Local discontinuous Galerkin methods for fractional ordinary differential equations (Q906946): Difference between revisions

A local discontinuous Galerkin (LDG) method is employed to solve initial value problems, which consists of fractional ordinary differential equations of the type \[ {}^C_aD^\alpha_t\,x(t)=f(x,t), \] where \(0<\alpha\leq1,\) together with an appropriate initial condition. The fractional derivative is of Caputo type. The basic idea in implementing the LDG scheme is to split the above equation into a system of a first-order ordinary differential equation and a fractional integral equation. Consequently, there is no need to add a penalty term or to introduce a numerical flux for the fractional integral equation. The authors introduce an \(L^\infty\)-stability of the scheme when \(f(x,t)\) is linear in \(x\), i.e., \(f(x,t)=Ax+B(t)\), \(A\) is a negative constant. They also investigate the convergence rates when \(f(x,t)\) is linear in \(x\) and give some numerical examples. The procedure is applied to approximately compute a generalized Mittag-Leffler function. In a similar manner, the method is applied to higher-order fractional initial value problems with illustrative examples, involving nonlinear functions \(f(x,t)\).