(n–) (t) + x(n–) (T) t ∞ a– (s) exp – ≥b T
v
exp
∞
× s
u
t
r(τ ) dτ t a(τ ) p– r(τ ) dτ q(u) du ds dv. a(τ ) s
Taking t → ∞, we obtain T
∞ ∞
a (s) exp –
v
≤b x
s
–
– (n–)
t
r(τ ) dτ a(τ )
∞ s
exp
u t
p– r(τ ) dτ q(u) du ds dv a(τ )
(T) < ∞,
which contradicts (.). Hence, l = m = n – and (.) holds. When n is odd, by Lemma ., l must be an even integer. By (.) and (.), we have either x (t) > or x (t) < . That means limt→∞ x(t) = c ≥ . We claim that if limt→∞ x(t) = , then l = m = n – . Otherwise, there is the even number l ≤ m – = n – such that (.) and (.) hold. By a similar argument to above, we can reach a contradiction to (.) or (.). This completes the proof. Lemma . Let conditions (H)-(H), (.), and either (.) or (.) hold. Assume x(t) is an eventually positive solution of (.) satisfying (.) eventually. Then there exists a T ∈ [t , ∞) such that for t ∈ [T, ∞), x (t) ≥ a(t) exp
t
t
r(τ ) dτ a(τ )
p–
x(n–) (t)βn– (t, T)
and x(t) ≥ a(t) exp
t
t
r(τ ) dτ a(τ )
p–
x(n–) (t)βn– (t, T).
Zhang et al. Advances in Difference Equations (2015) 2015:357
Page 9 of 16
Proof Since x(t) is an eventually positive solution of (.), by (H), there exists a T ∈ [t , ∞) such that x(t) > , x(g(t)) > , and (.) holds for t ≥ T. t r(τ ) Noting that (a(t) exp( t a(τ dτ )(x(n–) (t))p– ) < for t ≥ T, it follows from (.) that ) x
(n–)
(t) = x
(n–)
(a(s) exp(
t
r(τ ) dτ a(τ )
(T) + T
≥ a(t) exp
t
:= a(t) exp
t
t
s
r(τ ) t a(τ )
dτ )(x(n–) (s))p– ) p– ds s r(τ ) p– (a(s) exp( t a(τ dτ )) )
t
r(τ ) dτ a(τ )
p–
x
(n–)
t
(t) (a(s) exp(
T p–
s r(τ ) t a(τ )
ds
dτ )) p–
x(n–) (t)β (t, T).
Integrating the inequality above from T to t for t ≥ T, we get x
(n–)
(t) = x
(n–)
t
x(n–) (s) ds
(T) + T
≥ x(n–) (T) +
t a(s) exp T
≥ a(t) exp
t
t t
:= a(t) exp
s
t t
r(τ ) dτ a(τ ) r(τ ) dτ a(τ )
p–
r(τ ) dτ a(τ )
p–
x(n–) (s)β (s, T) ds
t
x(n–) (t)
β (s, T) ds T
p–
x(n–) (t)β (t, T).
It is easy to see by induction that x (t) ≥ a(t) exp
p– r(τ ) x(n–) (t)βn– (t, T), dτ t a(τ ) p– t r(τ ) dτ x(n–) (t)βn– (t, T). x(t) ≥ a(t) exp a(τ ) t
t
This completes the proof. Similar to [], Lemma ., we have the following lemma. γ +
Lemma . Let g(u) = Bu – Au γ , where A and γ are positive numbers, and B ≥ . Then g attains its maximum value on [, ∞) at u∗ = ( A(γBγ+) )γ , and max g = g u∗ =
u∈[,∞)
Bγ + γγ . γ + (γ + ) Aγ
5 Proofs of main results In this section, we give proofs for our main results by employing generalized Riccati techniques and integral averaging techniques.
Zhang et al. Advances in Difference Equations (2015) 2015:357
Page 10 of 16
Proof of Theorem . Suppose to the contrary that (.) has a nonoscillatory solution x(t). Without loss of generality, we may assume that x(t) is eventually positive. Then, by (H)(H), there exists a T ∈ [t , ∞) such that for t ≥ T, x(t) > , and Lemmas . and . hold. When n is odd, from Lemma . we see that (.) holds or limt→∞ x(t) = . If (.) holds, (.) reduces to
p– p– p– a(t) x(n–) (t) + r(t) xn– (t) + q(t) x g(t) = .
Multiplying by exp( a(t) exp
t
t
t
r(τ ) t a(τ )
(.)
dτ ) on (.), we have
t p– p– r(τ ) r(τ ) dτ x(n–) (t) dτ q(t) x g(t) + exp = . (.) a(τ ) t a(τ )
The rest of the proof is separated into three parts corresponding to conditions (a)-(c), respectively. Part I: Assume condition (a) holds. Define
v(t) :=
z(t) exp(
t
r(τ ) t a(τ )
dτ )a(t)(x(n–) (t))p–
xp– (t)
for t ≥ T.
Then v(t) > . From z(t) > , a(t) > , x(t) > , x(n–) (t) > , and (.), we have v (t) = a(t) exp
(n–) p– z(t) r(τ ) (t) dτ x xp– (t) t a(τ ) t (n–) p– r(τ ) z(t) dτ a(t) x (t) + exp xp– (t) t a(τ ) t r(τ ) – exp( t a(τ ) dτ )q(t)(x(g(t)))p– = z(t) xp– (t) t p– z (t)xp– (t) – z(t)(n)xp– (t)x (t) r(τ ) + a(t) exp dτ x(n–) (t) (xp– (t)) t a(τ ) t r(τ ) –z(t) exp( t a(τ dτ )q(t)(x(g(t)))p– ) ≤ xp– (t) t r(τ ) z+ (t)(a(t) exp( t a(τ dτ )(x(n–) (t))p– ) ) + xn (t) t r(τ ) (a(t) exp( t a(τ dτ )(x(n–) (t))p– )z(t)(p – )xp– (t)x (t) ) – (xp– (t)) t r(τ ) t r(τ ) –z(t) exp( t a(τ ) dτ )q(t)(x(g(t)))n z+ (t)(a(t) exp( t a(τ dτ )(x(n–) (t))p– ) ) + . ≤ xp– (t) xp– (t) t
By the Kiguradze lemma [], which shows that if a function y(t) satisfies y(i) (t) > , i = , , , . . . , k, and y(k+) (t) ≤ , then y(t)/y (t) ≥ t/k, we have x(t) t ≥ . x (t) n –
Zhang et al. Advances in Difference Equations (2015) 2015:357
Page 11 of 16
Thus, we see that x(t)/t n– is nonincreasing, and when g(t) ≤ t, we get x(g(t)) g n– (t) ≥ n– . x(t) t When g(t) > t, by x (t) > , we get x(g(t)) ≥ . x(t) It follows that
x(g(t)) x(t)
p– ≥ G(t).
(.)
Then, by Lemma ., we obtain
v (t) ≤ –z(t) exp
t
t
z (t) r(τ ) dτ q(t)G(t) + p–+ . a(τ ) βn– (t, T)
Integrating the above inequality from T to t for t ≥ T, we get t
z(s) exp
T
s t
z (s) r(τ ) ds ≤ v(T) – v(t) < v(T). dτ Q(s) – n + a(τ ) βn– (s, T)
Taking lim sup on both sides as t → ∞, we obtain a contradiction to condition (b). Therefore, every solution x(t) of (.) is either oscillatory or tends to zero as t → ∞. Part II: Assume condition (b) holds. Define w(t) :=
z(t)a(t)(x(n–) (t))n xp– (t)
for t ≥ T.
(.)
Then w(t) > . From z(t) > , x(t) > , x(n–) (t) > , we have p– w (t) = a(t) x(n–) (t)
(n–) n z(t) z(t) + a(t) x (t) xp– (t) xp– (t)
–r(t)(xn– (t))p– – q(t)(x(g(t)))p– xp– (t) p– z (t)xp– (t) – z(t)(n)xp– (t)x (t) + a(t) x(n–) (t) (xp– (t))
= z(t)
≤
–z(t)r(t)(x(n–) (t))p– – z(t)q(t)(x(g(t)))p– z+ (t)(a(t)(x(n–) (t))p– ) + xn (t) xp– (t) –
(a(t)(x(n–) (t))p– )z(t)(p – )xp– (t)x (t) . (xn (t))
(.)
From (.), we have z+ (t) r(t) w (t) ≤ – z(t)Q(t) + – w(t) z(t) a(t)
–
(a(t)(x(n–) (t))n )z(t)(p – )xp– (t)x (t) . (xp– (t))
(.)
Zhang et al. Advances in Difference Equations (2015) 2015:357
Page 12 of 16
From Lemma ., we get –
(a(t)(x(n–) (t))n )z(t)(p – )xp– (t)x (t) (xp– (t)) p
p
a(t)z(t)(p – ) a n (t)z p– (t)(x(n–) (t))p x (t) p p xp (t) x(n–) (t) a p– (t)z p– (t) t r(τ ) p p p– (t)z n (t)(x(n–) (t))p (p – )βn– (t, T) exp( p– t a(τ ) dτ ) a ≤– . xp (t) z p– (t) =–
Setting γ = p – , we have z+ (t) r(t) – w(t) z(t) a(t) t r(τ ) (p – )βn– (t, T) exp( p– t a(τ ) dτ )
w (t) ≤ –z(t)Q(t) + –
≤ –z(t)Q(t) + –
z
p–
(t)
z+ (t) r(t) – z(t) a(t)
z
γ + γ
(t)
(p – )βn– (t, T) exp( n p–
w
w(t) +
t
r(τ ) t a(τ )
dτ )
w
γ + γ
(t).
(.)
(t)
Let
z (t) r(t) B= + – z(t) a(t)
,
A=
(p – )βn– (t, T) exp( p–
z
+
p–
t
r(τ ) t a(τ )
dτ )
,
u = w(t).
(t)
Then by Lemma ., we obtain, for all t ≥ T, w (t) ≤ –z(t)Q(t) +
z+ (t) r(t) p z(t) – p– p p z(t) a(t) + βn– (t, T) exp( t
r(τ ) t a(τ )
dτ )
.
Integrating the above inequality from T to t for ≥ T, we get t z(s)Q(s) – T
z+ (s) r(s) p z(s) – p– p p z(s) a(s) + βn– (s, T) exp( s
r(τ ) t a(τ )
dτ )
ds
≤ w(T) – w(t) < w(T). By taking lim sup on both sides as t → ∞, we obtain a contradiction to condition (c). Therefore, every solution x(t) of (.) is either oscillatory or tends to zero as t → ∞. Part III: Assume condition (c) holds. From (.) we have, for H ∈ H∗ and t ≥ T, T
t
z+ (s) r(s) H(t, s)z(s)Q(s) ds ≤ – H(t, s)w (s) ds + H(t, s)w(s) – ds z(s) a(s) T T s r(τ ) t (p – )βn– (s, T) exp( p– t a(τ ) dτ ) γγ+ w (s) ds. H(t, s) – T z p– (s)
t
t
Zhang et al. Advances in Difference Equations (2015) 2015:357
Page 13 of 16
By integration by parts we obtain
t
–
H(t, s)w (s) ds = H(t, T)w(T) +
T
t T
Hs (t, s)w(s) ds.
It follows that
t
T
t z+ (s) r(s) Hs (t, s) + H(t, s) – H(t, s)z(s)Q(s) ds ≤ H(t, T)w(T) + w(s) ds z(s) a(s) + T s r(τ ) t (p – )βn– (s, T) exp( p– t a(τ ) dτ ) γγ+ – H(t, s) w (s) ds. T z p– (s)
Let γ = p – , u = w(s) and z (s) r(s) B = Hs (t, s) + H(t, s) + – , z(s) a(s) + s r(τ ) (p – )βn– (s, T) exp( p– t a(τ ) dτ ) , A = H(t, s) z p– (s) by Lemma . we obtain, for all t ≥ T,
t
H(t, s)z(s)Q(s) ds ≤ H(t, T)w(T) +
T
[Hs (t, s) + H(t, s)(
z+ (s) z(s)
r(s) p )] z(s) a(s) + s r(τ ) p– T βn– (s, T)H p– (t, s)pp exp( t a(τ ) dτ ) t
–
ds.
That is, H(t, T)
t H(t, s)z(s)Q(s) – T
r(s) p )] z(s) a(s) + ds ≤ w(T). s r(τ ) p– pp βn– (s, T)H p– (t, s) exp( t a(τ dτ ) ) [Hs (t, s) + H(t, s)(
z+ (s) z(s)
–
By taking lim sup on both sides as t → ∞, we obtain a contradiction to condition (d). Therefore, every solution x(t) of (.) is either oscillatory or tends to zero as t → ∞. When n is even, from Lemma . we see that only (.) holds. Similarly, we can show that (.) is oscillatory. We omit the details. The proof is complete. Proof of Theorem . Suppose to the contrary that (.) has a nonoscillatory solution x(t). Without loss of generality, we may assume that x(t) is eventually positive. Then, by (H)(H), there exists a T ∈ [t , ∞) such that for t ≥ T, x(t) > , and Lemmas . and . hold. When n is odd, from Lemma . we see that (.) holds or limt→∞ x(t) = . If (.) t r(τ ) dτ )(x(n–) (t))p– . Then φ(t) > and φ (t) < for t ≥ T, holds, we set φ(t) := a(t) exp( t a(τ ) and limt→∞ φ(t) = ζ ≥ . By (.), it follows that φ (t) + exp
t
t
p– r(τ ) dτ q(t) x g(t) = . a(τ )
(.)
Integrating both sides of (.) from t to ∞, we obtain
∞
ζ – φ(t) + t
exp
s t
p– r(τ ) dτ q(s) x g(s) ds = . a(τ )
(.)
Zhang et al. Advances in Difference Equations (2015) 2015:357
Page 14 of 16
When g(t) ≥ t, we have
ζ – φ(t) + x(t)
p–
∞
exp
t
s
t
r(τ ) dτ q(s) ds < , a(τ )
by which we have reached a contradiction if s r(τ ) ∞ t exp( t a(τ ) dτ )q(s) ds < ∞, we get p– φ(t) ≥ x(t)
∞
exp
t
s
t
∞ t
exp(
s
r(τ ) t a(τ )
dτ )q(s) ds = ∞. If
r(τ ) dτ q(s) ds. a(τ )
By Lemma ., we obtain exp
r(τ ) dτ q(s) ds t t a(τ ) s ∞ r(τ ) p– exp dτ q(s) ds ≤ , ≤ βn– (t, T) t t a(τ )
β∗p– (t, T)
∞
s
which is a contradiction to (.). When g(t) ≤ t and g (t) > , by (.), we get p– ζ – φ(t) + x g(t)
∞
exp
t
s
t
r(τ ) dτ q(s) ds < , a(τ )
by which we have reached a contradiction if s r(τ ) ∞ t exp( t a(τ ) dτ )q(s) ds < ∞, we get p– φ g(t) ≥ φ(t) ≥ x g(t)
t
∞
exp
s t
∞ t
exp(
s
r(τ ) t a(τ )
dτ )q(s) ds = ∞. If
r(τ ) dτ q(s) ds. a(τ )
By Lemma ., we obtain exp
r(τ ) dτ q(s) ds t t a(τ ) s ∞ r(τ ) p– dτ q(s) ds ≤ , exp ≤ βn– g(t), T t t a(τ )
β∗p– (t, T)
∞
s
which is a contradiction to (.). Therefore, every solution x(t) of (.) is either oscillatory or tends to zero as t → ∞. When n is even, from Lemma . we see that only (.) holds. Similarly, we can show that (.) is oscillatory. We omit the details. The proof is complete. Proof of Theorem . Suppose to the contrary that (.) has a nonoscillatory solution x(t). Without loss of generality, we may assume that x(t) is eventually positive. Then, by (H)(H), there exists a T ∈ [t , ∞) such that for t ≥ T, x(t) > , x(g(t)) > , and Lemmas . and . hold.
Zhang et al. Advances in Difference Equations (2015) 2015:357
Page 15 of 16
When n is odd, from Lemma . we see that either (.) holds or limt→∞ x(t) = . If (.) holds, the rest of the proof is separated into two parts corresponding to conditions (a) and (b), respectively. Part I: Assume condition (a) holds. Define w(t) as in (.). By x(t) > , and (.) we obtain (a(t)z(t)(x(n–) (t))p– ) (p – )xp– (t)x (t) z (t) r(t) – w(t) – z(t) a(t) (xp– (t)) z(t)a(t)(x(n–) (t))p– x (t) (p – )xp– (t) z (t) r(t) – w(t) – w(t) . ≤ –z(t)Q(t) + (n–) p– (n–) z(t) a(t) z(t)a(t)(x (t)) x (t)
w (t) ≤ –z(t)Q(t) +
From this and Lemma ., we get t (p – ) exp( t z (t) r(t) w (t) ≤ –z(t)Q(t) + w(t) – – z(t) a(t) z(t)
r(τ ) a(τ )
dτ )
β ∗ (t, T) w(t) . (.)
By completing the square for w(t) on the right-hand side, we have
(t) r(t) [ zz(t) – a(t) ] z(t) . w (t) ≤ –z(t)Q(t) + t r(τ ) (p – ) exp( t a(τ ) dτ )β ∗ (t, T)
Integrating the above inequality from T to t for t ≥ T, we get t T
(s) r(s) [ zz(s) – a(s) ] z(s) z(s)Q(s) – ds ≤ w(T) – w(t) < w(T). s r(τ ) (p – ) exp( t a(τ dτ )β ∗ (s, T) )
Taking lim sup on both sides as t → ∞, we obtain a contradiction to condition (a). Therefore, every solution x(t) of (.) is either oscillatory or tends to zero as t → ∞. Part II: Assume condition (b) holds. Based on (.), the proof is similar to those of Part III of Theorem . and Part I of Theorem ., and hence it is omitted. When n is even, from Lemma . we see that only (.) holds. Similarly, we can show that (.) is oscillatory and hence omit its proof. The proof is complete.
6 Conclusions In this paper, we have discussed the asymptotics and oscillation for nth-order nonlinear differential equation (.), where the deviation argument g(t) may be retarded, advanced, or mixed. Under certain assumptions, we have derived a complete characterization of an eventually positive solution x(t) of (.): there exists T ∈ R such that for all t ≥ T, (a(t)|x(n–) (t)|p– xn– (t)) ≤ , x(n) (t) < , and (i) x(j) (t) > (j = , , . . . , n – ) when n is even; (ii) either x(j) (t) > (j = , , . . . , n – ) holds or limt→∞ x(t) = when n is odd. By using generalized Riccati techniques and integral averaging techniques, we have proved that under a number of conditions: (i) every solution x(t) of (.) is oscillatory or tends to zero as t → ∞ when n is odd; (ii) equation (.) is oscillatory when n is even. Also, we have given two examples to illustrate the obtained results.
Zhang et al. Advances in Difference Equations (2015) 2015:357
Page 16 of 16
Competing interests The authors declare that they have no competing interests. Authors’ contributions All authors completed the paper together. All authors read and approve the final manuscript. Author details 1 Department of Mathematics, Guangdong University of Finance, Yingfu Road 527, Guangzhou, 510520, P.R. China. 2 School of Mathematics and Computational Science, Sun Yat-sen University, West Xingang Road 135, Guangzhou, 510275, P.R. China. 3 Department of Mathematics, Texas A&M University-Kingsville, 700 University Blvd., Kingsville, TX 78363-8202, USA. 4 Department of Mathematics, King Abdulaziz University, P.O. Box 80203, Jeddah, 21589, Saudi Arabia. Acknowledgements This work was supported by supported by the NNSF of China (No. 11271379), Foundation for Technology Innovation in Higher Education of Guangdong, China (No. 2013KJCX0136), and Foundation for Humanities and Social Science in Ministry of Education of P.R. China (No. 14YJC790141). Received: 16 July 2015 Accepted: 5 November 2015 References 1. Agarwal, RP, Bohner, M, Li, WT: Nonoscillation and Oscillation: Theory for Functional Differential Equations. Monographs and Textbooks in Pure and Applied Mathematics. Dekker, New York (2001) 2. Agarwal, RP, Grace, SR, O’Regan, D: Oscillation Theory for Difference and Functional Differential Equations. Kluwer Academic, Dordrecht (2000) 3. Agarwal, RP, Grace, SR, O’Regan, D: Oscillation criteria for certain nth order differential equations with deviating arguments. J. Math. Anal. Appl. 262, 601-622 (2001) 4. Agarwal, RP, Grace, SR, O’Regan, D: Oscillation Theory for Second Order Linear, Half-Linear, Superlinear and Sublinear Dynamic Equations. Kluwer Academic, Dordrecht (2002) 5. Agarwal, RP, Wang, QR: Oscillation and asymptotic behavior for second-order nonlinear perturbed differential equations. Math. Comput. Model. 39, 1477-1490 (2004) 6. Erbe, L, Kong, Q, Zhang, BG: Oscillation Theory for Functional Differential Equations. Dekker, New York (1995) 7. He, HJ, Wang, QR: Oscillation of certain even order nonlinear functional differential equations. Commun. Appl. Anal. 18, 281-296 (2014) 8. Kiguradze, IT, Chanturiya, TA: Asymptotic Properties of Solutions of Nonautonomous Ordinary Differential Equations. Kluwer Academic, Dordrecht (1993) 9. Liu, S, Zhang, Q, Yu, Y: Oscillation of even-order half-linear functional differential equations with damping. Comput. Math. Appl. 61, 2191-2196 (2011) 10. Philos, CG: A new criterion for the oscillatory and asymptotic behavior of delay differential equations. Bull. Acad. Pol. Sci., Sér. Sci. Math. 39, 61-64 (1981) 11. Wang, QR: Oscillation criteria for even order nonlinear damped differential equations. Acta Math. Hung. 95, 169-178 (2002) 12. Wu, HW, Wang, QR, Xu, YT: Oscillation criteria for certain even order nonlinear functional differential equations. Dyn. Syst. Appl. 13, 129-143 (2004) 13. Zhang, CH, Agarwal, RP, Bohner, M, Li, TX: New results for oscillatory behavior of even-order half-linear delay differential equations. Appl. Math. Lett. 26, 179-183 (2013) 14. Zhang, CH, Agarwal, RP, Li, TX: Oscillation and asymptotic behavior of higher-order delay differential equations with p-Laplacian like operators. J. Math. Anal. Appl. 409, 1093-1106 (2014) 15. Zhang, Q, Liu, S, Gao, L: Oscillation criteria for even-order half-linear functional differential equations with damping. Appl. Math. Lett. 24, 1709-1715 (2011) 16. Zhang, SY, Wang, QR: Oscillation of second-order nonlinear neutral dynamic equations on time scales. Appl. Math. Comput. 216, 2837-2848 (2010)