New numerical approach for fractional differential equations

Abdon Atangana,Kolade M. Owolabi
DOI: https://doi.org/10.48550/arXiv.1707.08177
2017-07-22
Abstract:In the present case, we propose the correct version of the fractional Adams-Bashforth methods which take into account the nonlinearity of the kernels including the power law for the Riemann-Liouville type, the exponential decay law for the Caputo-Fabrizio case and the Mittag-Leffler law for the Atangana-Baleanu scenario. The Adams-Bashforth method for fractional differentiation suggested and are commonly use in the literature nowadays is not mathematically correct and the method was derived without taking into account the nonlinearity of the power law kernel. Unlike the proposed version found in the literature, our approximation, in all the cases, we are able to recover the standard case whenever the fractional power $\alpha=1$.
Numerical Analysis
What problem does this paper attempt to address?
The problem that this paper attempts to solve is the mathematical incorrectness existing in the existing fractional - order Adams - Bashforth methods when dealing with fractional - order differential equations. Specifically, the current methods do not fully consider the nonlinear characteristics of the kernel function during the derivation process, especially for the power - law of the Riemann - Liouville type, the exponential decay law of the Caputo - Fabrizio type, and the Mittag - Leffler law of the Atangana - Baleanu type. Moreover, when the fractional - order \(\alpha = 1\), the existing methods cannot recover the classical Adams - Bashforth numerical scheme. To solve these problems, the author proposes a new fractional - order Adams - Bashforth method, which can correctly handle different types of fractional - order derivatives and can recover the classical Adams - Bashforth method when \(\alpha = 1\). This improvement makes the numerical solution more accurate and applicable to a wider range of fractional - order differential equation models. ### Specific problem description 1. **Deficiencies of existing methods**: - The existing fractional - order Adams - Bashforth methods ignore the nonlinearity of the fractional - order integral kernel during derivation. - When the fractional - order \(\alpha = 1\), the existing methods cannot recover the classical Adams - Bashforth method. 2. **Objectives**: - Propose a new fractional - order Adams - Bashforth method that can correctly handle different types of fractional - order derivatives (such as Riemann - Liouville, Caputo - Fabrizio, and Atangana - Baleanu). - Ensure that the new method can recover the classical Adams - Bashforth method when \(\alpha = 1\). 3. **Solutions**: - The new method takes into account the nonlinear characteristics of the kernel functions of different fractional - order derivatives. - Provide detailed numerical experiment results to verify the effectiveness and accuracy of the new method. Through these improvements, the author aims to provide a more accurate and more general numerical method for solving fractional - order differential equations, thereby promoting the further development of fractional - order calculus in practical applications.