Skip to main content

ORIGINAL RESEARCH article

Front. Appl. Math. Stat., 16 July 2024
Sec. Mathematical Physics

Oscillatory behavior of solutions of second-order non-linear differential equations with mixed non-linear neutral terms

  • 1Department of Mathematics, Faculty of Science, Benha University, Benha, Egypt
  • 2Section of Mathematics, International Telematic University Uninettuno, Rome, Italy
  • 3Department of Mathematics, Faculty of Science, Mansoura University, Mansoura, Egypt
  • 4Department of Statistics and Operations Research, College of Science, King Saud University, Riyadh, Saudi Arabia

This study primarily seeks to expand upon these developments by encompassing neutral differential equations of mixed type, incorporating both delay and advanced terms, particularly in the case of the canonical operator. The presented results are derived from the application of the comparison method, Riccati transformation, and integral averaging technique. These methodologies lead to substantial improvements and extensions of existing results found in the literature. Additionally, illustrative examples are provided to demonstrate the practical implications of the developed criteria.

1 Introduction

In recent years, scholars have shown a growing interest in studying the oscillatory behavior of solutions to second-order differential and dynamic equations. This attention is motivated by the significance of such behavior in real-life applications, such as steam turbine regulation, neural networks, and governing equations which describe the temporal variation of hormones; see [19]. Moreover, there exist numerous sophisticated equations that serve as applications or direct representations of problems reliant on both present and future rates of change for further applications in science and technology; see [10]. An advanced argument is characterized by its ability to characterize the impact of a hypothetical future acts. It is commonly observed in phenomena such as population dynamics and economic difficulties.

In this study, we discuss the oscillatory behavior of second-order non-linear neutral differential equations with mixed non-linear neutral terms of the form

(a(ζ)Φα(z(ζ)))+q(ζ)Φγ(x(σ(ζ)))+r(ζ)Φβ(x(φ(ζ)))=0,  ζζ0>0,    (1)

and

z(ζ)=Φη(x(ζ))+p1(ζ)Φλ(x(τ1(ζ)))+εp2(ζ)Φν(x(τ2(ζ))),

where Φ*(s)=|s|*-1for*>0 and ε = ±1. Throughout this article, we assume that the following hypotheses hold

(H1) α, η, λ, ν, γ and β are ratios of odd positive integers;

(H2) a, p1, p2, q and rC([ζ0, ∞), (0, ∞)), Equation (1) is in canonical form, i.e.,

ζ0a-1/α(s)ds=;    (2)

(H3) τ1, τ2, σ, ζ ∈ C([ζ0, ∞), ℝ) such that σ(ζ) ≤ ζ, τ1,τ2>0 and φ(ζ) ≥ ζ with

limζτ1(ζ)=limζτ2(ζ)=limζσ(ζ)=.

By a solution of Equation (1), we mean a non-trivial function xC([Tx, ∞), ℝ), Tx ≥ ζ0, which has the properties zC1([Tx,),), a(z)αC([Tx,),) and satisfies Equation (1) on [Tx, ∞). Our attention is restricted to those solutions x(ζ) of Equation (1) satisfying sup {|x(ζ)| : ζ ≥ T} > 0 for all TTx. We assume that Equation (1) possesses such a solution. A solution of Equation (1) is called oscillatory if it has arbitrarily large zeros on [Tx, ∞); otherwise, it is termed non-oscillatory. Equation (1) is said to be oscillatory if all solutions are oscillatory.

The literature extensively addresses the oscillation and asymptotic behavior of solutions across various classes of delay and advanced differential equations. Recently, Tunç and Özdemir [11] introduced novel sufficient conditions for solutions to second-order half-linear differential equations.

(a(ζ)((x(ζ)+p1(ζ)x(g1(ζ))+p2(ζ)x(g2(ζ))))α)+q(ζ)xα(h(ζ))=0,

where g1(ζ) < ζ, g2(ζ) > ζ, p1(ζ) ≥ 0 and p2(ζ) ≥ 1, p2(ζ) ≢ 1 for large ζ. In the study by Agwa et al. [12], developed new oscillation criteria for the second-order non-linear neutral dynamic equation with mixed arguments

    (a(ζ)((x(ζ)+p1(ζ)x(τ1(ζ))+p2(ζ)x(τ2(ζ)))Δ)α)Δ+f((ζ,x(σ(ζ)))+g(ζ,x(φ(ζ)))=0,

in the canonical and non-canonical cases. In the study by Moaaz et al. [9], investigated the oscillatory and asymptotic properties of a specific class of delay differential equations of mixed neutral type

(a(ζ)((x(ζ)+p1(ζ)x(τ1(ζ))+p2(ζ)x(τ2(ζ))))α)+q(ζ)xα(σ(ζ))+r(ζ)xα(φ(ζ))=0.

with the non-canonical operator.

Very recently, in the study by Grace et al. [13], introduced novel criteria for the oscillation of second-order non-linear differential equations featuring mixed non-linear neutral terms and mixed deviating arguments

(a(ζ)((x(ζ)+p1(ζ)xλ(τ1(ζ))-p2(ζ)xν(τ2(ζ))))α)+q(ζ)xα(σ(ζ))+r(ζ)xα(φ(ζ))=0.

Upon reviewing the literature, it becomes evident that numerous results pertain to the oscillation of second-order differential equations with linear neutral terms. In contrast, there is a paucity of articles dedicated to exploring differential equations featuring sublinear or superlinear neutral terms, as evidenced in the studies by Agarwal et al. [14], Bohner et al. [15], Džurina et al. [16], Grace et al. [17], Lin and Tang [18], and Muhib et al. [19]. Furthermore, there is a notable scarcity of results concerning equations incorporating both sublinear and superlinear neutral terms. Motivated by this gap, this study aims to establish oscillation criteria for a specific class of second-order mixed functional differential equations (Equation 1) characterized by sublinear and superlinear neutral terms under the conditions where either ε = −1 or ε = +1, η differs from 1. To the best of our knowledge, there does not appear to be any oscillation results for Equation (1) when η ≠ 1. Additionally, it is worth noting that the findings presented in this study are novel even in the linear case.

2 Preliminaries

In discussing oscillation results for Equation (1), we assume that any functional inequality holds for all large ζ. To lay the groundwork for our key results, we first define a few lemmas. To keep things simple, we will use these symbols

A(ζ,ζ1)=ζ1ζa-1/α(s)ds,   R1=(1-λ)p11/(1-λ)(ζ)λλ/(λ-1),                     R2=(ν-1)νν/(ν-1)p21/(1-ν)(ζ),   ρ+=max{ρ,0}
δi(ζ)={τ(i+1)/2-1(τ(5-i)/2(τ(i+1)/2-1(ζ))),i=1,3τi/2-1(τi/2-1(ζ)),i=2,4,   δ5=τ2-1(τ1-1(ζ)),
S1=ηλA(δ2(ζ)),ζ3)A(τ1-1(ζ),ζ3)p1η/λ(δ2(ζ))+νλp2(τ1-1(ζ))A(δ1(ζ),ζ3)A(τ1-1(ζ),ζ3)p1ν/λ(δ1(ζ))    +(1-ηλc*p1η/λ(δ2(ζ))+(1-νλ)p2(τ1-1(ζ))c*p1ν/λ(δ1(ζ)))S2=ηνp2η/ν(δ4(ζ))+λνp1(τ2-1(ζ))p2λ/ν(δ3(ζ))+(1-ην)c*p2η/ν(δ4(ζ))    +(1-λν)p1(τ2-1(ζ))c*p2λ/ν(δ3(ζ))S3=ηνp2η/ν(δ5(ζ))-νλp2(τ1-1(ζ))A(δ1(ζ),ζ3)p1ν/λ(δ1(ζ))A(τ1-1(ζ),ζ3)+(1-ην)c*p2η/ν(δ5(ζ))    -(1-νλ)p2(τ1-1(ζ))c*p1ν/λ(δ1(ζ)),  c*>0, ζ3[ζ0,).

Lemma 1. Baculíková [20] and Philos [21] Let q1:[ζ0, ∞) → (0, ∞), g1 : [ζ0, ∞) → ℝ and f : ℝ → ℝ are continuous functions, f is non-decreasing with xf(x) > 0 for x ≠ 0 and g1(ζ) → ∞ as ζ → ∞. If

(i) The first-order delay differential inequality (i.e., g1(ζ) ≤ ζ)

y1(ζ)+q1(ζ)f(y1(g1(ζ)))0

has an eventually positive solution, so does the corresponding delay differential equation.

(ii) The first-order advanced differential inequality (i.e., g1(ζ) ≥ ζ)

y1(ζ)-q1(ζ)f(y1(g1(ζ)))0

has an eventually positive solution, so does the corresponding advanced differential equation.

Lemma 2. Hardy et al. [22] If X, Y ≥ 0, then

Xξ+(ξ-1)Yξ-ξXYξ-10,  for ξ>1    (3)

and

Xξ-(1-ξ)Yξ-ξXYξ-10,  for 0<ξ<1,    (4)

where equalities hold if and only if X = Y.

Lemma 3. Hardy et al. [22] If B, L be non-negative numbers and if m, n > 1 are real numbers such that 1n+1m=1, then

BL1nBn+1mLm.

Equality holds if and only if Bn = Lm.

Lemma 4. Bohner et al. [23] Let G(U)=AU-B(U-R)μ+1μ, where B > 0, A and B are constants, μ is a ratio of odd positive integers. Then, G attains its maximum value at U*=R+(μA/((μ+1)B)μ) and

maxUG(U)=G(U*)=AR+μμ(μ+1)μ+1Aμ+1Bμ.    (5)

3 The case ε = −1

In this section, we investigate the oscillatory characteristics of solutions to Equation (1) under the conditions where ε = −1, τ1(ζ) = τ2(ζ) = τ(ζ), and either of the following conditions is satisfied either λ < 1 and ν > 1, or λ < ν ≤ 1.

Theorem 1. Suppose τ1(ζ) = τ2(ζ) = τ(ζ), λ < 1, ν > 1, and conditions (H1)-(H3) are satisfied. Additionally, assume that

limζ[R1(ζ)+R2(ζ)]=0,    (6)

and there exists a non-decreasing function ϱ(ζ) ∈ C([ζ0, ∞), ℝ) such that

τ-1(σ(ζ))ϱ(ζ)ζ,τ-1(φ(ζ))ζandτ-1(φ(ϱ(ϱ(ζ))))ζ,    (7)

for ζ ≥ ζ0. If there exist constant c2 ∈ (0, 1) such that the first-order differential equations

ϝ(ζ)+c2γq(ζ)Aγ/η(σ(ζ),ζ2)ϝγ/αη(σ(ζ))=0,    (8)
Ω(ζ)+q(ζ)(A(ϱ(ζ),τ-1(σ(ζ)))p2(τ-1(σ(ζ))))γ/νΩγ/αν(ϱ(ζ))=0,    (9)
z̄(ζ)-(1a(ζ)ϱ(ζ)ζr(s)p2β/ν(τ-1(φ(s)))ds)1/αz̄β/να(τ-1(φ(ϱ(ζ))))=0    (10)

are oscillatory for sufficiently large ζ2 > ζ1 ≥ ζ0, then every solution of Equation (1) is oscillatory.

Proof. Consider a non-oscillatory solution x(ζ) of Equation (1). Without loss of generality, let us assume that x(ζ) is eventually positive with limζx(ζ)0 for ζ ≥ ζ0. Hence, x(ζ) > 0, x(τ(ζ)) > 0, x(σ(ζ)) > 0, and x(φ(ζ)) > 0 for ζ ≥ ζ1 ≥ ζ0. It is worth noting that the proof for the case where x(ζ) is eventually negative follows a similar path and is therefore omitted. Now, from Equation (1), it can be inferred that

(a(ζ)(z(ζ))α)=-q(ζ)xγ(σ(ζ))-r(ζ)xβ(φ(ζ))<0.

Thus, (a(ζ)(z′(ζ))α) is decreasing and eventually of one sign. In other words, there exists ζ2 ≥ ζ1 such that z′(ζ) > 0 or z′(ζ) < 0 for ζ ≥ ζ2. Therefore, we will distinguish the following four cases:

Case(I):    z(ζ) > 0 and z′(ζ) < 0,

Case(II):   z(ζ) > 0 and z′(ζ) > 0,

Case(III):  z(ζ) < 0 and z′(ζ) > 0,

Case(IV):  z(ζ) < 0 and z′(ζ) < 0.

Initially, let us examine Case (I). Depending on the fact that z′(ζ) < 0 and (a(ζ)(z′(ζ))α) < 0, in accordance with Equation (2), we conclude that limζz(ζ)=-, contradicting the established condition z(ζ) > 0. Therefore, Case (I) is deemed impossible. For the second case, based on the definition of z(ζ), we have

z(ζ)=xη(ζ)-p2(ζ)xν(τ(ζ))+p1(ζ)xλ(τ(ζ)).    (11)

Utilizing inequality (5) with U = xη(ζ), A = 1, B = p2(ζ), R = xη(ζ)−x(τ(ζ)), and μ=1ν-1, we deduce

xη(ζ)-p2(ζ)xν(τ(ζ))xη(ζ)-x(τ(ζ))                                              +(1ν-1)1/(ν-1)p21/(ν-1)(ζ)(νν-1)ν/(ν-1)                                              xη(ζ)-x(τ(ζ))+(ν-1)νν/(ν-1)p21/(1-ν)(ζ).

This with Equation (11) leads to

xη(ζ)z(ζ)+x(τ(ζ))-p1(ζ)xλ(τ(ζ))-(ν-1)νν/(ν-1)p21/(1-ν)(ζ).    (12)

Applying the inequality (4) to [p1(ζ)xλ(τ(ζ))-x(τ(ζ))] with ξ = λ, X = x(τ(ζ)) and Y=1(λp1(ζ))1/(λ-1), we obtain

p1(ζ)xλ(τ(ζ))-x(τ(ζ))=p1(ζ)[xλ(τ(ζ))-λλp1(ζ)x(τ(ζ))](1-λ)p11/(1-λ)(ζ)λλ/(λ-1).

It is deduced from Equation (12) that

xη(ζ)z(ζ)-(ν-1)νν/(ν-1)p21/(1-ν)(ζ)-(1-λ)λλ/(λ-1)p11/(1-λ)(ζ)=(1-(R1(ζ)+R2(ζ))z(ζ))z(ζ).

As z(ζ) is positive and increasing, there exists a constant c > 0, ensuring z(ζ) ≥ c for ζ ≥ ζ2. This implies

xη(ζ)(1-(R1(ζ)+R2(ζ))c)z(ζ).    (13)

Therefore, based on Equations (6, 13), a constant c1 ∈ (0, 1) exists, such that

x(ζ)c11/ηz1/η(ζ)=c2z1/η(ζ);  c2=c11/η   for ζζ3ζ2.    (14)

Combining Equation (14) with Equation (1), we get

(a(ζ)(z(ζ))α)-c2γq(ζ)zγ/η(σ(ζ))-c2βr(ζ)zβ/η(φ(ζ)).    (15)

Consequently, since z(ζ) is increasing,

z(ζ)ζ3ζa1/α(s)z(s)a1/α(s)dsA(ζ,ζ3)(a1/α(ζ)z(ζ)).    (16)

This with Equation (15) leads to

(a(ζ)(z(ζ))α)+c2γq(ζ)Aγ/η(σ(ζ),ζ3)(a(σ(ζ))(z(σ(ζ)))α)γ/αη0.

Letting ϝ(ζ): = a(ζ)(z′(ζ))α, we have

ϝ(ζ)+c2γq(ζ)Aγ/η(σ(ζ),ζ3)ϝγ/αη(σ(ζ))0.    (17)

By applying Lemma 1 (i), the differential Equation (8) associated with the inequality (17) reveals the existence of a positive solution, leading to a contradiction.

Now, suppose the case (III) holds, we have z(ζ) < 0 and z′(ζ) > 0. Let

z̄(ζ)=-z(ζ)=-xη(ζ)-p1(ζ)xλ(τ(ζ))+p2(ζ)xν(τ(ζ))p2(ζ)xν(τ(ζ)),

it follows that

x(ζ)(z̄(τ-1(ζ))p2(τ-1(ζ)))1/ν.    (18)

It can be inferred from Equation (1) that

(a(ζ)(z̄(ζ))α)=q(ζ)xγ(σ(ζ))+r(ζ)xβ(φ(ζ))q(ζ)xγ(σ(ζ))q(ζ)(z̄(τ-1(σ(ζ)))p2(τ-1(σ(ζ))))γ/ν.    (19)

Since z̄(ζ)<0, we have

z̄(m)-z̄(n)=-mn(a(s)(z̄(s))α)1/αa1/α(s)dsA(n,m)(-a1/α(n)z̄(n)),ζ2mn.

Letting m = τ−1(σ(ζ)) and n = ϱ(ζ), we conclude

z̄(τ-1(σ(ζ)))A(ϱ(ζ),τ-1(σ(ζ)))(-a1/α(ϱ(ζ))z̄(ϱ(ζ))).    (20)

Combining Equations (19, 20) yields

(a(ζ)(z̄(ζ))α)-q(ζ)(A(ϱ(ζ),τ-1(σ(ζ)))p2(τ-1(σ(ζ))))γ/ν(a(ϱ(ζ))(z̄(ϱ(ζ)))α)γ/αν.

That is as follows:

Ω(ζ)+q(ζ)(A(ϱ(ζ),τ-1(σ(ζ)))p2(τ-1(σ(ζ))))γ/νΩγ/αν(ϱ(ζ))0,    (21)

where Ω(ζ):=(a(ζ)(z̄(ζ))α). By applying Lemma 1 (i), we conclude that the Equation (9) corresponding to the inequality (21) also possesses a positive solution, leading to a contradiction.

Ultimately, examine the case (IV) wherein z(ζ) < 0 and z′(ζ) < 0. As in the preceding case, let us take z̄(ζ)=-z(ζ). Utilizing Equation (18) in conjunction with Equation (1), we obtain

(a(ζ)(z̄(ζ))α)=q(ζ)xγ(σ(ζ))+r(ζ)xβ(φ(ζ))                               q(ζ)(z̄(τ-1(σ(ζ)))p2(τ-1(σ(ζ))))γ/ν                               +r(ζ)(z̄(τ-1(φ(ζ)))p2(τ-1(φ(ζ))))β/ν                               r(ζ)(z̄(τ-1(φ(ζ)))p2(τ-1(φ(ζ))))β/ν.

Integrating from ϱ(ζ) to ζ, we get

z̄(ζ)(1a(ζ)ϱ(ζ)ζr(s)p2β/ν(τ-1(φ(s)))ds)1/αz̄β/να(τ-1(φ(ϱ(ζ)))),    (22)

which has a positive solution z̄(ζ). It follows from Lemma 1 (ii) that the Equation 10 corresponds the inequality (22) also has a positive solution. This completes the proof.

Corollary 1. Assume that τ1(ζ) = τ2(ζ) = τ(ζ), λ < 1, ν > 1, and (H1)-(H3) hold. Furthermore, assume that Equation (6) holds and there exists a non-decreasing function ϱ(ζ) ∈ C([ζ0, ∞), ℝ) satisfies Equation (7). If

limζζ0ζq(s)Aγ/η(σ(s),s3)ds=    for γ<αη,
limζζ0ζq(s)(A(ϱ(s),τ-1(σ(s)))p2(τ-1(σ(s))))γ/νds=    for γ<αν,

and

limζζ0ζ(1a(s)ϱ(s)sr(u)p2β/ν(τ-1(φ(u)))du)1/αds=    for β>αν

hold for sufficiently large ζ2 > ζ1 ≥ ζ0, then every solution of Equation (1) is oscillatory.

Theorem 2. Assume that τ1(ζ) = τ2(ζ) = τ(ζ), λ < 1, ν > 1, β > αη, and (H1)-(H3) hold. Furthermore, assume that Equation (6) holds and there exists a non-decreasing function ϱ(ζ) ∈ C([ζ0, ∞), ℝ) satisfies Equation (7). If there exist function ρ(ζ)C1([ζ0,),(0,)) and c* > 0 such that

limsupζζ0ζ[c2βr(s)ρ(s)-(αηa1/α(s)βc*β/(ηα-1)ρ(s))α(ρ+(s)(α+1))α+1]ds=,    (23)

and the differential Equations (9) and (10) are oscillatory for sufficiently large ζ1 ≥ ζ0, then every solution of Equation (1) is oscillatory.

Proof. Let x(ζ) be a non-oscillatory solution of Equation (1). Following the same proof of Theorem 1, we arrive at Equation (16). Define the Riccati transformation by

ω(ζ)=ρ(ζ)a(ζ)(z(ζ))αzβ/η(ζ),    ζζ0.    (24)

Then, ω(ζ) > 0 and

ω(ζ)=ρ(ζ)ρ(ζ)ω(ζ)+ρ(ζ)(a(ζ)(z(ζ))α)zβ/η(ζ)          -ρ(ζ)a(ζ)(z(ζ))αβηzβ/η-1(ζ)z(ζ)z2β/η(ζ)          =ρ(ζ)ρ(ζ)ω(ζ)+ρ(ζ)(a(ζ)(z(ζ))α)zβ/η(ζ)-βηz(ζ)z(ζ)ω(ζ).

Using the monotonicity of z(ζ) with Equations (15, 16), we get

ω(ζ)ρ(ζ)ρ(ζ)ω(ζ)-c2βr(ζ)ρ(ζ)-βηz(ζ)z(ζ)ω(ζ).          ρ(ζ)ρ(ζ)ω(ζ)-c2βr(ζ)ρ(ζ)-βηzβ/(ηα)-1(ζ)ρ1/α(ζ)a1/α(ζ)ω(α+1)/α(ζ).

Given that z(ζ) > 0 and z′(ζ) > 0, it follows that there exists a positive constant c* > 0 such that z(ζ) ≥ c* for ζ ≥ ζ2. Consequently, we have

ω(ζ)ρ+(ζ)ρ(ζ)ω(ζ)-c2βr(ζ)ρ(ζ)-βηc*β/(ηα-1)ρ1/α(ζ)a1/α(ζ)ω(α+1)/α(ζ).

Applying the inequality (5), we get

ω(ζ)(αηρ1/α(ζ)a1/α(ζ)βc*β/(ηα-1))α(ρ+(ζ)(α+1)ρ(ζ))α+1-c2βr(ζ)ρ(ζ).

Integrating from (ζ3 > ζ2) to ζ, we obtain

ζ3ζ[c2βr(s)ρ(s)-(αηa1/α(s)βc*β/(ηα-1)ρ(s))α(ρ+(s)(α+1))α+1]dsω(ζ3).

This contradiction contradicts Equation (23). By completing the proof for the two cases (III) and (IV) in a manner similar to the proof of Theorem 1, we arrive at the conclusion of the theorem.

Theorem 3. Assume that τ1(ζ) = τ2(ζ) = τ(ζ), λ < ν ≤ 1, and (H1)-(H3) are satisfied. Additionally, assume that all other conditions of Theorem 1 are met, replacing Equation (6) with

limζ(ν-λλp1ν/(ν-λ)(ζ)p2λ/(λ-ν)(ζ))=0,    (25)

then the conclusion of Theorem 1 remains valid.

Proof. Let x(ζ) be a non-oscillatory solution of Equation (1). For convenience, assume, without loss of generality, that x(ζ) becomes eventually positive. Consequently, there exists ζ1 ≥ ζ0 such that x(ζ) > 0, x(τ(ζ)) > 0, x(σ(ζ)) > 0, and x(φ(ζ)) > 0 for ζ ≥ ζ1. Employing a similar approach as in the proof of Theorem 1, we analyze the four cases (I)–(IV) for z(ζ).

Firstly, consider case (I). As demonstrated in the proof of case (I) for Theorem 1, it is evident that this case is not feasible.

Next, let us examine case (II). Expressing [p1(ζ)xλ(τ(ζ))-p2(ζ)xν(τ(ζ))] in the form

p1(ζ)xλ(τ(ζ))-p2(ζ)xν(τ(ζ))=νλp2(ζ)[xλ(τ(ζ))λνp1(ζ)p2(ζ)-λν(xλ(τ(ζ)))ν/λ].

Applying inequality (3) with n=νλ, B = xλ(τ(ζ)), L=λνp1(ζ)p2(ζ) and m=νν-λ, we get

p1(ζ)xλ(τ(ζ))-p2(ζ)xν(τ(ζ))νλp2(ζ)[ν-λν](λνp1(ζ)p2(ζ))ν/(ν-λ)                                                              =ν-λλp1ν/(ν-λ)(ζ)p2λ/(λ-ν)(ζ).

It follows from definition of z(ζ) that

x(ζ)(1-ν-λλp1ν/(ν-λ)(ζ)p2λ/(λ-ν)(ζ)z(ζ))1/ηz1/η(ζ).

Therefore, considering Equation (25) and the positivity and increasing nature of z(ζ), it follows that there exists a constant c3 ∈ (0, 1) such that

x(ζ)c31/ηz1/η(ζ).    (26)

Concluding the proof in a manner similar to the proof of Theorem 1, replacing Equation (14) with Equation (26), leading to the conclusion of the theorem.

Remark 1. The outcomes of Corollary 1 and Theorem 2 can be directly applied to Theorem 3, replacing Equation (6) with Equation (25) and considering λ < ν ≤ 1 instead of λ < 1, ν > 1.

4 The case ε = +1

In this section, we investigate the oscillatory behavior of solutions to Equation (1) under the conditions ε = +1, τ1(ζ) ≤ ζ, and τ2(ζ) ≥ ζ. Specifically, we consider cases where either of the following three conditions holds ν < λ with η < λ, or λ < ν with η < ν, or η < ν < λ.

Theorem 4. Assume that τ1(ζ) ≤ ζ, τ2(ζ) ≥ ζ, ν < λ, η < λ, and conditions (H1)-(H3) are satisfied. Additionally, suppose that for any c* > 0

limζS1(ζ)=0.    (27)

If there exists a number k3 ∈ (0, 1) such that the first-order delay differential equation

ϝ(ζ)+k3γ/λq(ζ)Aγ/λ(σ(ζ),ζ3)p1γ/λ(τ1-1(σ(ζ)))ϝγ/αλ(σ(ζ))=0    (28)

is oscillatory for sufficiently large ζ2 > ζ1 ≥ ζ0, then every solution of Equation (1) is oscillatory.

Proof. Let x(ζ) be a non-oscillatory solution of Equation (1). Without loss of generality, assume that x(ζ) is eventually positive for ζ ≥ ζ0. Therefore, there exists ζ1 ≥ ζ0 such that x(ζ) > 0, x1(ζ)) > 0, x2(ζ)) > 0, x(σ(ζ)) > 0, and x(φ(ζ)) > 0 for ζ ≥ ζ1. It follows that z(ζ) > 0 for ζ ≥ ζ1, which means that cases (III) and (IV) mentioned before in the proof of Theorem 1 are impossible here. Consequently, we shall study the two cases, namely, case (I) and case (II) in detail.

From Equation (1), we have Equation (15). First, suppose that case (II) holds. Since z′(ζ) < 0 and (a(ζ)(z′(ζ))α) < 0, according to Equation (2), z(ζ) must be negative, which contradicts the positivity of z(ζ), making this case impossible.

Now, consider the possibility that case (II) holds. In this case, we have z′(ζ) > 0 for ζ ≥ ζ1. From the definition of z(ζ), we have

xλ(τ1(ζ))=1p1(ζ)[z(ζ)-xη(ζ)-p2(ζ)xν(τ2(ζ))].

It follows that

xλ(ζ)=1p1(τ1-1(ζ))[z(τ1-1(ζ))-xη(τ1-1(ζ))-p2(τ1-1(ζ))xν(τ2(τ1-1(ζ)))].    (29)

Hence,

xν(τ2(τ11(ζ)))=1p1νλ(δ1(ζ))[z(δ1(ζ))xη(δ1(ζ))                                 p2(δ1(ζ))xν(τ2(δ1(ζ)))]ν/λ    (30)

and

xη(τ11(ζ))=1p1η/λ(δ2(ζ))[z(δ2(ζ))xη(δ2(ζ))                              p2(δ2(ζ))xν(τ2(δ2(ζ)))]η/λ.    (31)

Combining Equations (30, 31) with Equation (29), we get

xλ(ζ)=1p1(τ11(ζ))[z(τ11(ζ))1p1η/λ(δ2(ζ))[z(δ2(ζ))               xη(δ2(ζ))p2(δ2(ζ))xν(τ2(δ2(ζ)))]η/λ               p2(τ11(ζ))p1ν/λ(δ1(ζ))[z(δ1(ζ))xη(δ1(ζ))               p2(δ1(ζ))xν(τ2(δ1(ζ)))]ν/λ].

Applying inequality (4) with Y = 1, we get

xλ(ζ)1p1(τ1-1(ζ))[z(τ1-1(ζ))-1-ηλp1η/λ(δ2(ζ))-(1-νλ)p2(τ1-1(ζ))p1ν/λ(δ1(ζ))-ηλp1η/λ(δ2(ζ))[z(δ2(ζ))-xη(δ2(ζ))-p2(δ2(ζ))xν(τ2(δ2(ζ)))]-νλp2(τ1-1(ζ))p1ν/λ(δ1(ζ))[z(δ1(ζ))-xη(δ1(ζ))-p2(δ1(ζ))xν(τ2(δ1(ζ)))]]1p1(τ1-1(ζ))[z(τ1-1(ζ))-ηλz(δ2(ζ))p1η/λ(δ2(ζ))-νλp2(τ1-1(ζ))z(δ1(ζ))p1ν/λ(δ1(ζ))-(1-ηλp1η/λ(δ2(ζ))+(1-νλ)p2(τ1-1(ζ))p1ν/λ(δ1(ζ)))].    (32)

Given Equation (16), which implies that z(ζ)A(ζ,ζ3) is decreasing, we can deduce that δ2(ζ)τ1-1(ζ) and δ1(ζ)τ2(τ1-1(ζ))τ1-1(ζ). Therefore, we have

z(δ2(ζ))A(δ2(ζ),ζ3)A(τ11(ζ),ζ3)z(τ11(ζ)  and       z(δ1(ζ))A(δ1(ζ)),ζ3)A(τ11(ζ),t3)z(τ11(ζ)).

Hence, Equation (32) takes the form

xλ(ζ)z(τ1-1(ζ))p1(τ1-1(ζ))[1-ηλA(δ2(ζ)),ζ3)A(τ1-1(ζ),ζ3)p1η/λ(δ2(ζ))               -νλp2(τ1-1(ζ))A(δ1(ζ),ζ3)A(τ1-1(ζ),ζ3)p1ν/λ(δ1(ζ))               -(1-ηλp1η/λ(δ2(ζ))+(1-νλ)p2(τ1-1(ζ))p1ν/λ(δ1(ζ)))1z(τ1-1(ζ))].

By virtue of the positivity and increasing fact of z(ζ), there a constant c* > 0 such that z(ζ) ≥ c* for ζ ≥ ζ4 ≥ ζ3 and consequently, we have

xλ(ζ)z(τ1-1(ζ))p1(τ1-1(ζ))[1-ηλA(δ2(ζ)),ζ3)A(τ1-1(ζ),ζ3)p1η/λ(δ2(ζ))               -νλp2(τ1-1(ζ))A(δ1(ζ),ζ3)A(τ1-1(ζ),ζ3)p1ν/λ(δ1(ζ))               -(1-ηλc*p1η/λ(δ2(ζ))+(1-νλ)p2(τ1-1(ζ))c*p1ν/λ(δ1(ζ)))].

It follows from Equation (27) that there exists a constant c4 ∈ (0, 1) such that

x(ζ)c41/λz1/λ(τ1-1(ζ))p11/λ(τ1-1(ζ))  for ζζ4.    (33)

This with Equation (15) leads to

(a(ζ)(z(ζ))α)-c4γ/λq(ζ)zγ/λ(τ1-1(σ(ζ)))p1γ/λ(τ1-1(σ(ζ)))                                    -c4β/λr(ζ)zβ/λ(τ1-1(φ(ζ)))p1β/λ(τ1-1(φ(ζ)))                              -c4γ/λq(ζ)zγ/λ(τ1-1(σ(ζ)))p1γ/λ(τ1-1(σ(ζ))).    (34)

Given that z(ζ) is an increasing function and τ1-1(σ(ζ))σ(ζ), we can conclude that z(τ1-1(σ(ζ)))z(σ(ζ)). Hence, Equation (34) takes the form

(a(ζ)(z(ζ))α)-c4γ/λq(ζ)zγ/λ(σ(ζ))p1γ/λ(τ1-1(σ(ζ))).    (35)

Since (a(ζ)(z′(ζ))α) is positive and decreasing, a conclusion analogous to the proof of Theorem 1 leads us to Equation (16). By substituting Equation (16) into Equation (35), we obtain

ϝ(ζ)+c4γ/λq(ζ)Aγ/λ(σ(ζ),ζ3)p1γ/λ(τ1-1(σ(ζ)))ϝγ/αλ(σ(ζ))0,    (36)

where ϝ(ζ): = a(ζ)(z′(ζ))α. According to Lemma 2 (i), the differential equation (27) associated with the inequality (36) also possesses a positive solution. However, this contradicts our earlier assertion. As a result, the proof is completed.

Theorem 5. Assume that τ1(ζ) ≤ ζ, τ2(ζ) ≥ ζ, λ < ν, η < ν, and conditions (H1)-(H3) are satisfied. Additionally, suppose that

limtS2(ζ)=0.    (37)

If there exists a number k4 ∈ (0, 1) such that the first-order delay differential equation

ϝ(ζ)+c5γ/νq(ζ)Aγ/ν(τ2-1(σ(ζ)),ζ3)p2γ/ν(τ2-1(σ(ζ)))ϝγ/αλ(σ(ζ))=0,    (38)

is oscillatory for sufficiently large ζ2 > ζ1 ≥ ζ0, then every solution of Equation (1) is also oscillatory.

Proof. Let x(ζ) be a non-oscillatory solution of Equation (1). Following the approach outlined in the proof of Theorem 4, we deduce that the possible case for z(ζ) is z(ζ) > 0, z′(ζ) > 0 and (a(ζ)(z′(ζ))α) < 0 holds for ζ ≥ ζ1 ≥ ζ0. From the definition of z(ζ), we have

xν(τ2(ζ))=1p2(ζ)[z(ζ)-xη(ζ)-p1(ζ)xλ(τ1(ζ))],

It follows that

xν(ζ)=1p2(τ2-1(ζ))[z(τ2-1(ζ))-xη(τ2-1(ζ))-p1(τ2-1(ζ))xλ(τ1(τ2-1(ζ)))].    (39)

Hence,

xλ(τ1(τ21(ζ)))=1p2λ/ν(δ3(ζ))[z(δ3(ζ))xη(δ3(ζ))                           p1(δ3(ζ))xλ(τ1(δ3(ζ)))]λ/ν,    (40)

and

x(τ21(ζ))=1p2η/ν(δ4(ζ))[z(δ4(ζ))xη(δ4(ζ))p1(δ4(ζ))xλ(τ1(δ4(ζ)))]η/ν.    (41)

Combining Equations (40, 41) with Equation (39), we get

xν(ζ)=1p2(τ21(ζ))[z(τ21(ζ))1p2η/ν(δ4(ζ))[z(δ4(ζ))            xη(δ4(ζ))p1(δ4(ζ))xλ(τ1(δ4(ζ)))]η/ν            p1(τ21(ζ))p2λ/ν(δ3(ζ))[z(δ3(ζ))xη(δ3(ζ))            p1(δ3(ζ))xλ(τ1(δ3(ζ)))]λ/ν].

Applying inequality (4) with Y = 1, we get

xν(ζ)1p2(τ21(ζ))[z(τ21(ζ))            ηνp2η/ν(δ4(ζ))[z(δ4(ζ))xη(δ4(ζ))            p1(δ4(ζ))xλ(τ1(δ4(ζ)))]               λνp1(τ21(ζ))p2λ/ν(δ3(ζ))[z(δ3(ζ))xη(δ3(ζ))            p1(δ3(ζ))xλ(τ1(δ3(ζ)))]                (1ην)p2η/ν(δ4(ζ))(1λν)p1(τ21(ζ))p2λ/ν(δ3(ζ))]        1p2(τ21(ζ))[z(τ21(ζ))ηνz(δ4(ζ))p2η/ν(δ4(ζ))            λνp1(τ21(ζ))z(δ3(ζ))p2λ/ν(δ3(ζ))            (1ην)p2η/ν(δ4(ζ))(1λν)p1(τ21(ζ))p2λ/ν(δ3(ζ))].    (42)

Since z(ζ)A(ζ,ζ3) is decreasing, δ4(ζ)τ2-1(ζ) and δ3(ζ)τ1(τ2-1(ζ))τ2-1(ζ), we have z(δ4(ζ))z(τ2-1(ζ)) and z(δ3(ζ))z(τ2-1(ζ)). Hence, Equation (42) takes the form

                    xν(ζ)z(τ2-1(ζ))p2(τ2-1(ζ))[1-(ηνp2η/ν(δ4(ζ))+λνp1(τ2-1(ζ))p2λ/ν(δ3(ζ))+((1-ην)p2η/ν(δ4(ζ))-(1-λν)p1(τ2-1(ζ))p2λ/ν(δ3(ζ)))1z(τ2-1(ζ)))].

By virtue of the positivity and increasing fact of z(ζ), there a constant c* > 0 such that z(ζ) ≥ c* for ζ ≥ ζ4 ≥ ζ3 and consequently, we have

xν(ζ)z(τ21(ζ))p2(τ21(ζ))[1(ηνp2η/ν(δ4(ζ))+λνp1(τ21(ζ))p2λ/ν(δ3(ζ))+(1ην)cp2η/ν(δ4(ζ))(1λν)p1(τ21(ζ))cp2λ/ν(δ3(ζ)))].

It follows from Equation (37) that there exists a constant c5 ∈ (0, 1) such that

x(ζ)c51/νz1/ν(τ2-1(ζ))p21/ν(τ2-1(ζ)).

This with Equation (15) leads to

(a(ζ)(z(ζ))α)-c4γ/νq(ζ)zγ/ν(τ2-1(σ(ζ)))p2γ/ν(τ2-1(σ(ζ)))                                    -c4βνr(ζ)zβ/ν(τ2-1(φ(ζ)))p2β/ν(τ2-1(φ(ζ)))                              -c4γ/νq(ζ)zγ/ν(τ2-1(σ(ζ)))p2γ/ν(τ2-1(σ(ζ))).    (43)

Since (a(ζ)(z′(ζ))α) is positive and decreasing, a conclusion analogous to the proof of Theorem 1 leads us to Equation (16). By substituting Equation (16) into Equation (43), we obtain

ϝ(ζ)+c5γ/νq(ζ)Aγ/ν(τ2-1(σ(ζ)),ζ3)p2γ/ν(τ2-1(σ(ζ)))ϝγ/αλ(σ(ζ))0,    (44)

where ϝ(ζ): = a(ζ)(z′(ζ))α. According to Lemma 2 (i), the differential equation (37) associated with the inequality (44) also possesses a positive solution. However, this contradicts our earlier assertion. As a result, the proof is completed.

Theorem 6. Assume that τ1(ζ) ≤ ζ, τ2(ζ) ≥ ζ, η < ν < λ, and conditions (H1)-(H3) are satisfied. Additionally, suppose that

limζS3(ζ)=0.    (45)

If there exists a number k3 ∈ (0, 1) such that the first-order delay differential equation (28) is oscillatory for sufficiently large ζ2 > ζ1 ≥ ζ0, then every solution of Equation (1) is also oscillatory.

Proof. Let x(ζ) be a non-oscillatory solution of Equation (1). Following the approach outlined in the proof of Theorem 4, we deduce that the possible case for z(ζ) is z(ζ) > 0, z′(ζ) > 0 and (a(ζ)(z′(ζ))α) < 0 holds for ζ ≥ ζ1 ≥ ζ0. From the definition of z(ζ), we have Equation (29). Hence,

xη(τ11(ζ))=1p2η/ν(δ5(ζ))[z(δ5(ζ))xη(δ5(ζ))                                  p1(δ5(ζ))xλ(τ1(δ5(ζ)))]η/ν,    (46)

and

xν(τ2(τ11(ζ)))=1p1ν/λ(δ1(ζ))[z(δ1(ζ))xη(δ1(ζ))                                   p2(δ1(ζ))xν(τ2(δ1(ζ)))]ν/λ.    (47)

Combining Equations (46, 47) with Equation (29), we get

xλ(ζ)=z(τ1-1(ζ))p1(τ1-1(ζ))                -1p1(τ1-1(ζ))[[z(δ5(ζ))-xη(δ5(ζ))-p1(δ5(ζ))xλ(τ1(δ5(ζ)))]η/νp2η/ν(δ5(ζ))]                -p2(τ1-1(ζ))p1(τ1-1(ζ))[[z(δ1(ζ))-xη(δ1(ζ))-p2(δ1(ζ))xν(τ2(δ1(ζ)))]ν/λp1ν/λ(δ1(ζ))].

Applying the inequality (4) with Y = 1, we obtain

xλ(ζ)=z(τ1-1(ζ))p1(τ1-1(ζ))-ηνz(δ5(ζ))p1(τ1-1(ζ))p2η/ν(δ5(ζ))            -νλp2(τ1-1(ζ))z(δ1(ζ))p1(τ1-1(ζ))p1ν/λ(δ1(ζ))                  -(1-ην)p1(τ1-1(ζ))p2η/ν(δ5(ζ))-(1-νλ)p2(τ1-1(ζ))p1(τ1-1(ζ))p1ν/λ(δ1(ζ)).    (48)

Since z(ζ)A(ζ,ζ3) is decreasing, δ5(ζ)τ1-1(ζ) and δ1(ζ)τ1-1(ζ), we have z(δ5(ζ))z(τ1-1(ζ)) and z(δ1(ζ))A(δ1(ζ),ζ3)A(τ1-1(ζ),ζ3)z(τ1-1(ζ)). Hence, Equation (48) takes the form

xλ(ζ)=z(τ1-1(ζ))p1(τ1-1(ζ))[1-ηνp2η/ν(δ5(ζ))-νλp2(τ1-1(ζ))A(δ1(ζ),ζ3)p1ν/λ(δ1(ζ))A(τ1-1(ζ),ζ3)                  -((1-ην)p2η/ν(δ5(ζ))+(1-νλ)p2(τ1-1(ζ))p1ν/λ(δ1(ζ)))1z(τ1-1(ζ))]            z(τ1-1(ζ))p1(τ1-1(ζ))[1-ηνp2η/ν(δ5(ζ))-νλp2(τ1-1(ζ))A(δ1(ζ),ζ3)p1ν/λ(δ1(ζ))A(τ1-1(ζ),ζ3)                  -(1-ην)c*p2η/ν(δ5(ζ))-(1-νλ)p2(τ1-1(ζ))c*p1ν/λ(δ1(ζ))]

for ζ ≥ ζ4 ≥ ζ3. Now from Equation (45), it follows that there exists a constant c6 ∈ (0, 1) such that

x(ζ)c61/λz1/λ(τ1-1(ζ))p11/λ(τ1-1(ζ))   for ζζ4    (49)

Completing the proof following the steps outlined in the proof of Theorem 4 and substituting Equation (33) with Equation (49), we arrive at the same conclusion as stated in the theorem. This completes the proof.

Corollary 2. Assume that all the hypotheses of Theorem 4 are satisfied, with the modification that the condition

limζζ0ζq(s)Aγ/λ(σ(s),ζ1)p1γ/λ(τ1-1(σ(s)))ds= for γ<αλ

is used instead of the condition in Equation (28), and then, the conclusion of Theorem 4 holds.

Corollary 3. Assume that all the hypotheses of Theorem 5 are satisfied, with the modification that the condition

limζζ0ζq(s)Aγ/ν(τ2-1(σ(s)),ζ1)p2γ/ν(τ2-1(σ(s)))ds= for γ<αν

is used instead of the condition in Equation (38), and then, the conclusion of Theorem 5 holds.

Now, we consider the following special case of Equation (1). We consider η = λ = ν = 1 and α = β = γ with p2(ζ) ≥ 0 and p1(ζ) ≥ 1 eventually, namely

(a(ζ)((x(ζ)+p1(ζ)x(τ1(ζ))+x(τ2(ζ))))α)+q(ζ)xα(σ(ζ))+r(ζ)xα(φ(ζ))=0,  ζζ0>0,    (50)

For the sake of notation, we define:

ϕ(ζ)=1p1(τ1-1(ζ))[1-A(δ2(ζ)),ζ3)A(τ1-1(ζ),ζ3)p1(δ2(ζ))-p2(τ1-1(ζ))A(δ1(ζ),ζ3)A(τ1-1(ζ),ζ3)p1(δ1(ζ))].

Theorem 7. Suppose that conditions (H1)- (H3) hold. If there exists a positive function ρC1([ζ0,),) such that

limsupζζ1ζ(ρ(s)r(s)ϕα(τ1-1(φ(s)))-a(s)(ρ +(s))α+1(α+1)α+1ρα(s))ds=,    (51)

then Equation (50) is oscillatory.

Proof. Let x(ζ) be a non-oscillatory solution of Equation (50). Employing a procedure akin to the proof of Theorem 4 with η = λ = ν = 1, Equation (32) can be expressed as follows:

x(ζ)z(τ1-1(ζ))p1(τ1-1(ζ))[1-A(δ2(ζ)),ζ3)A(τ1-1(ζ),ζ3)p1(δ2(ζ))    -p2(τ1-1(ζ))A(δ1(ζ),ζ3)A(τ1-1(ζ),ζ3)p1(δ1(ζ))]:=ϕ(ζ)z(τ1-1(ζ)).

Let us redefine the Riccati substitution ω according to Equation (24), setting α = β and η = 1, yielding

ω(ζ)=ρ(ζ)ρ(ζ)ω(ζ)+ρ(ζ)(a(ζ)(z(ζ))α)zα(ζ)-z(ζ)z(ζ)ω(ζ).

Utilizing the monotonicity properties of z(ζ) with Equation (16), we get

ω(ζ)ρ(ζ)ρ(ζ)ω(ζ)-ρ(ζ)r(ζ)ϕα(τ1-1(φ(ζ)))zα(τ1-1(φ(ζ)))zα(ζ)           -αρ1/α(ζ)a1/α(ζ)ω(α+1)/α(ζ)           ρ(ζ)ρ(ζ)ω(ζ)-ρ(ζ)r(ζ)ϕα(τ1-1(φ(ζ)))           -αρ1/α(ζ)a1/α(ζ)ω(α+1)/α(ζ).

Applying the inequality (5), we get

ω(ζ)a(ζ)(ρ +(ζ))α+1(α+1)α+1ρα(ζ)-ρ(ζ)r(ζ)ϕα(τ1-1(φ(ζ))).

Integrating from ζ1 to ζ, we get

ζ1ζ(ρ(s)r(s)ϕα(τ1-1(φ(s)))-a(s)(ρ +(s))α+1(α+1)α+1ρα(s))dsω(ζ1).

This contradicts Equation (51) and concludes the proof.

Example 1. Consider the second-order differential equation:

(((x2(ζ)+1ζx1/2(ζ/2)-ζ2x(ζ/2)))2)+q0x3(ζ/6)+r0x5(12ζ)=0,  ζ1    (52)

Here, a(ζ) = 1, η = 2, p1(ζ)=1ζ, λ=12, τ1(ζ)=τ2(ζ)=τ(ζ)=ζ2, ε = −1, p2(ζ)=ζ2, ν = 2, α = 2, q(ζ) = q0, γ = 2, σ(ζ) = ζ/6, r(ζ) = r0, β = 5 and φ(ζ) = 12ζ. It is clear that ζ0a-1/α(s)ds= and

limζ[R1(ζ)+R2(ζ)]=limζ[14ζ2-14ζ2]=0.

Choose ϱ(ζ)=ζ2, where τ−1(σ(ζ)) ≤ ϱ(ζ) ≤ ζ, τ−1(φ(ζ)) ≥ ζandτ−1(φ(ϱ(ϱ(ζ)))) ≥ ζ, and consequently, we have

limζζ0ζq(s)Aγ/η(s,s1)ds=limζζ0ζq0(s3/2-s13/2)ds=,
limζζ0ζq(s)(A(ϱ(s),τ-1(σ(s)))p2(τ-1(σ(s))))γ/νds=limζζ0ζq0(s/6s2/8)32ds=

and

limζζ0ζ(1a(s)ϱ(s)sr(u)p2β/ν(τ-1(φ(u)))du)1/αds=limζζ0ζ(s/2sr0(24u)5du)1/2ds=.

Therefore, all the assumptions stated in Corollary 1 are satisfied, indicating that every solution x(ζ) to Equation (52) exhibits oscillatory behavior.

Example 2. Consider the second-order differential equation

(ζ(x(ζ)+ζ110x1/5(ζ/3)-ζ1/3x1/3(ζ/3)))+ζ2x(ζ/5)+ζ5x5(4ζ)=0,  ζ1.    (53)

Here, a(ζ) = ζ2, η = 1, p1(ζ)=ζ1/10, λ=15, τ1(ζ)=τ2(ζ)=τ(ζ)=ζ3, ε = −1, p2(ζ)=ζ1/3, ν = 1/3, α = 1, q(ζ) = ζ2, γ = 1, σ(ζ) = ζ/5, r(ζ) = ζ5, β = 5 and φ(ζ) = 4ζ. It is clear that ζ0a-1/α(s)ds= and condition (25) becomes

limζ(ν-λλp1ν/(ν-λ)(ζ)p2λλ-ν(ζ))=limζ(1/3-1/51/5(ζ1/10)1/31/3-1/5(ζ1/3)1/51/5-1/3)=0.

Choose ϱ(ζ)=4ζ5, where τ−1(σ(ζ)) ≤ ϱ(ζ) ≤ ζ, τ−1(φ(ζ)) ≥ ζandτ−1(φ(ϱ(ϱ(ζ)))) ≥ ζ, consequently we have

limζζ0ζq(s)Aγ/η(s,s1)ds=limζζ0ζs2ln (s)ds=,
limζζ0ζq(s)(A(ϱ(s),τ-1(σ(s)))p2(τ-1(σ(s))))γ/νds=limζζ0ζs2(ln (4s/5)-ln(3s/5)(3s/5)1/3)3ds=,

and

limζζ0ζ(1a(s)ϱ(s)sr(u)p2β/ν(τ-1(φ(u)))du)1/αds=limζζ0ζ(1u4s/5su5(12u)5du)ds=.

It is evident that the requirements outlined in Corollary 1, transferred to Theorem 1 by substituting Equation (25) in place of Equation (6) and λ < ν ≤ 1 instead of λ < 1, ν > 1, are fulfilled. Consequently, each solution to Equation (53) demonstrates oscillatory behavior.

Remark 2. Note that oscillation results presented in the study by Grace et al. [13] fail to apply to the Equation (53), where 0 < λ, ν < 1 and γ ≠ β unlike in the study by Grace et al. [13].

Example 3. Consider the second-order differential equation

(ζ3((x1/5(ζ)+ζ5x5(ζ/2)+1ζx1/3(3ζ)))5)+ζ2x3(ζ/3)+1ζ2x3(2ζ)=0,  ζ1.    (54)

Here, a(ζ) = ζ3, η = 1/5, p1(ζ)=ζ5, λ = 5, τ1(ζ) = ζ/2, τ2(ζ) = 3ζ, ε = 1, p2(ζ)=1ζ, ν = 1/3, α = 5, q(ζ) = ζ2, γ = 3, σ(ζ) = ζ/3, r(ζ)=1ζ2, β = 3 and φ(ζ) = 2ζ. It is clear that ζ0a-1/α(s)ds= and limζS1(ζ)=0. Since

A(ζ,ζ1)=ζ1ζ(s3)-1/5ds=52(ζ2/5-ζ12/5),

we observe that

limζζ0ζq(s)Aγ/λ(σ(s),ζ1)p1γ/λ(τ1-1(σ(s)))ds=(52)3/5limζζ0ζs2((s/3)2/5-s12/5)35(2s/3)3ds=.

According to Corollary 2, every solution of Equation (54) is oscillatory.

Example 4. Consider the second-order differential equation

(ζ((x3(ζ)+1ζx(ζ/4)+5ζx4(6ζ)))3)+ζx3/2(ζ/3)+1ζx5(2ζ)=0,  ζ1.

Here, a(ζ) = ζ, η = 1/5, p1(ζ)=ζ5, λ = 5, τ1(ζ) = ζ/2, τ2(ζ) = 3ζ, ε = 1, p2(ζ)=1ζ, ν = 4, α = 5, q(ζ) = ζ, γ = 3/2, σ(ζ) = ζ/3, r(ζ)=1ζ, β = 5 and φ(ζ) = 2ζ. It is clear that ζ0a-1/α(s)ds= and limζS2(ζ)=0. Since

A(ζ,ζ1)=ζ1ζ(s)-1/3ds=32(ζ2/3-ζ12/3),

we observed that

limζζ0ζq(s)Aγ/ν(τ2-1(σ(s)),ζ1)p2γ/ν(τ2-1(σ(s)))ds=limζζ0ζs(32((s/18)2/3-s12/3))3/8(s/18)38ds=.

According to Corollary 3, every solution of Equation (54) is oscillatory.

Example 5. Consider the neutral differential equation

(x(ζ)+ζx(ζ-2π)+x(ζ+π))+2x(ζ-π2)+ζx(ζ+2π)=0,  ζ5.    (55)

Here, we have η = λ = ν = α = β = γ = 1, a(ζ) = 1, p1(ζ) = ζ, p2(ζ) = 1 τ1(ζ) = ζ−2π, τ2(ζ) = ζ+π, q(ζ) = 2, r(ζ) = ζ, σ(ζ)=ζ-π2 and φ(ζ) = ζ+2π. It is clear that A(ζ, ζ0) = ζ−5 and ϕ(ζ) > 0. Setting ρ(ζ) = 1, it becomes evident that condition (51) is satisfied. Thus, according to Theorem 7, every solution of Equation (55) is oscillatory. Indeed, x(ζ) = sin(ζ) is such a solution.

Remark 3. It is important to note that the oscillation results outlined in the study by Thandapani et al. [24] and Thandapani and Rama [25] can not directly apply to Equation (55) primarily because α = β = γ in Equation (55), whereas in the study by Thandapani et al. [24] and Thandapani and Rama [25], this condition does not hold.

5 Conclusion and discussion

The findings of this study are showcased in a fundamentally innovative and broadly applicable manner. These findings not only enrich but also complement the current literature study, as referenced in the studies by Agarwal et al. [14], Bohner et al. [15], Džurina et al. [16], Grace et al. [17], Lin and Tang [18], and Muhib et al. [19]. Furthermore, Equation (1) represents a more general formulation, where ε can take values of either −1 or 1, and (α, β, and γ) are distinct, with η differing from 1.

Data availability statement

The original contributions presented in the study are included in the article/supplementary material, further inquiries can be directed to the corresponding author.

Author contributions

AH: Conceptualization, Investigation, Methodology, Writing – original draft. OM: Conceptualization, Formal analysis, Methodology, Writing – original draft. SA: Formal analysis, Investigation, Methodology, Writing – review & editing. AA: Formal analysis, Investigation, Methodology, Writing – review & editing.

Funding

The author(s) declare financial support was received for the research, authorship, and/or publication of this article. This project was funded by King Saud University, Riyadh, Saudi Arabia, Project number (RSPD2024R533).

Conflict of interest

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Publisher's note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

References

1. Agarwal RP, Grace SR, O'Regan D. Oscillation Theory for Second Order Linear, Half-Linear, Superlinear and Sublinear Dynamic Equations. Heidelberg: Springer Science & Business Media (2013).

Google Scholar

2. Agarwal RP, Grace SR, O'Regan D. Oscillation Theory for Second Order Dynamic Equations. Boca Raton, FL: CRC Press (2002).

Google Scholar

3. Grace SR, Agarwal RP, Bohner M, O'Regan D. Oscillation of second-order strongly superlinear and strongly sublinear dynamic equations. Commun Nonlinear Sci Numer Simul. (2009) 14:3463–71. doi: 10.1016/j.cnsns.2009.01.003

Crossref Full Text | Google Scholar

4. Grace SR, Agarwal RP, Kaymakçalan B, Sae-Jie W. Oscillation theorems for second order nonlinear dynamic equations. J Appl Math Comp. (2010) 32:205. doi: 10.1007/s12190-009-0244-7

PubMed Abstract | Crossref Full Text | Google Scholar

5. Grace SR, Agarwal RP, O'Regan D. A selection of oscillation criteria for second-order differential inclusions. Appl Math Lett. (2009) 22:153–8. doi: 10.1016/j.aml.2008.01.006

Crossref Full Text | Google Scholar

6. Hassan A, Affan S. New oscillation criteria and some refinements for second-order neutral delay dynamic equations on time scales. J Math Comp Sci. (2023) 28:192–202. doi: 10.22436/jmcs.028.02.07

Crossref Full Text | Google Scholar

7. Hassana A, Affana S. Oscillation criteria for second-order delay dynamic equations with a sub-linear neutral term on time scales. Filomat. (2023) 37:7445–54. doi: 10.2298/FIL2322445H

Crossref Full Text | Google Scholar

8. Moaaz O, Muhib A, Santra SS. An oscillation test for solutions of second-order neutral differential equations of mixed type. Mathematics. (2021) 9:1634. doi: 10.3390/math9141634

Crossref Full Text | Google Scholar

9. Moaaz O, Anis M, Baleanu D, Muhib A. More effective criteria for oscillation of second-order differential equations with neutral arguments. Mathematics. (2020) 8:986. doi: 10.3390/math8060986

Crossref Full Text | Google Scholar

10. Hale JK. Functional differential equations. In: Analytic Theory of Differential Equations: The Proceedings of the Conference at Western Michigan University, Kalamazoo, from 30 April to 2 May 1970. Springer (2006). p. 9–22.

Google Scholar

11. Tunç E, Özdemir O. On the oscillation of second-order half-linear functional differential equations with mixed neutral term. J Taibah Univ Sci. (2019) 13:481–9. doi: 10.1080/16583655.2019.1595948

Crossref Full Text | Google Scholar

12. Agwa H, Khodier AM, Arafa HM. Oscillation of second-order nonlinear neutral dynamic equations with mixed arguments on time scales. J Basic Appl Res Int. (2016) 17:49–66.

Google Scholar

13. Grace SR, Graef JR, Kaleeswari S. New Criteria on the Oscillatory Behaviour of Second Order Nonlinear Differential Equations with Mixed Neutral Terms. Kraków: Tatra Mountains Mathematical Publications (2023).

Google Scholar

14. Agarwal RP, Bohner M, Li T, Zhang C. Oscillation of second-order differential equations with a sublinear neutral term. Carpath J Math. (2014) 30:1–6.

Google Scholar

15. Bohner M, Sudha B, Tangavelu K, Thandapani E. Oscillation criteria for second-order differential equations with superlinear neutral term. Nonlinear Stud. (2019) 26:425–34.

Google Scholar

16. Džurina J, Grace SR, Jadlovská I, Li T. Oscillation criteria for second-order Emden-Fowler delay differential equations with a sublinear neutral term. Mathematische Nachrichten. (2020) 293:910–22. doi: 10.1002/mana.201800196

Crossref Full Text | Google Scholar

17. Grace SR, Jadlovská I, Zafer A. On oscillation of second order delay differential equations with a sublinear neutral term. Mediterr J Math. (2020) 17:1–11. doi: 10.1007/s00009-020-01538-y

Crossref Full Text | Google Scholar

18. Lin X, Tang X. Oscillation of solutions of neutral differential equations with a superlinear neutral term. Appl Math Lett. (2007) 20:1016–22. doi: 10.1016/j.aml.2006.11.006

Crossref Full Text | Google Scholar

19. Muhib A, Elabbasy EM, Moaaz O. New oscillation criteria for differential equations with sublinear and superlinear neutral terms. Turk J Math. (2021) 45:919–28. doi: 10.3906/mat-2012-11

PubMed Abstract | Crossref Full Text | Google Scholar

20. Baculíková B. Properties of third-order nonlinear functional differential equations with mixed arguments. In: Abstract and Applied Analysis. Vol. 2011. New York, NY: Hindawi (2011).

Google Scholar

21. Philos CG. On the existence of nonoscillatory solutions tending to zero at ∞ for differential equations with positive delays. Archiv Mathematik. (1981) 36:168–78. doi: 10.1007/BF01223686

Crossref Full Text | Google Scholar

22. Hardy GH, Littlewood JE, Pólya G. Inequalities. Cambridge: Cambridge University Press (1952).

Google Scholar

23. Bohner M, Grace S, Jadlovská I. Oscillation criteria for second-order neutral delay differential equations. Electron J Qual Theory Differ Eq. (2017) 2017:1–12. doi: 10.14232/ejqtde.2017.1.60

Crossref Full Text | Google Scholar

24. Thandapani E, Selvarangam S, Vijaya M, Rama R. Oscillation results for second order nonlinear differential equation with delay and advanced arguments. Kyungpook Math J. (2016) 56:137–46. doi: 10.5666/KMJ.2016.56.1.137

Crossref Full Text | Google Scholar

25. Thandapani E, Rama R. Comparison and oscillation theorems for second order nonlinear neutral differential equations of mixed type. Serdica Math J. (2013) 39:1–16. doi: 10.14232/ejqtde.2012.1.75

PubMed Abstract | Crossref Full Text | Google Scholar

Keywords: neutral differential equations, second-order, oscillatory, mixed terms, Riccati transformation

Citation: Hassan AM, Moaaz O, Askar SS and Alshamrani AM (2024) Oscillatory behavior of solutions of second-order non-linear differential equations with mixed non-linear neutral terms. Front. Appl. Math. Stat. 10:1384559. doi: 10.3389/fams.2024.1384559

Received: 09 February 2024; Accepted: 25 June 2024;
Published: 16 July 2024.

Edited by:

Zailan Siri, University of Malaya, Malaysia

Reviewed by:

Jihad H. Asad, Palestine Technical University Kadoorie, Palestine
Abdullah Ozbekler, OSTIM Technical University, Türkiye

Copyright © 2024 Hassan, Moaaz, Askar and Alshamrani. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

*Correspondence: Osama Moaaz, b19tb2FheiYjeDAwMDQwO21hbnMuZWR1LmVn

Disclaimer: All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article or claim that may be made by its manufacturer is not guaranteed or endorsed by the publisher.