Fractional differential equations and numerical methods

Abstract

The increasing use of Fractional Calculus demands more accurate arid efficient methods for the numerical solution of fractional differential equations. We introduce the concepts of Fractional Calculus and give the definitions of fractional integrals and derivatives in the Riemann-Liouville and Caputo forms. We explore three existing Numerical Methods of solution of Fractional Differential Equations. 1. Diethelm's Backward Difference Form (BDF) method. 2. Lubich's Convolution Quadrature method. 3. Luchko and Diethelm's Operational Calculus (using the Mittag-Lefner function) method. We present useful recursive expressions we developed to compute the Taylor Series coefficients in the Operational Calculus method. These expressions are used in the calculation of the convolution and starting weights. We compare their accuracy and performance of the numerical methods, and conclude that the more complex methods produce the more accurate results.