A numerical Bernstein splines approach for nonlinear initial value problems with Hilfer fractional derivative

Abstract: The Hilfer fractional derivative interpolates the commonly used Riemann-Liouville and Caputo fractional derivative. In general, solutions to Hilfer fractional differential equations are singular for $t \downarrow 0$ and are difficult to approximate with high numerical accuracy. We propose a numerical Bernstein splines technique to approximate solutions to generalized nonlinear initial values problems with Hilfer fractional derivatives. Convergent approximations are obtained using an efficient vectorized solution setup with few convergence requirements for a wide range of nonlinear fractional differential equations. We demonstrate efficiency of the developed method by applying it to the fractional Van der Pol oscillator, a system with applications in control systems and electronic circuits.