This work advances the stability theory of fractional differential equations by establishing superstability criteria for a significant class of problems involving the Caputo-Katugampola derivative. Utilizing a generalized Taylor series as a foundational tool, we prove that these equations exhibit superstable behavior under specific conditions. Our results generalize a wide range of existing stability theorems, creating a unified framework that encompasses systems governed by the Caputo fractional derivative as special cases of the more general Katugampola operator.
A cornerstone of functional equations theory is the study of Hyers-Ulam stability (HUS). This area of inquiry originated with a foundational question posed by Ulam in 1940 concerning the stability of homomorphisms, specifically asking when an approximate linear mapping can be closely approximated by an exact one. A year later, Hyers provided a definitive solution to this problem [1]. The concept has since been significantly extended beyond functional equations to differential equations, leading to a substantial body of literature establishing HUS results for various classes of differential equations [2, 3].
Within this broader context, fractional differential equations (FDEs) have emerged as a particularly vibrant field of research. While the concept of the fractional derivative dates back to the correspondence between L’Hospital and Leibniz in 1695, its rigorous application has seen explosive growth in recent decades. Contemporary investigations into FDEs encompass a wide range of topics, including stability analysis [4– 7], finite-time stability [8], stabilization [9], observer design [9, 10], and fault estimation [11]. A specific and actively researched aspect is the Hyers-Ulam stability of FDEs, to which numerous significant contributions have been made [12– 15].
A specialized form of stability, known as superstability (SS), was introduced in [8] as a particular case of HUS. Superstability describes a situation where any approximate solution to a given equation is either an exact solution or is uniformly bounded by a constant multiple of the perturbation parameter. Initial research on this concept for differential equations can be found in [16, 17]. In [18], the author investigated the SS for fractional differential equations with Caputo derivative.
Our work focuses on establishing superstability for the following initial value problem. \[\label{fracdiffeqnb} {}^CD_{r}^{p\lambda,\rho} E(x)+A(x) E(x)=0, \tag{1}\] with initial conditions (IC): \[\label{it} E(r)={}^CD_{r}^{\lambda,\rho} E(r)={}^CD_{r}^{2\lambda,\rho} E(r)=…={}^CD_{r}^{(p-1)\lambda,\rho} E(r)=0, \tag{2}\] where \(p\in \mathbb{N}^*\), \({}^CD_{r}^{s\lambda,\rho} E \in C\big([r,r+u] \big)\), for each \(s\in \{0,1,…,p\}\), \(A\in C\big([r,r+u] \big)\), \(u>0\) and \({}^CD_{r}^{s\lambda,\rho} ={}^CD_{r}^{\lambda,\rho}.{}^CD_{r}^{\lambda,\rho}…{}^CD_{r}^{\lambda,\rho}\) (\(s-\)times).
Motivated by [16– 18], we introduce the following definition.
Definition 1. Suppose that \(E\) satisfies: \[\label{diffeqn} |\psi\big(A, E, {}^CD_{r}^{\lambda,\rho}E,{}^CD_{r}^{2\lambda,\rho}E,…,{}^CD_{r}^{p\lambda,\rho}E \big)|\leq \nu,\ \forall \omega\in [r,r+u], \tag{3}\] for some \(\nu\geq 0\) with IC therefore either \[|E(\omega)|\leq \vartheta \nu,\ \forall \omega\in [r,r+u],\] where \(\vartheta>0\), or \[\psi\big(A, E, {}^CD_{r}^{\lambda,\rho}E,{}^CD_{r}^{2\lambda,\rho}E,…,{}^CD_{r}^{p\lambda,\rho}E \big)=0.\]
Then, we say that (1) has SS with IC.
Definition 2.[19] Given \(0<l<1\), \(\rho>0\). The Caputo-Katugampola fractional derivative of an absolutely continuous function \(f\) is defined as, \[\label{dercap} ^CD_{c}^{l,\rho} f(s)=\displaystyle \frac{\rho^l}{\Gamma(1-l)} \displaystyle \int_{c}^s (s^\rho-\tau^\rho)^{-l}f'(\tau)d\tau. \tag{4}\]
Theorem 1 ([20] Generalized Taylor’s formulat).Let \(0<\eta< 1\), \(\rho>0\). Assume that \({}^CD_{r_1}^{t \eta,\rho} h \in C\big([r_1,r_2] \big)\), for each \(t\in \{0,1,…,s\}\), with \(s\in \mathbb{N}^*\), then we have \[h(x)=\displaystyle \sum_{t=0}^{s-1}{}^CD_{r_1}^{t\eta,\rho} h(r_1) \frac{(\frac{x^\rho-r_1^\rho}{\rho})^{t \eta}}{\Gamma(t \eta+1)}+{}^CD_{r_1}^{s\eta,\rho} h(c) \frac{(\frac{x^\rho-r_1^\rho}{\rho})^{s \eta}}{\Gamma(s \eta+1)},\] with \(c\in [r_1,x]\), for each \(x\in (r_1,r_2]\).
The main contribution of this work is outlined in this section.
Theorem 2. Assume that \(\displaystyle\sup_{\chi\in [r,r+u]}|A(\chi)|< \displaystyle\frac{\Gamma(p \lambda+1)}{\big(\frac{(u+r)^\rho-r^\rho}{\rho}\big)^{p \lambda}}\). Then, (1) has the SS with IC (2).
Proof. Let \(\nu>0\), and \(E\in C\big([r,r+u] \big)\) with \({}^CD_{r}^{t\lambda,\rho} E \in C\big([r,r+u] \big)\) for each \(t\in \{ 0,1,…p \}\), if \[|{}^CD_{r}^{p\lambda,\rho} E(x)+A(x) E(x)|\leq \nu,\] and \[E(r)={}^CD_{r}^{\lambda,\rho} E(r)={}^CD_{r}^{2\lambda,\rho} E(r)=…={}^CD_{r}^{(p-1)\lambda,\rho} E(r)=0.\]
It follows from Theorem 1 that \[E(x)=\displaystyle \sum_{t=0}^{p-1}{}^CD_{r}^{t\lambda,\rho} E(r) \frac{(\frac{x^\rho-r^\rho}{\rho})^{t \lambda}}{\Gamma(t \lambda+1)}+{}^CD_{r}^{p\lambda,\rho} E(c) \frac{(\frac{x^\rho-r^\rho}{\rho})^{p \lambda}}{\Gamma(p \lambda+1)},\] where \(c\in [r,x]\), for all \(x\in (r,r+u]\). Therefore \[\begin{aligned} |E(x)|&=&|{}^CD_{r}^{p\lambda,\rho} E(c) \frac{(\frac{x^\rho-r^\rho}{\rho})^{p \lambda}}{\Gamma(p \lambda+1)}| \notag \\ &\leq& \displaystyle\sup_{\chi\in [r,r+u]}|{}^CD_{r}^{p\lambda,\rho} E(\chi)| \frac{\big(\frac{(u+r)^\rho-r^\rho}{\rho}\big)^{p \lambda}}{\Gamma(p \lambda+1)}. \end{aligned} \tag{5}\]
Thus,
\[\begin{aligned} \displaystyle\sup_{\chi\in [r,r+u]}|E(\chi)|&\leq & \frac{\big(\frac{(u+r)^\rho-r^\rho}{\rho}\big)^{p \lambda}}{\Gamma(p \lambda+1)} \Big[ \displaystyle\sup_{\chi\in [r,r+u]}|{}^CD_{r}^{p\lambda,\rho}E(\chi)-A(\chi)E(\chi)|+\displaystyle\sup_{\chi\in [r,r+u]}|A(\chi)| \displaystyle\sup_{\chi\in [r,r+u]}|E(\chi)| \Big] \notag \\ &\leq& \frac{\big(\frac{(u+r)^\rho-r^\rho}{\rho}\big)^{p \lambda}}{\Gamma(p \lambda+1)} \nu + \frac{\big(\frac{(u+r)^\rho-r^\rho}{\rho}\big)^{p \lambda}}{\Gamma(p \lambda+1)} \displaystyle\sup_{\chi\in [r,r+u]}|A(\chi)| \displaystyle\sup_{\chi\in [r,r+u]}|E(\chi)| . \end{aligned} \tag{6}\]
Then, \[\displaystyle\sup_{\chi\in [r,r+u]}|E(\chi)|\Big( 1- \frac{\big(\frac{(u+r)^\rho-r^\rho}{\rho}\big)^{p \lambda}}{\Gamma(p \lambda+1)} \displaystyle\sup_{\chi\in [r,r+u]} |A(\chi)| \Big)\leq \frac{\big(\frac{(u+r)^\rho-r^\rho}{\rho}\big)^{p \lambda}}{\Gamma(p \lambda+1)} \nu.\]
Hence, there is \(K>0\) with \[|E(x)|\leq K \nu,\] for every \(x\in [r,r+u]\). ◻
Remark 1. It is important to note that in [17], the authors obtained SS results for DEs with integer-order derivatives. In [18], the SS is studied for the Caputo fractional derivative. In our work, however, the main result is obtained for the Caputo-Katugampola fractional-order derivative.
In conclusion, this work successfully establishes a unified superstability framework for fractional differential equations involving the Caputo-Katugampola derivative. By leveraging a generalized Taylor series, we have derived specific conditions under which these equations exhibit superstable behavior. Our findings not only generalize a wide spectrum of existing stability theorems but also consolidate systems governed by the standard Caputo derivative as special cases within this broader, more versatile theory.
Hyers, D. H. (1941). On the stability of the linear functional equation. Proceedings of the National Academy of Sciences, 27(4), 222-224.
Rezaei, H., Jung, S. M., & Rassias, T. M. (2013). Laplace transform and Hyers–Ulam stability of linear differential equations. Journal of Mathematical Analysis and Applications, 403(1), 244-251.
Rus, I. A. (2009). Ulam stability of ordinary differential equations. Studia Universitatis Babeş-Bolyai, Mathematica, 2009(4), 125–134.
Makhlouf, A. B., Hammami, M. A., & Sioud, K. (2017). Stability of fractional-order nonlinear systems depending on a parameter. Bulletin of the Korean Mathematical Society, 54(4), 1309-1321.
Naifar, O., Makhlouf, A. B., & Hammami, M. A. (2016). Comments on “Lyapunov stability theorem about fractional system without and with delay”. Communications in Nonlinear Science and Numerical Simulation, 30(1-3), 360-361.
Naifar, O., Makhlouf, A. B., & Hammami, M. A. (2017). Comments on “Mittag-Leffler stability of fractional order nonlinear dynamic systems [Automatica 45 (8)(2009) 1965–1969]”. Automatica, 75, 329.
Souahi, A., Naifar, O., Makhlouf, A. B., & Hammami, M. A. (2019). Discussion on Barbalat Lemma extensions for conformable fractional integrals. International Journal of Control, 92(2), 234-241.
Naifar, O., Nagy, A. M., Makhlouf, A. B., Kharrat, M., & Hammami, M. A. (2019). Finite‐time stability of linear fractional‐order time‐delay systems. International Journal of Robust and Nonlinear Control, 29(1), 180-187.
Naifar, O., Ben Makhlouf, A., Hammami, M. A., & Chen, L. (2018). Global practical Mittag Leffler stabilization by output feedback for a class of nonlinear fractional‐order systems. Asian Journal of Control, 20(1), 599-607.
Jmal, A., Naifar, O., Ben Makhlouf, A., Derbel, N., & Hammami, M. A. (2018). On observer design for nonlinear Caputo fractional‐order systems. Asian Journal of Control, 20(4), 1533-1540.
Jmal, A., Naifar, O., Ben Makhlouf, A., Derbel, N., & Hammami, M. A. (2017). Robust sensor fault estimation for fractional-order systems with monotone nonlinearities. Nonlinear Dynamics, 90(4), 2673-2685.
Almalahi, M. A., Abdo, M. S., & Panchal, S. K. (2021). Existence and Ulam–Hyers–Mittag-Leffler stability results of \(\psi\)-Hilfer nonlocal Cauchy problem. Rendiconti del Circolo Matematico di Palermo Series 2, 70(1), 57-77.
Boulares, S., Ben Makhlouf, A., & Khellaf, H. (2020). Generalized weakly singular integral inequalities with applications to fractional differential equations with respect to another function. https://projecteuclid.org/euclid.rmjm/1596037176.
Brzdęk, J., & Eghbali, N. (2016). On approximate solutions of some delayed fractional differential equations. Applied Mathematics Letters, 54, 31-35.
Saifia, O., Boucenna, D., & Chidouh, A. (2020). Study of Mainardi’s fractional heat problem. Journal of Computational and Applied Mathematics, 378, 112943.
GĂVRUŢĂ, P., Jung, S. M., & Li, Y. (2011). Hyers-ulam stability for second-order linear differential equations with boundary conditions. Electronic Journal of Differential Equations, 54, 1–5.
LI, Y. C. (2014). Distinction of turbulence from chaos–rough dependence on initial data. Electronic Journal of Differential Equations, 2014(104), 1-8.
Makhlouf, A. B. (2022). Superstability of higher-order fractional differential equations. Annals of the University of Craiova-Mathematics and Computer Science Series, 49(1), 11-14.
Kilbas, A. A. (2006). Theory and Applications of Fractional Differential Equations. North-Holland Mathematics Studies, 204.
Odibat, Z. M., & Shawagfeh, N. T. (2007). Generalized Taylor’s formula. Applied Mathematics and Computation, 186(1), 286-293.