The oscillation property of third-order differential equations with non-positive neutral coefficients is discussed. New sufficient conditions are provided to guarantee that every solution of the considered equation is almost oscillatory. Both the canonical and non-canonical cases are considered. Illustrative examples are introduced to support the obtained results.
Citation: A. A. El-Gaber, M. M. A. El-Sheikh, M. Zakarya, Amirah Ayidh I Al-Thaqfan, H. M. Rezk. On the oscillation of solutions of third-order differential equations with non-positive neutral coefficients[J]. AIMS Mathematics, 2024, 9(11): 32257-32271. doi: 10.3934/math.20241548
[1] | Maryam AlKandari . Nonlinear differential equations with neutral term: Asymptotic behavior of solutions. AIMS Mathematics, 2024, 9(12): 33649-33661. doi: 10.3934/math.20241606 |
[2] | Taher S. Hassan, Emad R. Attia, Bassant M. El-Matary . Iterative oscillation criteria of third-order nonlinear damped neutral differential equations. AIMS Mathematics, 2024, 9(8): 23128-23141. doi: 10.3934/math.20241124 |
[3] | Yun Xin, Hao Wang . Positive periodic solution for third-order singular neutral differential equation with time-dependent delay. AIMS Mathematics, 2020, 5(6): 7234-7251. doi: 10.3934/math.2020462 |
[4] | Zuhur Alqahtani, Insaf F. Ben Saud, Areej Almuneef, Belgees Qaraad, Higinio Ramos . New criteria for the oscillation of a class of third-order quasilinear delay differential equations. AIMS Mathematics, 2025, 10(2): 4205-4225. doi: 10.3934/math.2025195 |
[5] | Fahd Masood, Salma Aljawi, Omar Bazighifan . Novel iterative criteria for oscillatory behavior in nonlinear neutral differential equations. AIMS Mathematics, 2025, 10(3): 6981-7000. doi: 10.3934/math.2025319 |
[6] | Mohammed Ahmed Alomair, Ali Muhib . On the oscillation of fourth-order canonical differential equation with several delays. AIMS Mathematics, 2024, 9(8): 19997-20013. doi: 10.3934/math.2024975 |
[7] | M. Sathish Kumar, V. Ganesan . Asymptotic behavior of solutions of third-order neutral differential equations with discrete and distributed delay. AIMS Mathematics, 2020, 5(4): 3851-3874. doi: 10.3934/math.2020250 |
[8] | Ali Muhib, Hammad Alotaibi, Omar Bazighifan, Kamsing Nonlaopon . Oscillation theorems of solution of second-order neutral differential equations. AIMS Mathematics, 2021, 6(11): 12771-12779. doi: 10.3934/math.2021737 |
[9] | Duoduo Zhao, Kai Zhou, Fengming Ye, Xin Xu . A class of time-varying differential equations for vibration research and application. AIMS Mathematics, 2024, 9(10): 28778-28791. doi: 10.3934/math.20241396 |
[10] | Yibing Sun, Yige Zhao . Oscillatory and asymptotic behavior of third-order neutral delay differential equations with distributed deviating arguments. AIMS Mathematics, 2020, 5(5): 5076-5093. doi: 10.3934/math.2020326 |
The oscillation property of third-order differential equations with non-positive neutral coefficients is discussed. New sufficient conditions are provided to guarantee that every solution of the considered equation is almost oscillatory. Both the canonical and non-canonical cases are considered. Illustrative examples are introduced to support the obtained results.
In this article, we are concerned with the oscillatory behavior of solutions of a general class of third-order differential equations with non-positive neutral coefficients of the type
(d(l)(w′′(l))α)′+n∑i=1qi(l)yβi(τi(l))=0, l≥l0>0, | (1.1) |
where w(l)=y(l)−∑mj=1aj(l)y(δj(l)), α and βi are quotients of odd positive integers, and m,n are positive integers. We shall consider the following two cases: the canonical case
D(l0)=∫∞l01d1α(l)dl=∞, | (1.2) |
and the non-canonical case
D(l0)=∫∞l01d1α(l)dl<∞. | (1.3) |
Throughout the paper, we assume that
(H1) d(l)∈C([l0,∞),(0,∞)),aj(l)∈C([l0,∞)),0≤aj(l)≤a0j,∑mj=1a0j<1,j=1,2,...,m;
(H2) δj(l),τi(l),qi(l)∈C([l0,∞)),δj(l)≤l,liml→∞δj(l)=liml→∞τi(l)=∞, qi(l)>0, i=1,2,...,n,j=1,2,...,m.
Any nontrivial function y(l)∈C([ly,∞)),ly≥l0, such that w∈C2([ly,∞)), d(l)(w′′(l))α∈C1([ly,∞)) and y(l) satisfies (1.1) on [ly,∞) is called a solution of (1.1). Our attention is restricted to those solutions y(l) of (1.1) that satisfy sup{|y(l)|:l≥T}>0 for T≥ly. We tacitly suppose that (1.1) possesses such a solution. A solution y(l) of (1.1) is termed oscillatory if it has arbitrarily large zeros on [ly,∞); otherwise, it is said to be non-oscillatory. The equation itself is termed oscillatory if all its solutions oscillate.
In dynamical models, delay and oscillation effects are often formulated by means of external sources and/or nonlinear diffusion, perturbing the natural evolution of related systems (see, e.g., [1,2,3,4]). Recently, there has been considerable interest in studying the qualitative properties of solutions of various types of differential equations, see, e.g., [5,6,7], for the oscillation of second-order differential equations, while [8,9], for fourth-order differential equations, and the references [10,11,12], for the oscillation of nth-order differential equations. In fact, it is notable that the analysis of differential equations with non-positive neutral coefficients is more difficult in comparison with that of non-negative neutral coefficients. The same thing can be said also for the non-canonical case compared to the canonical case. Moreover, although there has been a lot of interest in the oscillatory behavior of solutions of third-order equations with non-negative neutral coefficients (see, e.g., [13,14,15,16]), for equations with non-positive neutral coefficients there are relatively fewer published results and so they are not very prevalent in the literature (see, e.g., [17,18,19,20]). For instance, we mention here some of the related works that motivate our work. In [20], Qiu established new oscillation criteria for the third-order nonlinear dynamic equation on time scales of the type,
(d1(l)([d2(l)((z(l))±b(l)z(γ(l))Δ)γ2]Δ)γ1)Δ+g(l,z(τ(l)))=0, | (1.4) |
with ∫∞l01d1γ11(l)Δl=∞,∫∞l01d1γ22(l)Δl=∞. The authors in [17] were concerned with the D.E.,
(d(l)[(v(l))±b(l)v(ξ(l))′′]α)′+g(l)vα(η(l))=0, l≥l0, | (1.5) |
and established numerous criteria for the so-called Hile and Nehari type under the assumptions 0≤b(l)≤1 and d′(l)≥0 in the canonical case (1.2). Meanwhile, Jiang et al. [19] were motivated by the work of Baculucova and Duzirina [17], and Li et al. [21], to establish an affirmative answer to the question about the possibility of obtaining asymptotic criteria for the D.E.,
(d(l)[(v(l)−∫bap(l,ξ)v(l,ζ)dξ)′′]α)′+∫dcg(l,ν)v(l,θ)dθ=0, | (1.6) |
where α>0 is a quotient of odd positive integers, 0≤∫bap(l,ζ)dζ≤p0<1, without the need for the restrictive condition d′(l)≥0. More recently Garce [22] studied the oscillatory behavior of solutions of the nonlinear differential equation
(v(l)−p(l)vα(τ(l)))′′′+g(l)vβ(σ(l))=0,l≥l0>0, | (1.7) |
with τ(l)≤l,σ(l)≤l,τ′(l)>0, and σ′(l)>0. Meanwhile, Garce et al. [18] were concerned with the oscillatory behavior of solutions of nonlinear differential equations of the type
(d(l)[(v(l)−p(l)vα(τ(l)))′′]γ)′+g(l)vβ(σ(l)=0,l≥l0>0, | (1.8) |
using comparison methods and integral conditions, with γ≥β,τ(l)≤l,σ(l)≤l,τ′(l)>0, and σ′(l)>0 in the canonical case (1.2).
The principal goal of this paper is to study the oscillatory behavior of solutions of the nonlinear third-order differential equation with non-positive neutral coefficients (1.1) in the two cases canonical (1.2) and non-canonical (1.3) by using Riccati transformation without the need for the restrictive condition d′(l)≥0. Moreover, we do not need specific restrictions on the functions τi(l), that is, τi(l) may be delayed or advanced; furthermore, we considered the two cases βi≥α and βi≤α.
This section is devoted to present some notations and lemmas needed for our results.
Define
D1(l,T)=∫lT1d1α(u)du, and D2(l,T)=∫lTD1(u,l1)du. |
We first start with the following two lemmas, which are very similar to Lemmas 2.1 and 2.2 of [19].
Lemma 2.1. Assume that y(l) is an eventually positive solution of (1.1), such that (1.2) be satisfied. Then there exists l1≥l0, such that for all l≥l1 the corresponding function w satisfies one of the following four cases:
i)w>0,w′>0,w′′>0,(d(w′′)α)′≤0, |
ii)w>0,w′<0,w′′>0,(d(w′′)α)′≤0, |
iii)w<0,w′<0,w′′>0,(d(w′′)α)′≤0, |
iv)w<0,w′<0,w′′<0,(d(w′′)α)′≤0. |
Lemma 2.2. If for any eventually positive solution y(l) of (1.1), the corresponding w(l) satisfies case (i) of Lemma 2.1, then for any l2>l1≥l0,
w(l)≥∫ll2∫vl1d−1α(h)dhdv∫ll1d−1α(h)dhw′(l), |
and w′(l)∫ll1d−1α(h)dh is nonincreasing eventually.
Now, we introduce the following preliminary result:
Lemma 2.3. If for any eventually positive solution y(l) of (1.1), the corresponding w(l) satisfies case (ii) of Lemma 2.1, and
∫∞l0∫∞v(1d(u)∫∞un∑i=1qi(s)ds)1αdudv=∞, | (2.1) |
then liml→∞y(l)=0.
Proof. Since w(l) satisfies property (ii), then there exists a finite constant M≥0 such that liml→∞w(l)=M. We claim that M=0. Otherwise, assume that M>0. By the definition of w, y(l)≥ w(l)>M. Consequently, by (1.1), we have
(d(l)(w′′(l))α)′=−n∑i=1qi(l)yβi(τi(l))≤−n∑i=1qi(l)wβi(τi(l))≤−n∑i=1Mβiqi(l)≤−Mκn∑i=1qi(l), | (2.2) |
where κ={minβiM≥1maxβiM<1. Integrating (2.2) from l to ∞, we obtain
w′′(l)≥Mκα(1d(l)∫∞ln∑i=1qi(s)ds)1α. |
Therefore, by integrating from l to ∞ and then integrating the result from l1 to ∞, it follows that
w(l1)≥Mκα∫∞l1∫∞v(1d(u)∫∞un∑i=1qi(s)ds)1αdudv. |
This contradicts (2.1). Hence, M=0 and liml→∞w(l)=0. Next, we claim that y(l) is bounded. If this is false, then there exists a sequence {lm} such that liml→∞lm=∞ and limm→∞y(lm)=∞, where y(lm)=max{y(s):l0≤s≤lm}. Since liml→∞δj(l)=∞,δj(lm)>l0 for sufficiently large m. By δj(l)≤l, we conclude that
y(δj(lm))=max{y(s):l0≤s≤δj(lm)}≤max{y(s):l0≤s≤lm}=y(lm), |
and so
w(lm)=y(lm)−m∑j=1aj(lm)y(δj(lm))≥y(lm)−m∑j=1aj(lm)y(lm)≥(1−m∑j=1a0j)y(lm), |
which yields liml→∞w(lm)=∞. This contradicts liml→∞w(l)=0, therefore y(l) is bounded, and hence we may suppose that limsupl→∞ y(l)=b0, where 0≤b0<∞. Then there exists a sequence {lk} such that liml→∞lk=∞ and liml→∞y(lk)=b0. Now assuming that b0>0, and letting
ϵ=b0(1−∑mj=1a0j)2∑mj=1a0j, |
we have y(δj(lk))<b0+ϵ eventually, and thus
0=limk→∞w(lk)≥limk→∞(y(lk)−m∑j=1a0j(b0+ϵ))=b0−m∑j=1a0j(b0+b0(1−∑mj=1a0j)2∑mj=1a0j)=b02(1−m∑j=1a0j)>0, |
which is a contradiction. Thus b0=0 and liml→∞y(l)=0. The proof is complete.
Theorem 3.1. Assume that βi≥α, i=1,...,n, (1.2) and (2.1) hold. Suppose that there exists I(l)∈C([l0,∞)) such that
I(l)≤inf{l,τi(l)}, liml→∞I(l)=∞. | (3.1) |
If there exist a function v(l)∈C1([l0,∞),(0,∞)), and a constant C1>0 such that, for all sufficiently large l1 ≥l0 and for some l3>l2>l1,
limsupl→∞∫ll3(v(u)n∑i=1qi(u)Cβi−α1N(u)−d(u)[v′(u)]α+1(α+1)α+1vα(u))du=∞, | (3.2) |
where
N(l)=[∫I(l)l2∫sl1d−1α(ϰ)dϰds∫ll1d−1α(ϰ)dϰ]α, | (3.3) |
then Eq (1.1) is almost oscillatory.
Proof. Assume that y(l) is an eventually positive solution of (1.1). Then there exists a l1≥l0 such that y(l)>0,y(δj(l))>0 and y(τi(l))>0 for l≥l1. It is clear by Lemma 2.1 that the function w(l) obeys one of four possible cases (i), (ii), (iii), or (iv). Assume first that case (i) is satisfied for l≥l1. Define the Riccati transformation ϕ(l) by
ϕ(l)=v(l)d(l)(w′′(l))α(w′(l))α, |
then ϕ(l)>0 for l≥l1, and
ϕ′(l)=v′(l)v(l)ϕ(l)+v(l)[d(l)(w′′(l))α]′(w′(l))α−αv(l)d(l)(w′′(l))α+1(w′(l))α+1. | (3.4) |
But since from (1.1) and the definition of w, we have
(d(l)(w′′(l))α)′=−n∑i=1qi(l)yβi(τi(l))≤−n∑i=1qi(l)wβi(τi(l)). |
Hence, since w′(l)>0 and τi(l)≥I(l), then
(d(l)(w′′(l))α)′≤−n∑i=1qi(l)wβi(I(l)). | (3.5) |
Then from (3.4), we have
ϕ′(l)≤v′(l)v(l)ϕ(l)−v(l)∑ni=1qi(l)wβi(I(l))(w′(l))α−α(ϕ(l))α+1α(d(l)v(l))1α. | (3.6) |
Now since w(l) is positive and increasing, then there exist a l2≥l1 and C1>0 such that
w(l)≥C1, l≥l2. | (3.7) |
This, with (3.6), leads to
ϕ′(l)≤v′(l)v(l)ϕ(l)−v(l)n∑i=1qi(l)Cβi−α1[w(I(l))w′(l)]α−α(ϕ(l))α+1α(d(l)v(l))1α. | (3.8) |
Since I(l)≤l, then by using the nonincreasing property of w′(l)∫ll1d−1α(h)dh (see Lemma 2.2), we obtain
w′(I(l))w′(l)≥∫I(l)l1d−1α(h)dh∫ll1d−1α(h)dh. | (3.9) |
Now by using Lemma 2.2, we have
[w(I(l))w′(l)]α=(w(I(l))w′(I(l))w′(I(l))w′(l))α≥N(l), | (3.10) |
and so, by substituting from (3.10) into (3.8), we obtain
ϕ′(l)≤v′(l)v(l)ϕ(l)−v(l)n∑i=1qi(l)Cβi−α1N(l)−α(ϕ(l))α+1α(d(l)v(l))1α. |
Applying the inequality
TV−RVα+1α≤ααTα+1(α+1)α+1Rα, R>0, | (3.11) |
with V=ϕ(l), R=α(d(l)v(l))1α and T= v′(l)v(l), we obtain
ϕ′(l)≤−v(l)n∑i=1qi(l)Cβi−α1N(l)+d(l)[v′(l)]α+1(α+1)α+1vα(l). |
By integrating from l3 (l3>l2) to l, we arrive at
∫ll3(v(u)n∑i=1qi(u)Cβi−α1N(u)−d(u)[v′(u)]α+1(α+1)α+1vα(u))du≤ϕ(l3), |
this contradicts (3.2). Now consider the case (ii), then by Lemma 2.3, liml→∞y(l)=0. In both cases (iii) and (iv), similar analysis to that in [19, Theorem 3.1], case (iii), and case (iv) can be used to arrive at the conclusion liml→∞y(l)=0. This completes the proof.
Theorem 3.2. Assume that βi≤α, i=1,...,n, (1.2) and (2.1) hold. Suppose that there exists I(l)∈C([l0,∞)) satisfies (3.1). Suppose further that there exist a function k(l)∈C1([l0,∞),(0,∞)), and a constant C2>0, for sufficiently large l1 ≥l0 and some l3>l2>l1. If
limsupl→∞∫ll3[k(s)n∑i=1qi(s)Cβi−α2Dβi2(I(s),l2)Dα2(s,l2)−1(α+1)α+1(k′(s))α+1Dα1(s,l1)kα(s)]ds=∞, | (3.12) |
then Eq (1.1) is almost oscillatory.
Proof. For the sake of contradiction, suppose that (1.1) has an eventually positive solution y(l). Then for any, l1≥l0, we have y(l)>0,y(δj(l))>0 and y(τi(l))>0,i=1,2,...,n,j=1,2,...,m. It is clear by Lemma 2.1, that the function w(l) obeys one of the four possible cases (i),(ii),(iii), or (iv). Assume first that case (i) is satisfied for l≥l1. Define
Ω(l)=k(l)d(l)(w′′(l))α(w(l))α, l≥l1, |
then Ω(l)>0, and by using (3.5), we obtain
Ω′(l)≤k′(l)k(l)Ω(l)−k(l)∑ni=1qi(l)wβi(I(l))(w(l))α−αk(l)d(l)(w′′(l))αw′(l)(w(l))α+1. | (3.13) |
But since w′(l)>0 and d(l)(w′′(l))α is nonincreasing, we obtain
w′(l)≥(d(l)(w′′(l))α)1αD1(l,l1). | (3.14) |
Substituting into (3.13), we obtain
Ω′(l)≤k′(l)k(l)Ω(l)−k(l)∑ni=1qi(l)wβi(I(l))(w(l))α−αk(l)D1(l,l1)d1+1α(l)(w′′(l))α+1(w(l))α+1. |
By the definition of Ω, we have
Ω′(l)≤k′(l)k(l)Ω(l)−k(l)∑ni=1qi(l)wβi(I(l))(w(l))α−αD1(l,l1)(Ω(l))α+1αk1α(l). |
By applying the inequality (3.11), with
R=αD1(l,l1)k1α(l),T=k′(l)k(l) and V=Ω(l), |
we obtain
Ω′(l)≤−k(l)∑ni=1qi(l)wβi(I(l))(w(l))α+1(α+1)α+1(k′(l))α+1Dα1(l,l1)kα(l). | (3.15) |
But since from (3.14), we have
(w′(l)D1(l,l1))′≤0, for l≥l2>l1, |
which leads to
w(l)≥∫ll2D1(s,l1)w′(s)D1(s,l1)ds≥w′(l)D1(l,l1)D2(l,l2). |
Hence
(w(l)D2(l,l2))′≤0. | (3.16) |
This with I(l)≤l, yields
w(I(l))w(l)≥D2(I(l),l2)D2(l,l2). | (3.17) |
Substituting from (3.17) into (3.15), we obtain
Ω′(l)≤−k(l)n∑i=1qi(l)wβi−α(I(l))[D2(I(l),l2)D2(l,l2)]α+1(α+1)α+1(k′(l))α+1Dα1(l,l1)kα(l). | (3.18) |
Now since by (3.16), w(l)D2(l,l2)is decreasing, there exists a constant C2>0 such that for l3>l2, we have
w(l)D2(l,l2)≤C2, for l≥l3. |
Substituting into (3.18), we obtain
Ω′(l)≤−k(l)n∑i=1qi(l)Cβi−α2Dβi2(I(l),l2)Dα2(l,l2)+1(α+1)α+1(k′(l))α+1Dα1(l,l1)kα(l). | (3.19) |
Integrating (3.19) from l3 to l, we obtain
∫ll3[k(s)n∑i=1qi(s)Cβi−α2Dβi2(I(s),l2)Dα2(s,l2)−1(α+1)α+1(k′(s))α+1Dα1(s,l1)kα(s)]ds≤Ω(l3). |
This contradicts (3.12). The proofs of the cases (ii)–(iv) are as in the proof of Theorem 3.1.
Now, we discuss the oscillatory behavior of Eq (1.1) in the non-canonical case (1.3).
Theorem 3.3. Assume that βi≥α, i=1,...,n, (1.3) and (2.1) hold. Assume that I(l) be as in Theorem 3.1, for sufficiently large l1 ≥l0 and for some l3>l2>l1, (3.2) is satisfied. Suppose further that there exist constants C3 >0 and 0<L<1, such that
limsupl→∞∫ll3[Dα(s)n∑i=1qi(s)Cβi−α3Lαsα−αL(I(s))αL−(αα+1)α+11D(s)d1α(s)]ds=∞, | (3.20) |
then Eq (1.1) is almost oscillatory.
Proof. For the sake of contradiction suppose that (1.1) has an eventually positive solution y(l). In view of (1.3), there exist six possible cases including (i)–(iv) (as in Lemma 2.1), and the two extra cases:
v)w>0,w′>0,w′′<0,(d(w′′)α)′≤0, |
vi)w<0,w′>0,w′′<0,(d(w′′)α)′≤0. |
The proofs of the four cases (i)–(iv) follow the same arguments of Theorem 3.1. Now consider the case(v). Since d(w′′)α is decreasing, then
d(s)(w′′(s))α≤d(l)(w′′(l))α, s≥l≥l1. | (3.21) |
Integrating from l to g and letting g→∞, we have
w′(l)≥−d1α(l)w′′(l)D(l). | (3.22) |
In view of case (v), since w(l)>0,w′(l)>0 and w′′(l)<0 on [l1,∞], for any constant L∈(0,1), we have
w(l)≥Llw′(l), for l≥l2≥l1. | (3.23) |
Now define
Φ(l)=d(l)[w′′(l)]α[w′(l)]α, | (3.24) |
then Φ(l)<0 for l≥l2 , and
Φ′(l)=[d(l)[w′′(l)]α]′[w′(l)]α−αd(l)[Φ(l)d(l)]α+1α. |
Thus, from (3.5), we have
Φ′(l)≤−∑ni=1qi(l)wβi(I(l))[w′(l)]α−αd(l)[Φ(l)d(l)]α+1α. | (3.25) |
Therefore, in view of (3.23), we obtain
(w(l)l1L)′≤0. | (3.26) |
But since I(l)≤l, then
w(I(l))w(l)≥[I(l)l]1L. | (3.27) |
This with (3.23), leads to
w(I(l))w′(l)≥LlI1L(l)l1L. |
Consequently, by substituting in (3.25), we have
Φ′(l)≤−n∑i=1qi(l)wβi−α(I(l))Lαlα−αL(I(l))αL−α[Φ(l)]α+1αd1α(l). | (3.28) |
Now since, from the positivity and increasing properties of w(l), there exists a constant C3>0 such that w(l)≥C3, then we have
Φ′(l)≤−n∑i=1qi(l)Cβi−α3Lαlα−αL(I(l))αL−α[Φ(l)]α+1αd1α(l). | (3.29) |
It is clear by (3.22) and (3.24) that
−Dα(l)Φ(l)≤1. | (3.30) |
Multiplying (3.29) by Dα(l) and integrating from l3 to l, we have
∫ll3[Dα(s)n∑i=1qi(s)Cβi−α3Lαsα−αL(I(s))αL+αDα(s)[Φ(s)]α+1αd1α(s)+αd−1α(s)Dα−1(s)Φ(s)]ds≤−Dα(l)Φ(l)+Dα(l3)Φ(l3). |
Set R=Dα(s)d1α(s),T=d−1α(s)Dα−1(s) and v=−Φ(s), then using the inequality (3.11), we have
∫ll3[Dα(s)n∑i=1qi(s)Cβi−α3Lαsα−αL(I(s))αL−(αα+1)α+11D(s)d1α(s)]ds≤1+Dα(l3)Φ(l3). |
This is a contradiction with (3.20). Assume case (vi) holds. Now, by using an argument similar to that used in Lemma 2.3, we arrive at the conclusion that liml→∞ y(l)=0. The proof is complete.
Theorem 3.4. Assume that βi≤α, i=1,...,n, (1.3) and (2.1) hold. Assume further that I(l) be as in Theorem 3.1, for sufficiently large l1 ≥l0 and for some l3>l2>l1, (3.12) is satisfied. If there exists a constant C4 >0, such that
limsupl→∞∫ll3[Dα(s)n∑i=1qi(s)Cβi−α4Lαsα−αL(I(s))βiL−(αα+1)α+11D(s)d1α(s)]ds=∞, | (3.31) |
then Eq (1.1) is almost oscillatory.
Proof. Let y(l) be a non-oscillatory solution of (1.1) such that y(l)>0, in view of (1.3), there exist six possible cases (i)−(vi) (as in Theorem 3.3). The proofs in the four cases (i)−(iv) are as in Theorem 3.2. Now assume that case (v) holds. Then following the same lines of the proof of Theorem 3.3, we arrive at (3.28), but since by (3.26), w(l)l1L is nonincreasing, then there exists a positive constant C4 such that
w(l)l1L≤C4, |
l≥l3≥l2. |
This with (3.28), leads to
Φ′(l)≤−n∑i=1qi(l)Cβi−α4Lαlα−αL(I(l))βiL−α[Φ(l)]α+1αd1α(l). | (3.32) |
Multiplying both sides of (3.32) by Dα(l) and integrating from l4(>l3) to l, and then applying the inequality (3.11), we obtain
∫ll4[Dα(s)n∑i=1qi(s)Cβi−α4Lαsα−αL(I(s))βiL−(αα+1)α+11D(s)d1α(s)]ds≤1+Dα(l4)Φ(l4). |
This contradicts (3.31). Assume that case (vi) holds. By a similar argument to that used in Lemma 2.3, we arrive at the conclusion that liml→∞ y(l)=0. The proof is complete.
Example 4.1. Consider the differential equation
(y(l)−1l2y(l5)−1l4y(l2))′′′+1l3y(l)+1l4y(2l)=0, l≥2. | (4.1) |
Here a1=1l2,a2=1l4,d(l)=1, q1(l)=1l3,q2(l)=1l4,δ1(l)=l5,δ2(l)=l2,τ1(l)=l,τ2(l)=2l, α=β1 =β2=1. Note that ∫∞l0d1α(s)ds=∫∞2ds=∞,
∫∞l0∫∞v(1d(u)∫∞un∑i=1qi(s)ds)1αdudv=∫∞2∫∞v∫∞u(1s3+1s4)dsdudv=∞. |
Choosing v(l)=l and I(l)=l, we have
N(l)=[∫I(l)l2∫sl1d−1α(h)dhds∫ll1d−1α(h)dh]α=∫ll2∫sl1dhds∫ll1dh=l2−2ll1+λ2(l−l1), λ=2l1l2−l22, |
and
limsupl→∞∫ll3(v(ι)n∑i=1qi(ι)Cβi−α1N(ι)−d(ι)[v′(ι)]α+1(α+1)α+1vα(ι))dι=limsupl→∞∫ll3(ι(1ι3+1ι4)(ι2−2ιl1+λ2(ι−l1))−14ι)dι=limsupl→∞∫ll3[12(ι−l1)−l1ι(ι−l1)+λ2ι2(ι−l1)+12ι(ι−l1)−l1ι2(ι−l1)+λ2ι3(ι−l1)−14ι]dι=∞. |
Thus, by Theorem 3.1, Eq (4.1) is almost oscillatory.
Example 4.2. Consider the differential equation
[l43(y(l)−12ly(l2)−17ly(l3))′′]′+1l53y3(l2)+1l4y5(l3)=0, l≥1. | (4.2) |
Here a1=12l,a2=17l,d(l)=l43, q1(l)=1l53, q2(l)=1l4,δ1(l)=l2, δ2(l)=l3,τ1(l)=l2, τ2(l)=l3, α=1,β1=3,β2=5. Note that ∫∞l0d1α(s)ds=∫∞1l−43ds<∞,
∫∞l0∫∞v(1d(u)∫∞un∑i=1qi(s)ds)1αdudv=∫∞1∫∞v1u43∫∞u(1s53+1s4)dsdudv=∞. |
Taking I(l)=l3, we have
N(l)=[∫I(l)l1∫sl2d−1α(h)dhds∫l1d−1α(h)dh]α=∫l3l2∫sl1h−43dhds∫ll1h−43dh=−33√32l+l43l−131+γl133l13l−131−3,γ=92l232−3l2l−131, |
choosing l1=1,l2=(1,2), l3>6 and v(l)=1l13, then
limsupl→∞∫ll3(v(ι)n∑i=1qi(ι)Cβi−α1N(ι)−d(ι)[v′(ι)]α+1(α+1)α+1vα(ι))dι=limsupl→∞∫ll3(1ι13(C21ι53+C41ι4)(−33√32ι+ι43l−131+γι133ι13l−131−3)−136ι)dι≥limsupl→∞∫ll3(1ι13(C21ι53+C41ι4)(−33√32ι+ι43l−131+γι133ι13l−131)−136ι)dι=limsupl→∞∫ll3(−3√3C212ι43+γC213ι2+C213ι+C41ι134(−33√32ι+ι43+γι133ι13)−136ι)dι=∞, |
for C_{1}^{2} > \frac{1}{12} , and
\begin{eqnarray*} &&\underset{l\rightarrow \infty }{\lim \sup }\int_{l_{3}}^{l}\left[ D^{\alpha }\left( s\right) \sum\limits_{i = 1}^{n}q_{i}\left( s\right) C_{3}^{\beta _{i}-\alpha }L^{\alpha }s^{\alpha -\frac{\alpha }{L}}\left( I\left( s\right) \right) ^{\frac{\alpha }{L}}-\left( \frac{\alpha }{\alpha +1}\right) ^{\alpha +1}\frac{1}{D\left( s\right) d^{\frac{1}{\alpha }}\left( s\right) } \right] ds \\ & = &\underset{l\rightarrow \infty }{\lim \sup }\int_{l_{3}}^{l}\left( \frac{3 }{s}C_{3}^{2}L\left( \frac{1}{3}\right) ^{\frac{1}{L}}+\frac{3}{s^{\frac{10}{ 3}}}C_{3}^{4}L\left( \frac{1}{3}\right) ^{\frac{1}{L}}-\frac{1}{12s}\right) ds = \infty , \end{eqnarray*} |
for C_{3}^{2}L\left(\frac{1}{3}\right) ^{\frac{1}{L}} > \frac{1}{36}. Thus, by Theorem 3.3, Eq (4.2) is almost oscillatory.
In this article, we discussed a general class of third-order differential equations with non-positive neutral coefficients (1.1) in the two cases of canonical and non-canonical conditions. Our criteria do not need to determine whether the functions \tau _{i}\left(l\right) are delayed or advanced. Moreover, our new criteria do not need the restrictive condition d^{\prime }\left(l\right) \geq 0 .
A. A. El-Gaber: Investigation, Software, Supervision, Writing-original draft; M. M. A. El-Sheikh: Investigation, Software, Supervision, Writing-original draft; M. Zakarya: Writing-review editing and Funding; A.A. I Al-Thaqfan: Writing-review editing and Funding; H. M. Rezk: Investigation, Softwar, Writing-original draft. All authors have read and agreed to the published version of the manuscript.
The authors extend their appreciation to the Deanship of Research and Graduate Studies at King Khalid University for funding this work through large Research Project under grant number RGP 2/190/45.
The authors declare that there are no conflicts of interest in this paper.
[1] |
A. Columbu, S. Frassu, G. Viglialoro, Refined criteria toward boundedness in an attraction-repulsion chemotaxis system with nonlinear productions, Appl. Anal., 103 (2023), 415–431. https://doi.org/10.1080/00036811.2023.2187789 doi: 10.1080/00036811.2023.2187789
![]() |
[2] |
T. Li, S. Frassu, G. Viglialoro, Combining effects ensuring boundedness in an attraction-repulsion chemotaxis model with production and consumption, Z. Angew. Math. Phys., 74 (2023), 109. https://doi.org/10.1007/s00033-023-01976-0 doi: 10.1007/s00033-023-01976-0
![]() |
[3] |
T. Li, N. Pintus, G. Viglialoro, Properties of solutions to porous medium problems with different sources and boundary conditions, Z. Angew. Math. Phys., 70 (2019), 86. https://doi.org/10.1007/s00033-019-1130-2 doi: 10.1007/s00033-019-1130-2
![]() |
[4] |
T. Li, G. Viglialoro, Boundedness for a nonlocal reaction chemotaxis model even in the attraction-dominated regime, Differ. Integral Equ., 34 (2021), 315–336. https://doi.org/10.57262/die034-0506-315 doi: 10.57262/die034-0506-315
![]() |
[5] |
R. P. Agarwal, C. Zhang, T. Li, Some remarks on oscillation of second order neutral differential equations, Appl. Math. Comput., 274 (2016), 178–181. https://doi.org/10.1016/j.amc.2015.10.089 doi: 10.1016/j.amc.2015.10.089
![]() |
[6] |
J. Džurina, S. R. Grace, I. Jadlovská, T. Li, Oscillation criteria for second-order Emden-Fowler delay differential equations with a sublinear neutral term, Math. Nachr., 293 (2020), 910–922. https://doi.org/10.1002/mana.201800196 doi: 10.1002/mana.201800196
![]() |
[7] |
M. M. A. El-Sheikh, Oscillation and nonoscillation criteria for second order nonlinear differential equations, I, J. Math. Anal. Appl., 179 (1993), 14–27. https://doi.org/10.1006/jmaa.1993.1332 doi: 10.1006/jmaa.1993.1332
![]() |
[8] |
A. A. El-Gaber, M. M. A. El-Sheikh, E. I. El-Saedy, Oscillation of super-linear fourth-order differential equations with several sub-linear neutral terms, Bound. Value Probl., 2022 (2022), 41. https://doi.org/10.1186/s13661-022-01620-2 doi: 10.1186/s13661-022-01620-2
![]() |
[9] |
A. A. El-Gaber, M. M. A. El-Sheikh, Oscillation of fourth-order neutral differential equations with distributed deviating arguments, J. Math. Comput. Sci., 28 (2023), 60–71. https://doi.org/10.22436/jmcs.028.01.06 doi: 10.22436/jmcs.028.01.06
![]() |
[10] | C. Dharuman, N. Prabaharan, E. Thandapani, E. Tunç, Oscillatory behavior of even-order functional differential equations with a superlinear neutral term, Palest. J. Math., 12 (2023), 722–731. |
[11] |
A. A. El-Gaber, Oscillatory criteria of noncanonical even-order differential equations with a superlinear neutral term, Bound. Value Probl., 2024 (2024), 67. https://doi.org/10.1186/s13661-024-01873-z doi: 10.1186/s13661-024-01873-z
![]() |
[12] |
C. Zhang, R. P. Agarwal, M. Bohner, T. Li, New results for oscillatory behavior of even-order half-linear delay differential equations, Appl. Math. Lett., 26 (2013), 179–183. https://doi.org/10.1016/j.aml.2012.08.004 doi: 10.1016/j.aml.2012.08.004
![]() |
[13] | B. Qaraad, O. Bazighifan, T. A. Nofal, A. H. Ali, Neutral differential equations with distribution deviating arguments: oscillation conditions, J. Ocean Eng. Sci., 2022. https://doi.org/10.1016/j.joes.2022.06.032 |
[14] |
A. A. El-Gaber, M. M. A. El-Sheikh, S. A. A. El-Marouf, New oscillation and non-oscillation criteria for third order neutral differential equations with distributed deviating arguments, J. Math. Comput. Sci., 32 (2024), 283–294. https://doi.org/10.22436/jmcs.032.04.01 doi: 10.22436/jmcs.032.04.01
![]() |
[15] |
T. Li, Y. V. Rogovchenko, On the asymptotic behavior of solutions to aclass of third-order nonlinear neutral differential equations, Appl. Math. Lett., 105 (2020), 1–7. https://doi.org/10.1016/j.aml.2020.106293 doi: 10.1016/j.aml.2020.106293
![]() |
[16] |
Q. Liu, S. R. Grace, I. Jadlovská, E. Tunč, T. Li, On the asymptotic behavior of noncanonical third-order Emden-Fowler delay differential equations with a superlinear neutral term, Mathematics, 10 (2022), 2902. https://doi.org/10.3390/math10162902 doi: 10.3390/math10162902
![]() |
[17] |
B. Baculková, J. Džurina, Oscillation of third-order neutral differential equations, Math. Comput. Model., 52 (2010), 215–226. https://doi.org/10.1016/j.mcm.2010.02.011 doi: 10.1016/j.mcm.2010.02.011
![]() |
[18] |
S. R. Grace, J. R. Graef, E. Tunç, Oscillatory behaviour of third order nonlinear differential equations with a nonlinear nonpositive neutral term, J. Taibah Univ. Sci., 13 (2019), 704–710. https://doi.org/10.1080/16583655.2019.1622847 doi: 10.1080/16583655.2019.1622847
![]() |
[19] |
C. Jiang, Y. Jiang, T. Li, Asymptotic behavior of third-order differential equations with nonpositive neutral coefficients and distributed deviating arguments, Adv. Differ. Equ., 2016 (2016), 105. https://doi.org/10.1186/s13662-016-0833-3 doi: 10.1186/s13662-016-0833-3
![]() |
[20] |
Y. C. Qiu, Oscillation criteria of third-order nonlinear dynamic equations with nonpositive neutral coefficients on time scales, Adv. Differ. Equ., 2015 (2015), 229. https://doi.org/10.1186/s13662-015-0636-y doi: 10.1186/s13662-015-0636-y
![]() |
[21] |
T. Li, C. Zhang, G. Xing, Oscillation of third-order neutral delay differential equations, Abstr. Appl. Anal., 2012 (2012), 569201. https://doi.org/10.1155/2012/569201 doi: 10.1155/2012/569201
![]() |
[22] |
S. R. Grace, Oscillation criteria for third order nonlinear differential equations with a nonpositive neutral terms, Adv. Anal., 4 (2019), 1–7. https://doi.org/10.22606/aan.2019.41001 doi: 10.22606/aan.2019.41001
![]() |
1. | Salah Benhiouna, Azzeddine Bellour, Reemah Alhuzally, Ahmad M. Alghamdi, Existence of Solutions for Generalized Nonlinear Fourth-Order Differential Equations, 2024, 12, 2227-7390, 4002, 10.3390/math12244002 |