On System of Mixed Fractional Hybrid Differential Equations
Abstract
In this article, we find the necessary conditions for the existence and uniqueness of solutions to a system of hybrid equations that contain mixed fractional derivatives (Caputo and Riemann-Liouville). We also verify the stability of these solutions using the Ulam-Hyers (U-H) technique. Finally, this study ends with applied examples that show how to proceed and verify the conditions of our theoretical results.
1. Introduction
Although the concept of fractional calculus was established 300 years ago, interest in this type of derivative appeared for a short period. So that it is no secret to anyone that the most important use of fractional derivatives is to find analytical solutions to differential equations if possible, or by using numerical analysis methods to find an approximation to these solutions. In this study, we will focus on the idea of studying theories that investigate the existence of a solution to a system of hybrid fractional equations that contain mixed fractional derivatives with boundary conditions attached to them.
As mentioned before, fractional calculus as a concept is not very recent. It is worth mentioning here the great names who have given a lot to this science, such as A.V. Letnikov, J. Hadamard, J. Liouville, B. Riemann M., and Caputo L. worked in this field. These names must be mentioned by way of example. To get acquainted with some of the names of scientists who have made great contributions to fractional calculus in the modern world, we ask the reader to look at [1].
Fractional derivatives have played a very important role in mathematical modeling in many diverse applied sciences, see [2, 3]. For example, the authors in [4] employed the fractional derivative of the Psi-Caputo type in modeling the logistic population equation, through which they were able to show that the model with the fractional derivative led to a better approximation of the variables than the classical model. In addition, the authors in [5] employed the fractional derivative of the Psi-Caputo type and used the kernel Rayleigh, to improve the model again in modeling the logistic population equation.
As a final example, the authors in [6] employed the fractional derivatives of the Caputo and Caputo-Fabrizio type by modeling the equation that gives the relationship between atmospheric pressure and altitude, and they were also able to show that the fractional equation gave less error in estimating atmospheric pressure at a certain altitude. There are many scientific papers in the literature that prove the superiority of fractional derivatives over classical ones.
There are a large number of manuscripts published in the literature that investigate the issue of the existence of a solution to fractional differential equations, whether they are sequential equations of type or nonsequential equations [7–14].
Here, D is the ordinary derivative, ψ : [0, 1] × ℝ⟶ℝ, λ ∈ ℝ + , δ is a real number such that δ ≠ ((λ + e−λ − 1)/(λη + e−λη − 1)).
where φ : [0, 1] × ℝ⟶ℝ, δ ∈ ℝ, ζ ∈ (0, 1).
It is known that fractional calculus and FDEs are used in different fields such as physics, signal and image processing, control theory, robotics, economics, biology, and metallurgy, see for example [16, 17] and references therein. On the other hand, recently, many researchers have paid much attention to hybrid differential equations of fractional order. This is because of the development and new advanced applications of fractional calculus. The fractional hybrid modeling is of great significance in different engineering fields, and it can be a unique idea for future combined research between various applied sciences, for example, see [18] in which fractional hybrid modeling of a thermostat is simulated, for some recent results on hybrid.
For FDEs, we refer to [19, 20]. Freshly, some authors have studied different characteristics of hybrid FDEs including the existence of solutions, see for some detail [21–29], and some go further and studied Hyers-Ulam stability for FDEs by different mathematical theories, see for some detail [26].
where RLDr is Riemann–Liouville fractional derivative, f : [0, T] × ℝ⟶ℝ/{0}, g : [0, 1] × ℝ⟶ℝ are assumed to be continuous.
in which f : [0, T] × ℝ⟶ℝ/{0}, g : [0, 1] × ℝ⟶ℝ are assumed to be continuous and a1 + a2 ≠ 0.
After this introductory section of this work, the manuscript is organized as the following hierarchical structure: Section 2 delivers the basic elements of fractional calculus definitions, Section 3 introduces the main results of the work, Section 4 introduces the (U-H) stability result for our problem, and the last section is arranged for a numerical example to support the theoretical results.
2. Preliminaries
In this part, we present some basic elements and definitions needed to find solutions to the main mathematical problem presented in this study.
Definition 1 (see [3].)The Riemann-Liouville (RL) fractional integral is defined by
Definition 2 (see [3].)The Caputo fractional derivative of order ν of a function ϑ : ℝ+⟶ℝ is given by
Theorem 3 (see [3], Banach’s contraction mapping principle.)Let (S, d) be a complete metric space; H : S⟶S is a contraction then
- (i)
H has a unique fixed point s ∈ S; that, is H(s) = s
- (ii)
∀s0 ∈ S, we have
Theorem 4 (see [3], nonlinear alternative of Leray-Schauder type.)Assume that V is an open subset of a Banach space U, 0 ∈ V, and be a contraction such that is bounded then either
- (i)
F has a fixed point in , or
- (ii)
∃μ ∈ (0, 1) and v ∈ ∂V such that v = μF(v) holds
Theorem 5 (see [2], Arzela-Ascoli theorem.)F ⊂ C(U, ℝ) is compact if and only if it is closed, bounded, and equicontinuous.
3. Main Results
Lemma 6. If h ∈ C([0, 1], ℝ), and
Proof. Taking to , then take to the resulting equation, we get
Substitution of χ(0) = 0 and χ′(0) = 0 in Equation (11) gives a2 = 0 and a0 = 0, respectively, and consequently, Equation (6) becomes
Inserting a1 in Equation (12) gives
Alternatively, we have
Equation (15) is equivalent to Equation (10), which makes the proof done.
To construct the necessary conditions for the results of uniqueness and existence of the problem (6), let us consider the following hypotheses.
Theorem 7. If both (C1) and (C2) are satisfied, and assume that [λℏΛ1(υ1 + υ2) + λƛΛ2(ℓ1 + ℓ2)] < 1. Then, the system in Equation (6) has a unique solution.
Proof. Define a closed ball with γ ≥ (λℏΛ1Nψ + λƛΛ2Nφ)/(1 − (λℏΛ1(υ1 + υ2) + λƛΛ2(ℓ1 + ℓ2))), where .
Observe that |ψ(τ, χ, ϑ)| = |ψ(τ, χ, ϑ) − ψ(τ, 0, 0) + ψ(τ, 0, 0)| ≤ υ1‖χ‖ + υ2‖ϑ‖ + Nψ ≤ (υ1 + υ2)γ + Nψ.
First, we show that . For any , we have
From Equation (25) and Equation (26), we deduce that ‖ϒ(χ, ϑ)‖ ≤ γ.
Next, for (χ1, ϑ1), (χ2, ϑ2) ∈ C × C, ∀τ ∈ [0, 1], we have
Similarly, we can find
Combining Equation (27) and Equation (28) yields
Equation (29) becomes ‖ϒ(χ1, ϑ1) − ϒ(χ2, ϑ2)‖ ≤ (‖χ1 − χ2‖ + ‖ϑ1 − ϑ2‖). That is, ϒ is a contraction; consequently, Banach fixed point theorem applies; thus, the uniqueness of solutions for Equation (6) holds on [0, 1].
Theorem 8. If (C1), (C3), and (C4) are satisfied, and if (λℏΛ1ω1 + λƛΛ2θ1) < 1 and (λℏΛ1ω2 + λƛΛ2θ2) < 1, then Equation (6) has at least one solution.
Proof. In the first step, we verify that the operator ϒ : C × C⟶C × C is completely continuous; obviously, the operator is continuous as a result that ℏ, ƛ, ψ, and φ are all assumed to be continuous.
With the aid of (C4), ∀(χ, ϑ) ∈ S, we have
Next, to verify the equicontinuity for the operator ϒ, we let τ1, τ2 ∈ [0, 1], (τ1 < τ2) then
The R.H.S for both inequalities (33) and (34) tend to zero as τ1⟶τ2, and they are both independent on (χ, ϑ). So, operator ϒ(χ, ϑ) is equicontinuous and yields; ϒ(χ, ϑ) is completely continuous.
Finally, we establish the bounded set given by ; then, ∀τ ∈ [0, 1]; the equation (χ, ϑ) = βϒ(χ, ϑ) gives
Using the hypothesis (C3), we get
Consequently, we have
Inequality (37) can be written as follows:
where Λ0 = min{1 − (λℏΛ1ω1 + λƛΛ2θ1), 1 − (λℏΛ1ω2 + λƛΛ2θ2)}.
Inequality (38) shows that Ω is bounded. Hence, Leray-Schauder alternative applies, implying the existence of the solution for Equation (6).
4. Stability
In this part, we address the issue of stability of solutions to the system of equations defined by Equation (6) via U-H definition.
Definition 9. The system of the coupled sequential fractional differential BVPs Equation (6) is stable in U–H sense if a real number c = max(c1, c2) > 0 exists so that, for any ε = max(ε1, ε2) > 0 and for any satisfying
It is clear that satisfies the inequalities (39) if there exists a function (h1, h2) ∈ C × C (which depends on ), such that
- (i)
|h1(τ)| < ε1 and |h2(τ)| < ε2, τ ∈ [0, 1]
- (ii)
For τ ∈ [0, 1]
Theorem 10. Suppose that (C2) is fulfilled. Moreover
Then, the system of coupled sequential fractional differential BVPs (6) is U–H stable.
Proof. Assume that for ε1, ε2 > 0 a couple satisfies the inequalities (39). Introduce the following operator
Then
From Equation (44) and Equation (45), we obtain
From Equation (27) and Equation (28), we obtain
Let (χ, ϑ) ∈ C × C be a solution of Equation (6). Thanks to Lemma 6, it is equivalent to the following integral equations:
By the same arguments in Theorem 7, we get
It follows that
Representing these inequalities as matrices, we get
Solving the above inequality, we get
Thus
For ε = max(ε1, ε2) and
Therefore, with the aid of Definition 9, the solution of the problem Equation (6) is U–H stable.
5. Example
Thus, the boundary value problem Equation (59) satisfies all the conditions of Theorem 7; consequently, the uniqueness of solution of Equation (59) is satisfied on [0, 1].
Finally, easy calculations with the data above give (λℏΛ1ω1 + λƛΛ2θ1) = 0.203275 < 1 and (λℏΛ1ω2 + λƛΛ2θ2) = 0.203 < 1; all conditions of Theorem 8 hold; that is, the problem (59) has at least one solution in [0, 1].
6. Conclusion
We have studied a coupled hybrid FDEs consisting of mixed fractional derivatives such as Caputo and Riemann-Liouville fractional derivatives and nonlocal boundary conditions. Existence/uniqueness results are established via a nonlinear alternative of the Leray-Schauder and Banach fixed point theorem. We also studied the Ulam-Hyers stability of these couple of hybrid FDEs. The obtained result is well illustrated by a numerical example. The result obtained in this paper is new and significantly contributes to the existing literature on the topic.
One possible direction in which to extend the results of this paper is toward different kinds of mixed fractional differential and mixed conformable fractional differential systems of higher order. Another challenge is to find out if similar results can be derived in the case of constant/variable delays in linear/nonlinear terms.
Conflicts of Interest
The authors declare that there is no conflict of interest regarding the publication of this paper.
Authors’ Contributions
M.A and N. M contributed to each part of this work equally and read and approved the final version of the manuscript.
Acknowledgments
This work was supported through the Annual Funding track by the Deanship of Scientific Research, Vice Presidency for Graduate Studies and Scientific Research, King Faisal University, Saudi Arabia (Project No. AN000501).
Open Research
Data Availability
No data were used to support this study.