A Predictor-Corrector Algorithm in the Framework of Conformable Fractional Differential Equations

Mohamed Echchehira,Youness Assebbane,Mustapha Atraoui,Mohamed Bouaouid
2024-06-24
Abstract:This work proposes a conformable fractional predictor-corrector algorithm for solving conformable fractional differential equations. Fractional calculus is finding applications in various scientific fields, but existing numerical methods might have limitations. This work addresses that gap by introducing a new algorithm specifically designed for the conformable fractional derivative using Adams-Bashforth and Adams-Moulton methods.
Numerical Analysis
What problem does this paper attempt to address?
The problem that this paper attempts to solve is to propose a new algorithm for solving Conformable Fractional Differential Equations (CFDEs). Specifically, the author introduced a Conformable Fractional Predictor - Corrector Algorithm based on the Adams - Bashforth and Adams - Moulton methods to improve the efficiency and accuracy of numerically solving CFDEs. ### Background and Motivation Fractional Calculus has been widely applied in various scientific fields in recent years. However, due to the possible limitations of existing numerical methods, new algorithms need to be developed to better handle fractional differential equations. The Conformable Fractional Derivative is a relatively new definition, which exhibits superior properties when dealing with certain types of fractional differential equations. ### Main Contributions of the Paper 1. **Proposal of a New Algorithm**: The paper proposes a new Conformable Fractional Predictor - Corrector Algorithm, which combines the explicit characteristics of the Adams - Bashforth method and the stability of the Adams - Moulton method. 2. **Theoretical Basis**: The paper elaborates on the basic concepts and definitions of Conformable Fractional Calculus and presents the theoretical basis of the algorithm. 3. **Numerical Application**: Through specific numerical examples, the effectiveness and accuracy of this algorithm in solving Conformable Fractional Differential Equations are demonstrated. ### Specific Problems The paper mainly solves the following specific problems: - **Limitations of Numerical Solution**: Existing numerical methods may not be efficient or accurate enough when solving Conformable Fractional Differential Equations. - **Applicability of the Algorithm**: Whether the proposed new algorithm can be effectively applied to various types of Conformable Fractional Differential Equations. - **Theoretical Verification**: Verify the effectiveness and stability of the new algorithm through theoretical analysis and numerical experiments. ### Mathematical Formulas Some of the key mathematical formulas involved in the paper are as follows: - **Definition of Conformable Fractional Derivative**: \[ T_\alpha f(t)=\lim_{\epsilon \to 0}\frac{f(t + \epsilon t^{1-\alpha})-f(t)}{\epsilon}, \quad t > 0 \] where \(\alpha\in(0, 1]\). - **Definition of Conformable Fractional Integral**: \[ I_\alpha f(t)=\int_0^t x^{\alpha - 1}f(x)\,dx, \quad t\geq0 \] - **Conformable Fractional Predictor - Corrector Algorithm**: \[ \begin{cases} y_0 = y(0),\\ y_p^{n + 1}=y_0+\frac{h^\alpha}{\alpha}\sum_{j = 0}^n\left((j + 1)^\alpha-j^\alpha\right)f(t_j,y_j),\\ y^{n + 1}=y_0+\frac{h^\alpha}{\alpha(\alpha + 1)}\sum_{j = 0}^n a_j f(t_j,y_j)+\frac{h^\alpha}{\alpha(\alpha + 1)}\left((\alpha + 1)(n + 1)^\alpha+n^{\alpha+1}-(n + 1)^{\alpha+1}\right)f(t_{n+1},y_p^{n+1}) \end{cases} \] where the coefficient \(a_j\) is defined as: