A fractional Traub method with $(2α+1)$th-order of convergence and its stability

Giro Candelario,Alicia Cordero,Juan R. Torregrosa

DOI: https://doi.org/10.48550/arXiv.1909.09076

2019-09-20

Abstract:Some fractional Newton methods have been proposed in order to find roots of nonlinear equations using fractional derivatives. In this paper we introduce a fractional Newton method with order $\alpha+1$ and compare with another fractional Newton method with order $2\alpha$. We also introduce a fractional Traub method with order $2\alpha+1$ and compare with its first step (fractional Newton method with order $\alpha+1$). Some tests and analysis of the dependence on the initial estimations are made for each case.

Numerical Analysis

What problem does this paper attempt to address?

The problem that this paper attempts to solve is: when finding the roots of nonlinear equations, use fractional - order derivatives to improve the convergence and stability of the Newton method and the Traub method. Specifically, the author introduced two new fractional - order methods - the fractional - order Traub method with \((2\alpha + 1)\)-order convergence, and compared it with the existing fractional - order Newton method.

### Core problems of the paper:
1. **Increasing the convergence rate**: By introducing fractional - order derivatives, iterative methods with higher - order convergence are designed, especially the fractional - order Traub method, whose convergence order is \(2\alpha + 1\).
2. **Enhancing stability**: Analyze the stability and convergence behavior of these new methods under different initial estimates to ensure that they can reliably find the roots of nonlinear equations in practical applications.
3. **Comparing with existing methods**: Compare the newly proposed method with the existing fractional - order Newton methods (with convergence orders of \(\alpha + 1\) and \(2\alpha\)) to verify the advantages of the new method.

### Specific research contents:
- **Caputo fractional - order derivative**: Define and use the Caputo fractional - order derivative to construct the fractional - order Newton method and the fractional - order Traub method.
- **Riemann - Liouville fractional - order derivative**: Similarly, define and use the Riemann - Liouville fractional - order derivative to construct the corresponding fractional - order Newton method and the fractional - order Traub method.
- **Convergence analysis**: Prove the local convergence order of these methods through mathematical derivation and give the error equations.
- **Numerical experiments**: Verify the effectiveness of these methods through multiple test functions (such as polynomials, complex - variable functions, etc.) and analyze the influence of initial estimate values on convergence.

### Conclusion:
The paper shows that the fractional - order Traub method has a higher convergence order (\(2\alpha + 1\)) compared with existing methods, and in some cases shows better stability and faster convergence speed. In addition, the numerical experiment results show that choosing appropriate initial estimate values can significantly affect the performance of the algorithm.

Through these studies, the paper provides a new and effective tool for solving nonlinear equations. Especially when dealing with complex or high - order nonlinear problems, fractional - order methods may show greater advantages.