In [
Citation: George L. Karakostas. On a conjecture for the difference equation xn+1=1+pxn−mx2n[J]. AIMS Mathematics, 2023, 8(10): 22714-22729. doi: 10.3934/math.20231156
[1] | Ibtisam Aldawish, Mohamed Jleli, Bessem Samet . Blow-up of solutions to fractional differential inequalities involving ψ-Caputo fractional derivatives of different orders. AIMS Mathematics, 2022, 7(5): 9189-9205. doi: 10.3934/math.2022509 |
[2] | Tao Yan, Ghulam Farid, Sidra Bibi, Kamsing Nonlaopon . On Caputo fractional derivative inequalities by using strongly (α,h−m)-convexity. AIMS Mathematics, 2022, 7(6): 10165-10179. doi: 10.3934/math.2022565 |
[3] | Muhammad Tariq, Asif Ali Shaikh, Sotiris K. Ntouyas, Jessada Tariboon . Some novel refinements of Hermite-Hadamard and Pachpatte type integral inequalities involving a generalized preinvex function pertaining to Caputo-Fabrizio fractional integral operator. AIMS Mathematics, 2023, 8(11): 25572-25610. doi: 10.3934/math.20231306 |
[4] | Hadjer Belbali, Maamar Benbachir, Sina Etemad, Choonkil Park, Shahram Rezapour . Existence theory and generalized Mittag-Leffler stability for a nonlinear Caputo-Hadamard FIVP via the Lyapunov method. AIMS Mathematics, 2022, 7(8): 14419-14433. doi: 10.3934/math.2022794 |
[5] | Choukri Derbazi, Hadda Hammouche . Caputo-Hadamard fractional differential equations with nonlocal fractional integro-differential boundary conditions via topological degree theory. AIMS Mathematics, 2020, 5(3): 2694-2709. doi: 10.3934/math.2020174 |
[6] | Deepak B. Pachpatte . On some ψ Caputo fractional Čebyšev like inequalities for functions of two and three variables. AIMS Mathematics, 2020, 5(3): 2244-2260. doi: 10.3934/math.2020148 |
[7] | M. A. Zaky, M. Babatin, M. Hammad, A. Akgül, A. S. Hendy . Efficient spectral collocation method for nonlinear systems of fractional pantograph delay differential equations. AIMS Mathematics, 2024, 9(6): 15246-15262. doi: 10.3934/math.2024740 |
[8] | M. J. Huntul . Inverse source problems for multi-parameter space-time fractional differential equations with bi-fractional Laplacian operators. AIMS Mathematics, 2024, 9(11): 32734-32756. doi: 10.3934/math.20241566 |
[9] | Imran Abbas Baloch, Thabet Abdeljawad, Sidra Bibi, Aiman Mukheimer, Ghulam Farid, Absar Ul Haq . Some new Caputo fractional derivative inequalities for exponentially (θ,h−m)–convex functions. AIMS Mathematics, 2022, 7(2): 3006-3026. doi: 10.3934/math.2022166 |
[10] | Muhammad Tariq, Hijaz Ahmad, Abdul Ghafoor Shaikh, Soubhagya Kumar Sahoo, Khaled Mohamed Khedher, Tuan Nguyen Gia . New fractional integral inequalities for preinvex functions involving Caputo-Fabrizio operator. AIMS Mathematics, 2022, 7(3): 3440-3455. doi: 10.3934/math.2022191 |
In [
Time-fractional differential equations arise in the mathematical modeling of a variety of real-world phenomena in many areas of sciences and engineering, such as elasticity, heat transfer, circuits systems, continuum mechanics, fluid mechanics, wave theory, etc. For more details, we refer the reader to [4,6,7,8,14,15,17,24] and the references therein. Consequently, the study of time-fractional differential equations attracted much attention of many researchers (see e.g. [1,5,9,10,19,22,23] and the references therein).
Multi-time differential equations arise, for example, in analyzing frequency and amplitude modulation in oscillators, see Narayan and Roychowdhury [18]. Some methods for solving Multi-time differential equations can be found in [20,21].
The study of blowing-up solutions to time-fractional differential equations was initiated by Kirane and his collaborators, see e.g. [3,11,12,13]. In particular, Kirane et al. [11] considered the two-times fractional differential equation
{CDα0|tu(t,s)+CDβ0|s|u|m(t,s)=|u|p(t,s),t,s>0,u(0,s)=u0(s),u(t,0)=u1(t),t,s>0, | (1.1) |
where p,m>1, 0<α,β<1, CDα0|t is the Caputo fractional derivative of order α with respect to the first time-variable t, and CDβ0|s is the Caputo fractional derivative of order β with respect to the second time-variable s. Namely, the authors provided sufficient conditions for which any solution to (1.1) blows-up in a finite time. In the same reference, the authors extended their study to the case of systems.
In this paper, we investigate the nonexistence of global solutions to two-times-fractional differential inequalities of the form
{HCDαa|tu(t,s)+CDβa|s|u|m(t,s)≥(s−a)γ(lnta)σ|u|p(t,s),t,s>a,u(a,s)=u0(s),u(t,a)=u1(t),t,s>a, | (1.2) |
where p>1, m≥1, γ,σ∈R, a>0, 0<α,β<1, HCDαa|t is the Hadamard-Caputo fractional derivative of order α with respect to the first time-variable t, and CDβa|s is the Caputo fractional derivative of order β with respect to the second time-variable s. Using the test function method (see e.g. [16]) and a judicious choice of a test function, we establish sufficient conditions ensuring the nonexistence of global solutions to (1.2). Our obtained conditions depend on the parameters α,β,p,m,γ,σ, and the initial values.
Our motivation for considering problems of type (1.2) is to study the combination effect of the two fractional derivatives of different nature HCDαa|t and CDβa|s on the nonexistence of global solutions to (1.2). As far as we know, the study of nonexistence of global solutions for time fractional differential equations (or inequalities) involving both Hadamard-Caputo and Caputo fractional derivatives, was never considered in the literature.
The rest of the paper is organized as follows: In Section 2, we recall some concepts from fractional calculus and provide some useful lemmas. In Section 3, we state our main results and provide some examples. Section 4 is devoted to the proofs of our main results.
Let a,T∈R be such that 0<a<T. The left-sided and right-sided Riemann-Liouville fractional integrals of order θ>0 of a function ϑ∈L1([a,T]), are defined respectively by (see [10])
(Iθaϑ)(t)=1Γ(θ)∫ta(t−τ)θ−1ϑ(τ)dτ |
and
(IθTϑ)(t)=1Γ(θ)∫Tt(τ−t)θ−1ϑ(τ)dτ, |
for almost everywhere t∈[a,T], where Γ is the Gamma function.
Notice that, if ϑ∈C([a,T]), then Iθaϑ,IθTϑ∈C([a,T]) with
(Iθaϑ)(a)=(IθTϑ)(T)=0. | (2.1) |
The Caputo fractional derivative of order θ∈(0,1) of a function ϑ∈AC([a,∞)), is defined by (see [10])
CDθaϑ(t)=(I1−θaϑ′)(t)=1Γ(1−θ)∫ta(t−τ)−θϑ′(τ)dτ, |
for almost everywhere t≥a.
Lemma 2.1. [see [10]]Let κ>0, p,q≥1, and 1p+1q≤1+κ (p≠1, q≠1, in the case 1p+1q=1+κ). Let ϑ∈Lp([a,T] and w∈Lq([a,T]). Then
∫Ta(Iκaϑ)(t)w(t)dt=∫Taϑ(t)(IκTw)(t)dt. |
The left-sided and right-sided Hadamard fractional integrals of order θ>0 of a function ϑ∈L1([a,T]), are defined respectively by (see [10])
(Jθaϑ)(t)=1Γ(θ)∫ta(lntτ)θ−1ϑ(τ)1τdτ |
and
(JθTϑ)(t)=1Γ(θ)∫Tt(lnτt)θ−1ϑ(τ)1τdτ, |
for almost everywhere t∈[a,T].
Notice that, if ϑ∈C([a,T]), then Jθaϑ,JθTϑ∈C([a,T]) with
(Jθaϑ)(a)=(JθTϑ)(T)=0. | (2.2) |
The Hadamard-Caputo fractional derivative of order θ∈(0,1) of a function ϑ∈AC([a,∞)), is defined by (see [2])
HCDθaϑ(t)=(J1−θaδϑ)(t)=1Γ(1−θ)∫ta(lntτ)−θδϑ(τ)1τdτ, |
for almost everywhere t≥a, where
δϑ(t)=tϑ′(t). |
We have the following integration by parts rule.
Lemma 2.2. Let κ>0, p,q≥1, and 1p+1q≤1+κ (p≠1, q≠1, in the case 1p+1q=1+κ). If ϑ∘exp∈Lp([lna,lnT]) and w∘exp∈Lq([lna,lnT]), then
∫Ta(Jκaϑ)(t)w(t)1tdt=∫Taϑ(t)(JκTw)(t)1tdt. |
Proof. Using the change of variable x=lnτ, we obtain
(Jκaϑ)(t)=1Γ(κ)∫ta(lntτ)κ−1ϑ(τ)1τdτ=1Γ(κ)∫lntlna(lnt−x)κ−1(ϑ∘exp)(x)dx, |
that is,
(Jκaϑ)(t)=(Iκlnaϑ∘exp)(lnt). | (2.3) |
Similarly, we have
(JκTw)(t)=(IκlnTw∘exp)(lnt). | (2.4) |
By (2.3), we obtain
∫Ta(Jκaϑ)(t)w(t)1tdt=∫Ta(Iκlnaϑ∘exp)(lnt)w(t)1tdt. |
Using the change of variable x=lnt, we get
∫Ta(Jκaϑ)(t)w(t)1tdt=∫lnTlna(Iκlnaϑ∘exp)(x)(w∘exp)(x)dx. |
Since ϑ∘exp∈Lp([lna,lnT]) and w∘exp∈Lq([lna,lnT]), by Lemma 2.1, we deduce that
∫Ta(Jκaϑ)(t)w(t)1tdt=∫lnTlna(ϑ∘exp)(x)(IκlnTw∘exp)(x)dx. |
Using again the change of variable x=lnt, there holds
∫Ta(Jκaϑ)(t)w(t)1tdt=∫Taϑ(t)(IκlnTw∘exp)(lnt)1tdt. |
Then, by (2.4), the desired result follows.
By elementary calculations, we obtain the following properties.
Lemma 2.3. For sufficiently large λ, let
ϕ1(t)=(lnTa)−λ(lnTt)λ,a≤t≤T. | (2.5) |
Let κ∈(0,1). Then
(JκTϕ1)(t)=Γ(λ+1)Γ(κ+λ+1)(lnTa)−λ(lnTt)κ+λ, | (2.6) |
(JκTϕ1)′(t)=−Γ(λ+1)Γ(κ+λ)(lnTa)−λ(lnTt)κ+λ−11t. | (2.7) |
Lemma 2.4. For sufficiently large λ, let
ϕ2(s)=(T−a)−λ(T−s)λ,a≤s≤T. | (2.8) |
Let κ∈(0,1). Then
(IκTϕ2)(s)=Γ(λ+1)Γ(κ+λ+1)(T−a)−λ(T−s)κ+λ, | (2.9) |
(IκTϕ2)′(s)=−Γ(λ+1)Γ(κ+λ)(T−a)−λ(T−s)κ+λ−1. | (2.10) |
First, let us define global solutions to (1.2). To do this, we need to introduce the functional space
Xa:={u∈C([a,∞)×[a,∞)):u(⋅,s)∈AC([a,∞)),|u|m(t,⋅)∈AC([a,∞))}. |
We say that u is a global solution to (1.2), if u∈Xa and u satisfies the fractional differential inequality
HCDαa|tu(t,s)+CDβa|s|u|m(t,s)≥(s−a)γ(lnta)σ|u|p(t,s) |
for almost everywhere t,s≥a, as well as the initial conditions
u(a,s)=u0(s),u(t,a)=u1(t),t,s>a. |
Now, we state our main results.
Theorem 3.1. Let u0∈L1([a,∞)), u1∈Lm([a,∞),1tdt), and u1≢0.Let
0<β<1m≤1,γ>max{m−11−mβ,m(σ+1)−1}β. | (3.1) |
If
mmax{γ+1,σ+1}<p<1+γβ, | (3.2) |
then, for all α∈(0,1), (1.2) admits no global solution.
Remark 3.1. Notice that by (3.1), the set of exponents p satisfying (3.2) is nonempty.
Theorem 3.2. Let u0∈L1([a,∞)), u1∈Lm([a,∞),1tdt), and u1≢0. Let
0<β<1m≤1,1−1m<α<1,σ>(m−1)(1−α)1−mβ−α. | (3.3) |
If
βmax{m−11−mβ,m(σ+1)−1}<γ<(σ+α)β1−α | (3.4) |
and
p=1+γβ, | (3.5) |
then (1.2) admits no global solution.
Remark 3.2. Notice that by (3.3), the set of real numbers γ satisfying (3.4) is nonempty.
We illustrate our obtained results by the following examples.
Example 3.1. Consider the fractional differential inequality
{HCDαa|tu(t,s)+CD14a|su2(t,s)≥(s−a)(lnta)−1|u|p(t,s),t,s>a,u(a,s)=(1+s2)−1,u(t,a)=exp(−t),t,s>a, | (3.6) |
where a>0 and 0<α<1. Observe that (3.6) is a special case of (1.2) with
β=14,m=2,σ=−1,γ=1,u0(s)=(1+s2)−1,u1(t)=exp(−t). |
Moreover, we have
0<β=14<12=1m<1,max{m−11−mβ,m(σ+1)−1}β=max{2,0}4=12<γ=1, |
and u0∈L1([a,∞)), u1∈Lm([a,∞),1tdt). Hence, condition (3.1) is satisfied. Then, by Theorem 3.1, we deduce that, if
mmax{γ+1,σ+1}<p<1+γβ, |
that is,
4<p<5, |
then (3.6) admits no global solution.
Example 3.2. Consider the fractional differential inequality
{HCD34a|tu(t,s)+CD12a|s|u|(t,s)≥(s−a)14(lnta)−12|u|32(t,s),t,s>a,u(a,s)=(1+s2)−1,u(t,a)=exp(−t),t,s>a, | (3.7) |
where a>0. Then (3.7) is a special case of (1.2) with
α=34,β=12,m=1,σ=−12,γ=14,p=32,u0(s)=(1+s2)−1,u1(t)=exp(−t). |
On the other hand, we have
0<β=12<1=1m,1−1m=0<α=34<1,σ=−12>−34=(m−1)(1−α)1−mβ−α, |
which shows that condition (3.3) is satisfied. Moreover, we have
βmax{m−11−mβ,m(σ+1)−1}=−14<γ=14<12=(σ+α)β1−α,p=32=1+γβ, |
which shows that conditions (3.4) and (3.5) are satisfied. Then, by Theorem 3.2, we deduce that (3.7) admits no global solution.
In this section, C denotes a positive constant independent on T, whose value may change from line to line.
Proof of Theorem 3.1. Suppose that u∈Xa is a global solution to (1.2). For sufficiently large T and λ, let
φ(t,s)=ϕ1(t)ϕ2(s),a≤t,s≤T, |
where ϕ1 and ϕ2 are defined respectively by (2.5) and (2.8). Multiplying the inequality in (1.2) by 1tφ and integrating over ΩT:=(a,T)×(a,T), we obtain
∫ΩT(s−a)γ(lnta)σ|u|pφ(t,s)1tdtds≤∫ΩTHCDαa|tuφ(t,s)1tdtds+∫ΩTCDβa|s|u|mφ(t,s)1tdtds. | (4.1) |
On the other hand, using Lemma 2.2, integrating by parts, using the initial conditions, and taking in consideration (2.2), we obtain
∫TaHCDαa|tuφ(t,s)1tdt=∫Ta(J1−αa|tt∂u∂t)(t,s)φ(t,s)1tdt=∫Ta∂u∂t(t,s)(J1−αT|tφ)(t,s)dt=[u(t,s)(J1−αT|tφ)(t,s)]Tt=a−∫Tau(t,s)∂(J1−αT|tφ)∂t(t,s)dt=−u0(s)(J1−αT|tφ)(a,s)−∫Tau(t,s)∂(J1−αT|tφ)∂t(t,s)dt. |
Integrating over (a,T), we get
∫ΩTHCDαa|tuφ(t,s)1tdtds=−∫Tau0(s)(J1−αT|tφ)(a,s)ds−∫ΩTu(t,s)∂(J1−αT|tφ)∂t(t,s)dtds. | (4.2) |
Similarly, using Lemma 2.1, integrating by parts, using the initial conditions, and taking in consideration (2.1), we obtain
∫TaCDβa|s|u|mφ(t,s)ds=∫Ta(I1−βa|s∂|u|m∂s(t,s))φ(t,s)ds=∫Ta∂|u|m∂s(t,s)(I1−βT|sφ)(t,s)ds=[|u|m(t,s)(I1−βT|sφ)(t,s)]Ts=a−∫Ta|u|m(t,s)∂(I1−βT|sφ)∂s(t,s)ds=−|u1(t)|m(I1−βT|sφ)(t,a)−∫Ta|u|m(t,s)∂(I1−βT|sφ)∂s(t,s)ds. |
Integrating over (a,T), there holds
∫ΩTCDβa|s|u|mφ(t,s)1tdtds=−∫Ta|u1(t)|m(I1−βT|sφ)(t,a)1tdt−∫ΩT|u|m(t,s)∂(I1−βT|sφ)∂s(t,s)1tdtds. | (4.3) |
It follows from (4.1)–(4.3) that
∫ΩT(s−a)γ(lnta)σ|u|pφ(t,s)1tdtds+∫Tau0(s)(J1−αT|tφ)(a,s)ds+∫Ta|u1(t)|m(I1−βT|sφ)(t,a)1tdt≤∫ΩT|u||∂(J1−αT|tφ)∂t|dtds+∫ΩT|u|m|∂(I1−βT|sφ)∂s|1tdtds. | (4.4) |
On the other hand, by Young's inequality, we have
∫ΩT|u||∂(J1−αT|tφ)∂t|dtds≤12∫ΩT(s−a)γ(lnta)σ|u|pφ(t,s)1tdtds+C∫ΩTt1p−1(s−a)−γp−1(lnta)−σp−1φ−1p−1(t,s)|∂(J1−αT|tφ)∂t|pp−1dtds. | (4.5) |
Similarly, since p>m, we have
∫ΩT|u|m|∂(I1−βT|sφ)∂s|1tdtds≤12∫ΩT(s−a)γ(lnta)σ|u|pφ(t,s)1tdtds+C∫ΩT1t(s−a)−γmp−m(lnta)−σmp−mφ−mp−m(t,s)|∂(I1−βT|sφ)∂s|pp−mdtds. | (4.6) |
Hence, combining (4.4)–(4.6), we deduce that
∫Tau0(s)(J1−αT|tφ)(a,s)ds+∫Ta|u1(t)|m(I1−βT|sφ)(t,a)1tdt≤C(K1+K2), | (4.7) |
where
K1=∫ΩTt1p−1(s−a)−γp−1(lnta)−σp−1φ−1p−1(t,s)|∂(J1−αT|tφ)∂t|pp−1dtds |
and
K2=∫ΩT1t(s−a)−γmp−m(lnta)−σmp−mφ−mp−m(t,s)|∂(I1−βT|sφ)∂s|pp−mdtds. |
By the definition of the function φ, we have
(J1−αT|tφ)(a,s)=ϕ2(s)(J1−αT|tϕ1)(a). |
Thus, using (2.6), we obtain
(J1−αT|tφ)(a,s)=Cϕ2(s)(lnTa)1−α. |
Integrating over (a,T), we get
∫Tau0(s)(J1−αT|tφ)(a,s)ds=C(lnTa)1−α∫Tau0(s)(T−a)−λ(T−s)λds. | (4.8) |
Similarly, by the definition of the function φ, we have
(I1−βT|sφ)(t,a)=ϕ1(t)(I1−βT|sϕ2)(a). |
Thus, using (2.9), we obtain
(I1−βT|sφ)(t,a)=Cϕ1(t)(T−a)1−β. |
Integrating over (a,T), we get
∫Ta|u1(t)|m(I1−βT|sφ)(t,a)1tdt=C(T−a)1−β∫Ta|u1(t)|m(lnTa)−λ(lnTt)λ1tdt. | (4.9) |
Combining (4.8) with (4.9), there holds
∫Tau0(s)(J1−αT|tφ)(a,s)ds+∫Ta|u1(t)|m(I1−βT|sφ)(t,a)1tdt=C(lnTa)1−α∫Tau0(s)(T−a)−λ(T−s)λds+C(T−a)1−β∫Ta|u1(t)|m(lnTa)−λ(lnTt)λ1tdt. |
Since u0∈L1([a,∞)), u1∈Lm([a,∞),1tdt), and u1≢0, by the dominated convergence theorem, we deduce that for sufficiently large T,
∫Tau0(s)(J1−αT|tφ)(a,s)ds+∫Ta|u1(t)|m(I1−βT|sφ)(t,a)1tdt≥C(T−a)1−β∫∞a|u1(t)|m1tdt. | (4.10) |
Now, we shall estimate the terms Ki, i=1,2. By the definition of the function φ, the term K1 can be written as
K1=(∫Ta(s−a)−γp−1ϕ2(s)ds)(∫Tat1p−1(lnta)−σp−1ϕ−1p−11(t)|(J1−αT|tϕ1)′(t)|pp−1dt). | (4.11) |
Next, by (2.8), we obtain
∫Ta(s−a)−γp−1ϕ2(s)ds=(T−a)−λ∫Ta(s−a)−γp−1(T−s)λds≤∫Ta(s−a)−γp−1ds. |
On the other hand, by (3.1) and (3.2), it is clear that γ<p−1. Thus, we deduce that
∫Ta(s−a)−γp−1ϕ2(s)ds≤C(T−a)1−γp−1. | (4.12) |
By (2.5) and (2.7), we have
∫Tat1p−1(lnta)−σp−1ϕ−1p−11(t)|(J1−αT|tϕ1)′(t)|pp−1dt=(lnTa)−λ∫Ta(lnTt)λ−αpp−1(lnta)−σp−11tdt≤(lnTa)−αpp−1∫Ta(lnta)−σp−11tdt. |
Notice that by (3.1) and (3.2), we have σ<p−1. Thus, we get
∫Tat1p−1(lnta)−σp−1ϕ−1p−11(t)|(J1−αT|tϕ1)′(t)|pp−1dt≤C(lnTa)1−αp+σp−1. | (4.13) |
Hence, it follows from (4.11)–(4.13) that
K1≤C(T−a)1−γp−1(lnTa)1−αp+σp−1. | (4.14) |
Similarly, we can write the term K2 as
K2=(∫Ta1t(lnta)−σmp−mϕ1(t)dt)(∫Ta(s−a)−γmp−mϕ−mp−m2(s)|(I1−βT|sϕ2)′(s)|pp−mds). | (4.15) |
By (2.5), we have
∫Ta1t(lnta)−σmp−mϕ1(t)dt=(lnTa)−λ∫Ta(lnta)−σmp−m(lnTt)λ1tdt≤∫Ta(lnta)−σmp−m1tdt. |
Notice that by (3.2), we have σm<p−m. Thus, we get
∫Ta1t(lnta)−σmp−mϕ1(t)dt≤C(lnTa)1−σmp−m. | (4.16) |
On the other hand, by (2.8) and (2.10), we have
∫Ta(s−a)−γmp−mϕ−mp−m2(s)|(I1−βT|sϕ2)′(s)|pp−mds=(T−a)−λ∫Ta(T−s)λ−βpp−m(s−a)−γmp−mds≤(T−a)−βpp−m∫Ta(s−a)−γmp−mds. |
Notice that by (3.2), we have p>m(γ+1). Therefore, we obtain
∫Ta(s−a)−γmp−mϕ−mp−m2(s)|(I1−βT|sϕ2)′(s)|pp−mds≤C(T−a)1−γm+βpp−m. | (4.17) |
Combining (4.16) with (4.17), there holds
K2≤C(lnTa)1−σmp−m(T−a)1−γm+βpp−m. | (4.18) |
Hence, it follows from (4.14) and (4.18) that
K1+K2≤C[(lnTa)1−αp+σp−1(T−a)1−γp−1+(lnTa)1−σmp−m(T−a)1−γm+βpp−m]. | (4.19) |
Thus, by (4.7), (4.10), and (4.19), we deduce that
∫∞a|u1(t)|m1tdt≤C[(lnTa)1−αp+σp−1(T−a)β−γp−1+(lnTa)1−σmp−m(T−a)β−γm+βpp−m]. | (4.20) |
Notice that by (3.1) and (3.2), we have
β−γp−1<0,β−γm+βpp−m<0. |
Hence, passing to the limit as T→∞ in (4.20), we obtain a contradiction with u1≢0. Consequently, (1.2) admits no global solution. The proof is completed.
Proof of Theorem 3.2. Suppose that u∈Xa is a global solution to (1.2). Notice that in the proof of Theorem 3.1, to obtain (4.20), we used that
p>m≥1,p>σ+1,p>m(σ+1),p>m(γ+1). |
On the other hand, by (3.3)–(3.5), it can be easily seen that the above conditions are satisfied. Thus, (4.20) holds. Hence, taking p=1+γβ in (4.20), we obtain
∫∞a|u1(t)|m1tdt≤C[(lnTa)1−αp+σp−1+(lnTa)1−σmp−m(T−a)β−γm+βpp−m]. | (4.21) |
On the other hand, by (3.3)–(3.5), we have
1−αp+σp−1<0,β−γm+βpp−m<0. |
Hence, passing to the limit as T→∞ in (4.21), we obtain a contradiction with u1≢0. This shows that (1.2) admits no global solution. The proof is completed.
The two-times fractional differential inequality (1.2) is investigated. Namely, using the test function method and a judicious choice of a test function, sufficient conditions ensuring the nonexistence of global solutions to (1.2) are obtained. Two cases are discussed separately: 1<p<1+γβ (see Theorem 3.1) and p=1+γβ (see Theorem 3.2). In the first case, no assumption is imposed on the fractional order α∈(0,1) of the Hadamard-Caputo fractional derivative, while in the second case, it is supposed that α>1−1m. About the initial conditions, in both cases, it is assumed that u0∈L1([a,∞)), u1∈Lm([a,∞),1tdt), and u1≢0.
Finally, it would be interesting to extend this study to two-times fractional evolution equations. For instance, the tow-times fractional semi-linear heat equation
{HCDαa|tu(t,s,x)+CDβa|s|u|m(t,s,x)≥(s−a)γ(lnta)σ|u|p(t,s,x),t,s>a,x∈RN,u(a,s,x)=u0(s,x),u(t,a,x)=u1(t,x),t,s>a,x∈RN, |
deserves to be studied.
The authors extend their appreciation to the Deanship of Scientific Research at Imam Mohammad Ibn Saud Islamic University for funding this work through Research Group no. RG-21-09-02.
The authors declare that they have no competing interests.
[1] |
R. M. Abu-Saris, R. DeVault, Global stability of yn+1=A+yn/yn−k, Appl. Math. Lett., 16 (2003), 173–178. https://doi.org/10.1016/S0893-9659(03)80028-9 doi: 10.1016/S0893-9659(03)80028-9
![]() |
[2] | R. Akerkar, Nonlinear functional analysis, Narosa Publishing House, New Delhi, 1999. |
[3] |
A. M. Amleh, E. A. Grove, G. Ladas, D. A. Georgiou, On the recursive sequence xn+1=α+xn−1/xn, J. Math. Anal. and Appl., 233 (1999), 790–798. https://doi.org/10.1006/jmaa.1999.6346 doi: 10.1006/jmaa.1999.6346
![]() |
[4] |
K. S. Berenhaut, S. Stević, A note on positive non-oscillatory solutions of the difference equation, xn+1=α+xpn−k/xpn, J. Differ. Equ. Appl., 12 (2006), 495–499. https://doi.org/10.1080/10236190500539543 doi: 10.1080/10236190500539543
![]() |
[5] | E. Camouzis, R. Devault, Asymptotic behaviour of solutions of xn+1=p+xn−1/xn, Proceedings of the 6th International Conference on Difference Equations, Augsburg, Germany, 2001,375–386. |
[6] | P. Cull, M. Flahive, R. Robson, Difference equations, from rabbits to chaos, New York: Springer, 2005. https://doi.org/10.1007/0-387-27645-9 |
[7] | H. M. El-Owaidy, A. M. Ahmed, M. S. Mousa, On asymptotic behaviour of the difference equation xn+1=α+xpn−k/xpn, J. Appl. Math. Comput., 12 (2003), 31–37. |
[8] | H. El-Metwally, E. A. Grove, G. Ladas, R. Levins, M. Radin, On the difference equation xn+1=α+βxn−1e−xn, Nonlinear Anal., 47 (2001), 4623–4634. |
[9] |
N. Fotiades, G. Papaschinopoulos, Existence, uniqueness and attractivity of prime period two solution for a difference equation of exponential form, Appl. Math. Comput., 218 (2012), 11648–11653. https://doi.org/10.1016/j.amc.2012.05.047 doi: 10.1016/j.amc.2012.05.047
![]() |
[10] | E. A. Grove, G. Ladas, Periodicity in nonlinear difference equations, Advances in Discrete Discrete Mathematics and Applications, Chapman and Hall/CRC, 2004. |
[11] |
A. E. Hamza, A. Morsy, On the recursive sequence xn+1=α+xn−1/xkn, Appl. Math. Lett., 22 (2009), 91–95. https://doi.org/10.1016/j.aml.2008.02.010 doi: 10.1016/j.aml.2008.02.010
![]() |
[12] |
W. S. He, W. T. Li, X. X. Yan, Global attractivity of the difference equation xn+1=p+xn−k/xn, Appl. Math. Appl., 151 (2004), 879–885. https://doi.org/10.1016/S0096-3003(03)00528-9 doi: 10.1016/S0096-3003(03)00528-9
![]() |
[13] |
G. L. Karakostas, Asymptotic 2-periodic difference equations with diagonally self-invertible responses, J. Differ. Equ. Appl., 6 (2000), 329–335. https://doi.org/10.1080/10236190008808232 doi: 10.1080/10236190008808232
![]() |
[14] |
G. L. Karakostas, S. Stević, On the recursive sequence xn+1=α+xn−kf(xn,⋯,xn−k+1), Demonstr. Math., 38 (2005), 595–610. https://doi.org/10.1515/dema-2005-0309 doi: 10.1515/dema-2005-0309
![]() |
[15] |
G. L. Karakostas, S. Stević, On the recursive sequence xn+1=α+xn−kα0xn+⋯+αk−1xn−k+1+γ, J. Differ. Equ. Appl., 10 (2004), 809–815. https://doi.org/10.1080/10236190410001659732 doi: 10.1080/10236190410001659732
![]() |
[16] | M. R. S. Kulenović, G. Ladas, Dynamics of second order rational difference equations, Chapman and Hall/CRC, Boca Raton, Fla, USA, 2002. |
[17] | J. P. LaSalle, The stability of dynamical systems, Society for Industrial and Applied Mathematics, 1976. |
[18] | S. Stević, On the recursive sequence xn+1=α+xn−1/xn, Dyn. Contin., Discrete Impuls. Syst. Ser. A, 10 (2003), 911–916. |
[19] |
S. Stević, On the difference equation xn+1=α+xn−1/xn, Comput. Math. Appl., 56 (2008), 1159–1171. https://doi.org/10.1016/j.camwa.2008.02.017 doi: 10.1016/j.camwa.2008.02.017
![]() |
[20] | S. Stević, On the recursive sequence xn+1=α+xpn−1/xpn, J. Appl. Math. Comput., 18 (2005), 229–234. |
[21] |
S. Stević, Asymptotics of some classes of higher-order difference equations, Discrete Dyn. Nat. Soc., 2007 (2007), 1–20. https://doi.org/10.1155/2007/56813 doi: 10.1155/2007/56813
![]() |
[22] |
S. Stević, On the recursive sequence xn+1=A+xpn−1/xpn, Discrete Dyn. Nat. Soc., 2007 (2007), 1–9. https://doi.org/10.1155/2007/34517 doi: 10.1155/2007/34517
![]() |
[23] |
S. Stević, On the recursive sequence xn+1=α+xpn/xrn−1, Discrete Dyn. Nat. Soc., 2007 (2007), 1–9. https://doi.org/10.1155/2007/40963 doi: 10.1155/2007/40963
![]() |
[24] | E. Tasdemir, M. Göcen, Y. Soykan, Global dynamical behaviours and periodicity of a certain quadratic-rational difference equation with delay, Milkolc Math. Notes, 23 (2022), 471–484. |
[25] |
I. Yalcinkaya, On the difference equation xn+1=a+xn−1/xkn, Discrete Dyn. Nat. Soc., 2008 (2008), 1–8. https://doi.org/10.1155/2008/805460 doi: 10.1155/2008/805460
![]() |
1. | Saeed M. Ali, Mohammed D. Kassim, Sufficient Conditions for the Non-Existence of Global Solutions to Fractional Systems with Lower-Order Hadamard-Type Fractional Derivatives, 2025, 13, 2227-7390, 1031, 10.3390/math13071031 |