We will study oscillation of bounded solutions of higher-order nonlinear neutral delay differential equations of the following type:
[y(t)+p(t)f(y(τ(t)))](n)+q(t)h(y(σ(t)))=0, t≥t0, t∈ℝ, where
p∈C([t0,∞),ℝ), limt→∞p(t)=0, q∈C([t0,∞),ℝ+), τ(t), σ(t)∈C([t0,∞),ℝ), τ(t), σ(t)<t, limt→∞τ(t), σ(t)=∞, and
f,h∈C(ℝ,ℝ). We obtain sufficient conditions for the oscillation of all solutions of this equation.
1. Introduction
In this paper, we are concerned with the oscillation of the solutions of a certain more general higher-order nonlinear neutral-type functional differential equation with an oscillating coefficient of the form
[y(t)+p(t)f(y(τ(t)))](n)+q(t)h(y(σ(t)))=0,t≥t0,t∈R,
wherep∈C([t0,∞),ℝ) is oscillatory and limt→∞p(t)=0, q∈C([t0,∞),ℝ+), τ(t), σ(t)∈C([t0,∞),ℝ), τ(t), σ(t)<t, limt→∞τ(t)=∞, limt→∞σ(t)=∞, and f,h∈C(ℝ,ℝ). As it is customary, a solution y(t) is said to be oscillatory if y(t) is not eventually positive or not eventually negative. Otherwise, the solution is called nonoscillatory. A differential equation is called oscillatory if all of its solutions oscillate. Otherwise, it is nonoscillatory. In this paper, we restrict our attention to real-valued solutions y.
In [1, 2], several authors have investigated the linear delay differential equation
x′(t)+q(t)x(σ(t))=0,t≥t0,
where q∈C([t0,∞),ℝ+) and σ(t)∈C([t0,∞),ℝ). A classical result is that every solution of (1.2) oscillates if
liminft→∞∫σ(t)tq(s)ds>1e.
In [3], Zein and Abu-Kaff have investigated the higher-order nonlinear delay differential equation
[x(t)+p(t)x(τ(t))](n)+f(t,x(t),x(σ(t)))=s(t),t≥t0,t∈R,
where p∈C([t0,∞),ℝ), limt→∞p(t)=0, τ(t), σ(t)∈C([t0,∞),ℝ), τ(t), σ(t)<t, limt→∞τ(t)=∞, limt→∞σ(t)=∞, f:ℝ+×ℝ×ℝ→ℝ is continuous, yf(t,x,y)>0 for xy>0, there exists an oscillatory function r∈Cn(ℝ+,ℝ), such that r(n)(t)=s(t), limt→∞r(t)=0.
In [4], Bolat and Akin have investigated the higher-order nonlinear differential equation
[y(t)+p(t)y(τ(t))](n)+∑i=1mqi(t)fi(y(σi(t)))=s(t),
where p(t), qi(t), τ(t), s(t)∈C([t0,∞),ℝ) for i=1,…,m, p(t) and s(t) are oscillating functions, qi(t)≥0 for i=1,…,m, σi(t)∈C1([t0,∞),ℝ), σi′(t)>0, σi(t)≤t, limt→∞σi(t)=∞ for i=1,…,m, limt→∞τ(t)=∞, fi(u)∈C((ℝ,ℝ)) is nondecreasing function, uf(u)>0 for u≠0, and i=1,…,m. If n is odd, limt→∞p(t)=0, limt→∞r(t)=0, and ∫t0∞νn-1q(ν)dν=∞ for i=1,…,m, then every bounded solution of (1.5) is either oscillatory or tends to zero as t→∞. If n is even, limt→∞p(t)=0, and limt→∞r(t)=0, there exists a continuously differentiable function φ(t)limsupt→∞∫t0tφ(ν)∑i=1mqi(ν)dν=∞,limsupt→∞∫t0t[[φ′(ν)]2φ(ν)σi′(ν)σin-2(ν)]dν<∞,
then every bounded solution of (1.5) is either oscillatory or tends to zero as t→∞.
Recently, many studies have been made on the oscillatory and asymptotic behaviour of solutions of higher-order neutral-type functional differential equations. Most of the known results which were studied are the cases when f(u)=I(u), where I is the identity function; see, for example, [1–15] and references cited there in.
The purpose of this paper is to study oscillatory behaviour of solutions of (1.1). For the general theory of differential equations, one can refer to [5, 6, 12–14]. Many references to some applications of the differential equations can be found in [2].
In this paper, the function z(t) is defined by
z(t)=y(t)+p(t)f(y(τ(t))).
2. Some Auxiliary LemmasLemma 2.1 (see [5]).
Let y be a positive and n-times differentiable function on [t0,+∞). If y(n)(t) is of constant sign and not identically zero in any interval [b,+∞), then there exist a t1≥t0 and an integer l, 0≤l≤n such that n+l is even, if y(n)(t) is nonnegative, or n+l odd, if y(n)(t) is nonpositive, and that, as t≥t1, if l>0, y(n)(t)>0 for k=0,1,2,…,l-1, and if l≤n-1, (-1)k+1y(n)(t)>0 for k=l,l+1,…,n-1.
Lemma 2.2 (see [5]).
Let y(t) be as in Lemma 2.1. In addition limt→∞y(t)≠0 and y(n-1)(t)y(n)(t)≤0 for every t≥ty; then for every λ, 0<λ<1, the following hold:
y(t)≥λ(n-1)!tn-1y(n-1)(t)for all larget.
3. Main ResultsTheorem 3.1.
Assume that n is even,
there exists a function H:ℝ→ℝ such that H is continuous and nondecreasing and satisfies the inequality
-H(-uv)≥H(uv)≥KH(u)H(v),foru,v>0,
where K is a positive constant, and
|h(u)|≥|H(u)|,H(u)u≥γ>0,H(u)>0,foru≠0,
limt→∞p(t)=0,
∫t0∞sn-1q(s)ds=∞
and every solution of the first-order delay differential equation
w′(t)+q(t)KγH(12λ(n-1)!σn-1(t))w(σ(t))=0
is oscillatory. Then every bounded solution of (1.1) is either oscillates or tends to zero as t→∞.
Proof.
Assume that (1.1) has a bounded nonoscillatory solution y. Without loss of generality, assume that y is eventually positive (the proof is similar when y is eventually negative). That is, y(t)>0, y(τ(t))>0, and y(σ(t))>0 for t≥t1≥t0. Further, suppose that y does not tend to zero as t→∞. By (1.1) and (1.7), we have
z(n)(t)=-q(t)h(y(σ(t)))≤0,t≥t1.
It follows that z(α)(t)(α=0,1,2,…,n-1) is strictly monotone and eventually of constant sign. Since y is bounded and does not tend to zero as t→∞, by virtue of (C2), limt→∞p(t)f(y(τ(t)))=0. Then we can find a t2≥t1 such that z(t)=y(t)+p(t)f(y(τ(t)))>0 eventually and z(t) is also bounded for sufficiently large t≥t2. Because n is even and (n+l) odd for z(n)(t)≤0 and z(t)>0 is bounded, by Lemma 2.1, since l=1 (otherwise, z(t) is not bounded), there exists a t3≥t2 such that for t≥t3(-1)k+1z(k)(t)>0(k=1,2,…,n-1).
In particular, since z′(t)>0 for t≥t3, z is increasing. Since y is bounded, limt→∞p(t)f(y(τ(t)))=0 by (C2). Then, there exists a t4≥t3 by (1.7),
y(t)=z(t)-p(t)f(y(τ(t)))≥12z(t)>0,
for t≥t4. We may find a t5≥t4 such that for t≥t5, we have
y(σ(t))≥12z(σ(t))>0.
From (3.4) and (3.7), we can obtain the result of
z(n)(t)+q(t)h(12z(σ(t)))≤0,
for t≥t5. Since z is defined for t≥t2, and z(t)>0 with z(n)(t)≤0 for t≥t2 and not identically zero, applying directly Lemma 2.2 (second part, since z is positive and increasing), it follows from Lemma 2.2 that
y(σ(t))≥12λ(n-1)!σn-1(t)y(n-1)(σ(t)).
Using (C1) and (3.7), we find for t≥t6≥t5,
h(y(σ(t)))≥H(y(σ(t)))≥H(12λ(n-1)!σn-1(t)z(n-1)(σ(t)))≥KH(12λ(n-1)!σn-1(t))H(z(n-1)(σ(t)))≥KγH(12λ(n-1)!σn-1(t))z(n-1)(σ(t)).
It follows from (3.4) and the above inequality that z(n-1)(t) is an eventually positive solution of
w′(t)+q(t)KγH(12λ(n-1)!σn-1(t))w(σ(t))≤0.
By a well-known result (see [14, Theorem 3.1]), the differential equation
w′(t)+q(t)KγH(12λ(n-1)!σn-1(t))w(σ(t))=0,t≥t7≥t6
has an eventually positive solution. This contradicts the fact that (1.1) is oscillatory, and the proof is completed.
Thus, from Theorem 3.1 and [11, Theorem 2.3] (see also [11, Example 3.1]), we can obtain the following corollary.
Corollary 3.2.
If
liminft→∞∫σ(t)tq(s)H(12λ(n-1)!σ(t)n-1)ds>1eKγ,
then every bounded solution of (1.1) is either oscillatory or tends to zero as t→∞.
Theorem 3.3.
Assume that n is odd and (C2), (C3) hold. Then, every bounded solution of (1.1) either oscillates or tends to zero as t→∞.
Proof.
Assume that (1.1) has a bounded nonoscillatory solution y. Without loss of generality, assume that y is eventually positive (the proof is similar when y is eventually negative). That is, y(t)>0, y(τ(t))>0, and y(σ(t))>0 for t≥t1≥t0. Further, we assume that y(t) does not tend to zero as t→∞. By (1.1) and (1.7), we have for t≥t1z(n)(t)=-q(t)h(y(σ(t)))≤0.
That is, z(n)(t)≤0. It follows that z(α)(t)(α=0,1,2,…,n-1) is strictly monotone and eventually of constant sign. Since limt→∞p(t)=0, there exists a t2≥t1, such that for t≥t2, we have z(t)>0. Since y is bounded, by virtue of (C2) and (1.7), there is a t3≥t2 such that z is also bounded, for t≥t3. Because n is odd and z is bounded, by Lemma 2.1, since l=0 (otherwise, z(t) is not bounded), there exists t4≥t3, such that for t≥t4, we have (-1)kz(k)(t)>0(k=1,2,…,n-1). In particular, since z′(t)<0 for t≥t4, z is decreasing. Since z is bounded, we may write limt→∞z(t)=L, (-∞<L<∞). Assume that 0≤L<∞. Let L>0. Then, there exist a constant c>0 and a t5 with t5≥t4, such that z(t)>c>0 for t≥t5. Since y is bounded, limt→∞p(t)f(y(τ(t)))=0 by (C1). Therefore, there exists a constant c1>0 and a t6 with t6≥t5, such that y(t)=z(t)-p(t)f(y(τ(t)))>c1>0 for t≥t6. So, we may find t7 with t7≥t6, such that y(σ(t))>c1>0 for t≥t7. From (3.14), we have
z(n)(t)≤-q(t)h(c1)(t≥t7).
If we multiply (3.15) by tn-1 and integrate from t7 to t, then we obtain
F(t)-F(t7)≤-h(c1)∫t7tq(s)sn-1ds,
where
F(t)=∫γ=2t(-1)γtn-1z(n-γ-1)(t+γ)dt.
Since (-1)kz(k)(t)>0, for k=1,2,…,n-1 and t≥t4, we have F(t)>0 for t≥t7. From (3.16), we have
-F(t7)≤-h(c1)∫t7tq(s)sn-1ds.
By (C3), we obtain
-F(t7)≤-h(c1)∫t7tq(s)sn-1ds=-∞,
as t→∞. This is a contradiction. So, L>0 is impossible. Therefore, L=0 is the only possible case. That is, limt→∞z(t)=0. Since y is bounded, by virtue of (C2) and (1.7), we obtain
limt→∞y(t)=limt→∞z(t)-limt→∞p(t)f(y(τ(t)))=0.
Now, let us consider the case of y(t)<0 for t≥t1. By (1.1) and (1.7),
z(n)(t)=-q(t)h(y(σ(t)))≥0(t≥t1).
That is, z(n)(t)≥0. It follow that z(α)(t)(α=0,1,2,…,n-1) is strictly monotone and eventually of constant sign. Since limt→∞p(t)=0, there exists a t2≥t1, such that for t≥t2, we have z(t)<0. Since y(t) is bounded, by virtue of (C2) and (1.7), there is a t3≥t2 such that z(t) is also bounded, for t≥t3. Assume that x(t)=-z(t). Then, x(n)(t)=-z(n)(t). Therefore, x(t)>0 and x(n)(t)≤0 for t≥t3. From this, we observe that x(t) is bounded. Because n is odd and x is bounded, by Lemma 2.1, since l=0 (otherwise, x is not bounded), there exists a t4≥t3, such that (-1)kx(k)(t)>0 for k=1,2,…,n-1 and t≥t4. That is, (-1)kz(k)(t)<0 for k=1,2,…,n-1 and t≥t4. In particular, for t≥t4, we have z′(t)>0. Therefore, z(t) is increasing. So, we can assume that limt→∞z(t)=L,(-∞<L≤0). As in the proof of y(t)>0, we may prove that L=0. As for the rest, it is similar to the case y(t)>0. That is, limt→∞y(t)=0. This contradicts our assumption. Hence, the proof is completed.
Example 3.4.
We consider difference equation of the form
[y(t)+1tsin(t)(y3(t-2)+y(t-2))](4)+1t2y3(t-3)=0,
where n=4, τ(t)=t-2, p(t)=(1/t)sin(t), q(t)=1/t2, σ(t)=t-3, h(y)=y3, and f(y)=y3+y. By taking H(u)=u,
liminft→∞∫t-3t1s21213!(s-323)3ds>1e,
we check that all the conditions of Theorem 3.1 are satisfied and that every bounded solution of (3.22) oscillates or tends to zero at infinity.
Example 3.5.
We consider difference equation of the form
[y(t)+coste-5t2[y5(t-5)+2y(t-5)]](3)+t2y2(t-3)=0,t≥2,
where n=3, q(t)=t2, σ(t)=t-3, τ(t)=t-5, and p(t)=coste-5t2, f(y)=y5-2y, h(y)=y2. Hence, we have
limt→∞p(t)=limt→∞1e5t2cost=0,∫t0∞sn-1q(s)ds=∫t0∞s4ds=∞.
Since Conditions (C2) and (C3) of Theorem 3.3 are satisfied, every bounded solution of (3.24) oscillates or tends to zero at infinity.
ErbeL. H.KongQ.ZhangB. G.1995190New York, NY, USAMarcel DekkerMonographs and Textbooks in Pure and Applied Mathematics1309905GyöriI.LadasG.1991Oxford, UKClarendon PressOxford Mathematical Monographs1168471ZeinA.Abu-KaffT.Bounded oscillation of higher order neutral differential equations with oscillating coefficients200661261312219160ZBL1161.34343BolatY.AkinÖ.Oscillatory behaviour of higher order neutral type nonlinear forced differential equation with oscillating coefficients2004290130230910.1016/j.jmaa.2003.09.0622032242AgarwalR. P.GraceS. R.O'ReganD.2000Dordrecht, The NetherlandsKluwer Academic Publishers1774732ZBL1160.34313AgarwalR. P.GraceS. R.The oscillation of higher-order differential equations with deviating arguments1999383-418519910.1016/S0898-1221(99)00193-51703416ZBL0935.34058AgarwalR. P.GraceS. R.Oscillation of certain functional-differential equations1999385-614315310.1016/S0898-1221(99)00221-71707833ZBL0945.34050GraceS. R.LalliB. S.Oscillation theorems for nth
-order delay differential equations198391235236610.1016/0022-247X(83)90157-9690876ParhiN.Oscillations of higher order differential equations of neutral type200050115517310.1023/A:10224057073491745469ZBL1045.34043YuecaiF.Oscillatory behaviour of higher order neutral nonlinear neutral functional differential equation with oscillating coefficients19993611BohnerM.KarpuzB.ÖcalanÖ.Iterated oscillation criteria for delay dynamic equations of first order2008124586872475963ZBL1160.39302BainovD.MishevD. P.1995SingaporeWorld ScientificLaddeG. S.LakshmikanthamV.ZhangB. G.1987110New York, NY, USAMarcel DekkerMonographs and Textbooks in Pure and Applied Mathematics1017244BainovD.MishevD. P.1991Oxford, UKClarendon PressGraceS. R.LalliB. S.Oscillation theorems for certain neutral differential equations1988384745753962917ZBL0705.34072